Man

ModelSim Users Manual
Version 6.1a Published: 12/Jul/05
DRAFT
Copyright Mentor Graphics Corporation 2005 All rights reserved.

This document contains information that is proprietary to Mentor Graphics Corporation. The original recipient of this document may duplicate this document in whole or in part for internal business purposes only, provided that this entire notice appears in all copies. In duplicating any part of this document, the recipient agrees to make every reasonable effort to prevent the unauthorized use and distribution of the proprietary information.
UM-2
This document is for information and instruction purposes. Mentor Graphics reserves the right to make changes in specifications and other information contained in this publication without prior notice, and the reader should, in all cases, consult Mentor Graphics to determine whether any changes have been made. The terms and conditions governing the sale and licensing of Mentor Graphics products are set forth in written agreements between Mentor Graphics and its customers. No representation or other affirmation of fact contained in this publication shall be deemed to be a warranty or give rise to any liability of Mentor Graphics whatsoever. MENTOR GRAPHICS MAKES NO WARRANTY OF ANY KIND WITH REGARD TO THIS MATERIAL INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. MENTOR GRAPHICS SHALL NOT BE LIABLE FOR ANY INCIDENTAL, INDIRECT, SPECIAL, OR CONSEQUENTIAL DAMAGES WHATSOEVER (INCLUDING BUT NOT LIMITED TO LOST PROFITS) ARISING OUT OF OR RELATED TO THIS PUBLICATION OR THE INFORMATION CONTAINED IN IT, EVEN IF MENTOR GRAPHICS CORPORATION HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. RESTRICTED RIGHTS LEGEND 03/97 U.S. Government Restricted Rights. The SOFTWARE and documentation have been developed entirely at private expense and are commercial computer software provided with restricted rights. Use, duplication or disclosure by the U.S. Government or a U.S. Government subcontractor is subject to the restrictions set forth in the license agreement provided with the software pursuant to DFARS 227.7202-3(a) or as set forth in subparagraph (c)(1) and (2) of the Commercial Computer Software - Restricted Rights clause at FAR 52.227-19, as applicable. Contractor/manufacturer is: Mentor Graphics Corporation 8005 S.W. Boeckman Road, Wilsonville, Oregon 97070-7777. Telephone: 503.685.7000 Toll-Free Telephone: 800.592.2210 Website: www.mentor.com SupportNet: www.mentor.com/supportnet Contact Your Technical Writer: www.mentor.com/supportnet/documentation/ reply_form.cfm
TRADEMARKS: The trademarks, logos and service marks ("Marks") used herein are the property of Mentor Graphics Corporation or other third parties. No one is permitted to use these Marks without the prior written consent of Mentor Graphics or the respective third-party owner. The use herein of a third-party Mark is not an
UM-3
attempt to indicate Mentor Graphics as a source of a product, but is intended to indicate a product from, or associated with, a particular third party. A current list of Mentor Graphics trademarks may be viewed at: www.mentor.com/ terms_conditions/trademarks.cfm.
UM-4
UM-5
Table of Contents
1 - Introduction (UM-19)
Tool structure and flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-20 Simulation task overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-21 Basic steps for simulation . . . . . . . . . . . . . Step 1 - Collecting files and mapping libraries . . . Step 2 - Compiling the design with vlog/vcom/sccom Step 3 - Loading the design for simulation . . . . Step 4 - Simulating the design . . . . . . . . . Step 5- Debugging the design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-22 . UM-22 . UM-23 . UM-24 . UM-24 . UM-24
Modes of operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-25 Command-line mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-25 Batch mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-26 Standards supported . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-27 Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-27 Sections in this document . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-28 What is an "object" Text conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-30 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-30 . . . . . . . . . . . . . . . . . . . . . . . . . UM-30
Installation directory pathnames
Where to find our documentation . . . . . . . . . . . . . . . . . . . . . . . . . UM-31 Technical support and updates . . . . . . . . . . . . . . . . . . . . . . . . . . UM-32 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-33
2 - Projects (UM-35)
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-36 Project conversion between versions . . . . . . . . . . . . . . . . . . . . . . UM-37 Getting started with projects . . . . . . Step 1 Creating a new project . . Step 2 Adding items to the project Step 3 Compiling the files . . . . Step 4 Simulating a design . . . Other basic project operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-38 . UM-38 . UM-39 . UM-41 . UM-42 . UM-42
The Project tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-43 Changing compile order . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-44 Grouping files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-45 Creating a Simulation Configuration . . . . . . . . . . . . . . . . . . . . . . . . UM-46 Organizing projects with folders . . . . . . . . . . . . . . . . . . . . . . . . . UM-48 Specifying file properties and project settings . . . . . . . . . . . . . . . . . . . . UM-50 Project settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-51 Accessing projects from the command line . . . . . . . . . . . . . . . . . . . . . UM-52
UM-6
Table of Contents
3 - Design libraries (UM-53)

Design library overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-54 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-56 . UM-57 . UM-58 . UM-59 . UM-59 . UM-60 . UM-60 . UM-61 . UM-61 . UM-62 . UM-62 . UM-63 . UM-63 . UM-64 . UM-64 . UM-64 Working with design libraries . . . . . . . Managing library contents . . . . . . . Assigning a logical name to a design library Moving a library . . . . . . . . . . . Setting up libraries for group use . . . .
Specifying the resource libraries . . . . . . . . . . . . Predefined libraries . . . . . . . . . . . . . . . . Alternate IEEE libraries supplied . . . . . . . . . . Rebuilding supplied libraries . . . . . . . . . . . . Regenerating your design libraries . . . . . . . . . . Maintaining 32-bit and 64-bit versions in the same library Referencing source files with location maps Using location mapping . . . . . . Pathname syntax . . . . . . . . . How location mapping works . . . . Mapping with Tcl variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Importing FPGA libraries . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-65 Protecting source code using -nodebug . . . . . . . . . . . . . . . . . . . . . . . UM-66
4 - VHDL simulation (UM-67)

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-69 Basic VHDL flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-69 Compiling VHDL files . . . . . . . . Invoking the VHDL compiler . . . Dependency checking . . . . . . . Range and index checking . . . . . Subprogram inlining . . . . . . . Differences between language versions Simulating VHDL designs . . Simulator resolution limit Default binding . . . . Delta delays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-70 . UM-70 . UM-70 . UM-71 . UM-71 . UM-72 . UM-75 . UM-75 . UM-76 . UM-77
Using the TextIO package . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-79 Syntax for file declaration . . . . . . . . . . . . . . . . . . . . . . . . . . UM-79 Using STD_INPUT and STD_OUTPUT within ModelSim . . . . . . . . . . . . . UM-80 TextIO implementation issues . . . . . . . Reading and writing hexadecimal numbers Dangling pointers . . . . . . . . . . The ENDLINE function . . . . . . . . The ENDFILE function . . . . . . . . Using alternative input/output files . . . Flushing the TEXTIO buffer . . . . . . Providing stimulus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-81 . UM-82 . UM-82 . UM-82 . UM-82 . UM-83 . UM-83 . UM-83
UM-7
VITAL specification and source code . . . . . . . . . . . . . . . . . . . . . . . UM-84 VITAL packages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-84 ModelSim VITAL compliance . . . . . . . . . . . . . . . . . . . . . . . . . . UM-84 VITAL compliance checking . . . . . . . . . . . . . . . . . . . . . . . . . UM-85 Compiling and simulating with accelerated VITAL packages . . . . . . . . . . . . . . UM-86 Util package . . . . get_resolution . . init_signal_driver() init_signal_spy() . signal_force() . . signal_release() . to_real() . . . . to_time() . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-87 . UM-87 . UM-88 . UM-88 . UM-88 . UM-88 . UM-89 . UM-90
Modeling memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-91 87 and 93 example . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-91 02 example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-94 Affecting performance by cancelling scheduled events . . . . . . . . . . . . . . . . UM-98 Converting an integer into a bit_vector . . . . . . . . . . . . . . . . . . . . . . . UM-99
5 - Verilog and SystemVerilog simulation (UM-101)

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-102 Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-102 Basic Verilog flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-102 Compiling Verilog files . . . . . . . . . . Creating a working library . . . . . . . Invoking the Verilog compiler . . . . . Incremental compilation . . . . . . . . Library usage . . . . . . . . . . . . Verilog-XL compatible compiler arguments Verilog-XL ùselib compiler directive . . Verilog configurations . . . . . . . . Verilog generate statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-103 UM-103 UM-103 UM-105 UM-107 UM-109 UM-110 UM-112 UM-113 UM-114 UM-114 UM-117 UM-119 UM-121 UM-121
Simulating Verilog designs . . . . . . . . . Simulator resolution limit . . . . . . . . Event ordering in Verilog designs . . . . . Debugging event order issues . . . . . . . Negative timing check limits . . . . . . . Verilog-XL compatible simulator arguments .
Cell libraries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-123 Delay modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-123 System tasks and functions . . . . . . . . . . . IEEE Std 1364 system tasks and functions . . . SystemVerilog system tasks and functions . . . System tasks and functions specific to ModelSim . Verilog-XL compatible system tasks and functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-125 UM-125 UM-127 UM-128 UM-129
UM-8
Table of Contents
Compiler directives . . . . . . . . . . . IEEE Std 1364 compiler directives . . . Compiler directives specific to ModelSim . Verilog-XL compatible compiler directives
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
UM-132 UM-132 UM-133 UM-134
Verilog PLI/VPI and SystemVerilog DPI . . . . . . . . . . . . . . . . . . . . . UM-135
6 - SystemC simulation (UM-137)

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-139 Supported platforms and compiler versions . . . . . . . . . . . . . . . . . . . . UM-140 Building gcc with custom configuration options . . . . . . . . . . . . . . . . UM-140 HP Limitations for SystemC . . . . . . . . . . . . . . . . . . . . . . . . UM-141 Usage flow for SystemC-only designs . . . . . . . . . . . . . . . . . . . . . . UM-142 Compiling SystemC files . . . . . . . . . . . . . Creating a design library . . . . . . . . . . . Modifying SystemC source code . . . . . . . . Code modification examples . . . . . . . . . . Invoking the SystemC compiler . . . . . . . . . Compiling optimized and/or debug code . . . . . Specifying an alternate g++ installation . . . . . . Maintaining portability between OSCI and ModelSim Restrictions on compiling with HP aCC . . . . . Switching platforms and compilation . . . . . . . Using sccom vs. raw C++ compiler . . . . . . . Issues with C++ templates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-143 UM-143 UM-143 UM-144 UM-146 UM-146 UM-147 UM-147 UM-148 UM-148 UM-149 UM-150
Linking the compiled source . . . . . . . . . . . . . . . . . . . . . . . . . . UM-151 sccom -link . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-151 Simulating SystemC designs . . . . . . . . . . . . . Loading the design . . . . . . . . . . . . . . . Running simulation . . . . . . . . . . . . . . SystemC time unit and simulator resolution . . . . . Initialization and cleanup of SystemC state-based code Debugging the design . . . Viewable SystemC objects Waveform compare . . Source-level debug . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-152 UM-152 UM-152 UM-153 UM-154 UM-155 UM-155 UM-156 UM-157 UM-159 UM-159 UM-159 UM-161
SystemC object and type display in ModelSim Support for Globals and Statics . . . . Support for aggregates . . . . . . . Viewing FIFOs . . . . . . . . . .
Differences between ModelSim and the OSCI simulator . . . . . . . . . . . . . . . UM-162 Fixed-point types . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-162 OSCI 2.1 feature implementation details . . . Support for OSCI TLM library . . . . . Phase callback . . . . . . . . . . . Accessing command-line arguments . . . Construction parameters for SystemC types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-163 UM-163 UM-163 UM-163 UM-164
UM-9
Troubleshooting SystemC errors . . . . . . . . . . . . . . . . . . . . . . . . UM-166 Unexplained behaviors during loading or runtime . . . . . . . . . . . . . . . . UM-166 Errors during loading . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-166
7 - Mixed-language simulation (UM-171)

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-172 Basic mixed-language flow . . . . . . . . . . . . . . . . . . . . . . . . UM-172 Separate compilers, common design libraries . . . . . . Access limitations in mixed-language designs . . . . Simulator resolution limit . . . . . . . . . . . . Runtime modeling semantics . . . . . . . . . . . Hierarchical references in mixed HDL/SystemC designs Mapping data types . . . . . . . . . . . . . . . Verilog to VHDL mappings . . . . . . . . . . VHDL to Verilog mappings . . . . . . . . . . Verilog and SystemC signal interaction and mappings VHDL and SystemC signal interaction and mappings VHDL: instantiating Verilog . . . . Verilog instantiation criteria . . Component declaration . . . . vgencomp component declaration Modules with unnamed ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-173 UM-173 UM-174 UM-174 UM-175 UM-176 UM-176 UM-178 UM-179 UM-184 UM-188 UM-188 UM-188 UM-189 UM-191 UM-192 UM-192 UM-192 UM-192 UM-192 UM-193 UM-194 UM-194 UM-194 UM-196 UM-197 UM-199 UM-199 UM-199 UM-199 UM-200 UM-202 UM-202 UM-202 UM-204 UM-205
Verilog: instantiating VHDL . . . . . . . . . . VHDL instantiation criteria . . . . . . . . Entity/architecture names and escaped identifiers Named port associations . . . . . . . . . . Generic associations . . . . . . . . . . . SDF annotation . . . . . . . . . . . . .
SystemC: instantiating Verilog . . . . . . . . . . . Verilog instantiation criteria . . . . . . . . . . SystemC foreign module declaration . . . . . . . Parameter support for SystemC instantiating Verilog Example of parameter use . . . . . . . . . . . Verilog: instantiating SystemC . . . . . . . . . . . SystemC instantiation criteria . . . . . . . . . . Exporting SystemC modules . . . . . . . . . . Parameter support for Verilog instantiating SystemC Example of parameter use . . . . . . . . . . . SystemC: instantiating VHDL . . . . . . . . . . VHDL instantiation criteria . . . . . . . . . SystemC foreign module declaration . . . . . . Generic support for SystemC instantiating VHDL Example of generic use . . . . . . . . . . . . . . . .
VHDL: instantiating SystemC . . . . . . . . . . . . . . . . . . . . . . . . . UM-208 SystemC instantiation criteria . . . . . . . . . . . . . . . . . . . . . . . . UM-208 Component declaration . . . . . . . . . . . . . . . . . . . . . . . . . . UM-208
UM-10
Table of Contents
vgencomp component declaration . . . . . . . Exporting SystemC modules . . . . . . . . . sccom -link . . . . . . . . . . . . . . . . Generic support for VHDL instantiating SystemC
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
UM-208 UM-209 UM-209 UM-209
8 - WLF files (datasets) and virtuals (UM-211)

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-212 Saving a simulation to a WLF file . . . . . . . . . . . . . . . . . . . . . . . . UM-213 WLF file parameter overview . . . . . . . . . . . . . . . . . . . . . . . UM-213 Opening datasets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-215 Viewing dataset structure . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-216 Structure tab columns . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-217 Managing multiple datasets . . . . . GUI . . . . . . . . . . . . . Command line . . . . . . . . Restricting the dataset prefix display . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-218 UM-218 UM-218 UM-219
Saving at intervals with Dataset Snapshot . . . . . . . . . . . . . . . . . . . . . UM-220 Collapsing time and delta steps . . . . . . . . . . . . . . . . . . . . . . . . . UM-221 Virtual Objects (User-defined buses, and more) Virtual signals . . . . . . . . . . . . Virtual functions . . . . . . . . . . . Virtual regions . . . . . . . . . . . Virtual types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-222 UM-222 UM-223 UM-224 UM-224
9 - Waveform analysis (UM-225)

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-227 Objects you can view . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-227 Wave window overview . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-228 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-232 UM-232 UM-232 UM-232 UM-232 UM-233 UM-233 UM-234 UM-235 List window overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-231 Adding objects to the Wave or List window . Adding objects with drag and drop . . Adding objects with a menu command . Adding objects with a command . . . Adding objects with a window format file
Measuring time with cursors in the Wave window Working with cursors . . . . . . . . . . Understanding cursor behavior . . . . . . Jumping to a signal transition . . . . . . .
Setting time markers in the List window . . . . . . . . . . . . . . . . . . . . . UM-236 Working with markers . . . . . . . . . . . . . . . . . . . . . . . . . . UM-236 Zooming the Wave window display . . . . . . . . . . . . . . . . . . . . . . . UM-237 Zooming with menu commands . . . . . . . . . . . . . . . . . . . . . . . UM-237 Zooming with toolbar buttons . . . . . . . . . . . . . . . . . . . . . . . UM-237
UM-11
Zooming with the mouse . . . . . . . . . . . . . . . . . . . . . . . . . UM-237 Saving zoom range and scroll position with bookmarks . . . . . . . . . . . . . UM-238 Searching in the Wave and List windows . . . . . . . Finding signal names . . . . . . . . . . . . . Searching for values or transitions . . . . . . . . Using the Expression Builder for expression searches Formatting the Wave window . . . . . . Setting Wave window display properties Formatting objects in the Wave window Dividing the Wave window . . . . . Splitting Wave window panes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-239 UM-239 UM-240 UM-241 UM-243 UM-243 UM-243 UM-244 UM-245
Formatting the List window . . . . . . . . . . . . . . . . . . . . . . . . . . UM-247 Setting List window display properties . . . . . . . . . . . . . . . . . . . . UM-247 Formatting objects in the List window . . . . . . . . . . . . . . . . . . . . UM-247 Saving the window format . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-249 Printing and saving waveforms in the Wave window Saving a .eps file and printing under UNIX . . Printing on Windows platforms . . . . . . . Printer page setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-250 UM-250 UM-250 UM-250
Saving List window data to a file . . . . . . . . . . . . . . . . . . . . . . . . UM-251 Combining objects/creating busses . . . . . . . . . . . . . . . . . . . . . . . UM-252 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-252 Configuring new line triggering in the List window . . . . . . . . . . . . . . . . . UM-253 Using gating expressions to control triggering . . . . . . . . . . . . . . . . . UM-254 Sampling signals at a clock change . . . . . . . . . . . . . . . . . . . . . UM-256 Miscellaneous tasks . . . . . . . . . . . . . Examining waveform values . . . . . . . . Displaying drivers of the selected waveform . . Setting signal breakpoints in the Wave window Waveform Compare . . . . . . . . . . . . Mixed-language waveform compare support . Three options for setting up a comparison . . Setting up a comparison with the GUI . . . Starting a waveform comparison . . . . . Adding signals, regions, and clocks . . . . Specifying the comparison method . . . . Setting compare options . . . . . . . . . Viewing differences in the Wave window . . Viewing differences in the List window . . Viewing differences in textual format . . . Saving and reloading comparison results . . Comparing hierarchical and flattened designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-257 UM-257 UM-257 UM-257 UM-258 UM-258 UM-258 UM-259 UM-260 UM-262 UM-264 UM-266 UM-267 UM-269 UM-270 UM-270 UM-271
10 - Tracing signals with the Dataflow window (UM-273)

Dataflow window overview . . . . . . . . . . . . . . . . . . . . . . . . . . UM-274 Objects you can view . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-274
UM-12
Table of Contents
Adding objects to the window . . . . . . . . . . . . . . . . . . . . . . . . . UM-275 Links to other windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-276 Exploring the connectivity of your design . . . . . . . . . . . . . . . . . . . . . UM-277 Tracking your path through the design . . . . . . . . . . . . . . . . . . . . UM-277 The embedded wave viewer . . . . . . . . . . . . . . . . . . . . . . . . . . UM-278 Zooming and panning . . . . . Zooming with toolbar buttons Zooming with the mouse . . Panning with the mouse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-279 UM-279 UM-279 UM-279
Tracing events (causality) . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-280 Tracing the source of an unknown (X) . . . . . . . . . . . . . . . . . . . . . . UM-281 Finding objects by name in the Dataflow window . . . . . . . . . . . . . . . . . UM-283 Saving the display . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-284 Saving a .eps file . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-284 Printing on Windows platforms . . . . . . . . . . . . . . . . . . . . . . . UM-285 Configuring page setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-286 Symbol mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-287 Configuring window options . . . . . . . . . . . . . . . . . . . . . . . . . . UM-289
11 - Measuring code coverage (UM-291)

Introduction . . . . . . . . . . . . . Usage flow for code coverage . . . . Supported types . . . . . . . . . . Important notes about coverage statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-292 UM-292 UM-293 UM-294
Enabling code coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-295 Viewing coverage data in the Main window . . . . . . . . . . . . . . . . . . . . UM-297 Viewing coverage data in the Source window . . . . . . . . . . . . . . . . . . . UM-298 Toggle coverage . . . . . . . . . . . . . . . . . Enabling toggle coverage . . . . . . . . . . . Viewing toggle coverage data in the Objects pane . Toggle coverage reporting . . . . . . . . . . . Port collapsing and toggle coverage . . . . . . . Excluding nodes from toggle coverage . . . . . . Excluding bus bits from toggle coverage . . . . . Excluding VHDL enum signals from toggle statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-300 UM-300 UM-301 UM-301 UM-302 UM-302 UM-303 UM-303 UM-305 UM-305 UM-305 UM-306 UM-307 UM-307 UM-308
Setting a coverage threshold . . . . . . . . . . . . . . . . . . . . . . . . . . UM-304 Excluding objects from coverage . . . . . . . . . . . Exclude lines/files via the GUI . . . . . . . . . . Exclude lines/files with pragmas . . . . . . . . . Exclude lines/files with a filter file . . . . . . . . Exclude lines/rows from UDP truth tables . . . . . Exclude lines/rows with the coverage exclude command Exclude nodes from toggle statistics . . . . . . . .
UM-13
Exclude VHDL enum signals from toggle statistics . . . . . . . . . . . . . . . UM-308 Reporting coverage data . . . . . Command example . . . . . . GUI example . . . . . . . . Setting a default coverage mode . XML output . . . . . . . . Sample reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-309 UM-309 UM-309 UM-310 UM-310 UM-311
Saving and reloading coverage data . . . . . . . . . . . . . . . . . . . . . . . UM-314 With the vcover utility . . . . . . . . . . . . . . . . . . . . . . . . . . UM-315 Coverage statistics details . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-316 Condition coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-316 Expression coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-317
12 - C Debug (UM-319)
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-320 Supported platforms and gdb versions . . . . . . . . . . . . . . . . . . . . . . UM-321 Running C Debug on Windows platforms . . . . . . . . . . . . . . . . . . . UM-321 Setting up C Debug . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-322 Running C Debug from a DO file . . . . . . . . . . . . . . . . . . . . . . UM-322 Setting breakpoints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-323 Stepping in C Debug . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-325 Known problems with stepping in C Debug . . . . . . . . . . . . . . . . . . UM-325 Finding function entry points with Auto find bp . . . . . . . . . . . . . . . . . . UM-326 Identifying all registered function calls . Enabling Auto step mode . . . . Example . . . . . . . . . . . Auto find bp versus Auto step mode Debugging functions during elaboration FLI functions in initialization mode PLI functions in initialization mode VPI functions in initialization mode Completing design load . . . . . C Debug command reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-327 UM-327 UM-328 UM-329 UM-330 UM-331 UM-331 UM-333 UM-333
Debugging functions when quitting simulation . . . . . . . . . . . . . . . . . . . UM-334 . . . . . . . . . . . . . . . . . . . . . . . . . UM-335
13 - Profiling performance and memory use (UM-337)

Introducing performance and memory profiling . . . . . . . . . . . . . . . . . . UM-338 A statistical sampling profiler . . . . . . . . . . . . . . . . . . . . . . . UM-338 A memory allocation profiler . . . . . . . . . . . . . . . . . . . . . . . . UM-338 Getting started . . . . . . . . . . . . . . . . . . . Enabling the statistical sampling profiler . . . . . . . Collecting memory allocation and performance data . . . Running the profiler on Windows with FLI/PLI/VPI code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-339 UM-341 UM-341 UM-342
UM-14
Table of Contents
Interpreting profiler data Viewing profiler results . The Ranked View . The Call Tree view . The Structural View
. . . . . . . . . . . . . . . . . . . . . . . . . . . UM-343 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-344 UM-344 UM-345 UM-347
Viewing profile details . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-348 Integration with Source windows . . . . . . . . . . . . . . . . . . . . . . . . UM-350 Analyzing C code performance . . . . . . . . . . . . . . . . . . . . . . . . . UM-351 Reporting profiler results . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-352
14 - Signal Spy (UM-355)

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-356 disable_signal_spy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-357 enable_signal_spy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-358 init_signal_driver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-359 init_signal_spy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-362 signal_force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-365 signal_release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-367 $disable_signal_spy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-369 $enable_signal_spy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-370 $init_signal_driver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-371 $init_signal_spy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-374 $signal_force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-377 $signal_release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-379
15 - Standard Delay Format (SDF) Timing Annotation (UM-381)

Specifying SDF files for simulation Instance specification . . . . SDF specification with the GUI Errors and warnings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-382 UM-382 UM-383 UM-383
VHDL VITAL SDF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-384 SDF to VHDL generic matching . . . . . . . . . . . . . . . . . . . . . . UM-384 Resolving errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-385 Verilog SDF . . . . . . . . . . . The $sdf_annotate system task . . SDF to Verilog construct matching Optional edge specifications . . . Optional conditions . . . . . . Rounded timing values . . . . . Interconnect delays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-386 UM-386 UM-387 UM-390 UM-391 UM-391
SDF for mixed VHDL and Verilog designs . . . . . . . . . . . . . . . . . . . . UM-392 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-393
UM-15
Disabling timing checks
. . . . . . . . . . . . . . . . . . . . . . . . . . . UM-393
Troubleshooting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-394 Mistaking a component or module name for an instance label . . . . . . . . . . . UM-395 Forgetting to specify the instance . . . . . . . . . . . . . . . . . . . . . . UM-395
16 - Value Change Dump (VCD) Files (UM-397)

Creating a VCD file . . . . . Flow for four-state VCD file Flow for extended VCD file Case sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-398 UM-398 UM-398 UM-398
Using extended VCD as stimulus . . . . . . . . . . . . . . . . . . . . . . . . UM-399 Simulating with input values from a VCD file . . . . . . . . . . . . . . . . . UM-399 Replacing instances with output values from a VCD file . . . . . . . . . . . . . UM-400 ModelSim VCD commands and VCD tasks . . . . . . . . . . . . . . . . . . . . UM-402 Compressing files with VCD tasks . . . . . . . . . . . . . . . . . . . . . . UM-403 A VCD file from source to output . . . . . . . . . . . . . . . . . . . . . . . . UM-404 VCD simulator commands . . . . . . . . . . . . . . . . . . . . . . . . . UM-404 VCD output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-405 Capturing port driver data . . . . . . . . . Driver states . . . . . . . . . . . . Driver strength . . . . . . . . . . . Identifier code . . . . . . . . . . . . Resolving values . . . . . . . . . . . Example VCD output from vcd dumpports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-408 UM-408 UM-409 UM-409 UM-409 UM-411
17 - Tcl and macros (DO files) (UM-413)

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-414 Tcl features within ModelSim . . . . . . . . . . . . . . . . . . . . . . . UM-414 Tcl References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-414 Tcl commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-415 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-416 UM-418 UM-418 UM-420 UM-420 UM-420 UM-420 UM-421 UM-421 Tcl command syntax . . . . . . . . if command syntax . . . . . . . Command substitution . . . . . Command separator . . . . . . Multiple-line commands . . . . . Evaluation order . . . . . . . . Tcl relational expression evaluation Variable substitution . . . . . . System commands . . . . . . .
List processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-422 ModelSim Tcl commands . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-422 ModelSim Tcl time commands . . . . . . . . . . . . . . . . . . . . . . . . . UM-423 Conversions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-423 Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-423
UM-16
Table of Contents
Arithmetic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-424 Tcl examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-425 Macros (DO files) . . . . . . . . . . . . . . . . Using Parameters with DO files . . . . . . . . . Deleting a file from a .do script . . . . . . . . . Making macro parameters optional . . . . . . . Useful commands for handling breakpoints and errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-429 UM-429 UM-429 UM-430 UM-431
A - Simulator variables (UM-433)

Variable settings report . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-434 Environment variables . . . . . . . . . . . . . . Creating environment variables in Windows . . . . Referencing environment variables within ModelSim Removing temp files (VSOUT) . . . . . . . . . Control variables located in INI files . . . . . [Library] library path variables . . . . . [vlog] Verilog compiler control variables . [vcom] VHDL compiler control variables . [sccom] SystemC compiler control variables [vsim] simulator control variables . . . . [msg_system] message system variables . Commonly used INI variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-434 UM-435 UM-436 UM-437 UM-438 UM-439 UM-439 UM-441 UM-442 UM-444 UM-449 UM-451
Variable precedence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-454 Simulator state variables . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-455 Referencing simulator state variables . . . . . . . . . . . . . . . . . . . . . UM-455 Special considerations for the now variable . . . . . . . . . . . . . . . . . . UM-456
B - Error and warning messages (UM-457)

Message system . . . . . . . . . Message format . . . . . . . Getting more information . . . Changing message severity level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-458 UM-458 UM-458 UM-458 UM-460 UM-460 UM-460 UM-460 UM-463 UM-463 UM-463 UM-463 UM-464 UM-464 UM-464 UM-466
Suppressing warning messages . . . . . Suppressing VCOM warning messages Suppressing VLOG warning messages Suppressing VSIM warning messages Miscellaneous messages . . . . . . Compilation of DPI export TFs error Empty port name warning . . . . Lock message . . . . . . . . . Metavalue detected warning . . . Sensitivity list warning . . . . . Tcl Initialization error 2 . . . . . Too few port connections . . . . . . . . . . . .
Exit codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-461
UM-17
VSIM license lost . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-467 Failed to find libswift entry . . . . . . . . . . . . . . . . . . . . . . . . UM-467 sccom error messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-468 Failed to load sc lib error: undefined symbol . . . . . . . . . . . . . . . . . . UM-468 Multiply defined symbols . . . . . . . . . . . . . . . . . . . . . . . . . UM-469 Enforcing strict 1076 compliance with -pedanticerrors . . . . . . . . . . . . . . . . UM-470
C - Verilog PLI / VPI / DPI (UM-473)

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-475 Registering PLI applications . . . . . . . . . . . . . . . . . . . . . . . . . . UM-476 Registering VPI applications . . . . . . . . . . . . . . . . . . . . . . . . . . UM-478 Registering DPI applications . . . . . . . . . . . . . . . . . . . . . . . . . . UM-480 DPI use flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-481 Steps in flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-481 Use model for read-only work libraries . . . . . . . . . . . . . . . . . . . . UM-482 Compiling and linking C applications for PLI/VPI/DPI . . . . . . . . . . . . . . . UM-484 For all UNIX platforms . . . . . . . . . . . . . . . . . . . . . . . . . . UM-484 Compiling and linking C++ applications for PLI/VPI/DPI . . . . . . . . . . . . . . UM-491 Specifying application files to load . . . . . . . . . PLI/VPI file loading . . . . . . . . . . . . . DPI file loading . . . . . . . . . . . . . . . Loading shared objects with global symbol visibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-497 UM-497 UM-497 UM-498
PLI example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-499 VPI example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-500 DPI example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-501 The PLI callback reason argument . . . . . . . . . . . . . . . . . . . . . . . UM-502 The sizetf callback function . . . . . . . . . . . . . . . . . . . . . . . . . . UM-504 PLI object handles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-505 Third party PLI applications . . . . . . . . . . . . . . . . . . . . . . . . . . UM-506 Support for VHDL objects . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-507 IEEE Std 1364 ACC routines IEEE Std 1364 TF routines . . . . . . . . . . . . . . . . . . . . . . . . . UM-508 . . . . . . . . . . . . . . . . . . . . . . . . . . UM-510 . . . . . . . . . . . . . . . . . . . . . . . . UM-512
SystemVerilog DPI access routines . . . . . . . . . . . . . . . . . . . . . . . UM-512 Verilog-XL compatible routines 64-bit support for PLI . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-513 Using 64-bit ModelSim with 32-bit Applications . . . . . . . . . . . . . . . . UM-513 PLI/VPI tracing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-514 Debugging PLI/VPI/DPI application code . . . . . . . . . . . . . . . . . . . . UM-516 HP-UX specific warnings . . . . . . . . . . . . . . . . . . . . . . . . . UM-516
UM-18
Table of Contents
D - ModelSim shortcuts (UM-517)

Command shortcuts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-517 Main and Source window mouse and keyboard shortcuts . . . . . . . . . . . . . . . UM-519 List window keyboard shortcuts . . . . . . . . . . . . . . . . . . . . . . . . UM-522 Wave window mouse and keyboard shortcuts . . . . . . . . . . . . . . . . . . . UM-523
E - System initialization (UM-525)

Files accessed during startup . . . . . . . . . . . . . . . . . . . . . . . . . . UM-526 Environment variables accessed during startup . . . . . . . . . . . . . . . . . . . UM-527 Initialization sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-529
F - Logic Modeling SmartModels (UM-531)

VHDL SmartModel interface . . . . . . . . Enabling the interface . . . . . . . . . . Creating foreign architectures with sm_entity Vector ports . . . . . . . . . . . . . Command channel . . . . . . . . . . . SmartModel Windows . . . . . . . . . Memory arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-532 UM-532 UM-533 UM-535 UM-536 UM-537 UM-538
Verilog SmartModel interface . . . . . . . . . . . . . . . . . . . . . . . . . UM-539 Linking the LMTV interface to the simulator . . . . . . . . . . . . . . . . . UM-539
Index
UM-19
1 - Introduction
Chapter contents
Tool structure and flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-20 . UM-21 . UM-22 . UM-25 . UM-25 . UM-26 . UM-27 . UM-27 . UM-28 . UM-30 . UM-30 . UM-30 . UM-31 . UM-32 Simulation task overview . Basic steps for simulation . Modes of operation. . . Command-line mode . Batch mode . . . Standards supported Assumptions . . . . . .
Sections in this document . What is an "object". Text conventions . . . . .
Installation directory pathnames . Where to find our documentation . Technical support and updates . .
This documentation was written for ModelSim for UNIX and Microsoft Windows. Not all versions of ModelSim are supported on all platforms. Contact your Mentor Graphics sales representative for details.
UM-20
1 - Introduction
Tool structure and flow

The diagram below illustrates the structure of the ModelSim tool, and the flow of that tool as it is used to verify a design.
VHDL Design Libraries
vlib vmap
local work library
Map libraries
Libraries
Vendor
Design files
vlog/ vcom/ sccom

Analyze/ Compile
.ini or .mpf file
HDL/SystemC Analyze/ Compile
compiled database
vsim
Interactive Debugging activities i.e. Simulation Output (e.g., vcd)
Simulate
Debug
Post-processing Debug
Simulation task overview UM-21
Simulation task overview

The following table provides a reference for the tasks required for compiling, loading, and simulating a design in ModelSim. Task Step 1: Map libraries Step 2: Compile the design Example command line entry vlib <library_name> vmap work <library_name> vlog file1.v file2.v ... (Verilog) vcom file1.vhd file2.vhd ... (VHDL) GUI menu pull-down a. File > New > Project b. Enter library name c. Add design files to project a. Compile > Compile or Compile > Compile All GUI icons N/A
Compile or Compile All icons:
Step 3: Load the design into the simulator
vsim <top> or vsim <opt_name>
a. Simulate > Start Simulation b. Click on top design module or optimized design unit name c. Click OK This action loads the design for simulation
Simulate icon:
Step 4: Run the simulation
run (CR-197) step (CR-213)
Simulate > Run
Run, or Run continue, or Run -all icons:
Step 5: Debug the design
Common debugging commands: bp (CR-61) describe (CR-123) drivers (CR-126) examine (CR-132) force (CR-141) log (CR-148) show (CR-209)
N/A
N/A
UM-22
1 - Introduction
Basic steps for simulation

This section provides further detail related to each step in the process of simulating your design using ModelSim.
Step 1 - Collecting files and mapping libraries

Files needed to run ModelSim on your design: design files (VHDL, Verilog, and/or SystemC), including stimulus for the design libraries, both working and resource modelsim.ini (automatically created by the library mapping command
Providing stimulus to the design

You can provide stimulus to your design in several ways: Language based testbench Tcl-based ModelSim interactive command, force (CR-141) VCD files / commands See "Using extended VCD as stimulus" (UM-399) and "Using extended VCD as stimulus"
(UM-399)
3rd party testbench generation tools
What is a library in ModelSim?

A library is a location where data to be used for simulation is stored. Libraries are ModelSims way of managing the creation of data before it is needed for use in simulation. It also serves as a way to streamline simulation invocation. Instead of compiling all design data each and every time you simulate, ModelSim uses binary pre-compiled data from these libraries. So, if you make a changes to a single Verilog module, only that module is recompiled, rather than all modules in the design.
Working and resource libraries

Design libraries can be used in two ways: 1) as a local working library that contains the compiled version of your design; 2) as a resource library. The contents of your working library will change as you update your design and recompile. A resource library is typically unchanging, and serves as a parts source for your design. Examples of resource libraries might be: shared information within your group, vendor libraries, packages, or previously compiled elements of your own working design. You can create your own resource libraries, or they may be supplied by another design team or a third party (e.g., a silicon vendor). For more information on resource libraries and working libraries, see "Working library versus resource libraries" (UM-54), "Managing library contents" (UM-57), "Working with design libraries" (UM-56), and "Specifying the resource libraries" (UM-60).
Creating the logical library - vlib

Before you can compile your source files, you must create a library in which to store the compilation results. You can create the logical library using the GUI, using File > New >
Basic steps for simulation UM-23
Library (see "Creating a library" (UM-56)), or you can use the vlib (CR-292) command. For example, the command:
vlib work
creates a library named work. By default, compilation results are stored in the work library.
Mapping the logical work to the physical work directory - vmap

VHDL uses logical library names that can be mapped to ModelSim library directories. If libraries are not mapped properly, and you invoke your simulation, necessary components will not be loaded and simulation will fail. Similarly, compilation can also depend on proper library mapping. By default, ModelSim can find libraries in your current directory (assuming they have the right name), but for it to find libraries located elsewhere, you need to map a logical library name to the pathname of the library. You can use the GUI ("Library mappings with the GUI" (UM-58), a command ("Library mappings with the GUI" (UM-58)), or a project ("Getting started with projects" (UM-38) to assign a logical name to a design library. The format for command line entry is:
vmap <logical_name> <directory_pathname>
This command sets the mapping between a logical library name and a directory.
Step 2 - Compiling the design with vlog/vcom/sccom

Designs are compiled with one of the three language compilers.
Compiling Verilog - vlog

ModelSims compiler for the Verilog modules in your design is vlog (CR-293). Verilog files may be compiled in any order, as they are not order dependent. See "Compiling Verilog files" (UM-103) for details.
Compiling VHDL - vcom

ModelSims compiler for VHDL design units is vcom (CR-246). VHDL files must be compiled according to the design requirements of the design. Projects may assist you in determining the compile order: for more information, see"Auto-generating compile order" (UM-44). See "Compiling VHDL files" (UM-70) for details. on VHDL compilation.
Compiling SystemC - sccom

ModelSims compiler for SystemC design units is sccom (CR-199), and is used only if you have SystemC components in your design. See "Compiling SystemC files" (UM-143) for details.
UM-24
1 - Introduction
Step 3 - Loading the design for simulation

vsim <top>
Your design is ready for simulation after it has been compiled You may then invoke vsim (CR-305) with the names of the top-level modules (many designs contain only one top-level module). For example, if your top-level modules are "testbench" and "globals", then invoke the simulator as follows:
vsim testbench globals
After the simulator loads the top-level modules, it iteratively loads the instantiated modules and UDPs in the design hierarchy, linking the design together by connecting the ports and resolving hierarchical references.
Using SDF
You can incorporate actual delay values to the simulation by applying SDF backannotation files to the design. For more information on how SDF is used in the design, see "Specifying SDF files for simulation" (UM-382).
Step 4 - Simulating the design

Once the design has been successfully loaded, the simulation time is set to zero, and you must enter a run command to begin simulation. For more information, see Verilog and SystemVerilog simulation (UM-101), VHDL simulation (UM-67), and SystemC simulation (UM-137). The basic simulator commands are: add wave (CR-49) force (CR-141) bp (CR-61) run (CR-197) step (CR-213)
Step 5- Debugging the design

Numerous tools and windows useful in debugging your design are available from the ModelSim GUI. In addition, several basic simulation commands are available from the command line to assist you in debugging your design: describe (CR-123) drivers (CR-126) examine (CR-132) force (CR-141) log (CR-148) show (CR-209)
Modes of operation UM-25
Modes of operation
Many users run ModelSim interactivelypushing buttons and/or pulling down menus in a series of windows in the GUI (graphical user interface). But there are really three modes of ModelSim operation, the characteristics of which are outlined in the following table.: ModelSim use mode GUI Characteristics interactive; has graphical windows, push-buttons, menus, and a command line in the transcript. Default mode. interactive command line; no GUI. non-interactive batch script; no windows or interactive command line. How ModelSim is invoked via a desktop icon or from the OS command shell prompt. Example:
OS> vsim
Command-line
with -c argument at the OS command prompt. Example:

OS> vsim -c
Batch
at OS command shell prompt using redirection of standard input. Example:

C:\ vsim vfiles.v <infile >outfile
The ModelSim Users Manual focuses primarily on the GUI mode of operation. However, this section provides an introduction to the Command-line and Batch modes.
Command-line mode
In command-line mode ModelSim executes any startup command specified by the Startup (UM-448) variable in the modelsim.ini file. If vsim (CR-305) is invoked with the -do "command_string" option, a DO file (macro) is called. A DO file executed in this manner will override any startup command in the modelsim.ini file. During simulation a transcript file is created containing any messages to stdout. A transcript file created in command-line mode may be used as a DO file if you invoke the transcript on command (CR-223) after the design loads (see the example below). The transcript on command writes all of the commands you invoke to the transcript file. For example, the following series of commands results in a transcript file that can be used for command input if top is re-simulated (remove the quit -f command from the transcript file if you want to remain in the simulator).
vsim -c top
library and design loading messages... then execute:

transcript on force clk 1 50, 0 100 -repeat 100 run 500 run @5000 quit -f
Rename transcript files that you intend to use as DO files. They will be overwritten the next time you run vsim if you dont rename them. Also, simulator messages are already commented out, but any messages generated from your design (and subsequently written to the transcript file) will cause the simulator to pause. A transcript file that contains only valid simulator commands will work fine; comment out anything else with a "#".
UM-26
1 - Introduction
Stand-alone tools pick up project settings in command-line mode if they are invoked in the project's root directory. If invoked outside the project directory, stand-alone tools pick up project settings only if you set the MODELSIM environment variable to the path to the project file (<Project_Root_Dir>/<Project_Name>.mpf).
Batch mode
Batch mode is an operational mode that provides neither an interactive command line nor interactive windows. In a Windows environment, vsim is run from a Windows command prompt and standard input and output are re-directed from and to files. Here is an example of a batch mode simulation using redirection of std input and output:
vsim counter < yourfile > outfile
where "yourfile" is a script containing various ModelSim commands. You can use the CNTL-C keyboard interrupt to break batch simulation in UNIX and Windows environments.
Standards supported UM-27
Standards supported
ModelSim VHDL implements the VHDL language as defined by IEEE Standards 1076-1987, 1076-1993, and 1076-2002. ModelSim also supports the 1164-1993 Standard Multivalue Logic System for VHDL Interoperability, and the 1076.2-1996 Standard VHDL Mathematical Packages standards. Any design developed with ModelSim will be compatible with any other VHDL system that is compliant with the 1076 specs. ModelSim Verilog implements the Verilog language as defined by the IEEE Std 1364-1995 and 1364-2005. ModelSim Verilog also supports a partial implementation of SystemVerilog P1800-2005 (see /<install_dir>/modeltech/docs/technotes/sysvlog.note for implementation details). Both PLI (Programming Language Interface) and VCD (Value Change Dump) are supported for ModelSim users. In addition, all products support SDF 1.0 through 4.0 (except the NETDELAY statement), VITAL 2.2b, VITAL95 IEEE 1076.4-1995, and VITAL 2000 IEEE 1076.4-2000. ModelSim implements the SystemC language based on the Open SystemC Initiative (OSCI) SystemC 2.1 reference simulator.
Assumptions
We assume that you are familiar with the use of your operating system and its graphical interface. We also assume that you have a working knowledge of VHDL, Verilog, and/or SystemC. Although ModelSim is an excellent tool to use while learning HDL concepts and practices, this document is not written to support that goal. Finally, we assume that you have worked the appropriate lessons in the ModelSim Tutorial and are familiar with the basic functionality of ModelSim. The ModelSim Tutorial is available from the ModelSim Help menu. The ModelSim Tutorial is also available from the Support page of our web site: www.model.com
UM-28
1 - Introduction
Sections in this document

In addition to this introduction, you will find the following major sections in this document: 2 - Projects (UM-35) This chapter discusses ModelSim "projects", a container for design files and their associated simulation properties. 3 - Design libraries (UM-53) To simulate an HDL design using ModelSim, you need to know how to create, compile, maintain, and delete design libraries as described in this chapter. 4 - VHDL simulation (UM-67) This chapter is an overview of compilation and simulation for VHDL within the ModelSim environment. 5 - Verilog and SystemVerilog simulation (UM-101) This chapter is an overview of compilation and simulation for Verilog and SystemVerilog within the ModelSim environment. 6 - SystemC simulation (UM-137) This chapter is an overview of preparation, compilation, and simulation for SystemC within the ModelSim environment. 7 - Mixed-language simulation (UM-171) This chapter outlines data mapping and the criteria established to instantiate design units between VHDL, Verilog, and SystemC. 8 - WLF files (datasets) and virtuals (UM-211) This chapter describes datasets and virtuals - both methods for viewing and organizing simulation data in ModelSim. 9 - Waveform analysis (UM-225) This chapter describes how to perform waveform analysis with the ModelSim Wave and List windows. 10 - Tracing signals with the Dataflow window (UM-273) This chapter describes how to trace signals and assess causality using the ModelSim Dataflow window. 11 - Measuring code coverage (UM-291) This chapter describes the Code Coverage feature. Code Coverage gives you graphical and report file feedback on how the source code is being executed. 12 - C Debug (UM-319) This chapter describes C Debug, a graphic interface to the gdb debugger that can be used to debug FLI/PLI/VPI/SystemC C/C++ source code.
Sections in this document UM-29
13 - Profiling performance and memory use (UM-337) This chapter describes how the ModelSim Performance Analyzer is used to easily identify areas in your simulation where performance can be improved. 14 - Signal Spy (UM-355) This chapter describes Signal Spy, a set of VHDL procedures and Verilog system tasks that let you monitor, drive, force, or release a design object from anywhere in the hierarchy of a VHDL or mixed design. 15 - Standard Delay Format (SDF) Timing Annotation (UM-381) This chapter discusses ModelSims implementation of SDF (Standard Delay Format) timing annotation. Included are sections on VITAL SDF and Verilog SDF, plus troubleshooting. 16 - Value Change Dump (VCD) Files (UM-397) This chapter explains Model Technologys Verilog VCD implementation for ModelSim. The VCD usage is extended to include VHDL designs. 17 - Tcl and macros (DO files) (UM-413) This chapter provides an overview of Tcl (tool command language) as used with ModelSim. A - Simulator variables (UM-433) This appendix describes environment, system, and preference variables used in ModelSim. B - Error and warning messages (UM-457) This appendix describes ModelSim error and warning messages. C - Verilog PLI / VPI / DPI (UM-473) This appendix describes the ModelSim implementation of the Verilog PLI and VPI. D - ModelSim shortcuts (UM-517) This appendix describes ModelSim keyboard and mouse shortcuts. E - System initialization (UM-525) This appendix describes what happens during ModelSim startup. F - Logic Modeling SmartModels (UM-531) This appendix describes the use of the SmartModel Library and SmartModel Windows with ModelSim.
UM-30
1 - Introduction
What is an "object"
Because ModelSim works with so many languages (SystemC, Verilog, VHDL, SystemVerilog, ), an object refers to any valid design element in those languages. The word "object" is used whenever a specific language reference is not needed. Depending on the context, object can refer to any of the following: VHDL Verilog SystemVerilog block statement, component instantiation, constant, generate statement, generic, package, signal, alias, or variable function, module instantiation, named fork, named begin, net, task, register, or variable In addition to those listed above for Verilog: class, package, program, interface, array, directive, property, or sequence SystemC PSL module, channel, port, variable, or aggregate property, sequence, directive, or endpoint
Text conventions
Text conventions used in this manual include: italic text bold text provides emphasis and sets off filenames, pathnames, and design unit names indicates commands, command options, menu choices, package and library logical names, as well as variables, dialog box selections, and language keywords monospace type is used for program and command examples is used to connect menu choices when traversing menus as in: File > Quit denotes file types used by ModelSim (e.g., DO, WLF, INI, MPF, PDF, etc.)
monospace type
The right angle (>) UPPER CASE
Installation directory pathnames

When referring to installation paths, this manual uses modeltech as a generic representation of the installation directory for all versions of ModelSim. The actual installation directory on your system may contain version information.
Where to find our documentation UM-31
Where to find our documentation

ModelSim documentation is available from our website at www.model.com/support or in the following for mats and locations:
Document ModelSim Installation & Licensing Guide ModelSim Quick Guide (command and feature quick-reference) ModelSim Tutorial ModelSim Users Manual ModelSim Command Reference ModelSim GUI Reference Std_DevelopersKit Users Manual
Format paper PDF paper PDF PDF, HTML PDF, HTML PDF, HTML PDF, HTML PDF
How to get it shipped with ModelSim select Help > Documentation; also available from the support site. shipped with ModelSim select Help > Documentation; also available from the support site. select Help > Documentation; also available from the support site. select Help > Documentation select Help > Documentation select Help > Documentation www.model.com/support/documentation/BOOK/sdk_um.pdf The Standard Developers Kit is for use with Mentor Graphics QuickHDL.
Command Help Error message help Tcl Man Pages (Tcl manual) Technotes
ASCII ASCII HTML HTML
type help
[command name] <msgNum>
at the prompt in the Transcript pane
type verror
at the Transcript or shell prompt
select Help > Tcl Man Pages, or find contents.htm in \modeltech\docs\tcl_help_html available from the support site.
Download a free PDF reader with Search

ModelSim PDF documentation requires an Adobe Acrobat Reader for viewing. The Reader is available without cost from Adobe at www.adobe.com. Be sure to download the Acrobat Reader with Search to take advantage of the index file supplied with our documentation; the index makes searching for keywords much faster.
UM-32
1 - Introduction
Technical support and updates

Support and Updates
Online and email technical support options, maintenance renewal, and links to international support contacts: www.model.com/support/default.asp
Updates
Access to the most current version of ModelSim: www.model.com/downloads/default.asp
Latest version email

Place your name on our list for email notification of news and updates: www.model.com/products/informant.asp
UM-33
UM-34
1 - Introduction
UM-35
2 - Projects
Chapter contents
Introduction . . . . . . . What are projects?. . . . . What are the benefits of projects?. Project conversion between versions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-36 UM-36 UM-36 UM-37 UM-38 UM-38 UM-39 UM-39 UM-39 UM-42
Getting started with projects . . . . Step 1 Creating a new project . . Step 2 Adding items to the project. Step 3 Compiling the files . . . Step 4 Simulating a design. . . Other basic project operations. . . The Project tab . . . . . Sorting the list . . . . Changing compile order . . . . . . . . . . . . . . . . . .
. UM-43 . UM-43 . UM-44 . UM-44 . UM-44 . UM-45 . UM-46 . UM-48 . UM-50 . UM-50 . UM-51 . UM-52
Changing compile order . . . Auto-generating compile order Grouping files . . . . .
Creating a Simulation Configuration . Organizing projects with folders . .
Specifying file properties and project settings . File compilation properties . . . . Project settings . . . . . . . Accessing projects from the command line .
This chapter discusses ModelSim projects. Projects simplify the process of compiling and simulating a design and are a great tool for getting started with ModelSim.
UM-36
2 - Projects
Introduction
What are projects?
Projects are collection entities for designs under specification or test. At a minimum, projects have a root directory, a work library, and "metadata" which are stored in a .mpf file located in a project's root directory. The metadata include compiler switch settings, compile order, and file mappings. Projects may also include: Source files or references to source files other files such as READMEs or other project documentation local libraries references to global libraries Simulation Configurations (see "Creating a Simulation Configuration" (UM-46)) Folders (see "Organizing projects with folders" (UM-48)) Important: Project metadata are updated and stored only for actions taken within the project itself. For example, if you have a file in a project, and you compile that file from the command line rather than using the project menu commands, the project will not update to reflect any new compile settings.
What are the benefits of projects?

Projects offer benefits to both new and advanced users. Projects simplify interaction with ModelSim; you dont need to understand the intricacies of compiler switches and library mappings eliminate the need to remember a conceptual model of the design; the compile order is maintained for you in the project. Compile order is maintained for HDL-only designs. remove the necessity to re-establish compiler switches and settings at each session; these are stored in the project metadata as are mappings to source files allow users to share libraries without copying files to a local directory; you can establish references to source files that are stored remotely or locally allow you to change individual parameters across multiple files; in previous versions you could only set parameters one file at a time enable "what-if" analysis; you can copy a project, manipulate the settings, and rerun it to observe the new results reload the initial settings from the project .mpf file every time the project is opened
Introduction UM-37
Project conversion between versions

Projects are generally not backwards compatible for either number or letter releases. When you open a project created in an earlier version, you will see a message warning that the project will be converted to the newer version. You have the option of continuing with the conversion or cancelling the operation. As stated in the warning message, a backup of the original project is created before the conversion occurs. The backup file is named <project name>.mpf.bak and is created in the same directory in which the original project is located.
UM-38
2 - Projects
Getting started with projects

This section describes the four basic steps to working with a project. Step 1 Creating a new project (UM-38) This creates a .mpf file and a working library. Step 2 Adding items to the project (UM-39) Projects can reference or include source files, folders for organization, simulations, and any other files you want to associate with the project. You can copy files into the project directory or simply create mappings to files in other locations. Step 3 Compiling the files (UM-41) This checks syntax and semantics and creates the pseudo machine code ModelSim uses for simulation. Step 4 Simulating a design (UM-42) This specifies the design unit you want to simulate and opens a structure tab in the Workspace pane.
Step 1 Creating a new project

Select File > New > Project to create a new project. This opens the Create Project dialog where you can specify a project name, location, and default library name. You can generally leave the Default Library Name set to "work." The name you specify will be used to create a working library subdirectory within the Project Location.
See "Create Project dialog" (GR-44) for more details on this dialog.
Getting started with projects UM-39
After selecting OK, you will see a blank Project tab in the Workspace pane of the Main window and the Add Items to the Project dialog.
workspace
The name of the current project is shown at the bottom left corner of the Main window.
Step 2 Adding items to the project

The Add Items to the Project dialog includes these options: Create New File Create a new VHDL, Verilog, SystemC, Tcl, or text file using the Source editor. See below for details. Add Existing File Add an existing file. See below for details. Create Simulation Create a Simulation Configuration that specifies source files and simulator options. See "Creating a Simulation Configuration" (UM-46) for details. Create New Folder Create an organization folder. See "Organizing projects with folders" (UM-48) for details.
UM-40
2 - Projects
Create New File

The Create New File command lets you create a new VHDL, Verilog, SystemC, Tcl, or text file using the Source editor. You can also access this command by selecting File > Add to Project > New File or right-clicking (2nd button in Windows; 3rd button in UNIX) in the Project tab and selecting Add to Project > New File.
Specify a name, file type, and folder location for the new file. See "Create Project File dialog" (GR-50) for additional details on this dialog. When you select OK, the file is listed in the Project tab.
Add Existing File

You can also access this command by selecting File > Add to Project > Existing File or by right-clicking (2nd button in Windows; 3rd button in UNIX) in the Project tab and selecting Add to Project > Existing File.
See "Add file to Project dialog" (GR-51) for details on this dialog. When you select OK, the file(s) is added to the Project tab.
Getting started with projects UM-41
Step 3 Compiling the files

The question marks in the Status column in the Project tab denote either the files havent been compiled into the project or the source has changed since the last compile. To compile the files, select Compile > Compile All or right click in the Project tab and select Compile > Compile All.
Once compilation is finished, click the Library tab, expand library work by clicking the "+", and you will see the compiled design units.
UM-42
2 - Projects
Step 4 Simulating a design

To simulate one of the designs, either double-click the name or right-click the name and select Simulate. A new tab named sim appears that shows the structure of the active simulation.
At this point you are ready to run the simulation and analyze your results. You often do this by adding signals to the Wave window and running the simulation for a given period of time. See the ModelSim Tutorial for examples.
Other basic project operations

Open an existing project
If you previously exited ModelSim with a project open, ModelSim automatically will open that same project upon startup. You can open a different project by selecting File > Open and choosing Project Files from the Files of type drop-down.
Close a project
Select File > Close > Project or right-click in the Project tab and select Close Project. This closes the Project tab but leaves the Library tab open in the workspace. Note that you cannot close a project while a simulation is in progress.
Delete a project
Select File > Delete > Project. You cannot delete a project while it is open.
The Project tab UM-43
The Project tab

The Project tab contains information about the objects in your project. By default the tab is divided into five columns.
Name The name of a file or object. Status Identifies whether a source file has been successfully compiled. Applies only to VHDL or Verilog files. A question mark means the file hasnt been compiled or the source file has changed since the last successful compile; an X means the compile failed; a check mark means the compile succeeded; a checkmark with a yellow triangle behind it means the file compiled but there were warnings generated. Type The file type as determined by registered file types on Windows or the type you specify when you add the file to the project. Order The order in which the file will be compiled when you execute a Compile All command. Modified The date and time of the last modification to the file. You can hide or show columns by right-clicking on a column title and selecting or deselecting entries.
Sorting the list

You can sort the list by any of the five columns. Click on a column heading to sort by that column; click the heading again to invert the sort order. An arrow in the column heading indicates which field the list is sorted by and whether the sort order is descending (down arrow) or ascending (up arrow).
UM-44
2 - Projects
Changing compile order

The Compile Order dialog box is functional for HDL-only designs. When you compile all files in a project, ModelSim by default compiles the files in the order in which they were added to the project. You have two alternatives for changing the default compile order: 1) select and compile each file individually; 2) specify a custom compile order. To specify a custom compile order, follow these steps: 1 Select Compile > Compile Order or select it from the context menu in the Project tab.
2 Drag the files into the correct order or use the up and down arrow buttons. Note that you can select multiple files and drag them simultaneously.
Auto-generating compile order

Auto Generate is supported for HDL-only designs. The Auto Generate button in the Compile Order dialog (see above) "determines" the correct compile order by making multiple passes over the files. It starts compiling from the top; if a file fails to compile due to dependencies, it moves that file to the bottom and then recompiles it after compiling the rest of the files. It continues in this manner until all files compile successfully or until a file(s) cant be compiled for reasons other than dependency. Files can be displayed in the Project tab in alphabetical or compile order (by clicking the column headings). Keep in mind that the order you see in the Project tab is not necessarily the order in which the files will be compiled.
Changing compile order UM-45
Grouping files
You can group two or more files in the Compile Order dialog so they are sent to the compiler at the same time. For example, you might have one file with a bunch of Verilog define statements and a second file that is a Verilog module. You would want to compile these two files together. To group files, follow these steps: 1 Select the files you want to group.
2 Click the Group button.
To ungroup files, select the group and click the Ungroup button.
UM-46
2 - Projects
Creating a Simulation Configuration

A Simulation Configuration associates a design unit(s) and its simulation options. For example, say you routinely load a particular design and you have to specify the simulator resolution, generics, and SDF timing files. Ordinarily you would have to specify those options each time you load the design. With a Simulation Configuration, you would specify the design and those options and then save the configuration with a name (e.g., top_config). The name is then listed in the Project tab and you can double-click it to load the design along with its options. To create a Simulation Configuration, follow these steps: 1 Select File > Add to Project > Simulation Configuration or select it from the context menu in the Project tab.
2 Specify a name in the Simulation Configuration Name field. 3 Specify the folder in which you want to place the configuration (see "Organizing projects with folders" (UM-48)).
Creating a Simulation Configuration UM-47
4 Select one or more design unit(s). Use the Control and/or Shift keys to select more than one design unit. The design unit names appear in the Simulate field when you select them. 5 Use the other tabs in the dialog to specify any required simulation options. See "Start Simulation dialog" (GR-76) for details on the available options. Click OK and the simulation configuration is added to the Project tab.
Double-click the Simulation Configuration to load the design.
UM-48
2 - Projects
Organizing projects with folders

The more files you add to a project, the harder it can be to locate the item you need. You can add "folders" to the project to organize your files. These folders are akin to directories in that you can have multiple levels of folders and sub-folders. However, no actual directories are created via the file systemthe folders are present only within the project file.
Adding a folder
To add a folder to your project, select File > Add to Project > Folder or right-click in the Project tab and select Add to Project > Folder.
Specify the Folder Name, the location for the folder, and click OK. The folder will be displayed in the Project tab.
Organizing projects with folders UM-49
You use the folders when you add new objects to the project. For example, when you add a file, you can select which folder to place it in.
If you want to move a file into a folder later on, you can do so using the Properties dialog for the file (right-click on the file and select Properties from the context menu).
On Windows platforms, you can also just drag-and-drop a file into a folder.
UM-50
2 - Projects
Specifying file properties and project settings

You can set two types of properties in a project: file properties and project settings. File properties affect individual files; project settings affect the entire project.
File compilation properties

The VHDL and Verilog compilers (vcom and vlog, respectively) have numerous options that affect how a design is compiled and subsequently simulated. You can customize the settings on individual files or a group of files. Important: Any changes you make to the compile properties outside of the project, whether from the command line, the GUI, or the modelsim.ini file, will not affect the properties of files already in the project. To customize specific files, select the file(s) in the Project tab, right click on the file names, and select Properties. The resulting Project Compiler Settings dialog varies depending on the number and type of files you have selected. If you select a single VHDL or Verilog file, you will see the General tab, Coverage tab, and the VHDL or Verilog tab, respectively. If you select a SystemC file, you will see only the General tab. On the General tab, you will see file properties such as Type, Location, and Size. If you select multiple files, the file properties on the General tab are not listed. Finally, if you select both a VHDL file and a Verilog file, you will see all tabs but no file information on the General tab.
See "Project Compiler Settings" (GR-56) for details on this dialog.
Specifying file properties and project settings UM-51
When setting options on a group of files, keep in mind the following: If two or more files have different settings for the same option, the checkbox in the dialog will be "grayed out." If you change the option, you cannot change it back to a "multi- state setting" without cancelling out of the dialog. Once you click OK, ModelSim will set the option the same for all selected files. If you select a combination of VHDL and Verilog files, the options you set on the VHDL and Verilog tabs apply only to those file types.
Project settings
To modify project settings, right-click anywhere within the Project tab and select Project Settings.
See "Project Settings dialog" (GR-63) for details on this dialog.
UM-52
2 - Projects
Accessing projects from the command line

Generally, projects are used from within the ModelSim GUI. However, standalone tools will use the project file if they are invoked in the project's root directory. If you want to invoke outside the project directory, set the MODELSIM environment variable with the path to the project file (<Project_Root_Dir>/<Project_Name>.mpf). You can also use the project command (CR-185) from the command line to perform common operations on projects.
UM-53
3 - Design libraries
Chapter contents
Design library overview . . . . . . Design unit information . . . . . Working library versus resource libraries . Archives . . . . . . . . . Working with design libraries . . . . . Creating a library . . . . . . . Managing library contents . . . . Assigning a logical name to a design library Moving a library . . . . . . . Setting up libraries for group use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-54 UM-54 UM-54 UM-55 UM-56 UM-56 UM-57 UM-58 UM-59 UM-59 UM-60 UM-60 UM-60 UM-60 UM-61 UM-61 UM-62 UM-62
Specifying the resource libraries . . . . . . . . Verilog resource libraries . . . . . . . . . VHDL resource libraries . . . . . . . . . Predefined libraries . . . . . . . . . . Alternate IEEE libraries supplied . . . . . . . Rebuilding supplied libraries . . . . . . . . Regenerating your design libraries . . . . . . Maintaining 32-bit and 64-bit versions in the same library Referencing source files with location maps . Importing FPGA libraries . . . . . . . . . . . . . . . . . . . .
. UM-63 . UM-65 . UM-66
Protecting source code using -nodebug
VHDL designs are associated with libraries, which are objects that contain compiled design units. SystemC, Verilog and System Verilog designs simulated within ModelSim are compiled into libraries as well.
UM-54
Design library overview

A design library is a directory or archive that serves as a repository for compiled design units. The design units contained in a design library consist of VHDL entities, packages, architectures, and configurations; and Verilog modules and UDPs (user-defined primitives); and SystemC modules. The design units are classified as follows: Primary design units Consist of entities, package declarations, configuration declarations, modules, and UDPs, and SystemC modules. Primary design units within a given library must have unique names. Secondary design units Consist of architecture bodies and package bodies. Secondary design units are associated with a primary design unit. Architectures by the same name can exist if they are associated with different entities or modules.
Design unit information

The information stored for each design unit in a design library is: retargetable, executable code debugging information dependency information
Working library versus resource libraries

Design libraries can be used in two ways: 1) as a local working library that contains the compiled version of your design; 2) as a resource library. The contents of your working library will change as you update your design and recompile. A resource library is typically static and serves as a parts source for your design. You can create your own resource libraries, or they may be supplied by another design team or a third party (e.g., a silicon vendor). Only one library can be the working library. In contrast any number of libraries can be resource libraries during a compilation. You specify which resource libraries will be used when the design is compiled, and there are rules to specify in which order they are searched (see "Specifying the resource libraries" (UM-60)). A common example of using both a working library and a resource library is one where your gate-level design and testbench are compiled into the working library, and the design references gate-level models in a separate resource library.
The library named "work"

The library named "work" has special attributes within ModelSim; it is predefined in the compiler and need not be declared explicitly (i.e. library work). It is also the library name used by the compiler as the default destination of compiled design units (i.e., it doesnt need to be mapped). In other words the work library is the default working library.
Design library overview UM-55
Archives
By default, design libraries are stored in a directory structure with a sub-directory for each design unit in the library. Alternatively, you can configure a design library to use archives. In this case each design unit is stored in its own archive file. To create an archive, use the -archive argument to the vlib command (CR-292). Generally you would do this only in the rare case that you hit the reference count limit on I-nodes due to the ".." entries in the lower-level directories (the maximum number of subdirectories on UNIX and Linux is 65533). An example of an error message that is produced when this limit is hit is:
mkdir: cannot create directory `65534': Too many links
Archives may also have limited value to customers seeking disk space savings. Note that GMAKE wont work with these archives on the IBM platform.
UM-56
Working with design libraries

The implementation of a design library is not defined within standard VHDL or Verilog. Within ModelSim, design libraries are implemented as directories and can have any legal name allowed by the operating system, with one exception; extended identifiers are not supported for library names.
Creating a library
When you create a project (see "Getting started with projects" (UM-38)), ModelSim automatically creates a working design library. If you dont create a project, you need to create a working design library before you run the compiler. This can be done from either the command line or from the ModelSim graphic interface. From the ModelSim prompt or a UNIX/DOS prompt, use this vlib command (CR-292):
vlib <directory_pathname>
To create a new library with the ModelSim graphic interface, select File > New > Library.
The options in this dialog are described under "Create a New Library dialog" (GR-45). When you click OK, ModelSim creates the specified library directory and writes a specially-formatted file named _info into that directory. The _info file must remain in the directory to distinguish it as a ModelSim library. The new map entry is written to the modelsim.ini file in the [Library] section. See "[Library] library path variables" (UM-439) for more information. Note: Remember that a design library is a special kind of directory; the only way to create a library is to use the ModelSim GUI or the vlib command (CR-292). Do not try to create libraries using UNIX, DOS, or Windows commands.
Working with design libraries UM-57
Managing library contents

Library contents can be viewed, deleted, recompiled, edited and so on using either the graphic interface or command line. The Library tab in the Workspace pane provides access to design units (configurations, modules, packages, entities, and architectures, and SystemC modules) in a library. Various information about the design units is displayed in columns to the right of the design unit name.
The Library tab has a context menu with various commands that you access by clicking your right mouse button (Windows2nd button, UNIX3rd button) in the Library tab. The context menu includes the following commands: Simulate Loads the selected design unit and opens structure and Files tabs in the workspace. Related command line command is vsim (CR-305). Simulate with Coverage Loads the selected design unit and collects code coverage data. Related command line command is vsim (CR-305) -coverage. Edit Opens the selected design unit in the Source window, or if a library is selected, opens the Edit Library Mapping dialog (see "Library mappings with the GUI" (UM-58)). Refresh Rebuilds the library image of the selected library without using source code. Related command line command is vcom (CR-246) or vlog (CR-293) with the -refresh argument. Recompile Recompiles the selected design unit. Related command line command is vcom (CR-246) or vlog (CR-293). Update Updates the display of available libraries and design units.
UM-58
Assigning a logical name to a design library

VHDL uses logical library names that can be mapped to ModelSim library directories. By default, ModelSim can find libraries in your current directory (assuming they have the right name), but for it to find libraries located elsewhere, you need to map a logical library name to the pathname of the library. You can use the GUI, a command, or a project to assign a logical name to a design library.
Library mappings with the GUI

To associate a logical name with a library, select the library in the workspace, right-click and select Edit from the context menu. This brings up a dialog box that allows you to edit the mapping.
The dialog box includes these options: Library Mapping Name The logical name of the library. Library Pathname The pathname to the library.
Library mapping from the command line

You can issue a command to set the mapping between a logical library name and a directory; its form is:
vmap <logical_name> <directory_pathname>
You may invoke this command from either a UNIX/DOS prompt or from the command line within ModelSim. The vmap (CR-304) command adds the mapping to the library section of the modelsim.ini file. You can also modify modelsim.ini manually by adding a mapping line. To do this, use a text editor and add a line under the [Library] section heading using the syntax:
<logical_name> = <directory_pathname>
Working with design libraries UM-59
More than one logical name can be mapped to a single directory. For example, suppose the modelsim.ini file in the current working directory contains following lines:
[Library] work = /usr/rick/design my_asic = /usr/rick/design
This would allow you to use either the logical name work or my_asic in a library or use clause to refer to the same design library.
Unix symbolic links

You can also create a UNIX symbolic link to the library using the host platform command:
ln -s <directory_pathname> <logical_name>
The vmap command (CR-304) can also be used to display the mapping of a logical library name to a directory. To do this, enter the shortened form of the command:
vmap <logical_name>
Library search rules

The system searches for the mapping of a logical name in the following order: First the system looks for a modelsim.ini file. If the system doesnt find a modelsim.ini file, or if the specified logical name does not exist in the modelsim.ini file, the system searches the current working directory for a subdirectory that matches the logical name. An error is generated by the compiler if you specify a logical name that does not resolve to an existing directory.
Moving a library
Individual design units in a design library cannot be moved. An entire design library can be moved, however, by using standard operating system commands for moving a directory or an archive.
Setting up libraries for group use

By adding an others clause to your modelsim.ini file, you can have a hierarchy of library mappings. If the ModelSim tools dont find a mapping in the modelsim.ini file, then they will search the library section of the initialization file specified by the others clause. For example:
[library] asic_lib = /cae/asic_lib work = my_work others = /usr/modeltech/modelsim.ini
Only one "others" clause can be entered in the library section.
UM-60
Specifying the resource libraries

Verilog resource libraries
ModelSim supports separate compilation of distinct portions of a Verilog design. The vlog (CR-293) compiler is used to compile one or more source files into a specified library. The library thus contains pre-compiled modules and UDPs that are referenced by the simulator as it loads the design. Important: Resource libraries are specified differently for Verilog and VHDL. For Verilog you use either the -L or -Lf argument to vlog (CR-293). See "Library usage" (UM107) for more information.
VHDL resource libraries

Within a VHDL source file, you use the VHDL library clause to specify logical names of one or more resource libraries to be referenced in the subsequent design unit. The scope of a library clause includes the text region that starts immediately after the library clause and extends to the end of the declarative region of the associated design unit. It does not extend to the next design unit in the file. Note that the library clause is not used to specify the working library into which the design unit is placed after compilation. The vcom command (CR-246) adds compiled design units to the current working library. By default, this is the library named work. To change the current working library, you can use vcom -work and specify the name of the desired target library.
Predefined libraries
Certain resource libraries are predefined in standard VHDL. The library named std contains the packages standard and textio, which should not be modified. The contents of these packages and other aspects of the predefined language environment are documented in the IEEE Standard VHDL Language Reference Manual, Std 1076. See also, "Using the TextIO package" (UM-79). A VHDL use clause can be specified to select particular declarations in a library or package that are to be visible within a design unit during compilation. A use clause references the compiled version of the packagenot the source. By default, every VHDL design unit is assumed to contain the following declarations:
LIBRARY std, work; USE std.standard.all
To specify that all declarations in a library or package can be referenced, add the suffix .all to the library/package name. For example, the use clause above specifies that all declarations in the package standard, in the design library named std, are to be visible to the VHDL design unit immediately following the use clause. Other libraries or packages are not visible unless they are explicitly specified using a library or use clause. Another predefined library is work, the library where a design unit is stored after it is compiled as described earlier. There is no limit to the number of libraries that can be referenced, but only one library is modified during compilation.
Specifying the resource libraries UM-61
Alternate IEEE libraries supplied

The installation directory may contain two or more versions of the IEEE library: ieeepure Contains only IEEE approved packages (accelerated for ModelSim). ieee Contains precompiled Synopsys and IEEE arithmetic packages which have been accelerated by Model Technology including math_complex, math_real, numeric_bit, numeric_std, std_logic_1164, std_logic_misc, std_logic_textio, std_logic_arith, std_logic_signed, std_logic_unsigned, vital_primitives, and vital_timing. You can select which library to use by changing the mapping in the modelsim.ini file. The modelsim.ini file in the installation directory defaults to the ieee library.
Rebuilding supplied libraries

Resource libraries are supplied precompiled in the modeltech installation directory. If you need to rebuild these libraries, the sources are provided in the vhdl_src directory; a macro file is also provided for Windows platforms (rebldlibs.do). To rebuild the libraries, invoke the DO file from within ModelSim with this command:
do rbldlibs.do
Make sure your current directory is the modeltech install directory before you run this file. Note: Because accelerated subprograms require attributes that are available only under the 1993 standard, many of the libraries are built using vcom (CR-246) with the -93 option. Shell scripts are provided for UNIX (rebuild_libs.csh and rebuild_libs.sh). To rebuild the libraries, execute one of the rebuild_libs scripts while in the modeltech directory.
UM-62
Regenerating your design libraries

Depending on your current ModelSim version, you may need to regenerate your design libraries before running a simulation. Check the installation README file to see if your libraries require an update. You can regenerate your design libraries using the Refresh command from the Library tab context menu (see "Managing library contents" (UM-57)), or by using the -refresh argument to vcom (CR-246) and vlog (CR-293). From the command line, you would use vcom with the -refresh option to update VHDL design units in a library, and vlog with the -refresh option to update Verilog design units. By default, the work library is updated; use -work <library> to update a different library. For example, if you have a library named mylib that contains both VHDL and Verilog design units:
vcom -work mylib -refresh vlog -work mylib -refresh
An important feature of -refresh is that it rebuilds the library image without using source code. This means that models delivered as compiled libraries without source code can be rebuilt for a specific release of ModelSim. In general, this works for moving forwards or backwards on a release. Moving backwards on a release may not work if the models used compiler switches or directives that do not exist in the older release. Note: You don't need to regenerate the std, ieee, vital22b, and verilog libraries. Also, you cannot use the -refresh option to update libraries that were built before the 4.6 release.
Maintaining 32-bit and 64-bit versions in the same library

It is possible with ModelSim to maintain 32-bit and 64-bit versions of a design in the same library. To do this, you must compile the design with the 32-bit version and then "refresh" the design with the 64-bit version. For example: Using the 32-bit version of ModelSim:
vlog file1.v file2.v -forcecode -work asic_lib
Next, using the 64-bit version of ModelSim:

vlog -work asic_lib -refresh
This allows you to use either version without having to do a refresh. Do not compile the design with one version, and then recompile it with the other. If you do this, ModelSim will remove the first module, because it could be "stale."
Referencing source files with location maps UM-63
Referencing source files with location maps

Pathnames to source files are recorded in libraries by storing the working directory from which the compile is invoked and the pathname to the file as specified in the invocation of the compiler. The pathname may be either a complete pathname or a relative pathname. ModelSim tools that reference source files from the library locate a source file as follows: If the pathname stored in the library is complete, then this is the path used to reference the file. If the pathname is relative, then the tool looks for the file relative to the current working directory. If this file does not exist, then the path relative to the working directory stored in the library is used. This method of referencing source files generally works fine if the libraries are created and used on a single system. However, when multiple systems access a library across a network, the physical pathnames are not always the same and the source file reference rules do not always work.
Using location mapping

Location maps are used to replace prefixes of physical pathnames in the library with environment variables. The location map defines a mapping between physical pathname prefixes and environment variables. ModelSim tools open the location map file on invocation if the MGC_LOCATION_MAP (UM-434) environment variable is set. If MGC_LOCATION_MAP is not set, ModelSim will look for a file named "mgc_location_map" in the following locations, in order: the current directory your home directory the directory containing the ModelSim binaries the ModelSim installation directory
Use these two steps to map your files:

1 Set the environment variable MGC_LOCATION_MAP to the path to your location map file. 2 Specify the mappings from physical pathnames to logical pathnames:
$SRC /home/vhdl/src /usr/vhdl/src $IEEE /usr/modeltech/ieee
UM-64
Pathname syntax
The logical pathnames must begin with $ and the physical pathnames must begin with /. The logical pathname is followed by one or more equivalent physical pathnames. Physical pathnames are equivalent if they refer to the same physical directory (they just have different pathnames on different systems).
How location mapping works

When a pathname is stored, an attempt is made to map the physical pathname to a path relative to a logical pathname. This is done by searching the location map file for the first physical pathname that is a prefix to the pathname in question. The logical pathname is then substituted for the prefix. For example, "/usr/vhdl/src/test.vhd" is mapped to "$SRC/ test.vhd". If a mapping can be made to a logical pathname, then this is the pathname that is saved. The path to a source file entry for a design unit in a library is a good example of a typical mapping. For mapping from a logical pathname back to the physical pathname, ModelSim expects an environment variable to be set for each logical pathname (with the same name). ModelSim reads the location map file when a tool is invoked. If the environment variables corresponding to logical pathnames have not been set in your shell, ModelSim sets the variables to the first physical pathname following the logical pathname in the location map. For example, if you don't set the SRC environment variable, ModelSim will automatically set it to "/home/vhdl/src".
Mapping with Tcl variables

Two Tcl variables may also be used to specify alternative source-file paths; SourceDir and SourceMap. You would define these variables in a modelsim.tcl file. See the "The modelsim.tcl file" (GR-253) for details.
Importing FPGA libraries UM-65
Importing FPGA libraries

ModelSim includes an import wizard for referencing and using vendor FPGA libraries. The wizard scans for and enforces dependencies in the libraries and determines the correct mappings and target directories. Important: The FPGA libraries you import must be pre-compiled. Most FPGA vendors supply pre-compiled libraries configured for use with ModelSim. To import an FPGA library, select File > Import > Library.
Follow the instructions in the wizard to complete the import.
UM-66
Protecting source code using -nodebug

The -nodebug argument for both vcom (CR-246) and vlog (CR-293) hides internal model data. This allows a model supplier to provide pre-compiled libraries without providing source code and without revealing internal model variables and structure. Note: -nodebug encrypts entire files. The Verilog `protect compiler directive allows you to encrypt regions within a file. See "Compiler directives specific to ModelSim" (UM-133) for details. When you compile with -nodebug, all source text, identifiers, and line number information are stripped from the resulting compiled object, so ModelSim cannot locate or display any information of the model except for the external pins. Specifically, this means that: a Source window will not display the design units source code a structure pane will not display the internal structure the Objects pane will not display internal signals the Active Processes pane will not display internal processes the Locals pane will not display internal variables none of the hidden objects may be accessed through the Dataflow window or with ModelSim commands You can access the design units comprising your model via the library, and you may invoke vsim (CR-305) directly on any of these design units and see the ports. To restrict even this access in the lower levels of your design, you can use the following -nodebug options when you compile: Command and switch vcom -nodebug=ports vlog -nodebug=ports vlog -nodebug=pli vlog -nodebug=ports+pli Result makes the ports of a VHDL design unit invisible makes the ports of a Verilog design unit invisible prevents the use of PLI functions to interrogate the module for information combines the functions of -nodebug=ports and -nodebug=pli
Dont use the =ports option on a design without hierarchy, or on the top level of a hierarchical design. If you do, no ports will be visible for simulation. Rather, compile all lower portions of the design with -nodebug=ports first, then compile the top level with -nodebug alone. Design units or modules compiled with -nodebug can only instantiate design units or modules that are also compiled -nodebug.
UM-67
4 - VHDL simulation
Chapter contents
Introduction . . . Basic VHDL flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-69 . UM-69 . . . . . . . . . . . UM-70 UM-70 UM-70 UM-70 UM-71 UM-71 UM-72 UM-75 UM-75 UM-76 UM-77
Compiling VHDL files. . . . . . Creating a design library . . . . Invoking the VHDL compiler . . . Dependency checking . . . . . Range and index checking . . . Subprogram inlining . . . . . Differences between language versions Simulating VHDL designs . . Simulator resolution limit . Default binding . . . Delta delays . . . . . . . . . . . . . . . .
Using the TextIO package . . . . . . . . . . . Syntax for file declaration. . . . . . . . . . Using STD_INPUT and STD_OUTPUT within ModelSim . TextIO implementation issues . . . . . Writing strings and aggregates . . . Reading and writing hexadecimal numbers Dangling pointers . . . . . . . The ENDLINE function . . . . . The ENDFILE function . . . . . Using alternative input/output files . . Providing stimulus . . . . . . VITAL specification and source code . VITAL packages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. UM-79 . UM-79 . UM-80 . . . . . . . . UM-81 UM-81 UM-82 UM-82 UM-82 UM-82 UM-83 UM-83
. UM-84 . UM-84 . UM-84 . UM-85 . UM-86 . UM-86 . . . . . . . . UM-87 UM-87 UM-88 UM-88 UM-88 UM-88 UM-89 UM-90
ModelSim VITAL compliance. . . . . . . . . . VITAL compliance checking . . . . . . . . . Compiling and simulating with accelerated VITAL packages Compiling and simulating with accelerated VITAL packages Util package . . . get_resolution . . init_signal_driver() init_signal_spy() . signal_force() . . signal_release() . to_real() . . . to_time() . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Modeling memory . . . 87 and 93 example .
. UM-91 . UM-91
UM-68
4 - VHDL simulation
02 example
. . .
. . .
. . .
. . .
. UM-94 . UM-98 . UM-99
Affecting performance by cancelling scheduled events Converting an integer into a bit_vector . . . .
Note: VHDL is not supported in ModelSim LE.
Introduction UM-69
Introduction
This chapter describes how to compile, optimize, and simulate VHDL designs in ModelSim. It also discusses using the TextIO package with ModelSim; ModelSims implementation of the VITAL (VHDL Initiative Towards ASIC Libraries) specification for ASIC modeling; and ModelSims special built-in utilities package. The TextIO package is defined within the VHDL Language Reference Manual, IEEE Std 1076; it allows human-readable text input from a declared source within a VHDL file during simulation.
Basic VHDL flow

Simulating VHDL designs with ModelSim includes four general steps: 1 Compile your VHDL code into one or more libraries using the vcom command (CR-246). See "Compiling VHDL files" (UM-70) for details. 2 Load your design with the vsim command (CR-305). See "Simulating VHDL designs" (UM-75) for details. 3 Run and debug your design.
UM-70
4 - VHDL simulation
Compiling VHDL files

Creating a design library
Before you can compile your source files, you must create a library in which to store the compilation results. Use vlib (CR-292) to create a new library. For example:
vlib work
This creates a library named work. By default, compilation results are stored in the work library. The work library is actually a subdirectory named work. This subdirectory contains a special file named _info. Do not create libraries using UNIX, MS Windows, or DOS commands always use the vlib command (CR-292). See "Design libraries" (UM-53) for additional information on working with libraries.
Invoking the VHDL compiler

246),
ModelSim compiles one or more VHDL design units with a single invocation of vcom (CRthe VHDL compiler. The design units are compiled in the order that they appear on the command line. For VHDL, the order of compilation is important you must compile any entities or configurations before an architecture that references them.
You can simulate a design containing units written with 1076 -1987, 1076 -1993, and 1076-2002 versions of VHDL. To do so you will need to compile units from each VHDL version separately. The vcom (CR-246) command compiles using 1076 -2002 rules by default; use the -87 or -93 argument to vcom (CR-246) to compile units written with version 1076-1987 or 1076 -1993, respectively. You can also change the default by modifying the VHDL93 variable in the modelsim.ini file (see "Control variables located in INI files" (UM438) for more information).
Dependency checking
Dependent design units must be reanalyzed when the design units they depend on are changed in the library. vcom (CR-246) determines whether or not the compilation results have changed. For example, if you keep an entity and its architectures in the same source file and you modify only an architecture and recompile the source file, the entity compilation results will remain unchanged and you will not have to recompile design units that depend on the entity.
Compiling VHDL files UM-71
Range and index checking

A range check verifies that a scalar value defined with a range subtype is always assigned a value within its range. An index check verifies that whenever an array subscript expression is evaluated, the subscript will be within the array's range. Range and index checks are performed by default when you compile your design. You can disable range checks (potentially offering a performance advantage) and index checks using arguments to the vcom (CR-246) command. Or, you can use the NoRangeCheck and NoIndexCheck variables in the modelsim.ini file to specify whether or not they are performed. See "Control variables located in INI files" (UM-438). Range checks in ModelSim are slightly more restrictive than those specified by the VHDL LRM. ModelSim requires any assignment to a signal to also be in range whereas the LRM requires only that range checks be done whenever a signal is updated. Most assignments to signals update the signal anyway, and the more restrictive requirement allows ModelSim to generate better error messages.
Subprogram inlining
ModelSim attempts to inline subprograms at compile time to improve simulation performance. This happens automatically and should be largely transparent. However, you can disable automatic inlining two ways: Invoke vcom (CR-246) with the -O0 or -O1 argument Use the mti_inhibit_inline attribute as described below Single-stepping through a simulation varies slightly depending on whether inlining occurred. When single-stepping to a subprogram call that has not been inlined, the simulator stops first at the line of the call, and then proceeds to the line of the first executable statement in the called subprogram. If the called subprogram has been inlined, the simulator does not first stop at the subprogram call, but stops immediately at the line of the first executable statement.
mti_inhibit_inline attribute
You can disable inlining for individual design units (a package, architecture, or entity) or subprograms with the mti_inhibit_inline attribute. Follow these rules to use the attribute: Declare the attribute within the design unit's scope as follows:
attribute mti_inhibit_inline : boolean;
Assign the value true to the attribute for the appropriate scope. For example, to inhibit inlining for a particular function (e.g., "foo"), add the following attribute assignment:
attribute mti_inhibit_inline of foo : procedure is true;
To inhibit inlining for a particular package (e.g., "pack"), add the following attribute assignment:
attribute mti_inhibit_inline of pack : package is true;
Do similarly for entities and architectures.
UM-72
4 - VHDL simulation
Differences between language versions

There are three versions of the IEEE VHDL 1076 standard: VHDL-1987, VHDL-1993, and VHDL-2002. The default language version for ModelSim is VHDL-2002. If your code was written according to the 87 or 93 version, you may need to update your code or instruct ModelSim to use the earlier versions rules. To select a specific language version, do one of the following: Select the appropriate version from the compiler options menu in the GUI Invoke vcom (CR-246) using the argument -87, -93, or -2002 Set the VHDL93 variable in the [vcom] section of the modelsim.ini file. Appropriate values for VHDL93 are: - 0, 87, or 1987 for VHDL-1987 - 1, 93, or 1993 for VHDL-1993 - 2, 02, or 2002 for VHDL-2002 The following is a list of language incompatibilities that may cause problems when compiling a design. The only major problem between VHDL-93 and VHDL-2002 is the addition of the keyword "PROTECTED". VHDL-93 programs which use this as an identifier should choose a different name. All other incompatibilities are between VHDL-87 and VHDL-93. VITAL and SDF It is important to use the correct language version for VITAL. VITAL2000 must be compiled with VHDL-93 or VHDL-2002. VITAL95 must be compiled with VHDL-87. A typical error message that indicates the need to compile under language version VHDL-87 is:
"VITALPathDelay DefaultDelay parameter must be locally static"
Purity of NOW In VHDL-93 the function "now" is impure. Consequently, any function that invokes "now" must also be declared to be impure. Such calls to "now" occur in VITAL. A typical error message:
"Cannot call impure function 'now' from inside pure function '<name>'"
Files File syntax and usage changed between VHDL-87 and VHDL-93. In many cases vcom issues a warning and continues:
"Using 1076-1987 syntax for file declaration."
In addition, when files are passed as parameters, the following warning message is produced:
"Subprogram parameter name is declared using VHDL 1987 syntax."
This message often involves calls to endfile(<name>) where <name> is a file parameter.
Compiling VHDL files UM-73
Files and packages Each package header and body should be compiled with the same language version. Common problems in this area involve files as parameters and the size of type CHARACTER. For example, consider a package header and body with a procedure that has a file parameter:
procedure proc1 ( out_file : out std.textio.text) ...
If you compile the package header with VHDL-87 and the body with VHDL-93 or VHDL-2002, you will get an error message such as:
"** Error: mixed_package_b.vhd(4): Parameter kinds do not conform between declarations in package header and body: 'out_file'."
Direction of concatenation To solve some technical problems, the rules for direction and bounds of concatenation were changed from VHDL-87 to VHDL-93. You won't see any difference in simple variable/signal assignments such as:
v1 := a & b;
But if you (1) have a function that takes an unconstrained array as a parameter, (2) pass a concatenation expression as a formal argument to this parameter, and (3) the body of the function makes assumptions about the direction or bounds of the parameter, then you will get unexpected results. This may be a problem in environments that assume all arrays have "downto" direction. xnor "xnor" is a reserved word in VHDL-93. If you declare an xnor function in VHDL-87 (without quotes) and compile it under VHDL-2002, you will get an error message like the following:
** Error: xnor.vhd(3): near "xnor": expecting: STRING IDENTIFIER
'FOREIGN attribute In VHDL-93 package STANDARD declares an attribute 'FOREIGN. If you declare your own attribute with that name in another package, then ModelSim issues a warning such as the following:
-- Compiling package foopack ** Warning: foreign.vhd(9): (vcom-1140) VHDL-1993 added a definition of the attribute foreign to package std.standard. The attribute is also defined in package 'standard'. Using the definition from package 'standard'.
Size of CHARACTER type In VHDL-87 type CHARACTER has 128 values; in VHDL-93 it has 256 values. Code which depends on this size will behave incorrectly. This situation occurs most commonly in test suites that check VHDL functionality. It's unlikely to occur in practical designs. A typical instance is the replacement of warning message:
"range nul downto del is null"
by
"range nul downto '' is null" -- range is nul downto y(umlaut)
UM-74
4 - VHDL simulation
bit string literals In VHDL-87 bit string literals are of type bit_vector. In VHDL-93 they can also be of type STRING or STD_LOGIC_VECTOR. This implies that some expressions that are unambiguous in VHDL-87 now become ambiguous is VHDL-93. A typical error message is:
** Error: bit_string_literal.vhd(5): Subprogram '=' is ambiguous. Suitable definitions exist in packages 'std_logic_1164' and 'standard'.
In VHDL-87 when using individual subelement association in an association list, associating individual sub-elements with NULL is discouraged. In VHDL-93 such association is forbidden. A typical message is:
"Formal '<name>' must not be associated with OPEN when subelements are associated individually."
Simulating VHDL designs UM-75
Simulating VHDL designs

A VHDL design is ready for simulation after it has been compiled with vcom. The simulator may then be invoked with the name of the configuration or entity/architecture pair. This section discusses simulation from the UNIX or Windows/DOS command line. You can also use a project to simulate (see "Getting started with projects" (UM-38)) or the Simulate dialog box (see "Start Simulation dialog" (GR-76)). This example invokes vsim (CR-305) on the entity my_asic and the architecture structure:
vsim my_asic structure
vsim (CR-305) is capable of annotating a design using VITAL compliant models with timing data from an SDF file. You can specify the min:typ:max delay by invoking vsim with the -sdfmin, -sdftyp, or -sdfmax option. Using the SDF file f1.sdf in the current work directory, the following invocation of vsim annotates maximum timing values for the design unit my_asic:
vsim -sdfmax /my_asic=f1.sdf my_asic
By default, the timing checks within VITAL models are enabled. They can be disabled with the +notimingchecks option. For example:
vsim +notimingchecks topmod
Simulator resolution limit

The simulator internally represents time as a 64-bit integer in units equivalent to the smallest unit of simulation time, also known as the simulator resolution limit. The default resolution limit is set to the value specified by the Resolution (UM-447) variable in the modelsim.ini file. You can view the current resolution by invoking the report command (CR-192) with the simulator state option.
Overriding the resolution

You can override ModelSims default resolution by specifying the -t option on the command line or by selecting a different Simulator Resolution in the Simulate dialog box. Available resolutions are: 1x, 10x, or 100x of fs, ps, ns, us, ms, or sec. For example this command chooses 10 ps resolution:
vsim -t 10ps topmod
Clearly you need to be careful when doing this type of operation. If the resolution set by -t is larger than a delay value in your design, the delay values in that design unit are rounded to the closest multiple of the resolution. In the example above, a delay of 4 ps would be rounded to 0 ps.
Choosing the resolution

You should choose the coarsest resolution limit possible that does not result in undesired rounding of your delays. The time precision should not be unnecessarily small because it
UM-76
4 - VHDL simulation
will limit the maximum simulation time limit, and it will degrade performance in some cases.
Default binding
305).
By default ModelSim performs default binding when you load the design with vsim (CRThe advantage of performing default binding at load time is that it provides more flexibility for compile order. Namely, entities don't necessarily have to be compiled before other entities/architectures which instantiate them. However, you can force ModelSim to perform default binding at compile time. This may allow you to catch design errors (e.g., entities with incorrect port lists) earlier in the flow. Use one of these two methods to change when default binding occurs: Specify the -bindAtCompile argument to vcom (CR-246) Set the BindAtCompile (UM-441) variable in the modelsim.ini to 1 (true)
Default binding rules

When looking for an entity to bind with, ModelSim searches the currently visible libraries for an entity with the same name as the component. ModelSim does this because IEEE 1076-1987 contained a flaw that made it almost impossible for an entity to be directly visible if it had the same name as the component. In short, if a component was declared in an architecture, any like-named entity above that declaration would be hidden because component/entity names cannot be overloaded. As a result we implemented the following rules for determining default binding: If performing default binding at load time, search the libraries specified with the -Lf argument to vsim. If a directly visible entity has the same name as the component, use it. If an entity would be directly visible in the absence of the component declaration, use it. If the component is declared in a package, search the library that contained the package for an entity with the same name. If none of these methods is successful, ModelSim will also do the following: Search the work library. Search all other libraries that are currently visible by means of the library clause. If performing default binding at load time, search the libraries specified with the -L argument to vsim. Note that these last three searches are an extension to the 1076 standard.
Disabling default binding

If you want default binding to occur only via configurations, you can disable ModelSims normal default binding methods by setting the RequireConfigForAllDefaultBinding (UM441) variable in the modelsim.ini to 1 (true).
Simulating VHDL designs UM-77
Delta delays
Event-based simulators such as ModelSim may process many events at a given simulation time. Multiple signals may need updating, statements that are sensitive to these signals must be executed, and any new events that result from these statements must then be queued and executed as well. The steps taken to evaluate the design without advancing simulation time are referred to as "delta times" or just "deltas." The diagram below represents the process for VHDL designs. This process continues until the end of simulation time.
Execute concurrent statements at current time
Advance delta time
Advance simulation time
No
Any transactions to process? Yes Any events to process? Yes Execute concurrent statements that are sensitive to events No
This mechanism in event-based simulators may cause unexpected results. Consider the following code snippet:
clk2 <= clk; process (rst, clk) begin if(rst = '0')then s0 <= '0'; elsif(clk'event and clk='1') then s0 <= inp; end if; end process; process (rst, clk2) begin if(rst = '0')then s1 <= '0'; elsif(clk2'event and clk2='1') then s1 <= s0; end if; end process;
UM-78
4 - VHDL simulation
In this example you have two synchronous processes, one triggered with clk and the other with clk2. To your surprise, the signals change in the clk2 process on the same edge as they are set in the clk process. As a result, the value of inp appears at s1 rather than s0. During simulation an event on clk occurs (from the testbench). From this event ModelSim performs the "clk2 <= clk" assignment and the process which is sensitive to clk. Before advancing the simulation time, ModelSim finds that the process sensitive to clk2 can also be run. Since there are no delays present, the effect is that the value of inp appears at s1 in the same simulation cycle. In order to get the expected results, you must do one of the following: Insert a delay at every output Make certain to use the same clock Insert a delta delay To insert a delta delay, you would modify the code like this:
process (rst, clk) begin if(rst = '0')then s0 <= '0'; elsif(clk'event and clk='1') then s0 <= inp; s0_delayed <= s0; end if; end process; process (rst, clk2) begin if(rst = '0')then s1 <= '0'; elsif(clk2'event and clk2='1') then s1 <= s0_delayed; end if; end process;
The best way to debug delta delay problems is observe your signals in the List window. There you can see how values change at each delta time.
Detecting infinite zero-delay loops

If a large number of deltas occur without advancing time, it is usually a symptom of an infinite zero-delay loop in the design. In order to detect the presence of these loops, ModelSim defines a limit, the iteration limit", on the number of successive deltas that can occur. When ModelSim reaches the iteration limit, it issues a warning message. The iteration limit default value is1000. If you receive an iteration limit warning, first increase the iteration limit and try to continue simulation. You can set the iteration limit from the Simulate > Runtime Options menu or by modifying the IterationLimit (UM-446) variable in the modelsim.ini. See "Control variables located in INI files" (UM-438) for more information on modifying the modelsim.ini file. If the problem persists, look for zero-delay loops. Run the simulation and look at the source code when the error occurs. Use the step button to step through the code and see which signals or variables are continuously oscillating. Two common causes are a loop that has no exit, or a series of gates with zero delay where the outputs are connected back to the inputs.
Using the TextIO package UM-79
Using the TextIO package

To access the routines in TextIO, include the following statement in your VHDL source code:
USE std.textio.all;
A simple example using the package TextIO is:

USE std.textio.all; ENTITY simple_textio IS END; ARCHITECTURE simple_behavior OF simple_textio IS BEGIN PROCESS VARIABLE i: INTEGER:= 42; VARIABLE LLL: LINE; BEGIN WRITE (LLL, i); WRITELINE (OUTPUT, LLL); WAIT; END PROCESS; END simple_behavior;
Syntax for file declaration

The VHDL87 syntax for a file declaration is: file
identifier : subtype_indication
is
[ mode ]
file_logical_name ;
where "file_logical_name" must be a string expression. In newer versions of the 1076 spec, syntax for a file declaration is: file
identifier_list : subtype_indication [ file_open_information ] ;
where "file_open_information" is:

[open file_open_kind_expression] is file_logical_name
You can specify a full or relative path as the file_logical_name; for example (VHDL87): file
filename : TEXT
is in
usr\rick\myfile;
Normally if a file is declared within an architecture, process, or package, the file is opened when you start the simulator and is closed when you exit from it. If a file is declared in a subprogram, the file is opened when the subprogram is called and closed when execution RETURNs from the subprogram. Alternatively, the opening of files can be delayed until the first read or write by setting the DelayFileOpen variable in the modelsim.ini file. Also, the number of concurrently open files can be controlled by the ConcurrentFileLimit variable. These variables help you manage a large number of files during simulation. See Appendix A - Simulator variables for more details.
UM-80
4 - VHDL simulation
Using STD_INPUT and STD_OUTPUT within ModelSim

The standard VHDL87 TextIO package contains the following file declarations: file file
input: TEXT is in "STD_INPUT"; output: TEXT is out "STD_OUTPUT";
Updated versions of the TextIO package contain these file declarations: file file
input: TEXT open read_mode is "STD_INPUT"; output: TEXT open write_mode is "STD_OUTPUT";
STD_INPUT is a file_logical_name that refers to characters that are entered interactively from the keyboard, and STD_OUTPUT refers to text that is displayed on the screen. In ModelSim, reading from the STD_INPUT file allows you to enter text into the current buffer from a prompt in the Transcript pane. The lines written to the STD_OUTPUT file appear in the Transcript.
TextIO implementation issues UM-81
TextIO implementation issues

Writing strings and aggregates
A common error in VHDL source code occurs when a call to a WRITE procedure does not specify whether the argument is of type STRING or BIT_VECTOR. For example, the VHDL procedure:
WRITE (L, "hello");
will cause the following error:

ERROR: Subprogram "WRITE" is ambiguous.
In the TextIO package, the WRITE procedure is overloaded for the types STRING and BIT_VECTOR. These lines are reproduced here:
procedure WRITE(L: inout LINE; VALUE: in BIT_VECTOR; JUSTIFIED: in SIDE:= RIGHT; FIELD: in WIDTH := 0); procedure WRITE(L: inout LINE; VALUE: in STRING; JUSTIFIED: in SIDE:= RIGHT; FIELD: in WIDTH := 0);
The error occurs because the argument "hello" could be interpreted as a string or a bit vector, but the compiler is not allowed to determine the argument type until it knows which function is being called. The following procedure call also generates an error:
WRITE (L, "010101");
This call is even more ambiguous, because the compiler could not determine, even if allowed to, whether the argument "010101" should be interpreted as a string or a bit vector. There are two possible solutions to this problem: Use a qualified expression to specify the type, as in:
WRITE (L, string("hello"));
Call a procedure that is not overloaded, as in:

WRITE_STRING (L, "hello");
The WRITE_STRING procedure simply defines the value to be a STRING and calls the WRITE procedure, but it serves as a shell around the WRITE procedure that solves the overloading problem. For further details, refer to the WRITE_STRING procedure in the io_utils package, which is located in the file <install_dir>/modeltech/examples/misc/ io_utils.vhd.
UM-82
4 - VHDL simulation
Reading and writing hexadecimal numbers

The reading and writing of hexadecimal numbers is not specified in standard VHDL. The Issues Screening and Analysis Committee of the VHDL Analysis and Standardization Group (ISAC-VASG) has specified that the TextIO package reads and writes only decimal numbers. To expand this functionality, ModelSim supplies hexadecimal routines in the package io_utils, which is located in the file <install_dir>/modeltech/examples/misc/io_utils.vhd. To use these routines, compile the io_utils package and then include the following use clauses in your VHDL source code:
use std.textio.all; use work.io_utils.all;
Dangling pointers
Dangling pointers are easily created when using the TextIO package, because WRITELINE de-allocates the access type (pointer) that is passed to it. Following are examples of good and bad VHDL coding styles: Bad VHDL (because L1 and L2 both point to the same buffer):
READLINE (infile, L1); L2 := L1; WRITELINE (outfile, L1); -- Read and allocate buffer -- Copy pointers -- Deallocate buffer
Good VHDL (because L1 and L2 point to different buffers):

READLINE (infile, L1); L2 := new string(L1.all); WRITELINE (outfile, L1); -- Read and allocate buffer -- Copy contents -- Deallocate buffer
The ENDLINE function

The ENDLINE function described in the IEEE Standard VHDL Language Reference Manual, IEEE Std 1076-1987 contains invalid VHDL syntax and cannot be implemented in VHDL. This is because access values must be passed as variables, but functions do not allow variable parameters. Based on an ISAC-VASG recommendation the ENDLINE function has been removed from the TextIO package. The following test may be substituted for this function:
(L = NULL) OR (LLENGTH = 0)
The ENDFILE function

In the VHDL Language Reference Manuals, the ENDFILE function is listed as:
-- function ENDFILE (L: in TEXT) return BOOLEAN;
As you can see, this function is commented out of the standard TextIO package. This is because the ENDFILE function is implicitly declared, so it can be used with files of any type, not just files of type TEXT.
TextIO implementation issues UM-83
Using alternative input/output files

You can use the TextIO package to read and write to your own files. To do this, just declare an input or output file of type TEXT. For example, for an input file: The VHDL87 declaration is: file
myinput : TEXT
is in
"pathname.dat";
The VHDL93 declaration is: file

myinput : TEXT
open
read_mode
is
"pathname.dat";
Then include the identifier for this file ("myinput" in this example) in the READLINE or WRITELINE procedure call.
Flushing the TEXTIO buffer

Flushing of the TEXTIO buffer is controlled by the UnbufferedOutput (UM-448) variable in the modelsim.ini file.
Providing stimulus
You can stimulate and test a design by reading vectors from a file, using them to drive values onto signals, and testing the results. A VHDL test bench has been included with the ModelSim install files as an example. Check for this file: <install_dir>/modeltech/examples/misc/stimulus.vhd
UM-84
4 - VHDL simulation
VITAL specification and source code

VITAL ASIC Modeling Specification
The IEEE 1076.4 VITAL ASIC Modeling Specification is available from the Institute of Electrical and Electronics Engineers, Inc.: IEEE Customer Service 445 Hoes Lane Piscataway, NJ 08854-1331 Tel: (732) 981-0060 Fax: (732) 981-1721 home page: http://www.ieee.org
VITAL source code

The source code for VITAL packages is provided in the /<install_dir>/vhdl_src/vital22b, /vital95, or /vital2000 directories.
VITAL packages
VITAL 1995 accelerated packages are pre-compiled into the ieee library in the installation directory. VITAL 2000 accelerated packages are pre-compiled into the vital2000 library. If you need to use the newer library, you either need to change the ieee library mapping or add a use clause to your VHDL code to access the VITAL 2000 packages. To change the ieee library mapping, issue the following command:
vmap ieee <modeltech>/vital2000
Or, alternatively, add use clauses to your code:

LIBRARY vital2000; USE vital2000.vital_primitives.all; USE vital2000.vital_timing.all; USE vital2000.vital_memory.all;
Note that if your design uses two libraries -one that depends on vital95 and one that depends on vital2000 - then you will have to change the references in the source code to vital2000. Changing the library mapping will not work.
ModelSim VITAL compliance

A simulator is VITAL compliant if it implements the SDF mapping and if it correctly simulates designs using the VITAL packages, as outlined in the VITAL Model Development Specification. ModelSim is compliant with the IEEE 1076.4 VITAL ASIC Modeling Specification. In addition, ModelSim accelerates the VITAL_Timing, VITAL_Primitives, and VITAL_memory packages. The optimized procedures are functionally equivalent to the IEEE 1076.4 VITAL ASIC Modeling Specification (VITAL 1995 and 2000).
ModelSim VITAL compliance UM-85
VITAL compliance checking

If you are using VITAL 2.2b, you must turn off the compliance checking either by not setting the attributes, or by invoking vcom (CR-246) with the option -novitalcheck.
UM-86
4 - VHDL simulation
Compiling and simulating with accelerated VITAL packages

vcom (CR-246) automatically recognizes that a VITAL function is being referenced from the ieee library and generates code to call the optimized built-in routines. Invoke with the -novital option if you do not want to use the built-in VITAL routines (when debugging for instance). To exclude all VITAL functions, use -novital all:
vcom -novital all design.vhd
To exclude selected VITAL functions, use one or more -novital <fname> options:
vcom -novital VitalTimingCheck -novital VitalAND design.vhd
The -novital switch only affects calls to VITAL functions from the design units currently being compiled. Pre-compiled design units referenced from the current design units will still call the built-in functions unless they too are compiled with the -novital option. ModelSim VITAL built-ins will be updated in step with new releases of the VITAL packages.
Util package UM-87
Util package
The util package serves as a container for various VHDL utilities. The package is part of the modelsim_lib library which is located in the modeltech tree and is mapped in the default modelsim.ini file. To access the utilities in the package, you would add lines like the following to your VHDL code:
library modelsim_lib; use modelsim_lib.util.all;
get_resolution
get_resolution returns the current simulator resolution as a real number. For example, 1 femtosecond corresponds to 1e-15.
Syntax
resval := get_resolution;
Returns
Name resval
Type real
Description The simulator resolution represented as a real
Arguments
None
Related functions
to_real() (UM-89) to_time() (UM-90)
Example
If the simulator resolution is set to 10ps, and you invoke the command:
resval := get_resolution;
the value returned to resval would be 1e-11.
UM-88
4 - VHDL simulation
init_signal_driver()
The init_signal_driver() procedure drives the value of a VHDL signal or Verilog net onto an existing VHDL signal or Verilog net. This allows you to drive signals or nets at any level of the design hierarchy from within a VHDL architecture (e.g., a testbench). See init_signal_driver (UM-359) in Chapter 14 - Signal Spy for complete details.
init_signal_spy()
The init_signal_spy() utility mirrors the value of a VHDL signal or Verilog register/net onto an existing VHDL signal or Verilog register. This allows you to reference signals, registers, or nets at any level of hierarchy from within a VHDL architecture (e.g., a testbench). See init_signal_spy (UM-362) in Chapter 14 - Signal Spy for complete details.
signal_force()
The signal_force() procedure forces the value specified onto an existing VHDL signal or Verilog register or net. This allows you to force signals, registers, or nets at any level of the design hierarchy from within a VHDL architecture (e.g., a testbench). A signal_force works the same as the force command (CR-141) with the exception that you cannot issue a repeating force. See signal_force (UM-365) in Chapter 14 - Signal Spy for complete details.
signal_release()
The signal_release() procedure releases any force that was applied to an existing VHDL signal or Verilog register or net. This allows you to release signals, registers, or nets at any level of the design hierarchy from within a VHDL architecture (e.g., a testbench). A signal_release works the same as the noforce command (CR-164). See signal_release (UM-367) in Chapter 14 - Signal Spy for complete details.
Util package UM-89
to_real()
to_real() converts the physical type time value into a real value with respect to the current simulator resolution. The precision of the converted value is determined by the simulator resolution. For example, if you were converting 1900 fs to a real and the simulator resolution was ps, then the real value would be 2.0 (i.e., 2 ps).
Syntax
realval := to_real(timeval);
Returns
Name realval
Type real
Description The time value represented as a real with respect to the simulator resolution
Arguments
Name timeval
Type time
Description The value of the physical type time
Related functions
get_resolution (UM-87) to_time() (UM-90)
Example
If the simulator resolution is set to ps, and you enter the following function:
realval := to_real(12.99 ns);
then the value returned to realval would be 12990.0. If you wanted the returned value to be in units of nanoseconds (ns) instead, you would use the get_resolution (UM-87) function to recalculate the value:
realval := 1e+9 * (to_real(12.99 ns)) * get_resolution();
If you wanted the returned value to be in units of femtoseconds (fs), you would enter the function this way:
realval := 1e+15 * (to_real(12.99 ns)) * get_resolution();
UM-90
4 - VHDL simulation
to_time()
to_time() converts a real value into a time value with respect to the current simulator resolution. The precision of the converted value is determined by the simulator resolution. For example, if you were converting 5.9 to a time and the simulator resolution was ps, then the time value would be 6 ps.
Syntax
timeval := to_time(realval);
Returns
Name timeval
Type time
Description The real value represented as a physical type time with respect to the simulator resolution
Arguments
Name realval
Type real
Description The value of the type real
Related functions
get_resolution (UM-87) to_real() (UM-89)
Example
If the simulator resolution is set to ps, and you enter the following function:
timeval := to_time(72.49);
then the value returned to timeval would be 72 ps.
Modeling memory UM-91
Modeling memory
As a VHDL user, you might be tempted to model a memory using signals. Two common simulator problems are the likely result: You may get a "memory allocation error" message, which typically means the simulator ran out of memory and failed to allocate enough storage. Or, you may get very long load, elaboration, or run times. These problems are usually explained by the fact that signals consume a substantial amount of memory (many dozens of bytes per bit), all of which needs to be loaded or initialized before your simulation starts. Modeling memory with variables or protected types instead provides some excellent performance benefits: storage required to model the memory can be reduced by 1-2 orders of magnitude startup and run times are reduced associated memory allocation errors are eliminated In the VHDL example below, we illustrate three alternative architectures for entity memory: Architecture bad_style_87 uses a vhdl signal to store the ram data. Architecture style_87 uses variables in the memory process Architecture style_93 uses variables in the architecture. For large memories, architecture bad_style_87 runs many times longer than the other two, and uses much more memory. This style should be avoided. Architectures style_87 and style_93 work with equal efficiently. However, VHDL 1993 offers additional flexibility because the ram storage can be shared between multiple processes. For example, a second process is shown that initializes the memory; you could add other processes to create a multi-ported memory. To implement this model, you will need functions that convert vectors to integers. To use it you will probably need to convert integers to vectors. Example functions are provided below in package "conversions". For completeness sake we also show an example using VHDL 2002 protected types, though in this example, protected types offer no advantage over shared variables.
87 and 93 example
library ieee; use ieee.std_logic_1164.all; use work.conversions.all; entity memory is generic(add_bits : integer := 12; data_bits : integer := 32); port(add_in : in std_ulogic_vector(add_bits-1 downto 0); data_in : in std_ulogic_vector(data_bits-1 downto 0); data_out : out std_ulogic_vector(data_bits-1 downto 0); cs, mwrite : in std_ulogic; do_init : in std_ulogic); subtype word is std_ulogic_vector(data_bits-1 downto 0);
UM-92
4 - VHDL simulation
constant nwords : integer := 2 ** add_bits; type ram_type is array(0 to nwords-1) of word; end; architecture style_93 of memory is -----------------------------shared variable ram : ram_type; -----------------------------begin memory: process (cs) variable address : natural; begin if rising_edge(cs) then address := sulv_to_natural(add_in); if (mwrite = '1') then ram(address) := data_in; end if; data_out <= ram(address); end if; end process memory; -- illustrates a second process using the shared variable initialize: process (do_init) variable address : natural; begin if rising_edge(do_init) then for address in 0 to nwords-1 loop ram(address) := data_in; end loop; end if; end process initialize; end architecture style_93; architecture style_87 of memory is begin memory: process (cs) ----------------------variable ram : ram_type; ----------------------variable address : natural; begin if rising_edge(cs) then address := sulv_to_natural(add_in); if (mwrite = '1') then ram(address) := data_in; end if; data_out <= ram(address); end if; end process; end style_87; architecture bad_style_87 of memory is ---------------------signal ram : ram_type; ---------------------begin memory: process (cs) variable address : natural := 0; begin if rising_edge(cs) then address := sulv_to_natural(add_in);
if (mwrite = '1') then ram(address) <= data_in; data_out <= data_in; else data_out <= ram(address); end if; end if; end process; end bad_style_87; ----------------------------------------------------------------------------------------------------------------------library ieee; use ieee.std_logic_1164.all; package conversions is function sulv_to_natural(x : std_ulogic_vector) return natural; function natural_to_sulv(n, bits : natural) return std_ulogic_vector; end conversions; package body conversions is function sulv_to_natural(x : std_ulogic_vector) return natural is variable n : natural := 0; variable failure : boolean := false; begin assert (x'high - x'low + 1) <= 31 report "Range of sulv_to_natural argument exceeds natural range" severity error; for i in x'range loop n := n * 2; case x(i) is when '1' | 'H' => n := n + 1; when '0' | 'L' => null; when others => failure := true; end case; end loop; assert not failure report "sulv_to_natural cannot convert indefinite std_ulogic_vector" severity error; if failure then return 0; else return n; end if; end sulv_to_natural; function natural_to_sulv(n, bits : natural) return std_ulogic_vector is variable x : std_ulogic_vector(bits-1 downto 0) := (others => '0'); variable tempn : natural := n; begin for i in x'reverse_range loop if (tempn mod 2) = 1 then x(i) := '1'; end if; tempn := tempn / 2;
UM-94
4 - VHDL simulation
end loop; return x; end natural_to_sulv; end conversions;
02 example
------------------------------------------------------------------------------ Source: sp_syn_ram_protected.vhd -- Component: VHDL synchronous, single-port RAM -- Remarks: Various VHDL examples: random access memory (RAM) ----------------------------------------------------------------------------LIBRARY ieee; USE ieee.std_logic_1164.ALL; USE ieee.numeric_std.ALL; ENTITY sp_syn_ram_protected IS GENERIC ( data_width : positive := 8; addr_width : positive := 3 ); PORT ( inclk : IN std_logic; outclk : IN std_logic; we : IN std_logic; addr : IN unsigned(addr_width-1 DOWNTO 0); data_in : IN std_logic_vector(data_width-1 DOWNTO 0); data_out : OUT std_logic_vector(data_width-1 DOWNTO 0) ); END sp_syn_ram_protected;
ARCHITECTURE intarch OF sp_syn_ram_protected IS TYPE mem_type IS PROTECTED PROCEDURE write ( data : IN std_logic_vector(data_width-1 downto 0); addr : IN unsigned(addr_width-1 DOWNTO 0)); IMPURE FUNCTION read ( addr : IN unsigned(addr_width-1 DOWNTO 0)) RETURN std_logic_vector; END PROTECTED mem_type; TYPE mem_type IS PROTECTED BODY TYPE mem_array IS ARRAY (0 TO 2**addr_width-1) OF std_logic_vector(data_width-1 DOWNTO 0); VARIABLE mem : mem_array; PROCEDURE write ( data : IN std_logic_vector(data_width-1 downto 0); addr : IN unsigned(addr_width-1 DOWNTO 0)) IS BEGIN mem(to_integer(addr)) := data; END; IMPURE FUNCTION read ( addr : IN unsigned(addr_width-1 DOWNTO 0)) RETURN std_logic_vector IS BEGIN
return mem(to_integer(addr)); END; END PROTECTED BODY mem_type; SHARED VARIABLE memory : mem_type; BEGIN ASSERT data_width <= 32 REPORT "### Illegal data width detected" SEVERITY failure; control_proc : PROCESS (inclk, outclk) BEGIN IF (inclk'event AND inclk = '1') THEN IF (we = '1') THEN memory.write(data_in, addr); END IF; END IF; IF (outclk'event AND outclk = '1') THEN data_out <= memory.read(addr); END IF; END PROCESS; END intarch; ------------------------------------------------------------------------------ Source: ram_tb.vhd -- Component: VHDL testbench for RAM memory example -- Remarks: Simple VHDL example: random access memory (RAM) ----------------------------------------------------------------------------LIBRARY ieee; USE ieee.std_logic_1164.ALL; USE ieee.numeric_std.ALL; ENTITY ram_tb IS END ram_tb; ARCHITECTURE testbench OF ram_tb IS -------------------------------------------- Component declaration single-port RAM ------------------------------------------COMPONENT sp_syn_ram_protected GENERIC ( data_width : positive := 8; addr_width : positive := 3 ); PORT ( inclk : IN std_logic; outclk : IN std_logic; we : IN std_logic; addr : IN unsigned(addr_width-1 DOWNTO 0); data_in : IN std_logic_vector(data_width-1 DOWNTO 0); data_out : OUT std_logic_vector(data_width-1 DOWNTO 0) );
UM-96
4 - VHDL simulation
END COMPONENT; -------------------------------------------- Intermediate signals and constants ------------------------------------------SIGNAL addr : unsigned(19 DOWNTO 0); SIGNAL inaddr : unsigned(3 DOWNTO 0); SIGNAL outaddr : unsigned(3 DOWNTO 0); SIGNAL data_in : unsigned(31 DOWNTO 0); SIGNAL data_in1 : std_logic_vector(7 DOWNTO 0); SIGNAL data_sp1 : std_logic_vector(7 DOWNTO 0); SIGNAL we : std_logic; SIGNAL clk : std_logic; CONSTANT clk_pd : time := 100 ns;
BEGIN ---------------------------------------------------- instantiations of single-port RAM architectures. -- All architectures behave equivalently, but they -- have different implementations. The signal-based -- architecture (rtl) is not a recommended style. --------------------------------------------------spram1 : entity work.sp_syn_ram_protected GENERIC MAP ( data_width => 8, addr_width => 12) PORT MAP ( inclk => clk, outclk => clk, we => we, addr => addr(11 downto 0), data_in => data_in1, data_out => data_sp1); -------------------------------------------- clock generator ------------------------------------------clock_driver : PROCESS BEGIN clk <= '0'; WAIT FOR clk_pd / 2; LOOP clk <= '1', '0' AFTER clk_pd / 2; WAIT FOR clk_pd; END LOOP; END PROCESS; -------------------------------------------- data-in process ------------------------------------------datain_drivers : PROCESS(data_in) BEGIN data_in1 <= std_logic_vector(data_in(7 downto 0)); END PROCESS; -------------------------------------------- simulation control process ------------------------------------------ctrl_sim : PROCESS
BEGIN FOR i IN 0 TO 1023 LOOP we <= '1'; data_in <= to_unsigned(9000 + i, data_in'length); addr <= to_unsigned(i, addr'length); inaddr <= to_unsigned(i, inaddr'length); outaddr <= to_unsigned(i, outaddr'length); WAIT UNTIL clk'EVENT AND clk = '0'; WAIT UNTIL clk'EVENT AND clk = '0'; data_in <= to_unsigned(7 + i, addr <= to_unsigned(1 + i, inaddr <= to_unsigned(1 + i, WAIT UNTIL clk'EVENT AND clk = WAIT UNTIL clk'EVENT AND clk = data_in'length); addr'length); inaddr'length); '0'; '0';
data_in <= to_unsigned(3, data_in'length); addr <= to_unsigned(2 + i, addr'length); inaddr <= to_unsigned(2 + i, inaddr'length); WAIT UNTIL clk'EVENT AND clk = '0'; WAIT UNTIL clk'EVENT AND clk = '0'; data_in <= to_unsigned(30330, addr <= to_unsigned(3 + i, inaddr <= to_unsigned(3 + i, WAIT UNTIL clk'EVENT AND clk = WAIT UNTIL clk'EVENT AND clk = data_in'length); addr'length); inaddr'length); '0'; '0';
we <= '0'; addr <= to_unsigned(i, addr'length); outaddr <= to_unsigned(i, outaddr'length); WAIT UNTIL clk'EVENT AND clk = '0'; WAIT UNTIL clk'EVENT AND clk = '0'; addr <= to_unsigned(1 + i, outaddr <= to_unsigned(1 + i, WAIT UNTIL clk'EVENT AND clk = WAIT UNTIL clk'EVENT AND clk = addr <= to_unsigned(2 + i, outaddr <= to_unsigned(2 + i, WAIT UNTIL clk'EVENT AND clk = WAIT UNTIL clk'EVENT AND clk = addr <= to_unsigned(3 + i, outaddr <= to_unsigned(3 + i, WAIT UNTIL clk'EVENT AND clk = WAIT UNTIL clk'EVENT AND clk = END LOOP; ASSERT false REPORT "### End of Simulation!" SEVERITY failure; END PROCESS; END testbench; addr'length); outaddr'length); '0'; '0'; addr'length); outaddr'length); '0'; '0'; addr'length); outaddr'length); '0'; '0';
UM-98
4 - VHDL simulation
Affecting performance by cancelling scheduled events

Performance will suffer if events are scheduled far into the future but then cancelled before they take effect. This situation will act like a memory leak and slow down simulation. In VHDL this situation can occur several ways. The most common are waits with time-out clauses and projected waveforms in signal assignments. The following code shows a wait with a time-out:
signals synch : bit := '0'; ... p: process begin wait for 10 ms until synch = 1; end process; synch <= not synch after 10 ns;
At time 0, process p makes an event for time 10ms. When synch goes to 1 at 10 ns, the event at 10 ms is marked as cancelled but not deleted, and a new event is scheduled at 10ms + 10ns. The cancelled events are not reclaimed until time 10ms is reached and the cancelled event is processed. As a result there will be 500000 (10ms/20ns) cancelled but un-deleted events. Once 10ms is reached, memory will no longer increase because the simulator will be reclaiming events as fast as they are added. For projected waveforms the following would behave the same way:
signals synch : bit := '0'; ... p: process(synch) begin output <= '0', '1' after 10ms; end process; synch <= not synch after 10 ns;
Converting an integer into a bit_vector UM-99
Converting an integer into a bit_vector

The following code demonstrates how to convert an integer into a bit_vector.
library ieee; use ieee.numeric_bit.ALL; entity test is end test; architecture only of test is signal s1 : bit_vector(7 downto 0); signal int : integer := 45; begin p:process begin wait for 10 ns; s1 <= bit_vector(to_signed(int,8)); end process p; end only;
UM-100 4 - VHDL simulation
UM-101
5 - Verilog and SystemVerilog simulation

Chapter contents
Introduction . . . Terminology . . Basic Verilog flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-102 UM-102 UM-102 UM-103 UM-105 UM-107 UM-109 UM-110 UM-112 UM-113 UM-114 UM-114 UM-117 UM-121 UM-121 UM-123 UM-123 UM-123 UM-125 UM-127 UM-128 UM-129 UM-132 UM-132 UM-133 UM-134 UM-135
Compiling Verilog files . . . . . . Incremental compilation . . . . . Library usage . . . . . . . . Verilog-XL compatible compiler arguments Verilog-XL ùselib compiler directive . Verilog configurations . . . . . Verilog generate statements . . . .
Simulating Verilog designs . . . . . . Simulator resolution limit . . . . . . Event ordering in Verilog designs . . . Negative timing check limits . . . . . Verilog-XL compatible simulator arguments . Cell libraries . . . . SDF timing annotation Delay modes . . . . . . . . . . . . . . . . . . . . .
System tasks and functions . . . . . . . SystemVerilog system tasks and functions . . System tasks and functions specific to ModelSim Verilog-XL compatible system tasks and functions Compiler directives . . . . . . . IEEE Std 1364 compiler directives . . Compiler directives specific to ModelSim Verilog-XL compatible compiler directives Verilog PLI/VPI and SystemVerilog DPI . . . . . . . . . . . .
UM-102 5 - Verilog and SystemVerilog simulation
Introduction
This chapter describes how to compile, optimize, and simulate Verilog and SystemVerilog designs with ModelSim. ModelSim implements the Verilog language as defined by the IEEE Standards 1364-1995 and 1364-2005. We recommend that you obtain these specifications for reference. The following functionality is partially implemented in ModelSim: Verilog Procedural Interface (VPI) (see /<install_dir>/modeltech/docs/technotes/ Verilog_VPI.note for details) IEEE Std P1800-2005 SystemVerilog (see /<install_dir>/modeltech/docs/technotes/ sysvlog.note for implementation details)
Terminology
This chapter uses the term Verilog to represent both Verilog and SystemVerilog, unless otherwise noted.
Basic Verilog flow

Simulating Verilog designs with ModelSim includes four general steps: 1 Compile your Verilog code into one or more libraries using the vlog command (CR-293). See "Compiling Verilog files" (UM-103) for details. 2 Load your design with the vsim command (CR-305). See "Simulating Verilog designs" (UM-114) for details. 3 Run and debug your design.
Compiling Verilog files UM-103
Compiling Verilog files

The first time you compile a design there is a two-step process: 1 Create a working library with vlib or select File > New > Library. 2 Compile the design using vlog or select Compile > Compile.
Creating a working library

Before you can compile your design, you must create a library in which to store the compilation results. Use the vlib (CR-292) command or select File > New > Library to create a new library. For example:
vlib work
This creates a library named work. By default compilation results are stored in the work library. The work library is actually a subdirectory named work. This subdirectory contains a special file named _info. Do not create libraries using UNIX commands always use the vlib command (CR-292). See "Design libraries" (UM-53) for additional information on working with libraries.
Invoking the Verilog compiler

The Verilog compiler, vlog, compiles Verilog source code into retargetable, executable code. The library format is compatible across all supported platforms, and you can simulate your design on any platform without having to recompile your design. As the design compiles, the resulting object code for modules and UDPs is generated into a library. As noted above, the compiler places results into the work library by default. You can specify an alternate library with the -work argument.
Example
Here is a sample invocation of vlog:
vlog top.v +libext+.v+.u -y vlog_lib
After compiling top.v, vlog will scan the vlog_lib library for files with modules with the same name as primitives referenced, but undefined in top.v. The use of +libext+.v+.u implies filenames with a .v or .u suffix (any combination of suffixes may be used). Only referenced definitions will be compiled.
Parsing SystemVerilog keywords

With standard Verilog files (<filename>.v), vlog will not automatically parse SystemVerilog keywords. SystemVerilog keywords are parsed when any of the following situations exists: any file within the design contains the .sv file extension, the -sv argument is used with the vlog command (CR-293),
the Use System Verilog option is selected in the Verilog tab of the Compiler Options dialog. Access this dialog by selecting Compile > Compile Options from the Main window menu bar.
Here are two examples of the vlog command that will enable SystemVerilog features and keywords in ModelSim:
vlog testbench.sv top.v memory.v cache.v vlog -sv testbench.v proc.v
In the first example, the .sv extension for testbench automatically instructs ModelSim to parse SystemVerilog keywords. The -sv option used in the second example enables SystemVerilog features and keywords. Though a primary goal of the SystemVerilog standardization efforts has been to ensure full backward compatibility with the IEEE 1364-2005 Verilog standard, there is an issue with keywords. SystemVerilog adds several new keywords to the Verilog language. If your design uses one of these keywords as a regular identifier for a variable, module, task, function, etc., your design will not compile in ModelSim.
Incremental compilation
ModelSim Verilog supports incremental compilation of designs. Unlike other Verilog simulators, there is no requirement that you compile the entire design in one invocation of the compiler. You are not required to compile your design in any particular order (unless you are using SystemVerilog packages; see note below) because all module and UDP instantiations and external hierarchical references are resolved when the design is loaded by the simulator. Note: Compilation order may matter when using SystemVerilog packages. As stated in the IEEE std p1800-2005 LRM, section entitled Referencing data in packages, which states: "Packages must exist in order for the items they define to be recognized by the scopes in which they are imported. Incremental compilation is made possible by deferring these bindings, and as a result some errors cannot be detected during compilation. Commonly, these errors include: modules that were referenced but not compiled, incorrect port connections, and incorrect hierarchical references.
Example
Contents of testbench.sv
module testbench; timeunit 1ns; timeprecision 10ps; bit d=1, clk = 0; wire q; initial for (int cycles=0; cycles < 100; cycles++) #100 clk = !clk; design dut(q, d, clk); endmodule
Contents of design.v:
module design(output bit q, input bit d, clk); timeunit 1ns; timeprecision 10ps; always @(posedge clk) q = d; endmodule
Compile the design incrementally as follows:

ModelSim> vlog testbench.sv . . # Top level modules: # testbench ModelSim> vlog -sv test1.v . . # Top level modules: # dut
Note that the compiler lists each module as a top-level module, although, ultimately, only testbench is a top-level module. If a module is not referenced by another module compiled in the same invocation of the compiler, then it is listed as a top-level module. This is just an informative message and can be ignored during incremental compilation. The message is more useful when you compile an entire design in one invocation of the compiler and need to know the top-level module names for the simulator. For example,
% vlog top.v -- Compiling -- Compiling -- Compiling and2.v module module module or2.v top and2 or2
Top level modules: top
Automatic incremental compilation with -incr

The most efficient method of incremental compilation is to manually compile only the modules that have changed. However, this is not always convenient, especially if your source files have compiler directive interdependencies (such as macros). In this case, you may prefer to compile your entire design along with the -incr argument. This causes the compiler to automatically determine which modules have changed and generate code only for those modules. The following is an example of how to compile a design with automatic incremental compilation:
% vlog -incr -- Compiling -- Compiling -- Compiling top.v and2.v or2.v module top module and2 module or2
Top level modules: top
Now, suppose that you modify the functionality of the or2 module:
% vlog -incr top.v and2.v or2.v -- Skipping module top -- Skipping module and2 -- Compiling module or2 Top level modules: top
The compiler informs you that it skipped the modules top and and2, and compiled or2. Automatic incremental compilation is intelligent about when to compile a module. For example, changing a comment in your source code does not result in a recompile; however, changing the compiler command line arguments results in a recompile of all modules. Note: Changes to your source code that do not change functionality but that do affect source code line numbers (such as adding a comment line) will cause all affected modules to be recompiled. This happens because debug information must be kept current so that ModelSim can trace back to the correct areas of the source code.
Library usage
All modules and UDPs in a Verilog design must be compiled into one or more libraries. One library is usually sufficient for a simple design, but you may want to organize your modules into various libraries for a complex design. If your design uses different modules having the same name, then you are required to put those modules in different libraries because design unit names must be unique within a library. The following is an example of how you may organize your ASIC cells into one library and the rest of your design into another:
% vlib work % vlib asiclib % vlog -work asiclib and2.v or2.v -- Compiling module and2 -- Compiling module or2 Top level modules: and2 or2 % vlog top.v -- Compiling module top Top level modules: top
Note that the first compilation uses the -work asiclib argument to instruct the compiler to place the results in the asiclib library rather than the default work library.
Library search rules

Since instantiation bindings are not determined at compile time, you must instruct the simulator to search your libraries when loading the design. The top-level modules are loaded from the library named work unless you prefix the modules with the <library>. option. All other Verilog instantiations are resolved in the following order: Search libraries specified with -Lf arguments in the order they appear on the command line. Search the library specified in the "Verilog-XL ùselib compiler directive" (UM-110). Search libraries specified with -L arguments in the order they appear on the command line. Search the work library. Search the library explicitly named in the special escaped identifier instance name.
Handling sub-modules with common names

Sometimes in one design you need to reference two different modules that have the same name. This situation can occur if you have hierarchical modules organized into separate libraries, and you have commonly-named sub-modules in the libraries that have different definitions. This may happen if you are using vendor-supplied libraries. For example, say you have the following:
top modA modB
lib1: modA cellX
lib2: modB cellX
The normal library search rules will fail in this situation. For example, if you load the design as follows:
vsim -L lib1 -L lib2 top
both instantiations of cellX resolve to the lib1 version of cellX. On the other hand, if you specify -L lib2 -L lib1, both instantiations of cellX resolve to the lib2 version of cellX. To handle this situation, ModelSim implements a special interpretation of the expression -L work. When you specify -L work first in the search library arguments you are directing vsim to search for the instantiated module or UDP in the library that contains the module that does the instantiation. In the example above you would invoke vsim as follows:
vsim -L work -L lib1 -L lib2 top
Verilog-XL compatible compiler arguments

The compiler arguments listed below are equivalent to Verilog-XL arguments and may ease the porting of a design to ModelSim. See the vlog command (CR-293) for a description of each argument.
+define+<macro_name>[=<macro_text>] +delay_mode_distributed +delay_mode_path +delay_mode_unit +delay_mode_zero -f <filename> +incdir+<directory> +mindelays +maxdelays +nowarn<mnemonic> +typdelays -u
Arguments supporting source libraries

The compiler arguments listed below support source libraries in the same manner as Verilog-XL. See the vlog command (CR-293) for a description of each argument. Note that these source libraries are very different from the libraries that the ModelSim compiler uses to store compilation results. You may find it convenient to use these arguments if you are porting a design to ModelSim or if you are familiar with these arguments and prefer to use them. Source libraries are searched after the source files on the command line are compiled. If there are any unresolved references to modules or UDPs, then the compiler searches the source libraries to satisfy them. The modules compiled from source libraries may in turn have additional unresolved references that cause the source libraries to be searched again. This process is repeated until all references are resolved or until no new unresolved references are found. Source libraries are searched in the order they appear on the command line.
-v <filename> -y <directory> +libext+<suffix> +librescan +nolibcell -R [<simargs>]
Verilog-XL ùselib compiler directive

The ùselib compiler directive is an alternative source library management scheme to the -v, -y, and +libext compiler arguments. It has the advantage that a design may reference different modules having the same name. You compile designs that contain ùselib directive statements using the -compile_uselibs argument (described below) to vlog (CR293). The syntax for the ùselib directive is:
ùselib <library_reference>...
where <library_reference> is:

dir=<library_directory> | file=<library_file> | libext=<file_extension> | lib=<library_name>
The library references are equivalent to command line arguments as follows:

dir=<library_directory> -y <library_directory> file=<library_file> -v <library_file> libext=<file_extension> +libext+<file_extension>
For example, the following directive

ùselib dir=/h/vendorA libext=.v
is equivalent to the following command line arguments:

-y /h/vendorA +libext+.v
Since the ùselib directives are embedded in the Verilog source code, there is more flexibility in defining the source libraries for the instantiations in the design. The appearance of a ùselib directive in the source code explicitly defines how instantiations that follow it are resolved, completely overriding any previous ùselib directives.
-compile_uselibs argument
Use the -compile_uselibs argument to vlog (CR-293) to reference ùselib directives. The argument finds the source files referenced in the directive, compiles them into automatically created object libraries, and updates the modelsim.ini file with the logical mappings to the libraries. When using -compile_uselibs, ModelSim determines into which directory to compile the object libraries by choosing, in order, from the following three values: The directory name specified by the -compile_uselibs argument. For example,
-compile_uselibs=./mydir
The directory specified by the MTI_USELIB_DIR environment variable (see "Environment variables" (UM-434)) A directory named mti_uselibs that is created in the current working directory
The following code fragment and compiler invocation show how two different modules that have the same name can be instantiated within the same design:
module top; ùselib dir=/h/vendorA libext=.v NAND2 u1(n1, n2, n3); ùselib dir=/h/vendorB libext=.v NAND2 u2(n4, n5, n6); endmodule vlog -compile_uselibs top
This allows the NAND2 module to have different definitions in the vendorA and vendorB libraries.
ùselib is persistent
As mentioned above, the appearance of a ùselib directive in the source code explicitly defines how instantiations that follow it are resolved. This may result in unexpected consequences. For example, consider the following compile command:
vlog -compile_uselibs dut.v srtr.v
Assume that dut.v contains a ùselib directive. Since srtr.v is compiled after dut.v, the ùselib directive is still in effect. When srtr is loaded it is using the ùselib directive from dut.v to decide where to locate modules. If this is not what you intend, then you need to put an empty ùselib at the end of dut.v to "close" the previous ùselib statement.
Verilog configurations
The Verilog 2001 specification added configurations. Configurations specify how a design is "assembled" during the elaboration phase of simulation. Configurations actually consist of two pieces: the library mapping and the configuration itself. The library mapping is used at compile time to determine into which libraries the source files are to be compiled. Here is an example of a simple library map file:
library library library library work rtlLib gateLib aLib ../top.v; lrm_ex_top.v; lrm_ex_adder.vg; lrm_ex_adder.v;
Here is an example of a library map file that uses -incdir:

library lib1 src_dir/*.v -incdir ../include_dir2, ../, my_incdir;
The name of the library map file is arbitrary. You specify the library map file using the -libmap argument to the vlog command (CR-293). Alternatively, you can specify the file name as the first item on the vlog command line, and the compiler will read it as a library map file. The library map file must be compiled along with the Verilog source files. Multiple map files are allowed but each must be preceded by the -libmap argument. The library map file and the configuration can exist in the same or different files. If they are separate, only the map file needs the -libmap argument. The configuration is treated as any other Verilog source file.
Configurations and the library named work

The library named work is treated specially by ModelSim (see "The library named "work"" (UM-54) for details) for Verilog configurations. Consider the following code example:
config cfg; design top; instance top.u1 use work.u1; endconfig
In this case, work.u1 indicates to load u1 from the current library.
Verilog generate statements

The Verilog 2001 rules for generate statements have numerous inconsistencies and ambiguities. As a result, ModelSim implements the rules that have been proposed for Verilog 2005. Most of the rules are backwards compatible, but there is one key difference related to name visibility.
Name visibility in generate statements

Consider the following code example:
module m; parameter p = 1; generate if (p) integer x = 1; else real x = 2.0; endgenerate initial $display(x); endmodule
This code sample is legal under 2001 rules. However, it is illegal under the proposed 2005 rules and will cause an error in ModelSim. Under the new rules, you cannot hierarchically reference a name in an anonymous scope from outside that scope. In the example above, x does not propagate its visibility upwards, and each condition alternative is considered to be an anonymous scope. To fix the code such that it will simulate properly in ModelSim, write it like this instead:
module m; parameter p = 1; if (p) begin:s integer x = 1; end else begin:s real x = 2.0; end initial $display(s.x); endmodule
Since the scope is named in this example, normal hierarchical resolution rules apply and the code is fine. Note too that the keywords generate - endgenerate are optional under the new rules and are excluded in the second example.
Simulating Verilog designs

A Verilog design is ready for simulation after it has been compiled with vlog. The simulator may then be invoked with the names of the top-level modules (many designs contain only one top-level module). For example, if your top-level modules are "testbench" and "globals", then invoke the simulator as follows:
vsim testbench globals
After the simulator loads the top-level modules, it iteratively loads the instantiated modules and UDPs in the design hierarchy, linking the design together by connecting the ports and resolving hierarchical references. By default all modules and UDPs are loaded from the library named work. Modules and UDPs from other libraries can be specified using the -L or -Lf arguments to vsim (see "Library usage" (UM-107) for details). On successful loading of the design, the simulation time is set to zero, and you must enter a run command to begin simulation. Commonly, you enter run -all to run until there are no more simulation events or until $finish is executed in the Verilog code. You can also run for specific time periods (e.g., run 100 ns). Enter the quit command to exit the simulator.

The simulator internally represents time as a 64-bit integer in units equivalent to the smallest unit of simulation time, also known as the simulator resolution limit. The resolution limit defaults to the smallest time precision found among all of the `timescale compiler directives in the design. Here is an example of a `timescale directive:
`timescale 1 ns / 100 ps
The first number is the time units and the second number is the time precision. The directive above causes time values to be read as ns and to be rounded to the nearest 100 ps. Time units and precision can also be specified with SystemVerilog keywords as follows:
timeunit 1 ns timeprecision 100 ps
Simulating Verilog designs UM-115
Modules without timescale directives

You may encounter unexpected behavior if your design contains some modules with timescale directives and others without. The time units for modules without a timescale directive default to the simulator resolution. For example, say you have the two modules shown in the table below: Module 1
`timescale 1 ns / 10 ps module mod1 (set); output set; reg set; parameter d = 1.55; initial begin set = 1'bz; #d set = 1'b0; #d set = 1'b1; end endmodule
Module 2
module mod2 (set); output set; reg set; parameter d = 1.55; initial begin set = 1'bz; #d set = 1'b0; #d set = 1'b1; end endmodule
If you invoke vsim as vsim mod2 mod1 then Module 1 sets the simulator resolution to 10 ps. Module 2 has no timescale directive, so the time units default to the simulator resolution, in this case 10 ps. If you watched /mod1/set and /mod2/set in the Wave window, youd see that in Module 1 it transitions every 1.55 ns as expected (because of the 1 ns time unit in the timescale directive). However, in Module 2, set transitions every 20 ps. Thats because the delay of 1.55 in Module 2 is read as 15.5 ps and is rounded up to 20 ps. In such cases ModelSim will issue the following warning message during elaboration:
** Warning: (vsim-3010) [TSCALE] - Module 'mod1' has a `timescale directive in effect, but previous modules do not.
If you invoke vsim as vsim mod1 mod2, the simulation results would be the same but ModelSim would produce a different warning message:
** Warning: (vsim-3009) [TSCALE] - Module 'mod2' does not have a `timescale directive in effect, but previous modules do.
These warnings should ALWAYS be investigated. If the design contains no `timescale directives, then the resolution limit and time units default to the value specified by the Resolution (UM-447) variable in the modelsim.ini file. (The variable is set to 1 ns by default.)
-timescale option
The -timescale option can be used with the vlog and vopt to specifies the default timescale for modules not having an explicit `timescale directive in effect during compilation. The format of the -timescale argument is the same as that of the `timescale directive
-timescale <time_units>/<time_precision>
The format for <time_units> and <time_precision> is <n><units>. The value of <n> must be 1, 10, or 100. The value of <units> must be fs, ps, ns, us, ms, or s. In addition, the <time_units> must be greater than or equal to the <time_precision>. For example:
-timescale "1ns / 1ps"
The argument above needs quotes because it contains white space.
Multiple timescale directives

As alluded to above, your design can have multiple timescale directives. The timescale directive takes effect where it appears in a source file and applies to all source files which follow in the same vlog (CR-293) command. Separately compiled modules can also have different timescales. The simulator determines the smallest timescale of all the modules in a design and uses that as the simulator resolution.
`timescale, -t, and rounding

The optional vsim argument -t sets the simulator resolution limit for the overall simulation. If the resolution set by -t is larger than the precision set in a module, the time values in that module are rounded up. If the resolution set by -t is smaller than the precision of the module, the precision of that module remains whatever is specified by the `timescale directive. Consider the following code:
`timescale 1 ns / 100 ps module foo; initial #12.536 $display
The list below shows three possibilities for -t and how the delays in the module would be handled in each case: -t not set The delay will be rounded to 12.5 as directed by the modules timescale directive. -t is set to 1 fs The delay will be rounded to 12.5. Again, the modules precision is determined by the timescale directive. ModelSim does not override the modules precision. -t is set to 1 ns The delay will be rounded to 12. The modules precision is determined by the -t setting. ModelSim has no choice but to round the modules time values because the entire simulation is operating at 1 ns.
Choosing the resolution

You should choose the coarsest resolution limit possible that does not result in undesired rounding of your delays. The time precision should not be unnecessarily small because it will limit the maximum simulation time limit, and it will degrade performance in some cases.
Event ordering in Verilog designs

Event-based simulators such as ModelSim may process multiple events at a given simulation time. The Verilog language is defined such that you cannot explicitly control the order in which simultaneous events are processed. Unfortunately, some designs rely on a particular event order, and these designs may behave differently than you expect.
Event queues
Section 5 of the IEEE Std 1364-1995 LRM defines several event queues that determine the order in which events are evaluated. At the current simulation time, the simulator has the following pending events: active events inactive events non-blocking assignment update events monitor events future events - inactive events - non-blocking assignment update events The LRM dictates that events are processed as follows 1) all active events are processed; 2) the inactive events are moved to the active event queue and then processed; 3) the non-blocking events are moved to the active event queue and then processed; 4) the monitor events are moved to the active queue and then processed; 5) simulation advances to the next time where there is an inactive event or a non-blocking assignment update event. Within the active event queue, the events can be processed in any order, and new active events can be added to the queue in any order. In other words, you cannot control event order within the active queue. The example below illustrates potential ramifications of this situation. Say you have these four statements: 1 always@(q) p = q; 2 always @(q) p2 = not q; 3 always @(p or p2) clk = p and p2; 4 always @(posedge clk) and current values as follows: q = 0, p = 0, p2=1
The tables below show two of the many valid evaluations of these statements. Evaluation events are denoted by a number where the number is the statement to be evaluated. Update events are denoted <name>(old->new) where <name> indicates the reg being updated and new is the updated value.
Table 1: Evaluation 1
Event being processed Active event queue q(0 1) q(0 1) 1 p(0 1) 3 clk(0 1) 4 2 p2(1 0) 3 clk(1 0) 1, 2 p(0 1), 2 3, 2 clk(0 1), 2 4, 2 2 p2(1 0) 3 clk(1 0) <empty>
Table 2: Evaluation 2
Event being processed Active event queue q(0 1) q(0 1) 1 2 p(0 1) p2(1 0) 3 1, 2 p(0 1), 2 p2(1 0), p(0 1) 3, p2(1 0) 3 <empty> (clk doesnt change)
Again, both evaluations are valid. However, in Evaluation 1, clk has a glitch on it; in Evaluation 2, clk doesnt. This indicates that the design has a zero-delay race condition on clk.
Controlling event queues with blocking/non-blocking assignments

The only control you have over event order is to assign an event to a particular queue. You do this via blocking or non-blocking assignments. Blocking assignments Blocking assignments place an event in the active, inactive, or future queues depending on what type of delay they have: a blocking assignment without a delay goes in the active queue a blocking assignment with an explicit delay of 0 goes in the inactive queue a blocking assignment with a non-zero delay goes in the future queue Non-blocking assignments A non-blocking assignment goes into either the non-blocking assignment update event queue or the future non-blocking assignment update event queue. (Non-blocking assignments with no delays and those with explicit zero delays are treated the same.) Non-blocking assignments should be used only for outputs of flip-flops. This insures that all outputs of flip-flops do not change until after all flip-flops have been evaluated. Attempting to use non-blocking assignments in combinational logic paths to remove race conditions may only cause more problems. (In the preceding example, changing all statements to non-blocking assignments would not remove the race condition.) This includes using non-blocking assignments in the generation of gated clocks. The following is an example of how to properly use non-blocking assignments.
gen1: always @(master) clk1 = master; gen2: always @(clk1) clk2 = clk1; f1 : always @(posedge clk1) begin q1 <= d1; end f2: always @(posedge clk2) begin q2 <= q1; end
If written this way, a value on d1 always takes two clock cycles to get from d1 to q2. If you change clk1 = master and clk2 = clk1 to non-blocking assignments or q2 <= q1 and q1 <= d1 to blocking assignments, then d1 may get to q2 is less than two clock cycles.
Debugging event order issues

Since many models have been developed on Verilog-XL, ModelSim tries to duplicate Verilog-XL event ordering to ease the porting of those models to ModelSim. However, ModelSim does not match Verilog-XL event ordering in all cases, and if a model ported to ModelSim does not behave as expected, then you should suspect that there are event order dependencies.
ModelSim helps you track down event order dependencies with the following compiler arguments: -compat, -hazards, and -keep_delta. See the vlog command (CR-293) for descriptions of -compat and -keep_delta.
Hazard detection
The -hazard argument to vsim (CR-305) detects event order hazards involving simultaneous reading and writing of the same register in concurrently executing processes. vsim detects the following kinds of hazards: WRITE/WRITE: Two processes writing to the same variable at the same time. READ/WRITE: One process reading a variable at the same time it is being written to by another process. ModelSim calls this a READ/WRITE hazard if it executed the read first. WRITE/READ: Same as a READ/WRITE hazard except that ModelSim executed the write first. vsim issues an error message when it detects a hazard. The message pinpoints the variable and the two processes involved. You can have the simulator break on the statement where the hazard is detected by setting the break on assertion level to Error. To enable hazard detection you must invoke vlog (CR-293) with the -hazards argument when you compile your source code and you must also invoke vsim with the -hazards argument when you simulate. Important: Enabling -hazards implicitly enables the -compat argument. As a result, using this argument may affect your simulation results.
Hazard detection and optimization levels

In certain cases hazard detection results are affected by the optimization level used in the simulation. Some optimizations change the read/write operations performed on a variable if the transformation is determined to yield equivalent results. Since the hazard detection algorithm doesnt know whether or not the read/write operations can affect the simulation results, the optimizations can result in different hazard detection results. Generally, the optimizations reduce the number of false hazards by eliminating unnecessary reads and writes, but there are also optimizations that can produce additional false hazards.
Limitations of hazard detection

Reads and writes involving bit and part selects of vectors are not considered for hazard detection. The overhead of tracking the overlap between the bit and part selects is too high. A WRITE/WRITE hazard is flagged even if the same value is written by both processes. A WRITE/READ or READ/WRITE hazard is flagged even if the write does not modify the variable's value. Glitches on nets caused by non-guaranteed event ordering are not detected. A non-blocking assignment is not treated as a WRITE for hazard detection purposes. This is because non-blocking assignments are not normally involved in hazards. (In fact, they should be used to avoid hazards.)
Negative timing check limits

Verilog supports negative limit values in the $setuphold and $recrem system tasks. These tasks have optional delayed versions of input signals to insure proper evaluation of models with negative timing check limits. Delay values for these delayed nets are determined by the simulator so that valid data is available for evaluation before a clocking signal.
Example
$setuphold(posedge clk, negedge d, 5, -3, Notifier,,, clk_dly, d_dly);
d violation region clk
3 0
ModelSim calculates the delay for signal d_dly as 4 time units instead of 3. It does this to prevent d_dly and clk_dly from occurring simultaneously when a violation isnt reported. ModelSim accepts negative limit checks by default, unlike current versions of Verilog-XL. To match Verilog-XL default behavior (i.e., zeroing all negative timing check limits), use the +no_neg_tcheck argument to vsim (CR-305).
Negative timing constraint algorithm

The algorithm ModelSim uses to calculate delays for delayed nets isnt described in IEEE Std 1364. Rather, ModelSim matches Verilog-XL behavior. The algorithm attempts to find a set of delays so the data net is valid when the clock net transitions and the timing checks are satisfied. The algorithm is iterative because a set of delays can be selected that satisfies all timing checks for a pair of inputs but then causes mis-ordering of another pair (where both pairs of inputs share a common input). When a set of delays that satisfies all timing checks is found, the delays are said to converge.
Verilog-XL compatible simulator arguments

The simulator arguments listed below are equivalent to Verilog-XL arguments and may ease the porting of a design to ModelSim. See the vsim command (CR-305) for a description of each argument.
+alt_path_delays -l <filename> +maxdelays +mindelays +multisource_int_delays +no_cancelled_e_msg +no_neg_tchk +no_notifier +no_path_edge +no_pulse_msg -no_risefall_delaynets +no_show_cancelled_e +nosdfwarn +nowarn<mnemonic> +ntc_warn +pulse_e/<percent> +pulse_e_style_ondetect
+pulse_e_style_onevent +pulse_int_e/<percent> +pulse_int_r/<percent> +pulse_r/<percent> +sdf_nocheck_celltype +sdf_verbose +show_cancelled_e +transport_int_delays +transport_path_delays +typdelays
Cell libraries UM-123
Cell libraries
Model Technology passed the ASIC Councils Verilog test suite and achieved the "Library Tested and Approved" designation from Si2 Labs. This test suite is designed to ensure Verilog timing accuracy and functionality and is the first significant hurdle to complete on the way to achieving full ASIC vendor support. As a consequence, many ASIC and FPGA vendors Verilog cell libraries are compatible with ModelSim Verilog. The cell models generally contain Verilog "specify blocks" that describe the path delays and timing constraints for the cells. See section 13 in the IEEE Std 1364-1995 for details on specify blocks, and section 14.5 for details on timing constraints. ModelSim Verilog fully implements specify blocks and timing constraints as defined in IEEE Std 1364 along with some Verilog-XL compatible extensions.
SDF timing annotation

ModelSim Verilog supports timing annotation from Standard Delay Format (SDF) files. See Chapter 15 - Standard Delay Format (SDF) Timing Annotation for details.
Delay modes
Verilog models may contain both distributed delays and path delays. The delays on primitives, UDPs, and continuous assignments are the distributed delays, whereas the portto-port delays specified in specify blocks are the path delays. These delays interact to determine the actual delay observed. Most Verilog cells use path delays exclusively, with the distributed delays set to zero. For example,
module and2(y, a, b); input a, b; output y; and(y, a, b); specify (a => y) = 5; (b => y) = 5; endspecify endmodule
In the above two-input "and" gate cell, the distributed delay for the "and" primitive is zero, and the actual delays observed on the module ports are taken from the path delays. This is typical for most cells, but a complex cell may require non-zero distributed delays to work properly. Even so, these delays are usually small enough that the path delays take priority over the distributed delays. The rule is that if a module contains both path delays and distributed delays, then the larger of the two delays for each path shall be used (as defined by the IEEE Std 1364). This is the default behavior, but you can specify alternate delay modes with compiler directives and arguments. These arguments and directives are compatible with Verilog-XL. Compiler delay mode arguments take precedence over delay mode directives in the source code.
Distributed delay mode

In distributed delay mode the specify path delays are ignored in favor of the distributed delays. Select this delay mode with the +delay_mode_distributed compiler argument or the `delay_mode_distributed compiler directive.
Path delay mode

In path delay mode the distributed delays are set to zero in any module that contains a path delay. Select this delay mode with the +delay_mode_path compiler argument or the `delay_mode_path compiler directive.
Unit delay mode

In unit delay mode the non-zero distributed delays are set to one unit of simulation resolution (determined by the minimum time_precision argument in all timescale directives in your design or the value specified with the -t argument to vsim), and the specify path delays and timing constraints are ignored. Select this delay mode with the +delay_mode_unit compiler argument or the `delay_mode_unit compiler directive.
Zero delay mode

In zero delay mode the distributed delays are set to zero, and the specify path delays and timing constraints are ignored. Select this delay mode with the +delay_mode_zero compiler argument or the `delay_mode_zero compiler directive.
System tasks and functions UM-125
System tasks and functions

ModelSim supports system tasks and functions as follows: All system tasks and functions defined in IEEE Std 1364 Some system tasks and functions defined in SystemVerilog IEEE std p1800-2005 LRM Several system tasks and functions that are specific to ModelSim Several non-standard, Verilog-XL system tasks The system tasks and functions listed in this section are built into the simulator, although some designs depend on user-defined system tasks implemented with the Programming Language Interface (PLI), Verilog Procedural Interface (VPI), or the SystemVerilog DPI (Direct Programming Interface). If the simulator issues warnings regarding undefined system tasks or functions, then it is likely that these tasks or functions are defined by a PLI/ VPI application that must be loaded by the simulator.
IEEE Std 1364 system tasks and functions

The following system tasks and functions are described in detail in the IEEE Std 1364.
Timescale tasks $printtimescale $timeformat
Simulator control tasks $finish $stop
Simulation time functions $realtime $stime $time
Command line input $test$plusargs $value$plusargs
Probabilistic distribution functions $dist_chi_square $dist_erlang $dist_exponential $dist_normal $dist_poisson $dist_t $dist_uniform $random
Conversion functions $bitstoreal $itor $realtobits $rtoi $signed $unsigned
Stochastic analysis tasks $q_add $q_exam $q_full $q_initialize $q_remove
Timing check tasks $hold $nochange $period $recovery $setup $setuphold $skew $widtha $removal $recrem
a.Verilog-XL ignores the threshold argument even though it is part of the Verilog spec. ModelSim does not ignore this argument. Be careful that you dont set the threshold argument greater-than-or-equal to the limit argument as that essentially disables the $width check. Note too that you cannot override the threshold argument via SDF annotation. Display tasks $display $displayb $displayh $displayo $monitor $monitorb $monitorh $monitoro $monitoroff $monitoron $strobe $strobeb $strobeh $strobeo $write $writeb $writeh $writeo PLA modeling tasks $async$and$array $async$nand$array $async$or$array $async$nor$array $async$and$plane $async$nand$plane $async$or$plane $async$nor$plane $sync$and$array $sync$nand$array $sync$or$array $sync$nor$array $sync$and$plane $sync$nand$plane $sync$or$plane $sync$nor$plane Value change dump (VCD) file tasks $dumpall $dumpfile $dumpflush $dumplimit $dumpoff $dumpon $dumpvars
File I/O tasks $fclose $fdisplay $fdisplayb $fdisplayh $fdisplayo $feof $ferror $fflush $fgetc $fgets $fmonitor $fmonitorb $fmonitorh $fmonitoro $fopen $fread $fscanf $fseek $fstrobe $fstrobeb $fstrobeh $fstrobeo $ftell $fwrite $fwriteb $fwriteh $fwriteo $readmemb $readmemh $rewind $sdf_annotate $sformat $sscanf $swrite $swriteb $swriteh $swriteo $ungetc
SystemVerilog system tasks and functions

The following ModelSim-supported system tasks and functions are described in detail in the SystemVerilog IEEE Std p1800-2005 LRM.
Elaboration-time function $typeof
Typename function $typename
Expression size function $bits
Range function $isunbounded
Shortreal conversions $shortrealbits $bitstoshortreal
Array querying functions $dimensions $left $right $low $high
Shortreal conversions
Array querying functions $increment $size
Reading packed data functions $readmemb $readmemh
Writing packed data functions $writememb $writememh
Other functions $root $unit
System tasks and functions specific to ModelSim

The following system tasks and functions are specific to ModelSim. They are not included in the IEEE Std 1364 nor are they likely supported in other simulators. Their use may limit the portability of your code.
$coverage_save(<filename>, [<instancepath>], [<xml_output>])
The $coverage_save() system function saves Code Coverage information to a file during a batch run that typically would terminate via the $finish call. It also returns a 1 to indicate that the coverage information was saved successfully or a 0 to indicate an error (unable to open file, instance name not found, etc.) If you dont specify <instancepath>, ModelSim saves all coverage data in the current design to the specified file. If you do specify <instancepath>, ModelSim saves data on that instance, and all instances below it (recursively), to the specified file. If set to 1, the [<xml_output>] argument specifies that the output be saved in XML format. See Chapter 11 - Measuring code coverage for more information on Code Coverage.
$init_signal_driver
The $init_signal_driver() system task drives the value of a VHDL signal or Verilog net onto an existing VHDL signal or Verilog net. This allows you to drive signals or nets at any level of the design hierarchy from within a Verilog module (e.g., a testbench). See $init_signal_driver (UM-371) in Chapter 14 - Signal Spy for complete details.
$init_signal_spy
The $init_signal_spy() system task mirrors the value of a VHDL signal or Verilog register/net onto an existing Verilog register or VHDL signal. This system task allows you to reference signals, registers, or nets at any level of hierarchy from within a Verilog module (e.g., a testbench). See $init_signal_spy (UM-374) in Chapter 14 - Signal Spy for complete details.
$signal_force
The $signal_force() system task forces the value specified onto an existing VHDL signal or Verilog register or net. This allows you to force signals, registers, or nets at any level of the design hierarchy from within a Verilog module (e.g., a testbench). A $signal_force
works the same as the force command (CR-141) with the exception that you cannot issue a repeating force. See $signal_force (UM-377) in Chapter 14 - Signal Spy for complete details.
$signal_release
The $signal_release() system task releases a value that had previously been forced onto an existing VHDL signal or Verilog register or net. A $signal_release works the same as the noforce command (CR-164). See $signal_release (UM-379) in Chapter 14 - Signal Spy.
$sdf_done
This task is a "cleanup" function that removes internal buffers, called MIPDs, that have a delay value of zero. These MIPDs are inserted in response to the -v2k_int_delay argument to the vsim command (CR-305). In general the simulator will automatically remove all zero delay MIPDs. However, if you have $sdf_annotate() calls in your design that are not getting executed, the zero-delay MIPDs are not removed. Adding the $sdf_done task after your last $sdf_annotate() will remove any zero-delay MIPDs that have been created.
Verilog-XL compatible system tasks and functions

ModelSim supports a number of Verilog-XL specific system tasks and functions.
Supported tasks and functions mentioned in IEEE Std 1364

The following supported system tasks and functions, though not part of the IEEE standard, are described in an annex of the IEEE Std 1364.
$countdrivers $getpattern $sreadmemb $sreadmemh
Supported tasks not described in the IEEE Std 1364

The following system tasks are also provided for compatibility with Verilog-XL, though they are not described in the IEEE Std 1364.
$deposit(variable, value);
This system task sets a Verilog register or net to the specified value. variable is the register or net to be changed; value is the new value for the register or net. The value remains until there is a subsequent driver transaction or another $deposit task for the same register or net. This system task operates identically to the ModelSim force -deposit command.
$disable_warnings(<keyword>[,<module_instance>...]);
This system task instructs ModelSim to disable warnings about timing check violations or triregs that acquire a value of X due to charge decay. <keyword> may be decay or timing. You can specify one or more module instance names. If you dont specify a module instance, ModelSim disables warnings for the entire simulation.
$enable_warnings(<keyword>[,<module_instance>...]);
This system task enables warnings about timing check violations or triregs that acquire a value of X due to charge decay. <keyword> may be decay or timing. You can specify one or more module instance names. If you dont specify a module_instance, ModelSim enables warnings for the entire simulation.
Supported tasks that have been extended

The following system tasks are extended to provide additional functionality for negative timing constraints and an alternate method of conditioning, as in Verilog-XL.
$recovery(reference event, data_event, removal_limit, recovery_limit, [notifier], [tstamp_cond], [tcheck_cond], [delayed_reference], [delayed_data])
The $recovery system task normally takes a recovery_limit as the third argument and an optional notifier as the fourth argument. By specifying a limit for both the third and fourth arguments, the $recovery timing check is transformed into a combination removal and recovery timing check similar to the $recrem timing check. The only difference is that the removal_limit and recovery_limit are swapped.
$setuphold(clk_event, data_event, setup_limit, hold_limit, [notifier], [tstamp_cond], [tcheck_cond], [delayed_clk], [delayed_data])
The tstamp_cond argument conditions the data_event for the setup check and the clk_event for the hold check. This alternate method of conditioning precludes specifying conditions in the clk_event and data_event arguments. The tcheck_cond argument conditions the data_event for the hold check and the clk_event for the setup check. This alternate method of conditioning precludes specifying conditions in the clk_event and data_event arguments. The delayed_clk argument is a net that is continuously assigned the value of the net specified in the clk_event. The delay is non-zero if the setup_limit is negative, zero otherwise. The delayed_data argument is a net that is continuously assigned the value of the net specified in the data_event. The delay is non-zero if the hold_limit is negative, zero otherwise. The delayed_clk and delayed_data arguments are provided to ease the modeling of devices that may have negative timing constraints. The model's logic should reference the delayed_clk and delayed_data nets in place of the normal clk and data nets. This ensures that the correct data is latched in the presence of negative constraints. The simulator automatically calculates the delays for delayed_clk and delayed_data such that the correct data is latched as long as a timing constraint has not been violated. See "Negative timing check limits" (UM-121) for more details.
Unsupported Verilog-XL system tasks

The following system tasks are Verilog-XL system tasks that are not implemented in ModelSim Verilog, but have equivalent simulator commands.
$input("filename")
This system task reads commands from the specified filename. The equivalent simulator command is do <filename>.
$list[(hierarchical_name)]
This system task lists the source code for the specified scope. The equivalent functionality is provided by selecting a module in the structure pane of the Workspace. The corresponding source code is displayed in a Source window.
$reset
This system task resets the simulation back to its time 0 state. The equivalent simulator command is restart.
$restart("filename")
This system task sets the simulation to the state specified by filename, saved in a previous call to $save. The equivalent simulator command is restore <filename>.
$save("filename")
This system task saves the current simulation state to the file specified by filename. The equivalent simulator command is checkpoint <filename>.
$scope(hierarchical_name)
This system task sets the interactive scope to the scope specified by hierarchical_name. The equivalent simulator command is environment <pathname>.
$showscopes
This system task displays a list of scopes defined in the current interactive scope. The equivalent simulator command is show.
$showvars
This system task displays a list of registers and nets defined in the current interactive scope. The equivalent simulator command is show.
Compiler directives
ModelSim Verilog supports all of the compiler directives defined in the IEEE Std 1364, some Verilog-XL compiler directives, and some that are proprietary. The SystemVerilog IEEE Std P1800-2005 version of the define and include compiler directives are not currently supported. Many of the compiler directives (such as `timescale) take effect at the point they are defined in the source code and stay in effect until the directive is redefined or until it is reset to its default by a `resetall directive. The effect of compiler directives spans source files, so the order of source files on the compilation command line could be significant. For example, if you have a file that defines some common macros for the entire design, then you might need to place it first in the list of files to be compiled. The `resetall directive affects only the following directives by resetting them back to their default settings (this information is not provided in the IEEE Std 1364):
`celldefine default_decay_time `default_nettype `delay_mode_distributed `delay_mode_path `delay_mode_unit `delay_mode_zero `protected `timescale ùnconnected_drive ùselib
ModelSim Verilog implicitly defines the following macro:

`define MODEL_TECH
IEEE Std 1364 compiler directives

The following compiler directives are described in detail in the IEEE Std 1364.
`celldefine `default_nettype `define èlse èlsif èndcelldefine èndif ìfdef ifndef ìnclude line `nounconnected_drive `resetall `timescale ùnconnected_drive ùndef
Compiler directives UM-133
Compiler directives specific to ModelSim

The following directives are specific to ModelSim and are not compatible with other simulators (see note below).
protect ... endprotect
This directive pair allows you to encrypt selected regions of your source code. The code in `protect regions has all debug information stripped out. This behaves exactly as if using the -nodebug argument except that it applies to selected regions of code rather than the whole file. This enables usage scenarios such as making module ports, parameters, and specify blocks publicly visible while keeping the implementation private. The `protect directive is ignored by default unless you use the +protect argument to vlog (CR-293). Once compiled, the original source file is copied to a new file in the current work directory. The name of the new file is the same as the original file with a "p" appended to the suffix. For example, "top.v" is copied to "top.vp". This new file can be delivered and used as a replacement for the original source file. A usage scenario might be that a vendor will use the `protect / ènd protect directives on a module or a portion of a module in a file named filename.v. They will compile it with vlog +protect filename.v to produce a new file named filename.vp. You can compile filename.vp just like any other verilog file. The protection is not compatible between tools, so the vendor mush ship you a different filename.vp than they ship to some who uses a different simulator. The +protect argument is not required when compiling .vp files because the `protect directives are converted to `protected directives which are processed even if +protect is omitted. `protect and `protected directives cannot be nested. If any ìnclude directives occur within a protected region, the compiler generates a copy of the include file with a ".vp" suffix and protects the entire contents of the include file. If errors are detected in a protected region, the error message always reports the first line of the protected block. Though other simulators have a `protect directive, the algorithm ModelSim uses to encrypt source files is different. Hence, even though an uncompiled source file with `protect is compatible with another simulator, once the source is compiled in ModelSim, you could not simulate it elsewhere.
Verilog-XL compatible compiler directives

The following compiler directives are provided for compatibility with Verilog-XL.
default_decay_time <time>
This directive specifies the default decay time to be used in trireg net declarations that do not explicitly declare a decay time. The decay time can be expressed as a real or integer number, or as "infinite" to specify that the charge never decays.
`delay_mode_distributed 123)
This directive disables path delays in favor of distributed delays. See "Delay modes" (UMfor details. This directive sets distributed delays to zero in favor of path delays. See "Delay modes" (UM-123) for details.
`delay_mode_path
`delay_mode_unit
This directive sets path delays to zero and non-zero distributed delays to one time unit. See "Delay modes" (UM-123) for details.
`delay_mode_zero
This directive sets path delays and distributed delays to zero. See "Delay modes" (UM123) for details.
ùselib
This directive is an alternative to the -v, -y, and +libext source library compiler arguments. See "Verilog-XL ùselib compiler directive" (UM-110) for details. The following Verilog-XL compiler directives are silently ignored by ModelSim Verilog. Many of these directives are irrelevant to ModelSim Verilog, but may appear in code being ported from Verilog-XL.
àccelerate àutoexpand_vectornets `disable_portfaults ènable_portfaults èxpand_vectornets `noaccelerate `noexpand_vectornets `noremove_gatenames `noremove_netnames `nosuppress_faults `remove_gatenames `remove_netnames `suppress_faults
The following Verilog-XL compiler directives produce warning messages in ModelSim Verilog. These are not implemented in ModelSim Verilog, and any code containing these directives may behave differently in ModelSim Verilog than in Verilog-XL.
`default_trireg_strength `signed ùnsigned
Verilog PLI/VPI and SystemVerilog DPI UM-135
Verilog PLI/VPI and SystemVerilog DPI

ModelSim supports the use of the Verilog PLI (Programming Language Interface) and VPI (Verilog Procedural Interface) and the SystemVerilog DPI (Direct Programming Interface). These three interfaces provide a mechanism for defining tasks and functions that communicate with the simulator through a C procedural interface. For more information on the ModelSim implementation, see Appendix C - Verilog PLI / VPI / DPI.
UM-137
6 - SystemC simulation
Chapter contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-139 UM-140 UM-140 UM-141 UM-142 UM-143 UM-143 UM-143 UM-146 UM-146 UM-147 UM-147 UM-149 UM-143 UM-151 UM-152 UM-152 UM-152 UM-153 UM-154 UM-155 UM-155 UM-157 UM-155 UM-159 UM-159 UM-161 UM-162 UM-162 UM-163 UM-163 UM-163 UM-163 UM-164 UM-166 UM-166 UM-166 Supported platforms and compiler versions . . . Building gcc with custom configuration options . HP Limitations for SystemC . . . . . . Usage flow for SystemC-only designs . . . . .
Compiling SystemC files . . . . . . . . . Creating a design library . . . . . . . . Modifying SystemC source code . . . . . . Invoking the SystemC compiler . . . . . . Compiling optimized and/or debug code . . . . Specifying an alternate g++ installation . . . . Maintaining portability between OSCI and ModelSim Using sccom vs. raw C++ compiler . . . . . Linking the compiled source . sccom -link . . . . Loading the design . . . . . . . . . . . . . . . . . . . . . . .
Simulating SystemC designs . . . . . . . . Running simulation . . . . . . . . . SystemC time unit and simulator resolution . . . Initialization and cleanup of SystemC state-based code Debugging the design . . . Viewable SystemC objects Source-level debug . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SystemC object and type display in ModelSim Support for Globals and Statics . . . Support for aggregates . . . . . Viewing FIFOs . . . . . . .
Differences between ModelSim and the OSCI simulator . Fixed-point types . . . . . . . . . . OSCI 2.1 feature implementation details . . Support for OSCI TLM library . . . Phase callback . . . . . . . . Accessing command-line arguments . . Construction parameters for SystemC types . . . . . . . . . . . . . . . . . .
Troubleshooting SystemC errors . . . . . . Unexplained behaviors during loading or runtime Errors during loading . . . . . . . .
UM-138 6 - SystemC simulation
Note: The functionality described in this chapter requires a systemc license feature in your ModelSim license file. Please contact your Mentor Graphics sales representative if you currently do not have such a feature.
Introduction UM-139
Introduction
This chapter describes how to compile and simulate SystemC designs with ModelSim. ModelSim implements the SystemC language based on the Open SystemC Initiative (OSCI) SystemC 2.1 reference simulator. It is recommended that you obtain the OSCI functional specification, or the latest version of the SystemC Language Reference Manual as a reference manual. Visit http://www.systemc.org for details. In addition to the functionality described in the OSCI specification, ModelSim for SystemC includes the following features: Single common Graphic Interface for SystemC and HDL languages. Extensive support for mixing SystemC, VHDL, and Verilog in the same design (SDF annotation for HDL only). For detailed information on mixing SystemC with HDL see Chapter 7 - Mixed-language simulation.
Supported platforms and compiler versions

SystemC runs on a subset of ModelSim supported platforms. The table below shows the currently supported platforms and compiler versions: Platform HP-UX 11.0 or later RedHat Linux 7.2 and 7.3 RedHat Linux Enterprise version 2.1 Solaris 2.6 and 7 Solaris 8, 9, and 10 Windows NT and other NT-based platforms (win2K, XP, etc.) Supported compiler versions aCC 3.45 with associated patches gcc 3.2.3 gcc 3.2 gcc 3.3 Minimalist GNU for Windows (MinGW) gcc 3.2.3
Important: ModelSim SystemC has been tested with the gcc versions available from the install tree. Customized versions of gcc may cause problems. We strongly encourage you to use the supplied gcc versions.
Building gcc with custom configuration options

We only test with our default options. If you use advanced gcc configuration options, we cannot guarantee that ModelSim will work with those options. To use a custom gcc build, set the CppPath variable in the modelsim.ini file. This variable specifies the pathname to the compiler binary you intend to use. When using a custom gcc, ModelSim requires that the custom gcc be built with several specific configuration options. These vary on a per-platform basis as shown in the following table: Platform Linux Solaris HP-UX Win32 (MinGW) Mandatory configuration options none --with-gnu-ld --with-ld=/path/to/binutils-2.14/bin/ld --with-gnu-as --with-as=/path/to/binutils-2.14/bin/as N/A --with-gnu-ld --with-gnu-as Do NOT build with the --enable-sjlj-exceptions option, as it can cause problems with catching exceptions thrown from SC_THREAD and SC_CTHREAD ld.exe and as.exe should be installed into the <install_dir>/bin before building gcc. ld and as are available in the binutils package. ModelSim uses binutils 2.13.90-20021006-2.
Supported platforms and compiler versions UM-141
If you don't have a GNU binutils2.14 assembler and linker handy, you can use the as and ld programs distributed with ModelSim. They are located inside the built-in gcc in directory <install_dir>/modeltech/gcc-3.2-<mtiplatform>/lib/gcc-lib/<gnuplatform>/3.2. By default ModelSim also uses the following options when configuring built-in gcc. --disable-nls --enable-languages=c,c++ These are not mandatory, but they do reduce the size of the gcc installation.
HP Limitations for SystemC

HP is supported for SystemC with the following limitations: variables are not supported aggregates are not supported objects must be explicitly named, using the same name as their object, in order to debug SystemC simulation objects such as modules, primitive channels, and ports can be explicitly named by passing a name to the constructors of said objects. If an object is not constructed with an explicit name, then the OSCI reference simulator generates an internal name for it, using names such as "signal_0", "signal_1", etc.
Required Patch for HP-UX 11.11

If you are running on HP-UX 11.11, you must have the following patch installed: B.11.11.0306 Gold Base Patches for HP-UX 11i, June 2003.
Usage flow for SystemC-only designs

ModelSim allows users to simulate SystemC, either alone or in combination with other VHDL/Verilog modules. The following is an overview of the usage flow for strictly SystemC designs. More detailed instructions are presented in the sections that follow. 1 Create and map the working design library with the vlib and vmap statements, as appropriate to your needs. 2 Modify the SystemC source code, including the following highlights: Replace sc_main() with an SC_MODULE, and potentially add a process to contain any testbench code Replace sc_start() by using the run (CR-197) command in the GUI Remove calls to sc_initialize() Export the top level SystemC design unit(s) using the SC_MODULE_EXPORT macro See "Modifying SystemC source code" (UM-143) for a complete list of all modifications. 3 Analyze the SystemC source using sccom (CR-199). sccom invokes the native C++ compiler to create the C++ object files in the design library. See "Using sccom vs. raw C++ compiler" (UM-149) for information on when you are required to use sccom vs. another C++ compiler. 4 Perform a final link of the C++ source using sccom -link (UM-151). This process creates a shared object file in the current work library which will be loaded by vsim at runtime. sccom -link must be re-run before simulation if any new sccom compiles were performed. 5 Simulate the design using the standard vsim command. 6 Simulate the design using the run command, entered at the vsim command prompt. 7 Debug the design using ModelSim GUI features, including the Source and Wave windows.
Compiling SystemC files UM-143
Compiling SystemC files

To compile SystemC designs, you must create a design library modify the SystemC source code run the sccom (CR-199) SystemC compiler run the sccom (CR-199) SystemC linker (sccom -link)
Creating a design library

Before you can compile your design, you must create a library in which to store the compilation results. Use vlib (CR-292) to create a new library. For example:
vlib work
This creates a library named work. By default, compilation results are stored in the work library. The work library is actually a subdirectory named work. This subdirectory contains a special file named _info. Do not create libraries using UNIX commands always use the vlib command (CR-292). See "Design libraries" (UM-53) for additional information on working with libraries.
Modifying SystemC source code

Several modifications must be applied to your original SystemC source code. To see example code containing the modifications listed below, see "Code modification examples" (UM-144).
Converting sc_main() to a module

In order for ModelSim to run the SystemC/C++ source code, the control function of sc_main() must be replaced by a constructor, SC_CTOR(), placed within a module at the top level of the design (see mytop in "Example 1" (UM-144)). In addition: any testbench code inside sc_main() should be moved to a process, normally an SC_THREAD process. all C++ variables in sc_main(), including SystemC primitive channels, ports, and modules, must be defined as members of sc_module. Therefore, initialization must take place in the SC_CTOR. For example, all sc_clock() and sc_signal() initializations must be moved into the constructor.
Replacing the sc_start() function with the run command and options
ModelSim uses the run command and its options in place of the sc_start() function. If sc_main() has multiple sc_start() calls mixed in with the testbench code, then use an SC_THREAD() with wait statements to emulate the same behavior. An example of this is shown below.
Removing calls to sc_initialize()

vsim calls sc_initialize() by default at the end of elaboration, so calls to sc_initialize() are unnecessary.
Exporting all top level SystemC modules

For SystemC designs, you must export all top level modules in your design to ModelSim. You do this with the SC_MODULE_EXPORT(<sc_module_name>) macro. SystemC templates are not supported as top level or boundary modules. See "Templatized SystemC modules" (UM-150). The sc_module_name is the name of the top level module to be simulated in ModelSim. You must specify this macro in a C++ source (.cpp) file. If the macro is contained in a header file instead of a C++ source file, an error may result.
For HP-UX: Explicitly naming signals, ports, and modules

Important: Verify that SystemC signals, ports, and modules are explicitly named to avoid port binding and debugging errors.
Code modification examples

Example 1
The following is a simple example of how to convert sc_main to a module and elaborate it with vsim. Original OSCI code #1 (partial)
int sc_main(int argc, char* argv[]) { sc_signal<bool> mysig; mymod mod("mod"); mod.outp(mysig); sc_start(100, SC_NS); }
Modified code #1 (partial)

SC_MODULE(mytop) { sc_signal<bool> mysig; mymod mod; SC_CTOR(mytop) : mysig("mysig"), mod("mod") { mod.outp(mysig); } }; SC_MODULE_EXPORT(mytop);
The run command equivalent to the sc_start(100,

run 100 ns
SC_NS)
statement is:
Example 2
This next example is slightly more complex, illustrating the use of sc_main() and signal assignments, and how you would get the same behavior using ModelSim. Original OSCI code #2 (partial)
int sc_main(int, char**) { sc_signal<bool> reset; counter_top top("top"); sc_clock CLK("CLK", 10, SC_NS, 0.5, 0.0, SC_NS, false); top.reset(reset); reset.write(1); sc_start(5, SC_NS); reset.write(0); sc_start(100, SC_NS); reset.write(1); sc_start(5, SC_NS); reset.write(0); sc_start(100, SC_NS); } void new_top::sc_main_body() { reset.write(1); wait(5, SC_NS); reset.write(0); wait(100, SC_NS); reset.write(1); wait(5, SC_NS); reset.write(0); wait(100, SC_NS); } SC_MODULE_EXPORT(new_top); SC_CTOR(new_top) : reset("reset"), top("top") CLK("CLK", 10, SC_NS, 0.5, 0.0, SC_NS, false) { top.reset(reset); SC_THREAD(sc_main_body); } };
Modified ModelSim code #2 (partial)

SC_MODULE(new_top) { sc_signal<bool> reset; counter_top top; sc_clock CLK; void sc_main_body();
Example 3
One last example illustrates the correct way to modify a design using an SCV transaction database. ModelSim requires that the transaction database be created before calling the constructors on the design subelements. The example is as follows: Original OSCI code # 3 (partial)
int sc_main(int argc, char* argv[]) { scv_startup(); scv_tr_text_init(); scv_tr_db db("my_db"); scv_tr_db db::set_default_db(&db); sc_clock clk ("clk",20,0.5,0,true); sc_signal<bool> rw; test t("t"); t.clk(clk);; t.rw(rw); sc_start(100); } }; SC_MODULE_EXPORT(new_top); }
Modified ModelSim code #3 (partial)

SC_MODULE(top) { sc_signal<bool>* rw; test* t; SC_CTOR(top) { scv_startup(); scv_tr_text_init() scv_tr_db* db = new scv_tr_db("my_db"); scv_tr_db::set_default_db(db):; clk = new sc_clock("clk",20,0.5,0,true); rw = new sc_signal<bool> ("rw"); t = new test("t");
Take care to preserve the order of functions called in sc_main() of the original code. Sub-elements cannot be placed in the initializer list, since the constructor body must be executed prior to their construction. Therefore, the sub-elements must be made pointer types, created with "new" in the SC_CTOR() module.
Invoking the SystemC compiler

ModelSim compiles one or more SystemC design units with a single invocation of sccom (CR-199), the SystemC compiler. The design units are compiled in the order that they appear on the command line. For SystemC designs, all design units must be compiled just as they would be for any C++ compilation. An example of an sccom command might be:
sccom -I ../myincludes mytop.cpp mydut.cpp
Compiling optimized and/or debug code

By default, sccom invokes the C++ compiler (g++ or aCC) without any optimizations. If desired, you can enter any g++/aCC optimization arguments at the sccom command line. Also, source level debug of SystemC code is not available by default in ModelSim. To compile your SystemC code for debug, use the g++/aCC -g argument on the sccom command line.
Specifying an alternate g++ installation

We recommend using the version of g++ that is shipped with ModelSim on its various supported platforms. However, if you want to use your own installation, you can do so by setting the CppPath variable in the modelsim.ini file to the g++ executable location. For example, if your g++ executable is installed in /u/abc/gcc-3.2/bin, then you would set the variable as follows:
CppPath /u/abc/gcc-3.2/bin/g++
Maintaining portability between OSCI and ModelSim

If you intend to simulate on both ModelSim and the OSCI reference simulator, you can use the MTI_SYSTEMC macro to execute the ModelSim specific code in your design only when running ModelSim. The MTI_SYSTEMC macro is defined in ModelSims systemc.h header file. When you #include this file in your SystemC code, you gain access to this macro. By including #ifdef/else statements in the code, you can then avoid having two copies of the design. Using the original and modified code shown in the example shown on page 144, you might write the code as follows:
#ifdef MTI_SYSTEMC //If using the ModelSim simulator, sccom compiles this SC_MODULE(mytop) { sc_signal<bool> mysig; mymod mod; SC_CTOR(mytop) : mysig("mysig"), mod("mod") { mod.outp(mysig); } }; SC_MODULE_EXPORT(top); #else //Otherwise, it compiles this int sc_main(int argc, char* argv[]) { sc_signal<bool> mysig; mymod mod("mod"); mod.outp(mysig); sc_start(100, SC_NS); } #endif
Restrictions on compiling with HP aCC

ModelSim uses the aCC -AA option by default when compiling C++ files on HP-UX. It does this so cout will function correctly in the systemc.so file. The -AA option tells aCC to use ANSI-compliant <iostream> rather than cfront-style <iostream.h>. Thus, all C++based objects in a program must be compiled with -AA. This means you must use <iostream> and "using" clauses in your code. Also, you cannot use the -AP option, which is incompatible with -AA.
Switching platforms and compilation

Compiled SystemC libraries are platform dependent. If you move between platforms, you must remove all SystemC files from the working library and then recompile your SystemC source files. To remove SystemC files from the working directory, use the vdel (CR-263) command with the -allsystemc argument. If you attempt to load a design that was compiled on a different platform, an error such as the following occurs:
# vsim work.test_ringbuf # Loading work/systemc.so # ** Error: (vsim-3197) Load of "work/systemc.so" failed: work/systemc.so: ELF file data encoding not little-endian. # ** Error: (vsim-3676) Could not load shared library work/systemc.so for SystemC module 'test_ringbuf'. # Error loading design
You can type verror 3197 at the vsim command prompt and get details about what caused the error and how to fix it.
Using sccom vs. raw C++ compiler

When compiling complex C/C++ testbench environments, it is common to compile code with many separate runs of the compiler. Often users compile code into archives (.a files), and then link the archives at the last minute using the -L and -l link options. When using ModelSim's SystemC, you may wish to compile a portion of your C design using raw g++ or aCC instead of sccom. Perhaps you have some legacy code or some nonSystemC utility code that you want to avoid compiling with sccom. You can do this, however, some caveats and rules apply.
Rules for sccom use

The rules governing when and how you must use sccom are as follows: 1 You must compile all code that references SystemC types or objects using sccom (CR199). 2 When using sccom, you should not use the -I compiler option to point the compiler at any search directories containing OSCI or any other vendor supplied SystemC header files. sccom does this for you accurately and automatically. 3 If you do use the raw C++ compiler to compile C/C++ functionality into archives or shared objects, you must then link your design using the -L and -l options with the sccom -link command. These options effectively pull the non-SystemC C/C++ code into a simulation image that is used at runtime. Failure to follow the above rules can result in link-time or elaboration-time errors due to mismatches between the OSCI or any other vendor supplied SystemC header files and the ModelSim SystemC header files.
Rules for using raw g++ to compile non-SystemC C/C++ code

If you use raw g++ to compile your non-systemC C/C++ code, the following rules apply: 1 The -fPIC option to g++ should be used during compilation with sccom. 2 For C++ code, you must use the built-in g++ delivered with ModelSim, or (if using a custom g++) use the one you built and specified with the CppPath .ini variable. Otherwise binary incompatibilities may arise between code compiled by sccom and code compiled by raw g++.
Rules for using raw HP aCC to compile non-SystemC C/C++ code

If you use HPs aCC compiler to compile your non-systemC C/C++ code, the following rules apply: 1 For C++ code, you should use the +Z and -AA options during compilation 2 You must use HP aCC version 3.45 or higher.
Issues with C++ templates

Templatized SystemC modules
Templatized SystemC modules are not supported for use at: the top level of the design the boundary between SystemC and higher level HDL modules (i.e. the top level of the SystemC branch) To convert a top level templatized SystemC module, you can either specialize the module to remove the template, or you can create a wrapper module that you can use as the top module. For example, lets say you have a templatized SystemC module as shown below:
template <class T> class top : public sc_module { sc_signal<T> sig1; . . . };
You can specialize the module by setting T = int, thereby removing the template, as follows:
class top : public sc_module { sc_signal<int> sig 1; . . . };
Or, alternatively, you could write a wrapper to be used over the template module:
class modelsim_top : public sc_module { top<int> actual_top; . . . }; SC_MODULE_EXPORT(modelsim_top);
Organizing templatized code

Suppose you have a class template, and it contains a certain number of member functions. All those member functions must be visible to the compiler when it compiles any instance of the class. For class templates, the C++ compiler generates code for each unique instance of the class template. Unless it can see the full implementation of the class template, it cannot generate code for it thus leaving the invisible parts as undefined. Since it is legal to have undefined symbols in a .so, sccom -link will not produce any errors or warnings. To make functions visible to the compiler, you must move them to the .h file.
Linking the compiled source UM-151
Linking the compiled source

Once the design has been compiled, it must be linked using the sccom (CR-199) command with the -link argument.
sccom -link
The sccom -link command collects the object files created in the different design libraries, and uses them to build a shared library (.so) in the current work library or the library specified by the -work option. If you have changed your SystemC source code and recompiled it using sccom, then you must relink the design by running sccom -link before invoking vsim. Otherwise, your changes to the code are not recognized by the simulator. Remember that any dependent .a or .o files should be listed on the sccom -link command line before the .a or .o on which it depends. For more details on dependencies and other syntax issues, see sccom (CR-199).
Simulating SystemC designs

After compiling the SystemC source code, you can simulate your design with vsim (CR305).
Loading the design

For SystemC, invoke vsim (CR-305) with the top-level module of the design. This example invokes vsim (CR-305) on a design named top:
vsim top
When the GUI comes up, you can expand the hierarchy of the design to view the SystemC modules. SystemC objects are denoted by green icons (see "Design object icons and their meaning" (GR-14) for more information).
To simulate from a command shell, without the GUI, invoke vsim with the -c option:
vsim -c <top_level_module>
Running simulation
Run the simulation using the run (CR-197) command or select one of the Simulate > Run options from the menu bar.
Simulating SystemC designs UM-153
SystemC time unit and simulator resolution

This section applies to SystemC only simulations. For simulations of mixed-language designs, the rules for how ModelSim interprets the resolution vary. See "Simulator resolution limit" (UM-174) for details on mixed-language simulations. Two related yet distinct concepts are involved with determining the simulation resolution: the SystemC time unit and the simulator resolution. The following table describes the concepts, lists the default values, and defines the methods for setting/overriding the values. Description SystemC time unit The unit of time used in your SystemC source code. You need to set this in cases where your SystemC default time unit is at odds with any other, non-SystemC segments of your design. Simulator resolution The smallest unit of time measured by the simulator. If your delays get truncated, set the resolution smaller; this value must be less than or equal to the UserTimeUnit (UM-448) Resolution
(UM-447)
Set by default as .ini file ScTimeUnit

(UM-447)
Default value 1ns
Override default by ScTimeUnit (UM-447) .ini file variable or sc_set_default_time_unit() function
1ns
-t argument to vsim (CR-305) (This overrides all other resolution settings.) or sc_set_time_resolution() function or GUI: Simulate > Start Simulation > Resolution
Available settings for both time unit and resolution are: 1x, 10x, or 100x of fs, ps, ns, us, ms, or sec. You can view the current simulator resolution by invoking the report command (CR-192) with the simulator state option.
Choosing your simulator resolution

You should choose the coarsest simulator resolution limit possible that does not result in undesired rounding of your delays. However, the time precision should also not be set unnecessarily small, because in some cases performance will be degraded. When deciding what to set the simulators resolution to, you must keep in mind the relationship between the simulators resolution and the SystemC time units specified in the source code. For example, with a time unit usage of:
sc_wait(10, SC_PS);
a simulator resolution of 10ps would be fine. No rounding off of the ones digits in the time units would occur. However, a specification of:
sc_wait(9, SC_PS);
would require you to set the resolution limit to 1ps in order to avoid inaccuracies caused by rounding.
Initialization and cleanup of SystemC state-based code

State-based code should not be used in Constructors and Destructors. Constructors and Destructors should be reserved for creating and destroying SystemC design objects, such as sc_modules or sc_signals. State-based code should also not be used in the elaboration phase callbacks before_end_of_elaboration() and end_of_elaboration(). The following virtual functions should be used to initialize and clean up state-based code, such as logfiles or the VCD trace functionality of SystemC. They are virtual methods of the following classes: sc_port_base, sc_module, sc_channel, and sc_prim_channel. You can think of them as phase callback routines in the SystemC language: before_end_of_elaboration () Called after all constructors are called, but before port binding. end_of_elaboration () Called at the end of elaboration after port binding. This function is available in the SystemC 2.1 reference simulator. start_of_simulation () Called before simulation starts. Simulation-specific initialization code can be placed in this function. end_of_simulation () Called before ending the current simulation session. The call sequence for these functions with respect to the SystemC object construction and destruction is as follows: 1 Constructors 2 before_end_of_elaboration () 3 end_of_elaboration () 4 start_of_simulation () 5 end_of_simulation () 6 Destructors
Usage of callbacks
The start_of_simulation() callback is used to initialize any state-based code. The corresponding cleanup code should be placed in the end_of_simulation() callback. These callbacks are only called during simulation by vsim and thus, are safe.
Debugging the design UM-155
Debugging the design

You can debug SystemC designs using all of ModelSims debugging features, with the exception of the Dataflow window.
Viewable SystemC objects

Objects which may be viewed in SystemC for debugging purposes are as shown in the following table. Channels sc_clock (a hierarchical channel) sc_event sc_export sc_mutex sc_fifo<type> sc_signal<type> sc_signal_rv<width> sc_signal_resolved tlm_fifo<type> Ports sc_in<type> sc_out<type> sc_inout<type> sc_in_rv<width> sc_out_rv<width> sc_inout_rv<width> sc_in_resolved sc_out_resolved sc_inout_resolved sc_in_clk sc_out_clk sc_inout_clk sc_fifo_in sc_fifo_out Variables Module member variables of all C++ and SystemC built-in types (listed in the Types list below) are supported. Aggregates Aggregates of SystemC signals or ports. Only three types of aggregates are supported for debug: struct class array
Types (<type>) of the objects which may be viewed for debugging are the following: Types bool, sc_bit sc_logic sc_bv<width> sc_lv<width> sc_int<width> sc_uint<width> sc_fix sc_fix_fast sc_fixed<W,I,Q,O,N> sc_fixed_fast<W,I,Q,O,N> sc_ufix sc_ufix_fast sc_ufixed sc_ufixed_fast sc_signed sc_unsigned char, unsigned char int, unsigned int short, unsigned short long, unsigned long sc_bigint<width> sc_biguint<width> sc_ufixed<W,I,Q,O,N> short, unsigned short long long, unsigned long long float double enum pointer class struct union
Waveform compare
Waveform compare supports the viewing of SystemC signals and variables. You can compare SystemC objects to SystemC, Verilog or VHDL objects. For pure SystemC compares, you can compare any two signals that match type and size exactly; for C/C++ types and some SystemC types, sign is ignored for compares. Thus, you can compare char to unsigned char or sc_signed to sc_unsigned. All SystemC fixed-point types may be mixed as long as the total number of bits and the number of integer bits match.
Debugging the design UM-157
Mixed-language compares are supported as listed in the following table: C/C++ types bool, char, unsigned char short, unsigned short int, unsigned int long, unsigned long sc_bit, sc_bv, sc_logic, sc_lv sc_int, sc_uint sc_bigint, sc_biguint sc_signed, sc_unsigned net, reg bit, bit_vector, boolean, std_logic, std_logic_vector
SystemC types
Verilog types VHDL types
The number of elements must match for vectors; specific indexes are ignored.
Source-level debug
In order to debug your SystemC source code, you must compile the design for debug using the -g C++ compiler option. You can add this option directly to the sccom (CR-199) command line on a per run basis, with a command such as:
sccom mytop -g
Or, if you plan to use it every time you run the compiler, you can specify it in the modelsim.ini file with the SccomCppOptions variable. See "[sccom] SystemC compiler control variables" (UM-442) for more information. The source code debugger, C Debug (UM-319), is automatically invoked when the design is compiled for debug in this way. You can set breakpoints in a Source window, and single-step through your SystemC/C++ source code. .
The gdb debugger has a known bug that makes it impossible to set breakpoints reliably in constructors or destructors. Try to avoid setting breakpoints in constructors of SystemC objects; it may crash the debugger. You can view and expand SystemC objects in the Objects pane and processes in the Active Processes pane.
SystemC object and type display in ModelSim UM-159
SystemC object and type display in ModelSim

This section contains information on how ModelSim displays certain objects and types, as they may differ from other simulators.
Support for Globals and Statics

Globals and statics are supported for ModelSim debugging purposes, however some additional naming conventions must be followed to make them viewable.
Naming requirement
In order to make a global viewable for debugging purposes, the name given must match the declared signal name. An example:
sc_signal<bool> clock("clock");
For statics to be viewable, the name given must be fully qualified, with the module name and declared name, as follows:
<module_name>::<declared_name>
For example, the static data member "count" is viewable in the following code excerpt:
SC_MODULE(top) { static sc_signal<floag> count; //static data member .... } sc_signal<float> top::count("top::count"); //static named in quotes
Viewing global and static signals

ModelSim translates C++ scopes into a hierarchical arrangement. Since globals and statics exist at a level above ModelSims scope, ModelSim must add a top level, sc_root, to all global and static signals. Thus, to view these static or global signals in ModelSim, you need to add sc_root to the hierarchical name for the signal. In the case of the above examples, the debugging statements for examining "top/count" (a static) and "clock" (a global) would be:
VSIM> examine /sc_root/top/count VSIM> examine /sc_root/clock
Support for aggregates

ModelSim supports aggregates of SystemC signals or ports. Three types of aggregates are supported: structures, classes, and arrays. Unions are not supported for debug. An aggregate of signals or ports will be shown as a signal of aggregate type. For example, an aggregate such as:
sc_signal <sc_logic> a[3];
is equivalent to:
sc_signal <sc_lv<3>> a;
for debug purposes. ModelSim shows one signal - object "a" - in both cases.
The following aggregate

sc_signal <float> fbus [6];
when viewed in the Wave window, would appear as follows:
SystemC object and type display in ModelSim UM-161
Viewing FIFOs
In ModelSim, the values contained in an sc_fifo appear in a definite order. The top-most or left-most value is always the next to be read from the FIFO. Elements of the FIFO that are not in use are not displayed. Example of a signal where the FIFO has five elements:
# examine f_char # {} VSIM 4> # run 10 VSIM 6> # examine f_char # A VSIM 8> # run 10 VSIM 10> # examine f_char # {A B} VSIM 12> # run 10 VSIM 14> # examine f_char # {A B C} VSIM 16> # run 10 VSIM 18> # examine f_char # {A B C D} VSIM 20> # run 10 VSIM 22> # examine f_char # {A B C D E} VSIM 24> # run 10 VSIM 26> # examine f_char # {B C D E} VSIM 28> # run 10 VSIM 30> # examine f_char # {C D E} VSIM 32> # run 10 VSIM 34> # examine f_char # {D E}
Differences between ModelSim and the OSCI simulator

ModelSim is based upon the 2.1 reference simulator provided by OSCI. However, there are some minor but key differences to understand: vsim calls sc_initialize() by default at the end of elaboration. The user has to explicitly call sc_initialize() in the reference simulator. You should remove calls to sc_initialize() from your code. The default time resolution of the reference simulator is 1ps. For vsim it is 1ns. The user can set the time resolution by using the vsim command with the -t option or by modifying the value of the Resolution (UM-447) variable in the modelsim.ini file. All SystemC processes without a dont_initialize() modifier are executed once at the end of elaboration. This can cause print messages to appear from user models before the first VSIM> prompt occurs. This behavior is normal and necessary in order to achieve compliance with both the SystemC and HDL LRMs. The run command in ModelSim is equivalent to sc_start(). In the reference simulator, sc_start() runs the simulation for the duration of time specified by its argument. In ModelSim the run command (CR-197) runs the simulation for the amount of time specified by its argument. The sc_cycle(), sc_start(), and sc_main() functions are not supported in ModelSim.
Fixed-point types
Contrary to OSCI, ModelSim compiles the SystemC kernel with support for fixed-point types. If you want to compile your own SystemC code to enable that support, youll need to define the compile time macro SC_INCLUDE_FX. You can do this in one of two ways: enter the g++/aCC argument -DSC_INCLUDE_FX on the sccom (CR-199) command line, such as:
sccom -DSC_INCLUDE_FX top.cpp
add a define statement to the C++ source code before the inclusion of the systemc.h, as shown below:
#define SC_INCLUDE_FX #include "systemc.h"
OSCI 2.1 feature implementation details UM-163
OSCI 2.1 feature implementation details

Support for OSCI TLM library
ModelSim includes the header files and examples from the OSCI SystemC TLM(Transaction Level Modeling) Library Standar version 1.0. The TLM library can be used with simulation, and requires no extra switches or files. TLM objects are not debuggable, with the exception of tlm_fifo. Examples and documentation are located in install_dir/examples/systemc/tlm. The TLM header files (tlm_*.h) are located in include/systemc.
Phase callback
The following functions are supported for phase callbacks: before_end_of_elaboration() start_of_simulation() end_of_simulation() For more information regarding the use of these functions, see "Initialization and cleanup of SystemC state-based code" (UM-154).
Accessing command-line arguments

The following global functions allow you to gain access to command-line arguments: sc_argc() Returns the number of arguments specified on the vsim (CR-305) command line with the -sc_arg argument. This function can be invoked from anywhere within SystemC code. sc_argv() Returns the arguments specified on the vsim (CR-305) command line with the -sc_arg argument. This function can be invoked from anywhere within SystemC code. Example: When vsim is invoked with the following command line:
vsim -sc_arg "-a" -c -sc_arg "-b -c" -t ns -sc_arg -d
sc_argc() and sc_argv() will behave as follows:

int argc; const char * const * argv; argc = sc_argc(); argv = sc_argv();
The number of arguments (argc) is now 4. argv[0] is "vsim" argv[1] is "-a" argv[2] is "-b -c" argv[3] is "-d"
Construction parameters for SystemC types

The material contained in this section applies only to SystemC signals, ports , variables or fifos using one of the following fixed-point types: sc_signed, sc_unsigned, sc_fix, sc_fix_fast, sc_ufix, sc_ufix_fast. These are the only SystemC types that have construction time parameters. If you require values other than the default parameters, you need to read this section. The default size for these types is 32. If you are using one of these types in a SystemC signal, port, fifo or an aggregate of one of these (e.g. array of sc_signal), you can not pass the size parameters to the type. This is a limitation imposed by the C++ language. Instead, SystemC provides a global default size (32) that you can control. For sc_signed and sc_unsigned, you need to use the two objects, sc_length_param and sc_length_context, and you need to use them in an unusual way. If you just want the default vector length, simply do this:
SC_MODULE(dut) { sc_signal<sc_signed> s1; sc_signal<sc_signed> s2; SC_CTOR(dut) : s1("s1"), s2("s2") { } }
For a single setting like using five-bit vectors, your module and its constructor would be something like:
SC_MODULE(dut) { sc_length_param l; sc_length_context c; sc_signal<sc_signed> s1; sc_signal<sc_signed> s2; SC_CTOR(dut) : l(5), c(l), s1("s1"), s2("s2") { } }
Notice that the constructor initialization list sets up the length parameter first, assigns the length parameter to the context object, and then constructs the two signals. You DO pass the name to the signal constructor, but the name is passed to the signal object, not to the underlying type. There is no way to reach the underlying type directly. Instead, the default constructors for sc_signed and sc_unsigned reach out to the global area and get the currently defined length parameter, the one you just set. If you need to have signals or ports with different vector sizes, you need to include a pair of parameter and context objects for each different size:
SC_MODULE(dut) { sc_length_param l1;
OSCI 2.1 feature implementation details UM-165
sc_length_context c1; sc_signal<sc_signed> s1; sc_signal<sc_signed> s2;
sc_length_param l2; sc_length_context c2; sc_signal<sc_signed> u1; sc_signal<sc_signed> u2; SC_CTOR(dut) : l1(5), c1(l1), s1("s1"), s2("s2"), l2(8), c2(l2), u1("u1"), u2("u2") { } }
With simple variables of this type, you reuse the context object. However, you must have the extra parameter and context objects when you are using them in a constructorinitialization list because the compiler does not allow repeated an item in that list. The four fixed-point types that use construction parameters work exactly the same way, except that they use the objects sc_fxnum_context and sc_fxnum_params to do the work. Also, their are more parameters you can set for fixed-point numbers. Assuming we just want to set the length of the number and the number of fractional bits, heres the example above modified for fixed point numbers:
SC_MODULE(dut) { sc_fxnum_params p1; sc_fxnum_contxt c1; sc_signal<sc_fix> s1; sc_signal<sc_fix> s2; sc_fxnum_params p2; sc_fxnum_contxt c2; sc_signal<sc_ufix> u1; sc_signal<sc_ufix> u2; SC_CTOR(dut) : p1(5,0), c1(p1), s1("s1"), s2("s2"), p2(8,5), c2(p2), u1("u1"), u2("u2") { } }
Troubleshooting SystemC errors

In the process of modifying your SystemC design to run on ModelSim, you may encounter several common errors. This section highlights some actions you can take to correct such errors.
Unexplained behaviors during loading or runtime

If your SystemC simulation behaves in otherwise unexplainable ways, you should determine whether you need to adjust the stack space ModelSim allocates for threads in your design. The required size for a stack depends on the depth of functions on that stack and the number of bytes they require for automatic (local) variables. By default the SystemC stack size is 10,000 bytes per thread. You may have one or more threads needing a larger stack size. If so, call the SystemC function set_stack_size() and adjust the stack to accommodate your needs. Note that you can ask for too much stack space and have unexplained behavior as well.
Errors during loading

When simulating your SystemC design, you might get a "failed to load sc lib" message because of an undefined symbol, looking something like this:
# Loading /home/cmg/newport2_systemc/chip/vhdl/work/systemc.so # ** Error: (vsim-3197) Load of "/home/cmg/newport2_systemc/chip/vhdl/work/ systemc.so" failed: ld.so.1: /home/icds_nut/modelsim/5.8a/sunos5/vsimk: fatal: relocation error: file /home/cmg/newport2_systemc/chip/vhdl/work/systemc.so: symbol _Z28host_respond_to_vhdl_requestPm: referenced symbol not found. # ** Error: (vsim-3676) Could not load shared library /home/cmg/ newport2_systemc/chip/vhdl/work/systemc.so for SystemC module 'host_xtor'.
Source of undefined symbol message

The causes for such an error could be: missing definition missing type using SystemC 2.1 header files from other vendors bad link order specified in sccom -link multiply-defined symbols
Missing definition
If the undefined symbol is a C function in your code or a library you are linking with, be sure that you declared it as an extern "C" function:
extern "C" void myFunc();
Troubleshooting SystemC errors UM-167
This should appear in any header files include in your C++ sources compiled by sccom. It tells the compiler to expect a regular C function; otherwise the compiler decorates the name for C++ and then the symbol can't be found. Also, be sure that you actually linked with an object file that fully defines the symbol. You can use the "nm" utility on Unix platforms to test your SystemC object files and any libraries you link with your SystemC sources. For example, assume you ran the following commands:
sccom test.cpp sccom -link libSupport.a
If there is an unresolved symbol and it is not defined in your sources, it should be correctly defined in any linked libraries:
nm libSupport.a | grep "mySymbol"
Missing type
When you get errors during design elaboration, be sure that all the items in your SystemC design hierarchy, including parent elements, are declared in the declarative region of a module. If not, sccom ignores them. For example, we have a design containing SystemC over VHDL. The following declaration of a child module "test" inside the constructor module of the code is not allowed and will produce an error:
SC_MODULE(Export) { SC_CTOR(Export) { test *testInst; testInst = new test("test"); } };
The error results from the fact that the SystemC parse operation will not see any of the children of "test". Nor will any debug information be attached to it. Thus, the signal has no type information and can not be bound to the VHDL port. The solution is to move the element declaration into the declarative region of the module.
Using SystemC 2.1 header files supplied by other vendors

SystemC 2.1 includes version control for SystemC header files. If you compile your SystemC design using a SystemC 2.1 header file that was distributed by other vendors, and then you run sccom -link to link the design, an error similar to the following may result upon loading the design:
** Error: (vsim-3197) Load of "work/systemc.so" failed: work/systemc.so: undefined symbol: _ZN20sc_api_version_2_1_0C1Ev.
To resolve the error, recompile the design using sccom (CR-199). Make sure any include paths read by sccom do not point to a SystemC 2.1 installation. By default, sccom automatically picks up the ModelSim SystemC header files.
Misplaced "-link" option

The order in which you place the -link option within the sccom -link command is critical. There is a big difference between the following two commands:
sccom -link liblocal.a
and
sccom liblocal.a -link
The first command ensures that your SystemC object files are seen by the linker before the library "liblocal.a" and the second command ensures that "liblocal.a" is seen first. Some linkers can look for undefined symbols in libraries that follow the undefined reference while others can look both ways. For more information on command syntax and dependencies, see sccom (CR-199).
Multiple symbol definitions

The most common type of error found during sccom -link operation is the multiple symbol definition error. The error message looks something like this:
work/sc/gensrc/test_ringbuf.o: In function `test_ringbuf::clock_generator(void)': work/sc/gensrc/test_ringbuf.o(.text+0x4): multiple definition of `test_ringbuf::clock_generator(void)' work/sc/test_ringbuf.o(.text+0x4): first defined here
This error arises when the same global symbol is present in more than one .o file. There are two common causes of this problem: A stale .o file in the working directory with conflicting symbol names. In this first case, just remove the stale files with the following command:
vdel -lib <lib_path> -allsystemc
Incorrect definition of symbols in header files. In the second case, if you have an out-of-line function (one that isnt preceded by the "inline" keyword) or a variable defined (i.e. not just referenced or prototyped, but truly defined) in a .h file, you can't include that .h file in more than one .cpp file. Text in .h files is included into .cpp files by the C++ preprocessor. By the time the compiler sees the text, it's just as if you had typed the entire text from the .h file into the .cpp file. So a .h file included into two .cpp files results in lots of duplicate text being processed by the C++ compiler when it starts up. Include guards are a common technique to avoid duplicate text problems. If an .h file has an out-of-line function defined, and that .h file is included into two .c files, then the out-of-line function symbol will be defined in the two corresponding. o files. This leads to a multiple symbol definition error during sccom -link. To solve this problem, add the "inline" keyword to give the function "internal linkage". This makes the function internal to the .o file, and prevents the function's symbol from colliding with a symbol in another .o file. For free functions or variables, you could modify the function definition by adding the "static" keyword instead of "inline", although "inline" is better for efficiency.
Troubleshooting SystemC errors UM-169
Sometimes compilers do not honor the "inline" keyword. In such cases, you should move your function(s) from a header file into an out-of-line implementation in a .cpp file.
UM-171
7 - Mixed-language simulation
Chapter contents
Introduction . . . . . Basic mixed-language flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-172 UM-172 UM-173 UM-173 UM-174 UM-174 UM-175 UM-176 UM-176 UM-178 UM-179 UM-184 UM-188 UM-188 UM-188 UM-189 UM-191 UM-192 UM-192 UM-193 UM-194 UM-194 UM-194 UM-196 UM-197 UM-199 UM-199 UM-199 UM-199 UM-199 UM-202 UM-202 UM-202 UM-204 UM-204 UM-208 UM-208 UM-208 UM-208 UM-209 UM-209 UM-209
Separate compilers, common design libraries . . . . . Access limitations in mixed-language designs . . . Simulator resolution limit . . . . . . . . . Runtime modeling semantics . . . . . . . . Hierarchical references in mixed HDL/SystemC designs. Mapping data types . . . . . . . . . . Verilog to VHDL mappings . . . . . . . VHDL to Verilog mappings . . . . . . . Verilog and SystemC signal interaction and mappings VHDL and SystemC signal interaction and mappings VHDL: instantiating Verilog . . . Verilog instantiation criteria . . Component declaration . . . vgencomp component declaration Modules with unnamed ports . . Verilog: instantiating VHDL . . VHDL instantiation criteria . SDF annotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SystemC: instantiating Verilog . . . . . . . Verilog instantiation criteria . . . . . . . SystemC foreign module declaration . . . . . Parameter support for SystemC instantiating Verilog Example of parameter use. . . . . . . . Verilog: instantiating SystemC . . . . . . . SystemC instantiation criteria . . . . . . . Exporting SystemC modules . . . . . . . Parameter support for Verilog instantiating SystemC Example of parameter use. . . . . . . . SystemC: instantiating VHDL . . . . . . . VHDL instantiation criteria . . . . . . SystemC foreign module declaration . . . . Generic support for SystemC instantiating VHDL Example of generic use . . . . . . . VHDL: instantiating SystemC . . . . . . . SystemC instantiation criteria . . . . . . Component declaration . . . . . . . vgencomp component declaration . . . . Exporting SystemC modules . . . . . . sccom -link . . . . . . . . . . Generic support for VHDL instantiating SystemC . . . . . . . . . . . .
UM-172 7 - Mixed-language simulation
Introduction
ModelSim single-kernel simulation allows you to simulate designs that are written in VHDL, Verilog, SystemVerilog, and SystemC (not all ModelSim versions support all languages). The boundaries between languages are enforced at the level of a design unit. This means that although a design unit itself must be entirely of one language type, it may instantiate design units from another language. Any instance in the design hierarchy may be a design unit from another language without restriction. Note: Mixed-language simulation is not supported in all versions of ModelSim. Contact your Mentor Graphics sales representative for more information.
Basic mixed-language flow

Simulating mixed-language designs with ModelSim includes these general steps: 1 Analyze HDL source code using vcom (CR-246) or vlog (CR-293) and SystemC C++ source code using sccom (CR-199). Analyze all modules in the design following orderof-analysis rules. For SystemC designs with HDL instances: Create a SystemC foreign module declaration for all Verilog and VHDL instances (see "SystemC foreign module declaration" (UM-194) or (UM-202)). For Verilog/VHDL designs with SystemC instances: Export any SystemC instances that will be directly instantiated by Verilog/VHDL using the SC_MODULE_EXPORT macro. Exported SystemC modules can be instantianted just as you would instantiate any Verilog/VHDL module or design unit. 2 For designs containing SystemC: Link all objects in the design using sccom (CR-199) -link. 3 Simulate the design with the vsim command (CR-305). 4 Run and debug your design.
Separate compilers, common design libraries UM-173
Separate compilers, common design libraries

VHDL source code is compiled by vcom (CR-246) and the resulting compiled design units (entities, architectures, configurations, and packages) are stored in the working library. Likewise, Verilog source code is compiled by vlog (CR-293) and the resulting design units (modules and UDPs) are stored in the working library. SystemC/C++ source code is compiled with the sccom command (CR-199). The resulting object code is compiled into the working library. Design libraries can store any combination of design units from any of the supported languages, provided the design unit names do not overlap (VHDL design unit names are changed to lower case). See "Design libraries" (UM-53) for more information about library management.
Access limitations in mixed-language designs

The Verilog language allows hierarchical access to objects throughout the design. This is not the case with VHDL or SystemC. You cannot directly read or change a VHDL or SystemC object (signal, variable, generic, etc.) with a hierarchical reference within a mixed-language design. Furthermore, you cannot directly access a Verilog object up or down the hierarchy if there is an interceding VHDL or SystemC block. You have two options for accessing VHDL objects or Verilog objects "obstructed" by an interceding block: 1) propagate the value through the ports of all design units in the hierarchy; 2) use the Signal Spy procedures or system tasks (see Chapter 14 - Signal Spy for details). To access obstructed SystemC objects, propagate the value through the ports of all design units in the hierarchy or use the control/observe functions. The following two member functions of sc_signal may be used to control and observe hierarchical signals in a design: control_foreign_signal() observe_foreign_signal() For more information on the use of control and observe, see "Hierarchical references in mixed HDL/SystemC designs" (UM-175).

In a mixed-language design with only one top, the resolution of the top design unit is applied to the whole design. If the root of the mixed design is VHDL, then VHDL simulator resolution rules are used (see "Simulator resolution limit" (UM-75) for VHDL details). If the root of the mixed design is Verilog, Verilog rules are used (see "Simulator resolution limit" (UM-114) for Verilog details). If the root is SystemC, then SystemC rules are used (see "SystemC time unit and simulator resolution" (UM-153) for SystemC details). In the case of a mixed-language design with multiple tops, the following algorithm is used: If VHDL or SystemC modules are present, then the Verilog resolution is ignored. An error is issued if the Verilog resolution is finer than the chosen one. If both VHDL and SystemC are present, then the resolution is chosen based on which design unit is elaborated first. For example:
vsim sc_top vhdl_top -do vsim.do
In this case, the SystemC resolution (default 1 ns) is chosen.

vsim vhdl_top sc_top -do vsim.do
In this case, the VHDL resolution is chosen. All resolutions specified in the source files are ignored if vsim is invoked with the -t option. When set, this overrides all other resolutions.
Runtime modeling semantics

The ModelSim simulator is compliant with all pertinent Language Reference Manuals. To achieve this compliance, the sequence of operations in one simulation iteration (i.e. delta cycle) is as follows: SystemC processes are run Signal updates are made HDL processes are run The above scheduling semantics are required to satisfy both the SystemC and the HDL LRM. Namely, all processes triggered by an event in a SystemC primitive channel shall wake up at the beginning of the following delta. All processes triggered by an event on an HDL signal shall wake up at the end of the current delta. For a signal chain that crosses the language boundary, this means that processes on the SystemC side get woken up one delta later than processes on the HDL side. Consequently, one delta of skew will be introduced between such processes. However, if the processes are communicating with each other, correct system behavior will still result.
Separate compilers, common design libraries UM-175
Hierarchical references in mixed HDL/SystemC designs

A SystemC signal (including sc_signal, sc_buffer, sc_signal_resolved, and sc_signal_rv) can control or observe an HDL signal using two member functions of sc_signal:
bool control_foreign_signal(const char* name); bool observe_foreign_signal(const char* name);
The argument (const char* name) is a full hierarchical path to an HDL signal or port. The return value is "true" if the HDL signal is found and its type is compatible with the SystemC signal type. See tables for Verilog "Data type mapping" (UM-180) and VHDL "Data type mapping" (UM-184) to view a list of types supported at the mixed language boundary. If it is a supported boundary type, it is supported for hierarchical references. If the function is called during elaboration time, when the HDL signal has not yet elaborated, the function always returns "true"; however, an error is issued before simulation starts.
Control
When a SystemC signal calls control_foreign_signal() on an HDL signal, the HDL signal is considered a fanout of the SystemC signal. This means that every value change of the SystemC signal is propagated to the HDL signal. If there is a pre-existing driver on the HDL signal which has been controlled, the value is changed to reflect the SystemC signals value. This value remains in effect until a subsequent driver transaction occurs on the HDL signal, following the semantics of the force -deposit command.
Observe
When a SystemC signal calls observe_foreign_signal() on an HDL signal, the SystemC signal is considered a fanout of the HDL signal. This means that every value change of the HDL signal is propagated to the SystemC signal. If there is a pre-existing driver on the SystemC signal which has been observed, the value is changed to reflect the HDL signals value. This value remains in effect until a subsequent driver transaction occurs on the SystemC signal, following the semantics of the force -deposit command. Once a SystemC signal executes a control or observe on an HDL signal, the effect stays throughout the whole simulation. Any subsequent control/observe on that signal will be an error. Example:
SC_MODULE(test_ringbuf) { sc_signal<bool> observe_sig; sc_signal<sc_lv<4> > control_sig; // HDL module instance ringbuf* ring_INST; SC_CTOR(test_ringbuf) { ring_INST = new ringbuf("ring_INST", "ringbuf"); ..... observe_sig.observe_foreign_signal("/test_ringbuf/ring_INST/ block1_INST/buffers(0)"); control_sig.control_foreign_signal("/test_ringbuf/ring_INST/ block1_INST/sig"); } };
Mapping data types

Cross-language (HDL) instantiation does not require any extra effort on your part. As ModelSim loads a design it detects cross-language instantiations made possible because a design unit's language type can be determined as it is loaded from a library and the necessary adaptations and data type conversions are performed automatically. SystemC and HDL cross-language instantiation requires minor modification of SystemC source code (addition of SC_MODULE_EXPORT, sc_foreign_module, etc.). A VHDL instantiation of Verilog may associate VHDL signals and values with Verilog ports and parameters. Likewise, a Verilog instantiation of VHDL may associate Verilog nets and values with VHDL ports and generics. The same holds true for SystemC and VHDL/Verilog ports. ModelSim automatically maps between the language data types as shown in the sections below.
Verilog to VHDL mappings

Verilog parameters
Verilog type integer real string VHDL type integer real string
The type of a Verilog parameter is determined by its initial value.
Verilog ports
The allowed VHDL types for ports connected to Verilog nets and for signals connected to Verilog ports are: Allowed VHDL types bit bit_vector std_logic std_logic_vector vl_logic vl_logic_vector The vl_logic type is an enumeration that defines the full state set for Verilog nets, including ambiguous strengths. The bit and std_logic types are convenient for most applications, but the vl_logic type is provided in case you need access to the full Verilog state set. For
Mapping data types UM-177
example, you may wish to convert between vl_logic and your own user-defined type. The vl_logic type is defined in the vl_types package in the pre-compiled verilog library. This library is provided in the installation directory along with the other pre-compiled libraries (std and ieee). The source code for the vl_types package can be found in the files installed with ModelSim. (See <install_dir>/modeltech/vhdl_src/verilog/vltypes.vhd.)
Verilog states
Verilog states are mapped to std_logic and bit as follows: Verilog HiZ Sm0 Sm1 SmX Me0 Me1 MeX We0 We1 WeX La0 La1 LaX Pu0 Pu1 PuX St0 St1 StX Su0 Su1 SuX std_logic 'Z' 'L' 'H' 'W' 'L' 'H' 'W' 'L' 'H' 'W' 'L' 'H' 'W' 'L' 'H' 'W' '0' '1' 'X' '0' '1' 'X' bit '0' '0' '1' '0' '0' '1' '0' '0' '1' '0' '0' '1' '0' '0' '1' '0' '0' '1' '0' '0' '1' '0'
For Verilog states with ambiguous strength: bit receives '0' std_logic receives 'X' if either the 0 or 1 strength component is greater than or equal to strong strength std_logic receives 'W' if both the 0 and 1 strength components are less than strong strength
VHDL to Verilog mappings

VHDL generics
VHDL type integer real time physical enumeration string Verilog type integer or real integer or real integer or real integer or real integer or real string literal
When a scalar type receives a real value, the real is converted to an integer by truncating the decimal portion. Type time is treated specially: the Verilog number is converted to a time value according to the timescale directive of the module. Physical and enumeration types receive a value that corresponds to the position number indicated by the Verilog number. In VHDL this is equivalent to T'VAL(P), where T is the type, VAL is the predefined function attribute that returns a value given a position number, and P is the position number. VHDL type bit is mapped to Verilog states as follows: bit '0' '1' Verilog St0 St1
VHDL type std_logic is mapped to Verilog states as follows: std_logic 'U' 'X' '0' Verilog StX StX St0
std_logic '1' 'Z' 'W' 'L' 'H' ''
Verilog St1 HiZ PuX Pu0 Pu1 StX
Verilog and SystemC signal interaction and mappings

SystemC has a more complex signal-level interconnect scheme than Verilog. Design units are interconnected via hierarchical and primitive channels. An sc_signal<> is one type of primitive channel. The following section discusses how various SystemC channel types map to Verilog wires when connected to each other across the language boundary.
Channel and Port type mapping

The following port type mapping table lists all channels. Three types of primitive channels and one hierarchical channel are supported on the language boundary (SystemC modules connected to Verilog modules). Channels sc_signal<T> Ports sc_in<T> sc_out<T> sc_inout<T> sc_in_rv<W> sc_out_rv<W> sc_inout_rv<W> sc_in_resolved sc_out_resolved sc_inout_resolved sc_in_clk sc_out_clk sc_inout_clk N/A sc_fifo_in sc_fifo_out N/A Verilog mapping Depends on type T. See table entitled "Data type mapping" (UM180). wire [W-1:0]
sc_signal_rv<W>
sc_signal_resolved
wire [W-1:0]
sc_clock
wire
sc_mutex sc_fifo sc_semaphore
Not supported on language boundary Not supported on language boundary Not supported on language boundary
Channels sc_buffer user-defined
Ports N/A user-defined
Verilog mapping Not supported on language boundary Not supported on language boundary
Data type mapping

SystemCs sc_signal<> types are mapped to Verilog types as follows: SystemC bool, sc_bit sc_logic sc_bv<W> sc_lv<W> sc_int<W>, sc_uint<W> sc_bigint<W>, sc_biguint<W> sc_fixed<W,I,Q,O,N>, sc_ufixed<W,I,Q,O,N> sc_fixed_fast<W,I,Q,O,N>, sc_ufixed_fast<W,I,Q,O,N>
a 1
Verilog wire wire wire [W-1:0] wire [W-1:0] wire [W-1:0] wire [W-1:0] wire [W-1:0] wire [W-1:0] wire [WL-1:0] wire [WL-1:0] wire[WL-1:0] wire [7:0] wire [15:0] wire [31:0] wire [31:0] wire [63:0] wire [31:0] wire [63:0]
sc_fix, sc_ufix
1sc_fix_fast, 1sc_ufix_fast b 2
sc_signed, sc_unsigned
char, unsigned char short, unsigned short int, unsigned int long, unsigned long long long, unsigned long long float double
SystemC enum pointers class struct union bit_fields
Verilog Not supported on language boundary Not supported on language boundary Not supported on language boundary Not supported on language boundary Not supported on language boundary Not supported on language boundary
a. WL (word length) is the total number of bits used in the type. It is specified during runtime. To make a port of type sc_fix, sc_ufix, sc_fix_fast, or sc_ufix_fast of word length other than the default(32), you must use sc_fxtype_params and sc_fxtype_context to set the word length. For more information, see "Construction parameters for SystemC types" (UM-164). b. To make a port of type sc_signed or sc_unsigned of word length other than the default (32), you must use sc_length_param and sc_length_context to set the word length. For more information, see "Construction parameters for SystemC types" (UM-164).
Port direction
Verilog port directions are mapped to SystemC as follows: Verilog input output inout SystemC sc_in<T>, sc_in_resolved, sc_in_rv<W> sc_out<T>, sc_out_resolved, sc_out_rv<W> sc_inout<T>, sc_inout_resolved, sc_inout_rv<W>
Verilog to SystemC state mappings

Verilog states are mapped to sc_logic, sc_bit, and bool as follows: Verilog HiZ Sm0 Sm1 SmX Me0 Me1 MeX We0 We1 WeX La0 La1 LaX Pu0 Pu1 PuX St0 St1 StX Su0 Su1 sc_logic 'Z' '0' '1' 'X' '0' '1' 'X' '0' '1' 'X' '0' '1' 'X' '0' '1' 'X' '0' '1' 'X' '0' '1' sc_bit '0' '0' '1' '0' '0' '1' '0' '0' '1' '0' '0' '1' '0' '0' '1' '0' '0' '1' '0' '0' '1' bool false false true false false true false false true false false true false false true false false true false false true
Verilog SuX
sc_logic 'X'
sc_bit '0'
bool false
For Verilog states with ambiguous strength: sc_bit receives '1' if the value component is 1, else it receives 0 bool receives true if the value component is 1, else it receives false sc_logic receives 'X' if the value component is X, H, or L sc_logic receives '0' if the value component is 0 sc_logic receives 1 if the value component is 1
SystemC to Verilog state mappings

SystemC type bool is mapped to Verilog states as follows: bool false true Verilog St0 St1
SystemC type sc_bit is mapped to Verilog states as follows: sc_bit '0' '1' Verilog St0 St1
SystemC type sc_logic is mapped to Verilog states as follows: sc_logic '0' '1' 'Z' 'X' Verilog St0 St1 HiZ StX
VHDL and SystemC signal interaction and mappings

SystemC has a more complex signal-level interconnect scheme than VHDL. Design units are interconnected via hierarchical and primitive channels. An sc_signal<> is one type of primitive channel. The following section discusses how various SystemC channel types map to VHDL types when connected to each other across the language boundary.
Port type mapping

The following port type mapping table lists all channels. Three types of primitive channels and one hierarchical channel are supported on the language boundary (SystemC modules connected to VHDL modules). Channels sc_signal<T> Ports sc_in<T> sc_out<T> sc_inout<T> sc_in_rv<W> sc_out_rv<W> sc_inout_rv<W> sc_in_resolved sc_out_resolved sc_inout_resolved sc_in_clk sc_out_clk sc_inout_clk N/A sc_fifo_in sc_fifo_out N/A N/A user-defined VHDL mapping Depends on type T. See table entitled "Data type mapping" (UM-184) below. std_logic_vector(W-1 downto 0)
sc_signal_rv<W>
sc_signal_resolved
std_logic
sc_clock
bit/std_logic/boolean
sc_mutex sc_fifo sc_semaphore sc_buffer user-defined
Not supported on language boundary Not supported on language boundary Not supported on language boundary Not supported on language boundary Not supported on language boundary
Data type mapping

SystemCs sc_signal types are mapped to VHDL types as follows SystemC bool, sc_bit sc_logic sc_bv<W> sc_lv<W> VHDL bit/std_logic/boolean std_logic bit_vector(W-1 downto 0) std_logic_vector(W-1 downto 0)
SystemC sc_bv<32>, sc_lv<32> sc_bv<64>, sc_lv<64> sc_int<W>, sc_uint<W> sc_bigint<W>, sc_biguint<W> sc_fixed<W,I,Q,O,N>, sc_ufixed<W,I,Q,O,N> sc_fixed_fast<W,I,Q,O,N>, sc_ufixed_fast<W,I,Q,O,N>
asc_fix, 1sc_ufix 1 1 b 2
VHDL integer real bit_vector(W-1 downto 0) std_logic_vector(W -1 downto 0) bit_vector(W-1 downto 0) std_logic_vector(W-1 downto 0) bit_vector(W-1 downto 0) std_logic_vector(W-1 downto 0) bit_vector(W-1 downto 0) std_logic_vector(W-1 downto 0) bit_vector(WL-1 downto 0) std_logic_vector(WL- 1 downto 0) bit_vector(WL-1 downto 0) std_logic_vector(WL- 1 downto 0) bit_vector(WL-1 downto 0) std_logic_vector(WL- 1 downto 0) bit_vector(7 downto 0) std_logic_vector(7 downto 0) bit_vector(15 downto 0) std_logic_vector(15 downto 0) bit_vector(31 downto 0) std_logic_vector(7 downto 0) bit_vector(31 downto 0) std_logic_vector(31 downto 0) bit_vector(63 downto 0) std_logic_vector(63 downto 0) bit_vector(31 downto 0) std_logic_vector(31 downto 0) bit_vector(63 downto 0) std_logic_vector(63 downto 0) Not supported on language boundary Not supported on language boundary Not supported on language boundary Not supported on language boundary
sc_fix_fast, sc_ufix_fast sc_signed, sc_unsigned
char, unsigned char short, unsigned short int, unsigned int long, unsigned long long long, unsigned long long float double enum pointers class structure
SystemC union bit_fields
VHDL Not supported on language boundary Not supported on language boundary
a. WL (word length) is the total number of bits used in the type. It is specified during runtime. To make a port of type sc_fix, sc_ufix, sc_fix_fast, or sc_ufix_fast of word length other than the default(32), you must use sc_fxtype_params and sc_fxtype_context to set the word length. For more information, see "Construction parameters for SystemC types" (UM-164). b. To make a port of type sc_signed or sc_unsigned of word length other than the default (32), you must use sc_length_param and sc_length_context to set the word length. For more information, see "Construction parameters for SystemC types" (UM-164).
Port direction mapping

VHDL port directions are mapped to SystemC as follows: VHDL in out inout buffer SystemC sc_in<T>, sc_in_resolved, sc_in_rv<W> sc_out<T>, sc_out_resolved, sc_out_rv<W> sc_inout<T>, sc_inout_resolved, sc_inout_rv<W> sc_out<T>, sc_out_resolved, sc_out_rv<W>
VHDL to SystemC state mapping

VHDL states are mapped to sc_logic, sc_bit, and bool as follows: std_logic 'U' 'X' '0' '1' 'Z' 'W' 'L' 'H' '-' sc_logic 'X' 'X' '0' '1' 'Z' 'X' '0' '1' 'X' sc_bit '0' '0' '0' '1' '0' '0' '0' '1' '0' bool false false false true false false false true false
SystemC to VHDL state mapping

SystemC type bool is mapped to VHDL boolean as follows: bool false true VHDL false true
SystemC type sc_bit is mapped to VHDL bit as follows: sc_bit '0' '1' VHDL '0' '1'
SystemC type sc_logic is mapped to VHDL std_logic states as follows: sc_logic '0' '1' 'Z' 'X' std_logic '0' '1' 'Z' 'X'
VHDL: instantiating Verilog

Once you have generated a component declaration for a Verilog module, you can instantiate the component just like any other VHDL component. You can reference a Verilog module in the entity aspect of a component configuration all you need to do is specify a module name instead of an entity name. You can also specify an optional secondary name for an optimized sub-module. Further, you can reference a Verilog configuration in the configuration aspect of a VHDL component configuration - just specify a Verilog configuration name instead of a VHDL configuration name.
Verilog instantiation criteria

A Verilog design unit may be instantiated within VHDL if it meets the following criteria: The design unit is a module or configuration. UDPs are not allowed. The ports are named ports (see "Modules with unnamed ports" (UM-191) below). The ports are not connected to bidirectional pass switches (it is not possible to handle pass switches in VHDL).
Component declaration
A Verilog module that is compiled into a library can be referenced from a VHDL design as though the module is a VHDL entity. Likewise, a Verilog configuration can be referenced as though it were a VHDL configuration. The interface to the module can be extracted from the library in the form of a component declaration by running vgencomp (CR-266). Given a library and module name, vgencomp (CR-266) writes a component declaration to standard output. The default component port types are: std_logic std_logic_vector Optionally, you can choose: bit and bit_vector vl_logic and vl_logic_vector
VHDL and Verilog identifiers

The VHDL identifiers for the component name, port names, and generic names are the same as the Verilog identifiers for the module name, port names, and parameter names. Except for the cases noted below, ModelSim leaves the Verilog identifier alone when it generates the entity. ModelSim converts Verilog identifiers to VHDL 1076-1993 extended identifiers in three cases: The Verilog identifier is not a valid VHDL 1076-1987 identifier. You compile the Verilog module with the -93 argument. One exception is a valid, lowercase identifier (e.g., topmod). Valid, lowercase identifiers will not be converted even if you compile with -93.
VHDL: instantiating Verilog UM-189
The Verilog identifier is not unique when case is ignored. For example, if you have TopMod and topmod in the same module, ModelSim will convert the former to \TopMod\).
vgencomp component declaration

vgencomp (CR-266) generates a component declaration according to these rules:
Generic clause
A generic clause is generated if the module has parameters. A corresponding generic is defined for each parameter that has an initial value that does not depend on any other parameters. The generic type is determined by the parameter's initial value as follows: Parameter value integer real string literal Generic type integer real string
The default value of the generic is the same as the parameter's initial value. Examples Verilog parameter parameter p1 = 1 - 3; parameter p2 = 3.0; parameter p3 = "Hello"; VHDL generic p1 : integer := -2; p2 : real := 3.000000; p3 : string := "Hello";
Port clause
A port clause is generated if the module has ports. A corresponding VHDL port is defined for each named Verilog port. You can set the VHDL port type to bit, std_logic, or vl_logic. If the Verilog port has a range, then the VHDL port type is bit_vector, std_logic_vector, or vl_logic_vector. If the range does not depend on parameters, then the vector type will be constrained accordingly, otherwise it will be unconstrained. Examples Verilog port input p1; output [7:0] p2; VHDL port p1 : in std_logic; p2 : out std_logic_vector(7 downto 0);
Verilog port output [4:7] p3; inout [W-1:0] p4;
VHDL port p3 : out std_logic_vector(4 to 7); p4 : inout std_logic_vector;
Configuration declarations are allowed to reference Verilog modules in the entity aspects of component configurations. However, the configuration declaration cannot extend into a Verilog instance to configure the instantiations within the Verilog module.
VHDL: instantiating Verilog UM-191
Modules with unnamed ports

Verilog allows modules to have unnamed ports, whereas VHDL requires that all ports have names. If any of the Verilog ports are unnamed, then all are considered to be unnamed, and it is not possible to create a matching VHDL component. In such cases, the module may not be instantiated from VHDL. Unnamed ports occur when the module port list contains bit-selects, part-selects, or concatenations, as in the following example:
module m(a[3:0], b[1], b[0], {c,d}); input [3:0] a; input [1:0] b; input c, d; endmodule
Note that a[3:0] is considered to be unnamed even though it is a full part-select. A common mistake is to include the vector bounds in the port list, which has the undesired side effect of making the ports unnamed (which prevents the user from connecting by name even in an all-Verilog design). Most modules having unnamed ports can be easily rewritten to explicitly name the ports, thus allowing the module to be instantiated from VHDL. Consider the following example:
module m(y[1], y[0], a[1], a[0]); output [1:0] y; input [1:0] a; endmodule
Here is the same module rewritten with explicit port names added:
module m(.y1(y[1]), .y0(y[0]), .a1(a[1]), .a0(a[0])); output [1:0] y; input [1:0] a; endmodule
"Empty" ports
Verilog modules may have "empty" ports, which are also unnamed, but they are treated differently from other unnamed ports. If the only unnamed ports are "empty", then the other ports may still be connected to by name, as in the following example:
module m(a, , b); input a, b; endmodule
Although this module has an empty port between ports "a" and "b", the named ports in the module can still be connected to from VHDL.
Verilog: instantiating VHDL

You can reference a VHDL entity or configuration from Verilog as though the design unit is a module or a configuration of the same name.
VHDL instantiation criteria

A VHDL design unit may be instantiated within Verilog if it meets the following criteria: The design unit is an entity/architecture pair or a configuration. The entity ports are of type bit, bit_vector, std_ulogic, std_ulogic_vector, vl_ulogic, vl_ulogic_vector, or their subtypes. The port clause may have any mix of these types. The generics are of type integer, real, time, physical, enumeration, or string. String is the only composite type allowed.
Entity/architecture names and escaped identifiers

An entity name is not case sensitive in Verilog instantiations. The entity default architecture is selected from the work library unless specified otherwise. Since instantiation bindings are not determined at compile time in Verilog, you must instruct the simulator to search your libraries when loading the design. See "Library usage" (UM-107) for more information. Alternatively, you can employ the escaped identifier to provide an extended form of instantiation:
\mylib.entity(arch) u1 (a, b, c); \mylib.entity u1 (a, b, c); \entity(arch) u1 (a, b, c);
If the escaped identifier takes the form of one of the above and is not the name of a design unit in the work library, then the instantiation is broken down as follows: library = mylib design unit = entity architecture = arch
Named port associations

Port associations may be named or positional. Use the same port names and port positions that appear in the entity. Named port associations are not case sensitive unless a VHDL port name is an extended identifier (1076-1993). If the VHDL port name is an extended identifier, the association is case sensitive and the VHDL identifiers leading and trailing backslashes are removed before comparison.
Generic associations
Generic associations are provided via the module instance parameter value list. List the values in the same order that the generics appear in the entity. Parameter assignment to generics is not case sensitive. The defparam statement is not allowed for setting generic values.
Verilog: instantiating VHDL UM-193
SDF annotation
A mixed VHDL/Verilog design can also be annotated with SDF. See "SDF for mixed VHDL and Verilog designs" (UM-392) for more information.
SystemC: instantiating Verilog

To instantiate Verilog modules into a SystemC design, you must first create a "SystemC foreign module declaration" (UM-194) for each Verilog module. Once you have created the foreign module declaration, you can instantiate the foreign module just like any other SystemC module.
Verilog instantiation criteria

A Verilog design unit may be instantiated within SystemC if it meets the following criteria: The design unit is a module (UDPs and Verilog primitives are not allowed). The ports are named ports (Verilog allows unnamed ports). The Verilog module name must be a valid C++ identifier. The ports are not connected to bidirectional pass switches (it is not possible to handle pass switches in SystemC). A Verilog module that is compiled into a library can be instantiated in a SystemC design as though the module were a SystemC module by passing the Verilog module name to the foreign module constructor. For an illustration of this, see "Example #1" (UM-195).
SystemC and Verilog identifiers

The SystemC identifiers for the module name and port names are the same as the Verilog identifiers for the module name and port names. Verilog identifiers must be valid C++ identifiers. SystemC and Verilog are both case sensitive.
SystemC foreign module declaration

In cases where you want to run a mixed simulation with SystemC and Verilog, you must generate and declare a foreign module that stands in for each Verilog module instantiated under SystemC. The foreign modules can be created in one of two ways: running scgenmod, a utility that automatically generates your foreign module declaration (much like vgencomp generates a component declaration) modifying your SystemC source code manually
Using scgenmod
After you have analyzed the design, you can generate a foreign module declaration with an scgenmod command (CR-203) similar to the following:
scgenmod mod1
where mod1 is a Verilog module. A foreign module declaration for the specified module is written to stdout.
SystemC: instantiating Verilog UM-195
Guidelines for manual creation

Apply the following guidelines to the creation of foreign modules. A foreign module: contains ports corresponding to Verilog ports. These ports must be explicitly named in the foreign modules constructor initializer list. must not contain any internal design elements such as child instances, primitive channels, or processes. must pass a secondary constructor argument denoting the modules HDL name to the sc_foreign_module base class constructor. For Verilog, the HDL name is simply the Verilog module name corresponding to the foreign module, or [<lib>].<module>. parameterized modules are allowed, see "Parameter support for SystemC instantiating Verilog" (UM-196) for details.
Example #1
A sample Verilog module to be instantiated in a SystemC design is:
module vcounter (clock, topcount, count); input clock; input topcount; output count; reg count; ... endmodule
The SystemC foreign module declaration for the above Verilog module is:
class counter : public sc_foreign_module { public: sc_in<bool> clock; sc_in<sc_logic> topcount; sc_out<sc_logic> count; counter(sc_module_name nm) : sc_foreign_module(nm, "lib.vcounter"), clock("clock"), topcount("topcount"), count("count") {} };
The Verilog module is then instantiated in the SystemC source as follows:

counter dut("dut");
where the constructor argument (dut) is the instance name of the Verilog module.
Example #2
Another variation of the SystemC foreign module declaration for the same Verilog module might be:
class counter : public sc_foreign_module { public: ... ... ... counter(sc_module_name nm, char* hdl_name) : sc_foreign_module(nm, hdl_name), clock("clock"), ... ... ... {} };
The instantiation of this module would be:

counter dut("dut", "lib.counter");
Parameter support for SystemC instantiating Verilog

Since the SystemC language has no concept of parameters, parameterized values must be passed from a SystemC parent to a Verilog child through the SystemC foreign module (sc_foreign_module). See "SystemC foreign module declaration" (UM-194) for information regarding the creation of sc_foreign_module.
Generic parameters in sc_foreign_module constructor

To instantiate a Verilog module containing parameterized values into the SystemC design, you must pass two parameters to the sc_foreign_module constructor: the number of parameters (int num_generics), and the parameter list (const char* generic_list). The generic_list is listed as an array of const char*. If you create your foreign module manually (see "Guidelines for manual creation" (UM195)), you must also pass the parameter information to the sc_foreign_module constructor.
If you use scgenmod to create the foreign module declaration, the parameter information is detected in the HDL child and is incorporated automatically.
Example
Following the example shown above (UM-196), lets see the parameter information that would be passed to the SystemC foreign module declaration:
class counter : public sc_foreign_module { public: sc_in<bool> clk; ... counter(sc_midule_name nm, char* hdl_name int num_generics, const char* generic_list) : sc_foreign_module(nm, hdl_name, num_generics, generic_list), {} };
SystemC: instantiating Verilog UM-197
Example of parameter use

Here is a complete example, ring buffer, including all files necessary for simulation.
// ringbuf.h #ifndef INCLUDED_RINGBUF #define INCLUDED_RINGBUF class ringbuf : public sc_foreign_module { public: sc_in<bool> clk; ... counter(sc_midule_name nm, char* hdl_name int num_generics, const char* generic_list) : sc_foreign_module(nm, hdl_name, num_generics, generic_list), {} }; #endif --------------------------------------------------------------------------// test_ringbuf.h #ifndef INCLUDED_TEST_RINGBUF #define INCLUDED_TEST_RINGBUF #include "ringbuf.h" #include "string.h" SC_MODULE(test_ringbuf) { sc_signal<sc_logic> iclock; ... ... // Verilog module instance ringbuf* chip; SC_CTOR(test_ringbuf) : iclock("iclock"), ... ... { const char* generic_list[3]; generic_list[0] = strdup("int_param=4"); generic_list[1] = strdup("real_param=2.6"); generic_list[2] = strdup("str_param=\"Hello\""); // Enclose the string // in double quotes // Cleanup the memory allocated for theh generic list for (int i = 0; i < 3; i++;) free((char*)generic_list[i]); // Create Verilog module instance. chip = new ringbuf("chip", "ringbuf", 3, generic_list);
// Connect ports chip->clock(iclock); ... ... } ~test_ringbuf() { delete chip; } }; #endif ------------------------------------------------------------------------------// test_ringbuf.cpp #include "test_ringbuf.h" SC_MODULE_EXPORT(test_ringbuf); -------------------------------------------------------------------// ringbuf.v `timescale 1ns / 1ns module ringbuf (clock, reset, txda, rxda, txc, outstrobe); // Design parameter parameter parameter Parameters Control Complete Design int_param = 0; real_param = 2.9; str_param = "Error";
// Define the I/O names input clock, txda, reset ; ... initial begin $display("int_param=%0d", int_param); $display("real_param=%g", real_param); $display("str_param=%s", str_param); end endmodule ---------------------------------------------------------------------------
To run the simulation, use the following commands:

vsim1> vsim2> vsim3> vsim4> vsim5> vsim6> vlib work vlog ringbuf.v scgenmod ringbuf > ringbuf.h sccom test_ringbuf.cpp sccom -link vsim -c test_ringbuf
The simulation returns:

# int_param=4 # real_param=2.6 # str_param=Hello
Verilog: instantiating SystemC UM-199
Verilog: instantiating SystemC

You can reference a SystemC module from Verilog as though the design unit is a module of the same name.
SystemC instantiation criteria

A SystemC module can be instantiated in Verilog if it meets the following criteria: SystemC module names are case sensitive. The module name at the SystemC instantiation site must match exactly with the actual SystemC module name. SystemC modules are exported using the SC_MODULE_EXPORT macro. See "Exporting SystemC modules" (UM-199). The module ports are as listed in the table shown in "Channel and Port type mapping" (UM-179). Port data type mapping must match exactly. See the table in "Data type mapping" (UM180). Port associations may be named or positional. Use the same port names and port positions that appear in the SystemC module declaration. Named port associations are case sensitive.
Exporting SystemC modules

To be able to instantiate a SystemC module from Verilog (or use a SystemC module as a top level module), the module must be exported. Assume a SystemC module named transceiver exists, and that it is declared in header file transceiver.h. Then the module is exported by placing the following code in a .cpp file:
#include "transceiver.h" SC_MODULE_EXPORT(transceiver);
Parameter support for Verilog instantiating SystemC

Passing parameters from Verilog to SystemC
To pass actual parameter values, simply use the native Verilog parameter override syntax. Parameters are passed to SystemC via the module instance parameter value list. In addition to int, real, and string, ModelSim supports parameters with a bit range. Named parameter association must be used for all Verilog modules that instantiate SystemC.
Retrieving parameter values

To retrieve parameter override information from Verilog, you can use the following functions:
void sc_get_param(const char* param_name, int& param_value); void sc_get_param(const char* param_name, double& param_value); void sc_get_param(const char* param_name, sc_string& param_value, char format_char = 'a');
The first argument to sc_get_param defines the parameter name, the second defines the parameter value. For retrieving string values, ModelSim also provides a third optional argument, format_char. It is used to specify the format for displaying the retrieved string. The format can be ASCII ("a" or "A"), binary ("b" or "b"), decimal ("d" or "d"), octal ("o" or "O"), or hexadecimal ("h" or "H"). ASCII is the default. Alternatively, you can use the following forms of the above functions in the constructor initializer list:
int sc_get_int_param(const char* param_name); double sc_get_real_param(const char* param_name); sc_string sc_get_string_param(const char* param_name, char format_char = 'a');
Example of parameter use

Here is a complete example, ring buffer, including all files necessary for simulation.
// test_ringbuf.v `timescale 1ns / 1ps module test_ringbuf(); reg clock; ... parameter int_param = 4; parameter real_param = 2.6; parameter str_param = "Hello World"; parameter [7:0] reg_param = 'b001100xz; // Instantiate SystemC module ringbuf #(.int_param(int_param), .real_param(real_param), .str_param(str_param), .reg_param(reg_param)) chip(.clock(clock), ... ... }; endmodule --------------------------------------------------------------------------// ringbuf.h #ifndef INCLUDED_RINGBUF #define INCLUDED_RINGBUF #include SC_MODULE(ringbuf) { public: // Module ports sc_in clock; ... ... SC_CTOR(ringbuf) : clock("clock"), ... ... { cout << "int_param="
Verilog: instantiating SystemC UM-201
<< sc_get_int_param("int_param") << endl; cout << "real_param=" << sc_get_real_param("real_param") << endl; cout << "str_param=" << (const char*)sc_get_string_param("str_param", 'a') << endl; cout << "reg_param=" << (const char*)sc_get_string_param("reg_param", 'b') << endl; } ~ringbuf() {} }; #endif --------------------------------------------------------------------------// ringbuf.cpp #include "ringbuf.h" SC_MODULE_EXPORT(ringbuf);
To run the simulation, you would enter the following commands:

vsim1> vsim1> vsim1> vsim1> vsim1> vlib work sccom ringbuf.cpp vlog test_ringbuf.v sccom -link vsim test_ringbuf
The simulation would return the following:

# # # # int_param=4 real_param=2.6 str_param=Hello World reg_param=001100xz
SystemC: instantiating VHDL

To instantiate VHDL design units into a SystemC design, you must first generate a SystemC foreign module declaration (UM-194) for each VHDL design unit you want to instantiate. Once you have generated the foreign module declaration, you can instantiate the foreign module just like any other SystemC module.
VHDL instantiation criteria

A VHDL design unit may be instantiated from SystemC if it meets the following criteria: The design unit is an entity/architecture pair or a configuration. The entity ports are of type bit, bit_vector, std_logic, std_logic_vector, std_ulogic, std_ulogic_vector, or their subtypes. The port clause may have any mix of these types. Only locally static subtypes are allowed. Also, array types must be locally static constrained array subtypes (e.g. std_logic_vector(8 downto 0), rather than std_logic_vector(4*2 downto 0), which is not supported.) Port associations may be named or positional. Use the same port names and port positions that appear in the entity.
SystemC foreign module declaration

In cases where you want to run a mixed simulation with SystemC and VHDL, you must create and declare a foreign module that stands in for each VHDL design unit instantiated under SystemC. The foreign modules can be created in one of two ways: running scgenmod, a utility that automatically generates your foreign module declaration (much like vgencomp generates a component declaration) modifying your SystemC source code manually Only locally static subtypes are allowed with the use of scgenmod. Also, array types must be locally static constrained array subtypes (e.g. std_logic_vector(4*2 downto 0). A declaration such as std_logic_vector(generic_val downto 0), where generic_val is a generic, is not supported, since generics are not locally static.
Using scgenmod
After you have analyzed the design, you can generate a foreign module declaration with an scgenmod command similar to the following:
scgenmod mod1
Where mod1 is a VHDL entity. A foreign module declaration for the specified entity is written to stdout.
Guidelines for manual creation

Apply the following guidelines to the creation of foreign modules. A foreign module: contains ports corresponding to VHDL ports. These ports must be explicitly named in the foreign modules constructor initializer list. must not contain any internal design elements such as child instances, primitive channels, or processes.
SystemC: instantiating VHDL UM-203
must pass a secondary constructor argument denoting the modules HDL name to the sc_foreign_module base class constructor. For VHDL, the HDL name can be in the format [<lib>.]<primary>[(<secondary>)] or [<lib>.]<conf>. generics are supported for VHDL instantiations in SystemC designs. See "Generic support for SystemC instantiating VHDL" (UM-204) for more information.
Example
A sample VHDL design unit to be instantiated in a SystemC design is:
entity counter is port (count : buffer bit_vector(8 downto 1); clk : in bit; reset : in bit); end; architecture only of counter is ... ... end only;
The SystemC foreign module declaration for the above VHDL module is:
class counter : public sc_foreign_module { public: sc_in<bool> clk; sc_in<bool> reset; sc_out<sc_logic> count; counter(sc_module_name nm) : sc_foreign_module(nm, "work.counter(only)"), clk("clk"), reset("reset"), count("count") {} };
The VHDL module is then instantiated in the SystemC source as follows:

counter dut("dut");
where the constructor argument (dut) is the VHDL instance name.
Generic support for SystemC instantiating VHDL

Since the SystemC language has no concept of generics, generic values must be passed from a SystemC parent to an HDL child through the SystemC foreign module (sc_foreign_module). See "SystemC foreign module declaration" (UM-194) for information regarding the creation of sc_foreign_module.
Generic parameters in sc_foreign_module constructor

To instantiate a VHDL entity containing generics into the SystemC design, you must pass two parameters to the sc_foreign_module constructor: the number of generics (int num_generics), and the generic list (const char* generics_list). The generic_list is listed as an array of const char*.
195)), you
If you create your foreign module manually (see "Guidelines for manual creation" (UMmust also pass the generic information to the sc_foreign_module constructor. If you use scgenmod to create the foreign module declaration, the generic information is detected in the HDL child and is incorporated automatically.
Example
Following the example shown above (UM-204), lets see the generic information that would be passed to the SystemC foreign module declaration. The generic parameters passed to the constructor are shown in magenta color:
class counter : public sc_foreign_module { public: sc_in<bool> clk; ... counter(sc_midule_name nm, char* hdl_name int num_generics, const char* generic_list) : sc_foreign_module(nm, hdl_name, num_generics, generic_list), {} };
The instantiation is:

dut = new counter ("dut", "work.counter", 9, generic_list);
Example of generic use

Here is another example, a ring buffer, complete with all files necessary for the simulation.
// ringbuf.h #ifndef INCLUDED_RINGBUF #define INCLUDED_RINGBUF class ringbuf : public sc_foreign_module { public: sc_in<bool> clk; ... counter(sc_midule_name nm, char* hdl_name int num_generics, const char* generic_list) : sc_foreign_module(nm, hdl_name, num_generics, generic_list), {} }; #endif
--------------------------------------------------------------------------// test_ringbuf.h #ifndef INCLUDED_TEST_RINGBUF #define INCLUDED_TEST_RINGBUF #include "ringbuf.h" SC_MODULE(test_ringbuf) { sc_signal<T> iclock; ... ...
// VHDL module instance ringbuf* chip; SC_CTOR(test_ringbuf) : iclock("iclock"), ... ... { const char* generic_list[9]; generic_list[0] generic_list[1] generic_list[2] generic_list[3] generic_list[4] generic_list[5] generic_list[6] generic_list[7] generic_list[8] = = = = = = = = = strdup("int_param=4"); strdup("real_param=2.6"); strdup("str_param=\"Hello\""); strdup("bool_param=false"); strdup("char_param=Y"); strdup("bit_param=0"); strdup("bv_param=010"); strdup("logic_param=Z"); strdup("lv_param=01XZ");
// Cleanup the memory allocated for the generic list for (int = 0; i < 9; i++;) free((char*)generic_list[i]); // Create VHDL module instance. chip = new ringbuf("chip", "ringbuf", 9, generic_list); }; #endif ---------------------------------------------------------------------------- test_ringbuf.cpp #include "test_ringbuf.h" SC_MODULE_EXPORT(test_ringbuf); ---------------------------------------------------------------------------- ringbuf.vhd LIBRARY ieee; USE ieee.std_logic_1164.all; USE std.textio.all; ENTITY ringbuf IS generic ( int_param : integer; real_param : real; str_param : string; bool_param : boolean; char_param : character; bit_param : bit; bv_param : bit_vector(0 to 2); logic_param : std_logic; lv_param : std_logic_vector(3 downto 0)); PORT ( clock : IN std_logic; .. ...
); END ringbuf;
ARCHITECTURE RTL OF ringbuf IS BEGIN print_param: PROCESS variable line_out: Line; BEGIN write(line_out, string'("int_param="), left); write(line_out, int_param); writeline(OUTPUT, line_out); write(line_out, string'("real_param="), left); write(line_out, real_param); writeline(OUTPUT, line_out); write(line_out, string'("str_param="), left); write(line_out, str_param); writeline(OUTPUT, line_out); write(line_out, string'("bool_param="), left); write(line_out, bool_param); writeline(OUTPUT, line_out); write(line_out, string'("char_param="), left); write(line_out, char_param); writeline(OUTPUT, line_out); write(line_out, string'("bit_param="), left); write(line_out, bit_param); writeline(OUTPUT, line_out); write(line_out, string'("bv_param="), left); write(line_out, bv_param); writeline(OUTPUT, line_out); WAIT FOR 20 NS; END PROCESS; END RTL;
To run the parameterized design, use the following commands.

vsim1> vlib work vsim2> vcom ringbuf.vhd vsim3> scgenmod ringbuf > ringbuf.h //creates the sc_foreign_module including generic mapping info vsim4> sccom test_ringbuf.cpp vsim5> sccom -link vsim6> vsim -c test_ringbuf
The simulation returns the following:

# # # # # # # int_param=4 real_param=2.600000e+00 str_param=Hello bool_param=FALSE char_param=Y bit_param=0 bv_param=010
VHDL: instantiating SystemC

To instantiate SystemC in a VHDL design, you must create a component declaration for the SystemC module. Once you have generated the component declaration, you can instantiate the SystemC component just like any other VHDL component.
SystemC instantiation criteria

A SystemC design unit may be instantiated within VHDL if it meets the following criteria: SystemC module names are case sensitive. The module name at the SystemC instantiation site must match exactly with the actual SystemC module name. The SystemC design unit is exported using the SC_MODULE_EXPORT macro. The module ports are as listed in the table in "Data type mapping" (UM-184) Port data type mapping must match exactly. See the table in "Port type mapping" (UM184). Port associations may be named or positional. Use the same port names and port positions that appear in the SystemC module. Named port associations are case sensitive.
Component declaration
A SystemC design unit can be referenced from a VHDL design as though it is a VHDL entity. The interface to the design unit can be extracted from the library in the form of a component declaration by running vgencomp. Given a library and a SystemC module name, vgencomp writes a component declaration to standard output. The default component port types are: std_logic std_logic_vector Optionally, you can choose: bit and bit_vector
VHDL and SystemC identifiers

The VHDL identifiers for the component name and port names are the same as the SystemC identifiers for the module name and port names. If a SystemC identifier is not a valid VHDL 1076-1987 identifier, it is converted to a VHDL 1076-1993 extended identifier (in which case you must compile the VHDL with the -93 or later switch).
vgencomp component declaration

vgencomp (CR-266) generates a component declaration according to these rules:
Port clause
A port clause is generated if the module has ports. A corresponding VHDL port is defined for each named SystemC port. You can set the VHDL port type to bit or std_logic. If the SystemC port has a range, then the VHDL port type is bit_vector or std_logic_vector.
VHDL: instantiating SystemC UM-209
Examples SystemC port sc_in<sc_logic>p1; sc_out<sc_lv<8>>p2; sc_inout<sc_lv<8>>p3; VHDL port p1 : in std_logic; p2 : out std_logic_vector(7 downto 0); p3 : inout std_logic_vector(7 downto 0)
Configuration declarations are allowed to reference SystemC modules in the entity aspects of component configurations. However, the configuration declaration cannot extend into a SystemC instance to configure the instantiations within the SystemC module.
Exporting SystemC modules

To be able to instantiate a SystemC module within VHDL (or use a SystemC module as a top level module), the module must be exported. Assume a SystemC module named transceiver exists, and that it is declared in header file transceiver.h. Then the module is exported by placing the following code in a .cpp file:
#include "transceiver.h" SC_MODULE_EXPORT(transceiver);
sccom -link
The sccom -link command collects the object files created in the work library, and uses them to build a shared library (.so) in the current work library. If you have changed your SystemC source code and recompiled it using sccom, then you must run sccom -link before invoking vsim. Otherwise your changes to the code are not recognized by the simulator.
Generic support for VHDL instantiating SystemC

Support for generics is available in a workaround flow for the current release. For workaround flow details, please refer to systemc_generics.note located in the <install_dir>/modeltech/docs/technotes directory.
UM-211
8 - WLF files (datasets) and virtuals

Chapter contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-212 UM-213 UM-215 UM-216 UM-217 UM-218 UM-218 UM-218 UM-219 UM-220 UM-221 UM-222 UM-222 UM-223 UM-224 UM-224 Saving a simulation to a WLF file Opening datasets . . . . . .
Viewing dataset structure . Structure tab columns .
Managing multiple datasets . . . GUI . . . . . . . . Command line . . . . . Restricting the dataset prefix display Saving at intervals with Dataset Snapshot Collapsing time and delta steps . .
Virtual Objects (User-defined buses, and more) Virtual signals . . . . . . . Virtual functions . . . . . . . Virtual regions . . . . . . . Virtual types . . . . . . . .
A ModelSim simulation can be saved to a wave log format (WLF) file for future viewing or comparison to a current simulation. We use the term "dataset" to refer to a WLF file that has been reopened for viewing. You can open more than one WLF file for simultaneous viewing. You can also create virtual signals that are simple logical combinations of, or logical functions of, signals from different datasets.
UM-212 8 - WLF files (datasets) and virtuals
Introduction
Wave Log Format (WLF) files are recordings of simulation runs. The WLF file is written as an archive file in binary format and is used to drive the debug windows at a later time. The files contain data from logged objects (e.g., signals and variables) and the design hierarchy in which the logged objects are found. You can record the entire design or choose specific objects. The WLF file provides you with precise in-simulation and post-simulation debugging capability. Any number of WLF files can be reloaded for viewing or comparing to the active simulation. A dataset is a previously recorded simulation that has been loaded into ModelSim. Each dataset has a logical name to let you indicate the dataset to which any command applies. This logical name is displayed as a prefix. The current, active simulation is prefixed by "sim:", while any other datasets are prefixed by the name of the WLF file by default. Two datasets are displayed in the Wave window below. The current simulation is shown in the top pane and is indicated by the "sim" prefix. A dataset from a previous simulation is shown in the bottom pane and is indicated by the "gold" prefix.
The simulator resolution (see "Simulator resolution limit" (UM-114) or (UM-75)) must be the same for all datasets you are comparing, including the current simulation. If you have a WLF file that is in a different resolution, you can use the wlfman command (CR-336) to change it.
Saving a simulation to a WLF file UM-213
Saving a simulation to a WLF file

If you add objects to the Dataflow, List, or Wave windows, or log objects with the log command, the results of each simulation run are automatically saved to a WLF file called vsim.wlf in the current directory. If you run a new simulation in the same directory, the vsim.wlf file is overwritten with the new results. If you want to save the WLF file and not have it be overwritten, select the dataset tab in the Workspace and then select File > Save. Or, you can use the -wlf <filename> argument to the vsim command (CR-305) or the dataset save command (CR-119). Important: If you do not use dataset save or dataset snapshot, you must end a simulation session with a quit or quit -sim command in order to produce a valid WLF file. If you dont end the simulation in this manner, the WLF file will not close properly, and ModelSim may issue the error message "bad magic number" when you try to open an incomplete dataset in subsequent sessions. If you end up with a "damaged" WLF file, you can try to "repair" it using the wlfrecover command (CR-340).
WLF file parameter overview

There are a number of WLF file parameters that you can control via the modelsim.ini file or a simulator argument. This section summarizes the various parameters. Feature WLF Filename WLF Size Limit WLF Time Limit WLF Compression WLF Optimizationa WLF Delete on Quita WLF Cache Sizea WLF Collapse Mode vsim argument -wlf <filename> -wlfslim <n> -wlftlim <t> -wlfcompress -wlfnocompress -wlfopt -wlfnoopt -wlfdeleteonquit -wlfnodeleteonquit -wlfcachesize <n> -wlfnocollapse -wlfcollapsedelta -wlfcollapsetime modelsim.ini WLFFilename=<filename> WLFSizeLimit = <n> WLFTimeLimit = <t> WLFCompress = 0|1 WLFOptimize = 0|1 WLFDeleteOnQuit = 0|1 WLFCacheSize = <n> WLFCollapseModel = 0|1|2 Default vsim.wlf no limit no limit 1 (-wlfcompress) 1 (-wlfopt) 0 0 1
a.These parameters can also be set using the dataset config command (CR-114). WLF Filename Specify the name of the WLF file. WLF Size Limit Limit the size of a WLF file to <n> megabytes by truncating from the front of the file as necessary.
WLF Time Limit Limit the size of a WLF file to <t> time by truncating from the front of the file as necessary. WLF Compression Compress the data in the WLF file. WLF Optimization Write additional data to the WLF file to improve draw performance at large zoom ranges. Optimization results in approximately 15% larger WLF files. Disabling WLF optimization also prevents ModelSim from reading a previously generated WLF file that contains optimized data. WLF Delete on Quit Delete the WLF file automatically when the simulation exits. Valid for current simulation dataset (vsim.wlf) only. WLF Cache Size Specify the size in megabytes of the WLF reader cache. WLF reader caching caches blocks of the WLF file to reduce redundant file I/O. If the cache is made smaller or disabled, least recently used data will be freed to reduce the cache to the specified size. WLF Collapse Mode WLF event collapsing has three settings: disabled, delta, time: - When disabled, all events and event order are preserved. - Delta mode records an object's value at the end of a simulation delta (iteration) only. Default. - Time mode records an object's value at the end of a simulation time step only.
Opening datasets UM-215
Opening datasets
To open a dataset, do one of the following: Select File > Open and choose Log Files or use the dataset open command (CR-117).
The Open Dataset dialog includes the following options: Dataset Pathname Identifies the path and filename of the WLF file you want to open. Logical Name for Dataset This is the name by which the dataset will be referred. By default this is the name of the WLF file.
Viewing dataset structure

Each dataset you open creates a structure tab in the Main window workspace. The tab is labeled with the name of the dataset and displays a hierarchy of the design units in that dataset. The graphic below shows three structure tabs: one for the active simulation (sim) and one each for two datasets (test and gold).
Click here to scroll tabs
If you have too many tabs to display in the available space, you can scroll the tabs left or right by clicking the arrow icons at the bottom right-hand corner of the window.
Viewing dataset structure UM-217
Structure tab columns

Each structure tab displays three columns by default: Column name Instance Design unit Design unit type Description the name of the instance the name of the design unit the type (e.g., Module, Entity, etc.) of the design unit
Aside from the three columns listed above, there are numerous columns related to code coverage that can be displayed in structure tabs. You can hide or show columns by rightclicking a column name and selecting the name on the list.
Managing multiple datasets

GUI
When you have one or more datasets open, you can manage them using the Dataset Browser. To open the browser, select View > Datasets.
See "Dataset Browser dialog" (GR-55) for details on this dialog.
Command line
You can open multiple datasets when the simulator is invoked by specifying more than one vsim -view <filename> option. By default the dataset prefix will be the filename of the WLF file. You can specify a different dataset name as an optional qualifier to the vsim -view switch on the command line using the following syntax:
-view <dataset>=<filename>
For example: vsim -view foo=vsim.wlf ModelSim designates one of the datasets to be the "active" dataset, and refers all names without dataset prefixes to that dataset. The active dataset is displayed in the context path at the bottom of the Main window. When you select a design unit in a datasets structure tab, that dataset becomes active automatically. Alternatively, you can use the Dataset Browser or the environment command (CR-131) to change the active dataset. Design regions and signal names can be fully specified over multiple WLF files by using the dataset name as a prefix in the path. For example:
sim:/top/alu/out view:/top/alu/out golden:.top.alu.out
Managing multiple datasets UM-219
Dataset prefixes are not required unless more than one dataset is open, and you want to refer to something outside the active dataset. When more than one dataset is open, ModelSim will automatically prefix names in the Wave and List windows with the dataset name. You can change this default by selecting Tools > Window Preferences (Wave and List windows). ModelSim also remembers a "current context" within each open dataset. You can toggle between the current context of each dataset using the environment command (CR-131), specifying the dataset without a path. For example:
env foo:
sets the active dataset to foo and the current context to the context last specified for foo. The context is then applied to any unlocked windows. The current context of the current dataset (usually referred to as just "current context") is used for finding objects specified without a path. The Objects pane can be locked to a specific context of a dataset. Being locked to a dataset means that the pane will update only when the content of that dataset changes. If locked to both a dataset and a context (e.g., test: /top/foo), the pane will update only when that specific context changes. You specify the dataset to which the pane is locked by selecting File > Environment.
Restricting the dataset prefix display

The default for dataset prefix viewing is set with a variable in pref.tcl, PrefMain(DisplayDatasetPrefix). Setting the variable to 1 will display the prefix, setting it to 0 will not. It is set to 1 by default. Either edit the pref.tcl file directly or use the Tools > Edit Preferences command to change the variable value. Additionally, you can restrict display of the dataset prefix if you use the environment -nodataset command to view a dataset. To display the prefix use the environment command (CR-131) with the -dataset option (you wont need to specify this option if the variable noted above is set to 1). The environment command line switches override the pref.tcl variable.
Saving at intervals with Dataset Snapshot

Dataset Snapshot lets you periodically copy data from the current simulation WLF file to another file. This is useful for taking periodic "snapshots" of your simulation or for clearing the current simulation WLF file based on size or elapsed time. Once you have logged the appropriate objects, select Tools > Dataset Snapshot (Wave window).
See "Dataset Snapshot dialog" (GR-233) for details on this dialog.
Collapsing time and delta steps UM-221
Collapsing time and delta steps

By default ModelSim collapses delta steps. This means each logged signal that has events during a simulation delta has its final value recorded to the WLF file when the delta has expired. The event order in the WLF file matches the order of the first events of each signal. You can configure how ModelSim collapses time and delta steps using arguments to the vsim command (CR-305) or by setting the WLFCollapseMode (UM-449) variable in the modelsim.ini file. The table below summarizes the arguments and how they affect event recording. vsim argument -wlfnocollapse effect All events for each logged signal are recorded to the WLF file in the exact order they occur in the simulation. Each logged signal which has events during a simulation delta has its final value recorded to the WLF file when the delta has expired. Default. Same as delta collapsing but at the timestep granularity. modelsim.ini setting WLFCollapseMode = 0
-wlfdeltacollapse
WLFCollapseMode = 1
-wlftimecollapse
WLFCollapseMode = 2
When a run completes that includes single stepping or hitting a breakpoint, all events are flushed to the WLF file regardless of the time collapse mode. Its possible that single stepping through part of a simulation may yield a slightly different WLF file than just running over that piece of code. If particular detail is required in debugging, you should disable time collapsing.
Virtual Objects (User-defined buses, and more)

Virtual objects are signal-like or region-like objects created in the GUI that do not exist in the ModelSim simulation kernel. ModelSim supports the following kinds of virtual objects: Virtual signals (UM-222) Virtual functions (UM-223) Virtual regions (UM-224) Virtual types (UM-224) Virtual objects are indicated by an orange diamond as illustrated by bus below:
Virtual signals
Virtual signals are aliases for combinations or subelements of signals written to the WLF file by the simulation kernel. They can be displayed in the Objects, List, and Wave windows, accessed by the examine command, and set using the force command. You can create virtual signals using the Tools > Combine Signals (Wave and List windows) menu selections or by using the virtual signal command (CR-287). Once created, virtual signals can be dragged and dropped from the Objects pane to the Wave and List windows. Virtual signals are automatically attached to the design region in the hierarchy that corresponds to the nearest common ancestor of all the elements of the virtual signal. The virtual signal command has an -install <region> option to specify where the virtual signal should be installed. This can be used to install the virtual signal in a user-defined region in
Virtual Objects (User-defined buses, and more) UM-223
order to reconstruct the original RTL hierarchy when simulating and driving a post-synthesis, gate-level implementation. A virtual signal can be used to reconstruct RTL-level design buses that were broken down during synthesis. The virtual hide command (CR-278) can be used to hide the display of the broken-down bits if you don't want them cluttering up the Objects pane. If the virtual signal has elements from more than one WLF file, it will be automatically installed in the virtual region virtuals:/Signals. Virtual signals are not hierarchical if two virtual signals are concatenated to become a third virtual signal, the resulting virtual signal will be a concatenation of all the scalar elements of the first two virtual signals. The definitions of virtuals can be saved to a macro file using the virtual save command
(CR-285). By default, when quitting, ModelSim will append any newly-created virtuals (that
have not been saved) to the virtuals.do file in the local directory. If you have virtual signals displayed in the Wave or List window when you save the Wave or List format, you will need to execute the virtuals.do file (or some other equivalent) to restore the virtual signal definitions before you re-load the Wave or List format during a later run. There is one exception: "implicit virtuals" are automatically saved with the Wave or List format.
Implicit and explicit virtuals

An implicit virtual is a virtual signal that was automatically created by ModelSim without your knowledge and without you providing a name for it. An example would be if you expand a bus in the Wave window, then drag one bit out of the bus to display it separately. That action creates a one-bit virtual signal whose definition is stored in a special location, and is not visible in the Objects pane or to the normal virtual commands. All other virtual signals are considered "explicit virtuals".
Virtual functions
Virtual functions behave in the GUI like signals but are not aliases of combinations or elements of signals logged by the kernel. They consist of logical operations on logged signals and can be dependent on simulation time. They can be displayed in the Objects, Wave, and List windows and accessed by the examine command (CR-132), but cannot be set by the force command (CR-141). Examples of virtual functions include the following: a function defined as the inverse of a given signal a function defined as the exclusive-OR of two signals a function defined as a repetitive clock a function defined as "the rising edge of CLK delayed by 1.34 ns" Virtual functions can also be used to convert signal types and map signal values. The result type of a virtual function can be any of the types supported in the GUI expression syntax: integer, real, boolean, std_logic, std_logic_vector, and arrays and records of these types. Verilog types are converted to VHDL 9-state std_logic equivalents and Verilog net strengths are ignored.
Virtual functions can be created using the virtual function command (CR-275). Virtual functions are also implicitly created by ModelSim when referencing bit-selects or part-selects of Verilog registers in the GUI, or when expanding Verilog registers in the Objects, Wave, or List window. This is necessary because referencing Verilog register elements requires an intermediate step of shifting and masking of the Verilog "vreg" data structure.
Virtual regions
User-defined design hierarchy regions can be defined and attached to any existing design region or to the virtuals context tree. They can be used to reconstruct the RTL hierarchy in a gate-level design and to locate virtual signals. Thus, virtual signals and virtual regions can be used in a gate-level design to allow you to use the RTL test bench. Virtual regions are created and attached using the virtual region command (CR-284).
Virtual types
User-defined enumerated types can be defined in order to display signal bit sequences as meaningful alphanumeric names. The virtual type is then used in a type conversion expression to convert a signal to values of the new type. When the converted signal is displayed in any of the windows, the value will be displayed as the enumeration string corresponding to the value of the original signal. Virtual types are created using the virtual type command (CR-290).
UM-225
9 - Waveform analysis
Chapter contents
Introduction . . . . Objects you can view . Wave window overview List window overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-227 UM-227 UM-228 UM-228 UM-232 UM-233 UM-233 UM-234 UM-235 UM-236 UM-236 UM-237 UM-238 UM-239 UM-239 UM-240 UM-241 UM-243 UM-243 UM-243 UM-247 UM-244 UM-245 UM-247 UM-247 UM-247 UM-249 UM-250 UM-250 UM-250 UM-250 UM-251 UM-252 UM-252 UM-253 UM-254 UM-256
Adding objects to the Wave or List window .
Measuring time with cursors in the Wave window Working with cursors . . . . . . . Understanding cursor behavior . . . . Jumping to a signal transition . . . . . Setting time markers in the List window . Working with markers . . . . . . . .
Zooming the Wave window display . . . . . . . Saving zoom range and scroll position with bookmarks . Searching in the Wave and List windows . . . . . Finding signal names . . . . . . . . . Searching for values or transitions . . . . . Using the Expression Builder for expression searches Formatting the Wave window . . . . . Setting Wave window display properties . Formatting objects in the Wave window . Changing radix (base) . . . . . . Dividing the Wave window . . . . Splitting Wave window panes. . . . Formatting the List window . . . . Setting List window display properties Formatting objects in the List window Saving the window format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Printing and saving waveforms in the Wave window Saving a .eps file and printing under UNIX . Printing on Windows platforms . . . . Printer page setup . . . . . . . . Saving List window data to a file . Combining objects/creating busses Example . . . . . . . . . . . . . . . . . .
Configuring new line triggering in the List window . Using gating expressions to control triggering . Sampling signals at a clock change . . . .
UM-226 9 - Waveform analysis
Miscellaneous tasks . . . . . . . . Examining waveform values . . . . . Displaying drivers of the selected waveform . Setting signal breakpoints in the Wave window Examining waveform values . . . . . Waveform Compare . . . . . . . . Three options for setting up a comparison . Setting up a comparison with the GUI . . Starting a waveform comparison . . . . Adding signals, regions, and clocks . . . Specifying the comparison method . . . Setting compare options . . . . . . Viewing differences in the Wave window . Viewing differences in the List window . . Viewing differences in textual format . . Saving and reloading comparison results . . Comparing hierarchical and flattened designs
. . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . .
UM-257 UM-257 UM-257 UM-257 UM-257 UM-257 UM-258 UM-259 UM-260 UM-262 UM-264 UM-266 UM-267 UM-269 UM-270 UM-270 UM-271
Introduction UM-227
Introduction
When your simulation finishes, you will often want to analyze waveforms to assess and debug your design. Designers typically use the Wave window for waveform analysis. However, you can also look at waveform data in a textual format in the List window. To analyze waveforms in ModelSim, follow these steps: 1 Compile your files. 2 Load your design. 3 Add objects to the Wave or List window.
add wave <object_name> add list <object_name>
4 Run the design.
Objects you can view

The list below identifies the types of objects can be viewed in the Wave or List window.
VHDL objects
(indicated by dark blue diamond in the Wave window) signals, aliases, process variables, and shared variables
Verilog objects
(indicated by light blue diamond in the Wave window) nets, registers, variables, and named events
SystemC objects
(indicated by a green diamond in the Wave window) primitive channels and ports
Virtual objects
(indicated by an orange diamond in the Wave window) virtual signals, buses, and functions, see; "Virtual Objects (User-defined buses, and more)" (UM-222) for more information
Comparisons
(indicated by a yellow triangle) comparison regions and comparison signals; see Waveform Compare (UM-258) for more information
Wave window overview

The Wave window opens by default in the MDI frame of the Main window as shown below. The window can be undocked from the main window by pressing the Undock button in the window header or by using the view -undock wave command. The preference variable PrefWave(ViewUnDocked) can be used to control this default behavior. By setting the value of this variable to 1, the Wave Window will open undocked.
Undock button
Wave window overview UM-229
Here is an example of a Wave window that is undocked from the MDI frame. All menus and icons associated with Wave window functions now appear in the menu and toolbar areas of the Wave window.
Dock button
Undock button
If the Wave window is docked into the Main window MDI frame, all menus and icons that were in the standalone version of the Wave window move into the Main window menu bar and toolbar. See "Main window menu bar" (GR-23) for more information.
The Wave window is divided into a number of window panes. All window panes in the Wave window can be resized by clicking and dragging the bar between any two panes.
pathnames values waveforms
cursors names and values
cursors
List window overview UM-231
List window overview

The List window displays simulation results in tabular format. Common tasks that people use the window for include: Using gating expressions and trigger settings to focus in on particular signals or events. See "Configuring new line triggering in the List window" (UM-253). Debugging delta delay issues. See "Delta delays" (UM-77) for more information. The window is divided into two adjustable panes, which allows you to scroll horizontally through the listing on the right, while keeping time and delta visible on the left.
Adding objects to the Wave or List window

You can add objects to the Wave or List window in several ways.
Adding objects with drag and drop

You can drag and drop objects into the Wave or List window from the Workspace, Active Processes, Memory, Objects, Source, or Locals panes. You can also drag objects from the Wave window to the List window and vice versa. Select the objects in the first window, then drop them into the Wave window. Depending on what you select, all objects or any portion of the design can be added.
Adding objects with a menu command

The Add menu in the Main windows let you add objects to the Wave window, List window, or Log file.
Adding objects with a command

Use the add list command (CR-44) or add wave command (CR-49) to add objects from the command line. For example:
VSIM> add wave /proc/a
Adds signal /proc/a to the Wave window.

VSIM> add list *
Adds all the objects in the current region to the List window.
VSIM> add wave -r /*
Adds all objects in the design to the Wave window.
Adding objects with a window format file

Select File > Open > Format and specify a previously saved format file. See "Saving the window format" (UM-249) for details on how to create a format file.
Measuring time with cursors in the Wave window UM-233
Measuring time with cursors in the Wave window

ModelSim uses cursors to measure time in the Wave window. Cursors extend a vertical line over the waveform display and identify a specific simulation time. Multiple cursors can be used to measure time intervals, as shown in the graphic below. When the Wave window is first drawn, there is one cursor located at time zero. Clicking anywhere in the waveform display brings that cursor to the mouse location. The selected cursor is drawn as a bold solid line; all other cursors are drawn with thin lines. As shown in the graphic below, three window panes relate to cursors: the cursor name pane on the bottom left, the cursor value pane in the bottom middle, and the cursor pane with horizontal "tracks" on the bottom right.
right-click here to name a cursor
cursor value pane
interval measurement
locked cursor is red
Working with cursors

The table below summarizes common cursor actions. Action Add cursor Menu command Insert > Cursor Toolbar button
Action Delete cursor
Menu command Edit > Delete Cursor
Toolbar button
Lock cursor Name cursor Select cursor
Edit > Edit Cursor Edit > Edit Cursor View > Cursors
NA NA NA
Shortcuts for working with cursors

There are a number of useful keyboard and mouse shortcuts related to the actions listed above: Select a cursor by clicking the cursor name. Jump to a "hidden" cursor (one that is out of view) by double-clicking the cursor name. Name a cursor by right-clicking the cursor name and entering a new value. Press <Enter> after you have typed the new name. Move a locked cursor by holding down the <shift> key and then clicking-and-dragging the cursor. Move a cursor to a particular time by right-clicking the cursor value and typing the value to which you want to scroll. Press <Enter> after you have typed the new value.
Understanding cursor behavior

The following list describes how cursors "behave" when you click in various panes of the Wave window: If you click in the waveform pane, the cursor closest to the mouse position is selected and then moved to the mouse position. Clicking in a horizontal "track" in the cursor pane selects that cursor and moves it to the mouse position. Cursors "snap" to a waveform edge if you click or drag a cursor along the selected waveform to within ten pixels of a waveform edge. You can set the snap distance in the Window Preferences dialog. Select Tools > Options > Wave Preferences when the Wave window is docked in the Main window MDI frame. Select Tools > Window Preferences when the Wave window is a stand-alone, undocked window. You can position a cursor without snapping by dragging in the cursor pane below the waveforms.
Measuring time with cursors in the Wave window UM-235
Jumping to a signal transition

You can move the active cursor to the next or previous transition on the selected signal using these two buttons on the toolbar: Find Previous Transition locate the previous signal value change for the selected signal Find Next Transition locate the next signal value change for the selected signal
Setting time markers in the List window

Time markers in the List window are similar to cursors in the Wave window. Time markers tag lines in the data table so you can quickly jump back to that time. Markers are indicated by a thin box surrounding the marked line.
Working with markers

The table below summarizes actions you can take with markers. Action Add marker Delete marker Goto marker Method Select a line and then select Edit > Add Marker Select a tagged line and then select Edit > Delete Marker Select View > Goto > <time>
Zooming the Wave window display UM-237
Zooming the Wave window display

Zooming lets you change the simulation range in the waveform pane. You can zoom using the context menu, toolbar buttons, mouse, keyboard, or commands.
Zooming with menu commands

You can access Zoom commands from the View menu on the toolbar or by clicking the right mouse button in the waveform pane.
Zooming with toolbar buttons

These zoom buttons are available on the toolbar: Zoom In 2x zoom in by a factor of two from the current view Zoom Full zoom out to view the full range of the simulation from time 0 to the current time Zoom Out 2x zoom out by a factor of two from current view Zoom Mode change mouse pointer to zoom mode; see below
Zooming with the mouse

To zoom with the mouse, first enter zoom mode by selecting View > Mouse Mode > Zoom Mode. The left mouse button then offers 3 zoom options by clicking and dragging in different directions: Down-Right or Down-Left: Zoom Area (In) Up-Right: Zoom Out Up-Left: Zoom Fit
Also note the following about zooming with the mouse: The zoom amount is displayed at the mouse cursor. A zoom operation must be more than 10 pixels to activate. You can enter zoom mode temporarily by holding the <Ctrl> key down while in select mode. With the mouse in the Select Mode, the middle mouse button will perform the above zoom operations.
Saving zoom range and scroll position with bookmarks

Bookmarks save a particular zoom range and scroll position. This lets you return easily to a specific view later. You save the bookmark with a name and then access the named bookmark from the Bookmark menu. Bookmarks are saved in the Wave format file (see "Adding objects with a window format file" (UM-232)) and are restored when the format file is read.
Managing bookmarks
The table below summarizes actions you can take with bookmarks. Action Add bookmark View bookmark Delete bookmark Menu Edit > Insert Bookmark View > Bookmark > <name> Tools > Bookmarks Command bookmark add wave (CR-57) bookmark goto wave (CR-59) bookmark delete wave (CR-58)
Adding bookmarks
To add a bookmark, follow these steps: 1 Zoom the wave window as you see fit using one of the techniques discussed in "Zooming the Wave window display" (UM-237). 2 Select Edit > Insert Bookmark.
3 Give the bookmark a name and click OK.
Editing Bookmarks
Once a bookmark exists, you can change its properties by selecting Tools > Bookmarks. See "Modify Breakpoints dialog" (GR-231) for more details.
Searching in the Wave and List windows UM-239
Searching in the Wave and List windows

The Wave and List windows provide two methods for locating objects: Finding signal names Select Edit > Find or use the find command (CR-137) to search for the name of a signal. Search for values or transitions Select Edit > Search to locate transitions or signal values. The search feature is not available in all versions of ModelSim.
Finding signal names

The Find command is used primarily to locate a signal name in the Wave or List window. When you select Edit > Find, the Find dialog appears:
The Find dialog gives various options that are discussed further under "Find in .wave dialog" (GR-215). One option of note is the "Exact" checkbox. Check Exact if you only want to find objects that match your search exactly. For example, searching for "clk" without Exact will find /top/clk and clk1. There are two differences between the Wave and List windows as it relates to the Find feature: In the Wave window you can specify a value to search for in the values pane. The find operation works only within the active pane in the Wave window.
Searching for values or transitions

Available in some versions of ModelSim, the Search command lets you search for transitions or values on selected signals. When you select Edit > Search, the Signal Search dialog appears:
The Search dialog gives various options that are discussed further under "Wave Signal Search dialog" (GR-216). One option of note is Search for Expression. The expression can involve more than one signal but is limited to signals currently in the window. Expressions can include constants, variables, and DO files. See "Expression syntax" (CR-24) for more information.
Searching in the Wave and List windows UM-241
Using the Expression Builder for expression searches

The Expression Builder is a feature of the Wave and List Signal Search dialog boxes, and the List trigger properties dialog box. It aids in building a search expression that follows the "GUI_expression_format" (CR-23). To locate the Builder: select Edit > Search (List or Wave window) select the Search for Expression option in the resulting dialog box select the Builder button
The Expression Builder dialog box provides an array of buttons that help you build a GUI expression. For instance, rather than typing in a signal name, you can select the signal in the associated Wave or List window and press Insert Selected Signal. All Expression Builder buttons correspond to the "Expression syntax" (CR-24).
Saving an expression to a Tcl variable

Clicking the Save button will save the expression to a Tcl variable. Once saved this variable can be used in place of the expression. For example, say you save an expression to the variable "foo". Here are some operations you could do with the saved variable: Read the value of foo with the set command:
set foo
Put $foo in the Expression: entry box for the Search for Expression selection. Issue a searchlog command using foo:
searchlog -expr $foo 0
To search for when a signal reaches a particular value

Select the signal in the Wave window and click Insert Selected Signal and ==. Then, click the value buttons or type a value.
To evaluate only on clock edges

Click the && button to AND this condition with the rest of the expression. Then select the clock in the Wave window and click Insert Selected Signal and rising. You can also select the falling edge or both edges.
Operators
Other buttons will add operators of various kinds (see "Expression syntax" (CR-24)), or you can type them in.
Formatting the Wave window UM-243
Formatting the Wave window

Setting Wave window display properties
You can set display properties of the Wave window by selecting Tools > Options > Wave Preferences (when the window is docked in the MDI frame) or Tools > Window Preferences (when the window is undocked). These commands open the Window Preferences dialog (see "Window Preferences dialog" (GR-237) for details on the dialog).
Hiding/showing path hierarchy

You can set how many elements of the object path display by changing the Display Signal Path value in the Window Preferences dialog. Zero indicates the full path while a non-zero number indicates the number of path elements to be displayed.
Formatting objects in the Wave window

You can adjust various object properties to create the view you find most useful. Select one or more objects and then select View > Signal Properties (see "Wave Signal Properties dialog" (GR-219) for details on the dialog) or use the selections in the Format menu.
Changing radix (base)

One common adjustment is changing the radix (base) of an object. When you select View > Signal Properties, the Wave Signal Properties dialog appears:
The default radix is symbolic, which means that for an enumerated type, the value pane lists the actual values of the enumerated type of that object. For the other radixes - binary, octal,
decimal, unsigned, hexadecimal, or ASCII - the object value is converted to an appropriate representation in that radix. Aside from the Wave Signal Properties dialog, there are three other ways to change the radix: Change the default radix for the current simulation using Simulate > Runtime Options (Main window) Change the default radix for the current simulation using the radix command (CR-190). Change the default radix permanently by editing the DefaultRadix (UM-446) variable in the modelsim.ini file.
Dividing the Wave window

Dividers serve as a visual aid for debugging, allowing you to separate signals and waveforms for easier viewing. In the graphic below, two datasets have been separated with a divider called "gold."
To insert a divider, follow these steps: 1 Select the signal above which you want to place the divider. 2 If the Wave pane is docked in MDI frame of the Main window, select Add > Divider from the Main window menu bar. If the Wave window stands alone, undocked from the Main window, select Insert > Divider from the Wave window menu bar.
Formatting the Wave window UM-245
3 Specify the divider name in the Wave Divider Properties dialog. The default name is New Divider. Unnamed dividers are permitted. Simply delete "New Divider" in the Divider Name field to create an unnamed divider. 4 Specify the divider height (default height is 17 pixels) and then click OK. You can also insert dividers with the -divider argument to the add wave command (CR-49).
Working with dividers

The table below summarizes several actions you can take with dividers: Action Move a divider Change a dividers name or size Delete a divider Method Click-and-drag the divider to the desired location Right-click the divider and select Divider Properties Right-click the divider and select Delete
Splitting Wave window panes

The pathnames, values, and waveforms panes of the Wave window display can be split to accommodate signals from one or more datasets. For more information on viewing multiple simulations, see Chapter 8 - WLF files (datasets) and virtuals. To split the window, select Insert > Window Pane.
In the illustration below, the top split shows the current active simulation with the prefix "sim," and the bottom split shows a second dataset with the prefix "gold".
The active split

The active split is denoted with a solid white bar to the left of the signal names. The active split becomes the target for objects added to the Wave window.
Formatting the List window UM-247
Formatting the List window

Setting List window display properties
Before you add objects to the List window, you can set the windows display properties. To change when and how a signal is displayed in the List window, select Tools > Window Preferences. See "Modify Display Properties dialog" (GR-145) for more details.
Formatting objects in the List window

You can adjust various properties of objects to create the view you find most useful. Select one or more objects and then select View > Signal Properties. This dialog is described in detail under "List Signal Properties dialog" (GR-142).
Changing radix (base)

One common adjustment is changing the radix (base) of an object. When you select View > Signal Properties, the List Signal Properties dialog appears:
The default radix is symbolic, which means that for an enumerated type, the window lists the actual values of the enumerated type of that object. For the other radixes - binary, octal, decimal, unsigned, hexadecimal, or ASCII - the object value is converted to an appropriate representation in that radix.
Changing the radix can make it easier to view information in the List window. Compare the image below (with decimal values) with the image on page UM-231 (with symbolic values).
Aside from the List Signal Properties dialog, there are three other ways to change the radix: Change the default radix for the current simulation using Simulate > Runtime Options (Main window) Change the default radix for the current simulation using the radix command (CR-190). Change the default radix permanently by editing the DefaultRadix (UM-446) variable in the modelsim.ini file.
Saving the window format UM-249
Saving the window format

By default all Wave and List window information is forgotten once you close the windows. If you want to restore the windows to a previously configured layout, you must save a window format file. Follow these steps: 1 Add the objects you want to the Wave or List window. 2 Edit and format the objects to create the view you want. 3 Save the format to a file by selecting File > Save > Format. To use the format file, start with a blank Wave or List window and run the DO file in one of two ways: Invoke the do command (CR-125) from the command line:
VSIM> do <my_format_file>
Select File > Load. Note: Window format files are design-specific. Use them only with the design you were simulating when they were created.
Printing and saving waveforms in the Wave window

You can print the waveform display or save it as an encapsulated postscript (EPS) file. The printing dialogs are described in detail in the GUI Reference appendix.
Saving a .eps file and printing under UNIX

Select File > Print Postscript (Wave window) to print all or part of the waveform in the current Wave window in UNIX, or save the waveform as a .eps file on any platform (see also the write wave command (CR-350)).
Printing on Windows platforms

Select File > Print (Wave window) to print all or part of the waveform in the current Wave window, or save the waveform as a printer file (a Postscript file for Postscript printers).
Printer page setup

Select File > Page setup or click the Setup button in the Write Postscript or Print dialog box to define how the printed page will appear. The Page Setup dialog is described in detail in the GUI Reference appendix.
Saving List window data to a file UM-251
Saving List window data to a file

Select File > Write List in the List window to save the data in one of these formats: Tabular writes a text file that looks like the window listing
ns 0 0 2 delta +0 +1 +0 /a X 0 0 /b X 1 1 /cin U 0 0 /sum X X X /cout U U U
Events writes a text file containing transitions during simulation

@0 +0 /a X /b X /cin U /sum X /cout U @0 +1 /a 0 /b 1 /cin 0
TSSI writes a file in standard TSSI format; see also, the write tssi command (CR-348)
0 00000000000000010????????? 2 00000000000000010???????1? 3 00000000000000010??????010 4 00000000000000010000000010 100 00000001000000010000000010
You can also save List window output using the write list command (CR-343).
Combining objects/creating busses

You can combine signals in the Wave or List window into busses. A bus is a collection of signals concatenated in a specific order to create a new virtual signal with a specific value. A virtual compare signal (the result of a comparison simulation) is not supported for combination with any other signal. To combine signals into a bus, use one of the following methods: Select two or more signals in the Wave or List window and then choose Tools > Combine Signals from the menu bar. A virtual signal that is the result of a comparison simulation is not supported for combining with any other signal. Use the virtual signal command (CR-287) at the Main window command prompt. The Combine Signals dialog in the List and Wave windows differ slightly. See "Combine Selected Signals dialog" (GR-144) for the List window and "Combine Selected Signals dialog" (GR-235) for the Wave window.
Example
In the illustration below, three signals have been combined to form a new bus called "bus". Note that the component signals are listed in the order in which they were selected in the Wave window. Also note that the value of the bus is made up of the values of its component signals, arranged in a specific order.
Configuring new line triggering in the List window UM-253
Configuring new line triggering in the List window

New line triggering refers to what events cause a new line of data to be added to the List window. By default ModelSim adds a new line for any signal change including deltas within a single unit of time resolution. You can set new line triggering on a signal-by-signal basis or for the whole simulation. To set for a single signal, select View > Signal Properties and modify the Triggers setting (see "List Signal Properties dialog" (GR-142) for details). Individual signal settings override global settings. To modify new line triggering for the whole simulation, select Tools > Window Preferences or use the configure command (CR-98). When you select Tools > Window Preferences, the Modify Display Properties dialog appears:
The dialog gives various options that are discussed further under "Modify Display Properties dialog" (GR-145). The following table summaries the options: Option Deltas Description Choose between displaying all deltas (Expand Deltas), displaying the value at the final delta (Collapse Delta), or hiding the delta column all together (No Delta)
Option Strobe trigger Trigger gating
Description Specify an interval at which you want to trigger data display Use a gating expression to control triggering; see "Using gating expressions to control triggering" (UM-254) for more details
Using gating expressions to control triggering

Trigger gating controls the display of data based on an expression. Triggering is enabled once the gating expression evaluates to true. This setup behaves much like a hardware signal analyzer that starts recording data on a specified setup of address bits and clock edges. Here are some points about gating expressions: Gating expressions affect the display of data but not acquisition of the data. The expression is evaluated when the List window would normally have displayed a row of data (given the other trigger settings). The duration determines for how long triggering stays enabled after the gating expression returns to false (0). The default of 0 duration will enable triggering only while the expression is true (1). The duration is expressed in x number of default timescale units. Gating is level-sensitive rather than edge-triggered.
Trigger gating example using the Expression Builder

This example shows how to create a gating expression with the ModelSim Expression Builder. Here is the procedure: 1 Select Tools > Window Preferences to access the Triggers tab.
Configuring new line triggering in the List window UM-255
2 Check the Use Gating Expression check box and click Use Expression Builder.
3 Select the signal in the List window that you want to be the enable signal by clicking on its name in the header area of the List window. 4 Click Insert Selected Signal and then 'rising in the Expression Builder. 5 Click OK to close the Expression Builder. You should see the name of the signal plus "'rising" added to the Expression entry box of the Modify Display Properties dialog box. 6 Click OK to close the dialog. If you already have simulation data in the List window, the display should immediately switch to showing only those cycles for which the gating signal is rising. If that isn't quite what you want, you can go back to the expression builder and play with it until you get it the way you want it. If you want the enable signal to work like a "One-Shot" that would display all values for the next, say 10 ns, after the rising edge of enable, then set the On Duration value to 10 ns.
Trigger gating example using commands

The following commands show the gating portion of a trigger configuration statement:
configure list -usegating 1 configure list -gateduration 100 configure list -gateexpr {/test_delta/iom_dd'rising}
See the configure command (CR-98) for more details.
Sampling signals at a clock change

You easily can sample signals at a clock change using the add list command (CR-44) with the -notrigger argument. The -notrigger argument disables triggering the display on the specified signals. For example:
add list clk -notrigger a b c
When you run the simulation, List window entries for clk, a, b, and c appear only when clk changes. If you want to display on rising edges only, you have two options: 1 Turn off the List window triggering on the clock signal, and then define a repeating strobe for the List window. 2 Define a "gating expression" for the List window that requires the clock to be in a specified state. See above.
Miscellaneous tasks UM-257
Miscellaneous tasks
Examining waveform values
You can use your mouse to display a dialog that shows the value of a waveform at a particular time. You can do this two ways: Rest your mouse pointer on a waveform. After a short delay, a dialog will pop-up that displays the value for the time at which your mouse pointer is positioned. If youd prefer that this popup not display, it can be toggled off in the display properties. See "Setting Wave window display properties" (UM-243). Right-click a waveform and select Examine. A dialog displays the value for the time at which you clicked your mouse. This method works in the List window as well.
Displaying drivers of the selected waveform

You can automatically display in the Dataflow window the drivers of a signal selected in the Wave window. You can do this three ways: Select a waveform and click the Show Drivers button on the toolbar. Select a waveform and select Show Drivers from the shortcut menu Double-click a waveform edge (you can enable/disable this option in the display properties dialog; see "Setting Wave window display properties" (UM-243)) This operation opens the Dataflow window and displays the drivers of the signal selected in the Wave window. The Wave pane in the Dataflow window also opens to show the selected signal with a cursor at the selected time. The Dataflow window shows the signal(s) values at the current cursor position.
Sorting a group of objects in the Wave window

Select View > Sort to sort the objects in the pathname and values panes.
Setting signal breakpoints in the Wave window

You can set signal breakpoints in the Wave window. When a signal breakpoint is hit, a message appears in the Main window Transcript stating which signal caused the breakpoint. To insert a signal breakpoint, right-click a signal and select Insert Breakpoint. A breakpoint will be set on the selected signal. See "Creating and managing breakpoints" (GR240) for more information.
Waveform Compare
The ModelSim Waveform Compare feature allows you to compare simulation runs. Differences encountered in the comparison are summarized and listed in the Main window transcript. Differences are also shown in the Wave and List windows, and you can write a list of the differences to a file using the compare info command (CR-81). Note: The Waveform Compare feature is available as an add-on to the PELE version. Contact Model Technology sales for more information. The basic steps for running a comparison are as follows: 1 Run one simulation and save the dataset. For more information on saving datasets, see "Saving a simulation to a WLF file" (UM-213). 2 Run a second simulation. 3 Setup and run a comparison. 4 Analyze the differences in the Wave or List window.
Mixed-language waveform compare support

Mixed-language compares are supported as listed in the following table: C/C++ types bool, char, unsigned char short, unsigned short int, unsigned int long, unsigned long sc_bit, sc_bv, sc_logic, sc_lv sc_int, sc_uint sc_bigint, sc_biguint sc_signed, sc_unsigned net, reg bit, bit_vector, boolean, std_logic, std_logic_vector
SystemC types
Verilog types VHDL types
The number of elements must match for vectors; specific indexes are ignored.
Three options for setting up a comparison

There are three options for setting up a comparison: Comparison Wizard A series of dialogs that "walk" you through the process GUI Use various dialogs to "manually" configure the comparison Comparison commands Use a series of compare commands
Waveform Compare UM-259
Comparison Wizard
The simplest method for setting up a comparison is using the Wizard. The wizard is a series of dialogs that walks you through the process. To start the Wizard, select Tools > Waveform Compare > Comparison Wizard from either the Wave or Main window. The graphic below shows the first dialog in the Wizard. As you can see from this example, the dialogs include instructions on the left-hand side.
Comparison graphic interface

You can also set up a comparison via the GUI without using the Wizard. The steps of this process are described further in "Setting up a comparison with the GUI" (UM-259).
Comparison commands
There are numerous commands that give you complete control over a comparison. These commands can be entered in the Main window transcript or run via a DO file. The commands are detailed in the ModelSim Command Reference, but the following example shows the basic sequence:
compare start gold.wlf vsim.wlf compare add /* compare run
Setting up a comparison with the GUI

To setup a comparison with the GUI, follow these steps: 1 Initiate the comparison by specifying the reference and test datasets. See "Starting a waveform comparison" (UM-260) for details.
2 Add objects to the comparison. See "Adding signals, regions, and clocks" (UM-262) for details. 3 Specify the comparison method. See "Specifying the comparison method" (UM-264) for details. 4 Configure comparison options. See "Setting compare options" (UM-266) for details. 5 Run the comparison by selecting Tools > Waveform Compare > Run Comparison. 6 View the results. See "Viewing differences in the Wave window" (UM-267), "Viewing differences in the List window" (UM-269), and "Viewing differences in textual format" (UM-270)for details. Waveform Compare is initiated from either the Main or Wave window by selecting Tools >Waveform Compare > Start Comparison.
Starting a waveform comparison

Select Tools >Waveform Compare > Start Comparison to initiate the comparison. The Start Comparison dialog box allows you define the Reference and Test datasets.
Reference Dataset
The Reference Dataset is the .wlf file to which the test dataset will be compared. It can be a saved dataset, the current simulation dataset, or any part of the current simulation dataset.
Test Dataset
The Test Dataset is the .wlf file that will be compared against the Reference Dataset. Like the Reference Dataset, it can be a saved dataset, the current simulation dataset, or any part of the current simulation dataset. Once you click OK in the Start Comparison dialog box, ModelSim adds a Compare tab to the Main window.
After adding the signals, regions, and/or clocks you want to use in the comparison (see "Adding signals, regions, and clocks" (UM-262)), you will be able to drag compare objects from this tab into the Wave and List windows.
Adding signals, regions, and clocks

To designate the signals, regions, or clocks to be used in the comparison, click Tools > Waveform Compare > Add.
Adding signals
Clicking Tools > Waveform Compare > Add > Compare by Signal in the Wave window opens the structure_browser window, where you can specify signals to be used in the comparison.
Adding regions
Rather than comparing individual signals, you can also compare entire regions of your design. Select Tools > Waveform Compare > Add > Compare by Region to open the Add Comparison by Region dialog. The dialog has several options which are detailed in the GUI reference appendix.
Adding clocks
You add clocks when you want to perform a clocked comparison. See "Specifying the comparison method" (UM-264) for details.
Specifying the comparison method

The Waveform Compare feature provides two comparison methods: Continuous comparison Test signals are compared to reference signals at each transition of the reference. Timing differences between the test and reference signals are shown with rectangular red markers in the Wave window and yellow markers in the List window. Clocked comparisons Signals are compared only at or just after an edge on some signal. In this mode, you define one or more clocks. The test signal is compared to a reference signal and both are sampled relative to the defined clock. The clock can be defined as the rising or falling edge (or either edge) of a particular signal plus a user-specified delay. The design need not have any events occurring at the specified clock time. Differences between test signals and the clock are highlighted with red diamonds in the Wave window. To specify the comparison method, select Tools > Waveform Compare > Options and select the Comparison Method tab.
Continuous comparison
Continuous comparisons are the default. You have the option of specifying leading and trailing tolerances and a when expression that must evaluate to "true" or 1 at the signal edge for the comparison to become effective. See "Add Signal Options dialog" (GR-226) for more details on this dialog.
Clocked comparison
To specify a clocked comparison you must define a clock in the Add Clock dialog. You can access this dialog via the Clocks button in the Comparison Method tab or by selecting Tools > Waveform Compare > Add > Clocks.
See "Add Clocks dialog" (GR-228) for details on this dialog.
Setting compare options

There are a few "global" options that you can set for a comparison. Select Tools > Waveform Compare > Options.
Options in this dialog include setting the maximum number of differences allowed before the comparison terminates, specifying signal value matching rules, and saving or resetting the defaults. See "Comparison Options dialog" (GR-229) for more details.
Viewing differences in the Wave window

The Wave window provides a graphic display of comparison results. Pathnames of all test signals included in the comparison are denoted by yellow triangles. Test signals that contain timing differences when compared with the reference signals are denoted by a red X over the yellow triangle.
The names of the comparison objects take the form:

<path>/\refSignalName<>testSignalName\
If you compare two signals from different regions, the signal names include the uncommon part of the path. In comparisons of signals with multiple bits, you can display them in "buswise" or "bitwise" format. Buswise format lists the busses under the compare object whereas bitwise format lists each individual bit under the compare object. To select one format or the other, click your right mouse button on the plus sign (+) next to a compare object. Timing differences are also indicated by red bars in the vertical and horizontal scroll bars of the waveform display, and by red difference markers on the waveforms themselves. Rectangular difference markers denote continuous differences. Diamond difference markers denote clocked differences. Placing your mouse cursor over any difference marker
will initiate a popup display that provides timing details for that difference. You can toggle this popup on and off in the "Window Preferences dialog" (GR-237).
Pathnames
Values
Waveform display
difference details
difference markers
The values column of the Wave window displays the words "match" or "diff" for every test signal, depending on the location of the selected cursor. "Match" indicates that the value of the test signal matches the value of the reference signal at the time of the selected cursor. "Diff" indicates a difference between the test and reference signal values at the selected cursor.
Annotating differences
You can tag differences with textual notes that are included in the difference details popup and comparison reports. Click a difference with the right mouse button, and select Annotate Diff. Or, use the compare annotate (CR-73) command.
Compare icons
The Wave window includes six comparison icons that let you quickly jump between differences. From left to right, the icons do the following: find first difference, find previous annotated difference, find previous difference, find next difference, find next annotated difference, find last difference. Use these icons to move the selected cursor. These buttons cycle through differences on all signals. To view differences for just the selected signal, press <tab> and <shift - tab> on your keyboard. Note: If you have differences on individual bits of a bus, the compare icons will stop on those differences but <tab> and <shift - tab> will not. The compare icons cycle through comparison objects in all open Wave windows. If you have two Wave windows displayed, each containing different comparison objects, the compare icons will cycle through the differences displayed in both windows.
Viewing differences in the List window

Compare objects can be displayed in the List window too. Differences are highlighted with a yellow background. Tabbing on selected columns moves the selection to the next difference (actually difference edge). Shift-tabbing moves the selection backwards.
Right-clicking on a yellow-highlighted difference gives you three options: Diff Info, Annotate Diff, and Ignore/Noignore diff. With these options you can elect to display difference information, you can ignore selected differences or turn off ignore, and you can annotate individual differences.
Viewing differences in textual format

You can also view text output of the differences either in the Transcript pane of the Main window or in a saved file. To view them in the transcript, select Tools > Waveform Compare > Differences > Show. To save them to a text file, select Tools > Waveform Compare > Differences > Write Report.
Saving and reloading comparison results

To save comparison results for future use, you must save both the comparison setup rules and the comparison differences. To save the rules, select Tools > Waveform Compare > Rules > Save. This file will contain all rules for reproducing the comparison. The default file name is "compare.rul." To save the differences, select Tools > Waveform Compare > Differences > Save. The default file name is "compare.dif". To reload the comparison results at a later time, select Tools > Waveform Compare > Reload and specify the rules and difference files.
Comparing hierarchical and flattened designs

If you are comparing a hierarchical RTL design simulation against a flattened synthesized design simulation, you may have different hierarchies, different signal names, and the buses may be broken down into one-bit signals in the gate-level design. All of these differences can be handled by ModelSims Waveform Compare feature. If the test design is hierarchical but the hierarchy is different from the hierarchy of the reference design, you can use the compare add command (CR-69) to specify which region path in the test design corresponds to that in the reference design. If the test design is flattened and test signal names are different from reference signal names, the compare add command (CR-69) allows you to specify which signal in the test design will be compared to which signal in the reference design. If, in addition, buses have been dismantled, or "bit-blasted", you can use the -rebuild option of the compare add command (CR-69) to automatically rebuild the bus in the test design. This will allow you to look at the differences as one bus versus another. If signals in the RTL test design are different in type from the synthesized signals in the reference design registers versus nets, for example the Waveform Compare feature will automatically do the type conversion for you. If the type differences are too extreme (say integer versus real), Waveform Compare will let you know.
UM-273
10 - Tracing signals with the Dataflow window

Chapter contents
Dataflow window overview Objects you can view . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-274 UM-274 UM-275 UM-276 UM-277 UM-277 UM-278 UM-279 UM-279 UM-279 UM-279 UM-280 UM-281 UM-283 UM-284 UM-284 UM-285 UM-286 UM-287 UM-289
Adding objects to the window . Links to other windows . .
Exploring the connectivity of your design . Tracking your path through the design The embedded wave viewer . . . . . . . . . . . . . . .
Zooming and panning . . . . Zooming with toolbar buttons. Zooming with the mouse . . Panning with the mouse . . Tracing events (causality) . . .
Tracing the source of an unknown (X).
Finding objects by name in the Dataflow window. Saving the display . . . . . . Saving a .eps file . . . . . Printing on Windows platforms . Configuring page setup Symbol mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Configuring window options .
This chapter discusses how to use the Dataflow window for tracing signal values and browsing the physical connectivity of your design.
UM-274 10 - Tracing signals with the Dataflow window
Dataflow window overview

The Dataflow window allows you to explore the "physical" connectivity of your design; to trace events that propagate through the design; and to identify the cause of unexpected outputs. Note: ModelSim versions operating without a dataflow license feature have limited Dataflow functionality. Without the license feature, the window will show only one process and its attached signals or one signal and its attached processes. Contact your Mentor Graphics sales representative if you currently dont have a dataflow feature.
Objects you can view

The Dataflow window displays processes; signals, nets, and registers; and interconnect. The window has built-in mappings for all Verilog primitive gates (i.e., AND, OR, etc.). For components other than Verilog primitives, you can define a mapping between processes and built-in symbols. See "Symbol mapping" (UM-287) for details. You cannot view SystemC objects in the Dataflow window; however, you can view HDL regions from mixed designs that include SystemC.
Adding objects to the window UM-275
Adding objects to the window

You can use any of the following methods to add objects to the Dataflow window: drag and drop objects from other windows use the Navigate menu options in the Dataflow window use the add dataflow command (CR-43) double-click any waveform in the Wave window display The Navigate menu offers four commands that will add objects to the window. The commands include: View region clear the window and display all signals from the current region Add region display all signals from the current region without first clearing window View all nets clear the window and display all signals from the entire design Add ports add port symbols to the port signals in the current region When you view regions or entire nets, the window initially displays only the drivers of the added objects in order to reduce clutter. You can easily view readers by selecting an object and invoking Navigate > Expand net to readers. A small circle above an input signal on a block denotes a trigger signal that is on the process sensitivity list.
Links to other windows

The Dataflow window has links to other windows as described below: Window Main window (GR-16) Link select a signal or process in the Dataflow window, and the structure tab updates if that object is in a different design unit select a process in either window, and that process is highlighted in the other select a design object in either window, and that object is highlighted in the other trace through the design in the Dataflow window, and the associated signals are added to the Wave window move a cursor in the Wave window, and the values update in the Dataflow window Source window (GR-182) select an object in the Dataflow window, and the Source window updates if that object is in a different source file
Active Processes pane (GR-107) Objects pane (GR-167) Wave window (GR-194)
Exploring the connectivity of your design UM-277
Exploring the connectivity of your design

A primary use of the Dataflow window is exploring the "physical" connectivity of your design. One way of doing this is by expanding the view from process to process. This allows you to see the drivers/receivers of a particular signal, net, or register. You can expand the view of your design using menu commands or your mouse. To expand with the mouse, simply double click a signal, register, or process. Depending on the specific object you click, the view will expand to show the driving process and interconnect, the reading process and interconnect, or both. Alternatively, you can select a signal, register, or net, and use one of the toolbar buttons or menu commands described below: Expand net to all drivers display driver(s) of the selected signal, net, or register Expand net to all drivers and readers display driver(s) and reader(s) of the selected signal, net, or register Expand net to all readers display reader(s) of the selected signal, net, or register Navigate > Expand net to drivers
Navigate > Expand net
Navigate > Expand net to readers
As you expand the view, note that the "layout" of the design may adjust to best show the connectivity. For example, the location of an input signal may shift from the bottom to the top of a process.
Tracking your path through the design

You can quickly traverse through many components in your design. To help mark your path, the objects that you have expanded are highlighted in green.
You can clear this highlighting using the Edit > Erase highlight command.
The embedded wave viewer

Another way of exploring your design is to use the Dataflow windows embedded wave viewer. This viewer closely resembles, in appearance and operation, the stand-alone Wave window (see Chapter 9 - Waveform analysis for more information). The wave viewer is opened using the View > Show Wave command. One common scenario is to place signals in the wave viewer and the Dataflow panes, run the design for some amount of time, and then use time cursors to investigate value changes. In other words, as you place and move cursors in the wave viewer pane (see "Measuring time with cursors in the Wave window" (UM-233) for details), the signal values update in the Dataflow pane.
Another scenario is to select a process in the Dataflow pane, which automatically adds to the wave viewer pane all signals attached to the process. See "Tracing events (causality)" (UM-280) for another example of using the embedded wave viewer.
Zooming and panning UM-279
Zooming and panning

The Dataflow window offers several tools for zooming and panning the display.
Zooming with toolbar buttons

These zoom buttons are available on the toolbar: Zoom In zoom in by a factor of two from the current view Zoom Full zoom out to view the entire schematic Zoom Out zoom out by a factor of two from current view
Zooming with the mouse

To zoom with the mouse, you can either use the middle mouse button or enter Zoom Mode by selecting View > Zoom and then use the left mouse button. Four zoom options are possible by clicking and dragging in different directions: Down-Right: Zoom Area (In) Up-Right: Zoom Out (zoom amount is displayed at the mouse cursor) Down-Left: Zoom Selected Up-Left: Zoom Full
The zoom amount is displayed at the mouse cursor. A zoom operation must be more than 10 pixels to activate.
Panning with the mouse

You can pan with the mouse in two ways: 1) enter Pan Mode by selecting View > Pan and then drag with the left mouse button to move the design; 2) hold down the <Ctrl> key and drag with the middle mouse button to move the design.
Tracing events (causality)

One of the most useful features of the Dataflow window is tracing an event to see the cause of an unexpected output. This feature uses the Dataflow windows embedded wave viewer (see "The embedded wave viewer" (UM-278) for more details). In short you identify an output of interest in the Dataflow pane and then use time cursors in the wave viewer pane to identify events that contribute to the output. The process for tracing events is as follows: 1 Log all signals before starting the simulation (add log -r /*). 2 After running a simulation for some period of time, open the Dataflow window and the wave viewer pane. 3 Add a process or signal of interest into the Dataflow window (if adding a signal, find its driving process). Select the process and all signals attached to the selected process will appear in the wave viewer pane. 4 Place a time cursor on an edge of interest; the edge should be on a signal that is an output of the process.
5 Select Trace > Trace next event. A second cursor is added at the most recent input event. 6 Keep selecting Trace > Trace next event until you've reached an input event of interest. Note that the signals with the events are selected in the wave pane.
7 Now select Trace > Trace event set. The Dataflow display "jumps" to the source of the selected input event(s). The operation follows all signals selected in the wave viewer pane. You can change which signals are followed by changing the selection. 8 To continue tracing, go back to step 5 and repeat. If you want to start over at the originally selected output, select Trace > Trace event reset.
Tracing the source of an unknown (X) UM-281
Tracing the source of an unknown (X)

Another useful debugging option is locating the source of an unknown (X). Unknown values are most clearly seen in the Wave windowthe waveform displays in red when a value is unknown.
The procedure for tracing an unknown is as follows: 1 Load your design. 2 Log all signals in the design or any signals that may possibly contribute to the unknown value (log -r /* will log all signals in the design). 3 Add signals to the Wave window or wave viewer pane, and run your design the desired length of time. 4 Put a cursor on the time at which the signal value is unknown. 5 Add the signal of interest to the Dataflow window, making sure the signal is selected. 6 Select Trace > TraceX, Trace > TraceX Delay, Trace > ChaseX, or Trace > ChaseX Delay.
These commands behave as follows: TraceX / TraceX Delay Step back to the last driver of an X value. TraceX Delay works similarly but it steps back in time to the last driver of an X value. TraceX should be used for RTL designs; TraceX Delay should be used for gate-level netlists with back annotated delays. ChaseX / ChaseX Delay "Jumps" through a design from output to input, following X values. ChaseX Delay acts the same as ChaseX but also moves backwards in time to the point where the output value transitions to X. ChaseX should be used for RTL designs; ChaseX Delay should be used for gate-level netlists with back annotated delays.
Finding objects by name in the Dataflow window UM-283
Finding objects by name in the Dataflow window

Select Edit > Find to search for signal, net, or register names or an instance of a component.
See "Find in dataflow dialog" (GR-132) for more details.
Saving the display

Saving a .eps file
Select File > Print Postscript to save the waveform as a .eps file.
See "Print Postscript dialog" (GR-130) for more details.
Saving the display UM-285
Printing on Windows platforms

Select File > Print to print the Dataflow display or to save the display to a file.
See "Print dialog" (GR-128) for more details.
Configuring page setup

Clicking the Setup button in the Print Postscript or Print dialog box allows you to define the following options (this is the same dialog that opens via File > Page setup).
See "Dataflow Page Setup dialog" (GR-131) for more details.
Symbol mapping UM-287
Symbol mapping
The Dataflow window has built-in mappings for all Verilog primitive gates (i.e., AND, OR, etc.). For components other than Verilog primitives, you can define a mapping between processes and built-in symbols. This is done through a file containing name pairs, one per line, where the first name is the concatenation of the design unit and process names, (DUname.Processname), and the second name is the name of a built-in symbol. For example:
xorg(only).p1 XOR org(only).p1 OR andg(only).p1 AND
Entities and modules are mapped the same way:

AND1 AND AND2 AND # A 2-input and gate AND3 AND AND4 AND AND5 AND AND6 AND xnor(test) XNOR
Note that for primitive gate symbols, pin mapping is automatic. The Dataflow window looks in the current working directory and inside each library referenced by the design for the file dataflow.bsm (.bsm stands for "Built-in Symbol Map"). It will read all files found.
User-defined symbols
You can also define your own symbols using an ASCII symbol library file format for defining symbol shapes. This capability is delivered via Concept Engineerings NlviewTM widget Symlib format. For more specific details on this widget, see www.model.com/ support/documentation/BOOK/nlviewSymlib.pdf. The Dataflow window will search the current working directory, and inside each library referenced by the design, for the file dataflow.sym. Any and all files found will be given to the Nlview widget to use for symbol lookups. Again, as with the built-in symbols, the DU name and optional process name is used for the symbol lookup. Here's an example of a symbol for a full adder:
symbol adder(structural) * DEF \ port a in -loc -12 -15 0 -15 \ pinattrdsp @name -cl 2 -15 8 \ port b in -loc -12 15 0 15 \ pinattrdsp @name -cl 2 15 8 \ port cin in -loc 20 -40 20 -28 \ pinattrdsp @name -uc 19 -26 8 \ port cout out -loc 20 40 20 28 \ pinattrdsp @name -lc 19 26 8 \ port sum out -loc 63 0 51 0 \ pinattrdsp @name -cr 49 0 8 \ path 10 0 0 7 \ path 0 7 0 35 \ path 0 35 51 17 \ path 51 17 51 -17 \ path 51 -17 0 -35 \ path 0 -35 0 -7 \
path 0 -7 10 0
Port mapping is done by name for these symbols, so the port names in the symbol definition must match the port names of the Entity|Module|Process (in the case of the process, its the signal names that the process reads/writes). Important: When you create or modify a symlib file, you must generate a file index. This index is how the Nlview widget finds and extracts symbols from the file. To generate the index, select Tools > Create symlib index (Dataflow window) and specify the symlib file. The file will be rewritten with a correct, up-to-date index.
Configuring window options UM-289
Configuring window options

You can configure several options that determine how the Dataflow window behaves. The settings affect only the current session. Select Tools > Options to open the Dataflow Options dialog box.
See "Dataflow Options dialog" (GR-133) for more details.
UM-291
11 - Measuring code coverage

Chapter contents
Introduction . . . . . . . . . Usage flow for code coverage. . . . Supported types . . . . . . . Important notes about coverage statistics . Enabling code coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-292 UM-292 UM-293 UM-294 UM-295 UM-297 UM-298 UM-300 UM-300 UM-302 UM-301 UM-301 UM-304 UM-305 UM-305 UM-305 UM-306 UM-307 UM-308 UM-309 UM-310 UM-311 UM-314 UM-314 UM-314 UM-315 UM-316 UM-316 UM-317
Viewing coverage data in the Main window . Viewing coverage data in the Source window .
Toggle coverage . . . . . . . . . . Enabling toggle coverage . . . . . . . Excluding nodes from toggle coverage . . . Viewing toggle coverage data in the Objects pane Toggle coverage reporting . . . . . . Setting a coverage threshold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Excluding objects from coverage . . . . Exclude lines/files via the GUI . . . Exclude lines/files with pragmas . . . Exclude lines/files with a filter file . . Exclude lines/rows from UDP truth tables Exclude nodes from toggle statistics . . Reporting coverage data XML output . . Sample reports. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Saving and reloading coverage data From the command line . . From the graphic interface . With the vcover utility . . Coverage statistics details . Condition coverage . Expression coverage . . . . . . .
Note: The functionality described in this chapter requires a coverage license feature in your ModelSim license file. Please contact your Mentor Graphics sales representative if you currently do not have such a feature.
UM-292 11 - Measuring code coverage
Introduction
ModelSim code coverage gives you graphical and report file feedback on which statements, branches, conditions, and expressions in your source code have been executed. It also measures bits of logic that have been toggled during execution. With coverage enabled, ModelSim counts how many times each executable statement, branch, condition, expression, and logic node in each instance is executed during simulation. Statement coverage counts the execution of each statement on a line individually, even if there are multiple statements in a line. Branch coverage counts the execution of each conditional "if/then/else" and "case" statement and indicates when a true or false condition has not executed. Condition coverage analyzes the decision made in "if" and ternary statements and is an extension to branch coverage. Expression coverage analyzes the expressions on the right hand side of assignment statements, and is similar to condition coverage. Toggle coverage counts each time a logic node transitions from one state to another. Coverage statistics are displayed in the Main, Objects, and Source windows and also can be output in different text reports (see "Reporting coverage data" (UM-309)). Raw coverage data can be saved and recalled, or merged with coverage data from the current simulation (see "Saving and reloading coverage data" (UM-314)). ModelSim code coverage offers these benefits: It is totally non-intrusive because its integrated into the ModelSim engine it doesnt require instrumented HDL code as do third-party coverage products. It has very little impact on simulation performance (typically 10 to 20 percent). It allows you to merge sets of coverage data without requiring elaboration of the design or a simulation license.
Usage flow for code coverage

The following is an overview of the usage flow for simulating with code coverage. More detailed instructions are presented in the sections that follow. 1 Compile the design using the -cover bcest argument to vcom (CR-246) or vlog (CR-293). 2 Simulate the design using the -coverage argument to vsim (CR-305). 3 Run the design. 4 Analyze coverage statistics in the Main, Objects, and Source windows. 5 Edit the source code to improve coverage. 6 Re-compile, re-simulate, and re-analyze the statistics and design.
Introduction UM-293
Supported types
ModelSim code coverage supports the following data types.
VHDL
Supported types for toggle coverage are boolean, bit, std_logic, enum, integer and arrays of these types. Counts are recorded for each enumeration value and a signal is considered toggled if all the enumerations have non-zero counts. For VHDL integers, a record is kept of each value the integer assumes and an associated count. The maximum number of values recorded is determined by a limit variable that can be changed on a per-signal basis. The default is 100 values. The limit variable can be turned off completely with the toggleNoIntegers option for the vsim command (CR-305). The limit variable can be increased by setting the vsim command line option -toggleMaxIntValues, setting ToggleMaxIntValues in the modelsim.ini file, or setting the Tcl variable ToggleMaxIntValues. Condition and expression coverage supports bit and boolean types. Arbitrary types are supported when they involve a relational operator with a boolean result. These types of subexpressions are treated as an external expression that is first evaluated and then used as a boolean input to the full condition. The subexpression can look like:
(var <relop> const)
or:
(var1 <relop> var2)
where var, var1 and var2 may be of any type; <relop> is a relational operator (e.g.,==,<,>,>=); and const is a constant of the appropriate type. Logical operators (e.g.,and,or,xor) are supported for std_logic/std_ulogic, bit and boolean variable types.
Verilog
Supported types for toggle coverage are net, register, bit, enum and integer (which includes shortint, int, longint, byte, integer and time). For condition and expression coverage, as in VHDL, arbitrary types are supported when they involve a relational operator with a boolean result. Logical operator (e.g.,&&,||,^) are supported for one-bit net, logic and reg types.
SystemC
Code coverage does not work on SystemC design units.
Important notes about coverage statistics

You should be aware of the following special circumstances related to calculating coverage statistics: When ModelSim optimizes a design, it "removes" unnecessary lines of code (e.g., code in a procedure that is never called). The lines that are optimized away aren't counted in the coverage data, and this may cause misleading results. As a result, when you compile with coverage enabled, ModelSim disables certain optimizations depending on which coverage types you choose. This produces more accurate statistics but also may slow simulation. The table below shows the coverage types and what ModelSim does to optimizations. Coverage type statement branch condition expression toggle Effect on optimizations optimizations not disabled automatically; specify -O0 to get most accurate statistics case statement optimizations are disabled automatically optimizations not disabled automatically all optimizations disabled automatically optimizations not disabled automatically
Package bodies are not instance-specific: ModelSim sums the counts for all invocations no matter who the caller is. Also, all standard and accelerated packages are ignored for coverage statistics calculation. Design units compiled with -nodebug are ignored, as if they were excluded. When condition coverage is enabled, expression short circuiting in VHDL is disabled. This can result in simulation errors. For example:
if ( (i /= 0) AND (10/i > 1) ) then
will never produce a divide-by-0 error in normal VHDL because the test (i/=0) will be FALSE if i is 0 and 10/i will never be done. With condition coverage enabled both expressions (i /= 0) and (10/i > 1) will always be evaluated and if i is 0, there will be a divide-by-zero error reported. Coverage data is not collected for generate blocks. You may find that design units or instances excluded from code coverage will appear in toggle coverage statistics reports. This happens when ports of the design unit or instance are connected to nets that have toggle coverage turned on elsewhere in the design.
Enabling code coverage UM-295
Enabling code coverage

Enabling code coverage is a two-step process: 1 Use the -cover argument to vcom or vlog when you compile your design. This argument tells ModelSim which coverage statistics to collect. For example:
vlog top.v proc.v cache.v -cover bcesx
Each character after the -cover argument identifies a type of coverage statistic: "b" indicates branch, "c" indicates condition, "e" indicates expression, "s" indicates statement, "t" indicates 2-transition toggle, and "x" indicates extended 6-transition toggle coverage (t and x are mutually exclusive). See "Enabling toggle coverage" (UM-300) for details on two other methods for enabling toggle coverage. You can use graphic interface to perform the same task. Select Compile > Compile Options and select the Coverage tab. Alternatively, if you are using a project, right-click on a selected design object (or objects) and select Properties.
If you check Ignore VHDL Subprograms, code coverage collection is disabled for VHDL subprograms.
2 Use the -coverage argument to vsim when you simulate your design. For example:
vsim -coverage work.top
Or, use the graphic interface. Select Simulate > Start Simulation and select the design unit to be simulated in the Design tab. Then select the Others tab and check Enable code coverage box as shown below.
Viewing coverage data in the Main window UM-297
Viewing coverage data in the Main window

When you simulate a design with code coverage enabled, coverage data is displayed in the Main, Source, and Objects windows. In the Main window, coverage data displays in five window panes: Workspace, Missed Coverage, Current Exclusions, Instance Coverage, and Details.
Workspace
Missed Coverage
Current Exclusions
Instance Coverage
Details
The table below summarizes the Main window coverage panes. For further details, see "Code coverage panes" (GR-109). Coverage pane Workspace Missed Coverage Current exclusionsa Instance coverage Details Description displays coverage data and graphs for each design object or file displays missed coverage for the selected design object or file lists all files and lines that are excluded from the current analysis displays coverage statistics for each instance in a flat format displays details of missed coverage and toggle coverage
a.The Current Exclusions pane does not display by default. Select View > Code Coverage > Current Exclusions to display the pane.
Viewing coverage data in the Source window

The Source window (GR-182) includes two columns for code coverage statistics the Hits column and the BC (Branch Coverage) column. These columns provide an immediate visual indication about how your source code is executing. The default code coverage indicators are check marks and Xs. A green check mark indicates that the statements and/or branches in a particular line have executed. A red X indicates that a statement or branch was not executed. An XT indicates the true branch of an conditional statement was not executed. An XF indicates the false branch was not executed. A green "E" indicates a line of code that has been excluded from code coverage statistics.
When you hover the cursor over a line of code (see line 58 in the illustration above), the number of statement and branch executions, or "hits," will be displayed in place of the check marks and Xs. If you prefer, you can display only numbers by selecting Tools > Code Coverage > Show Coverage Numbers. Also, when you click in either the Hits or BC column, the Details pane in the Main window updates to display information on that line. You can skip to "missed lines" three ways: select Edit > Previous Coverage Miss and Edit > Next Coverage Miss from the menu bar; click the Previous zero hits and Next zero hits icons on the toolbar; or press <Shift> - <Tab> (previous miss) or Tab (next miss).
Viewing coverage data in the Source window UM-299
Controlling data display in a Source window

The Tools > Code Coverage menu contains several commands for controlling coverage data display in a Source window. Hide/Show coverage data toggles the Hits column off and on. Hide/Show branch coverage toggles the BC column off and on. Hide/Show coverage numbers displays the number of executions in the Hits and BC columns rather than checkmarks and Xs. When multiple statements occur on a single line an ellipsis ("...") replaces the Hits number. In such cases, hover the cursor over each statement to highlight it and display the number of executions for that statement. Show coverage By Instance displays only the number of executions for the currently selected instance in the Main window workspace.
Toggle coverage
Toggle coverage is the ability to count and collect changes of state on specified nodes, including Verilog nets and registers and the following VHDL signal types: bit, bit_vector, std_logic, and std_logic_vector. Toggle coverage is integrated as a metric into the coverage tool so that the use model and reporting are the same as the other coverage metrics. There are two modes of toggle coverage operation - standard and extended. Standard toggle coverage only counts Low or 0 <--> High or 1 transitions. Extended toggle coverage counts these transitions plus the following: Z <--> 1 or H Z <--> 0 or L Extended coverage allows a more detailed view of testbench effectiveness and is especially useful for examining coverage of tri-state signals. It helps to ensure, for example, that a bus has toggled from high 'Z' to a '1' or '0', and a '1' or '0' back to a high 'Z'. Toggle coverage ignores zero-delay glitches.
Enabling toggle coverage

In the Enabling code coverage (UM-295) section we explained that toggle coverage could be enabled during compile by using the t or x arguments with vcom -cover or vlog -cover. This section describes two other methods for enabling toggle coverage: 1 using the toggle add command (CR-216) 2 using the Tools > Toggle Coverage > Add or Tools > Toggle Coverage > Extended selections in the Main window menu.
Using the toggle add command

The toggle add command allows you to initiate toggle coverage at any time from the command line. (See the Command Reference (CR-216) for correct syntax and arguments.) Upon the next running of the simulation, toggle coverage data will be collected according to the arguments employed (i.e., the -full argument enables collection of extended toggle coverage statistics for the transitions mentioned above). If you use a toggle add command (CR-216) on a group of signals to get standard toggle coverage, then try to convert to extended toggle coverage with the toggle add -full command on the same signals, nothing will change. The only way to change the internal toggle triggers from standard to extended toggle coverage is to restart vsim and use the correct command.
Using the Main window menu selections

You can enable toggle coverage by selecting Tools > Toggle Coverage > Add or Tools > Toggle Coverage > Extended from the Main window menu. These selections allow you to enable toggle coverage for Selected Signals, Signals in Region, or Signals in Design. After making a selection, toggle coverage statistics will be captured the next time you run the simulation.
Toggle coverage UM-301
Viewing toggle coverage data in the Objects pane

To view toggle coverage data in the Objects pane right-click in the pane to open a context popup menu the Toggle Coverage selection. When highlighted, this selection allows you to display toggle coverage data for the Selected Signals, the Signals in Region, or the Signals in Design. Toggle coverage data is displayed in the Objects pane in multiple columns, as shown below. There is a column for each of the six transition types. Right click any column name to toggle that column on or off. See "Objects pane toggle coverage" (GR-118) for more details on each column.
Toggle coverage reporting

The toggle report command (CR-220) displays a list of all nodes and the counts for how many times they toggled for each state transition type. Also displayed is a summary of the number of nodes checked, the number that toggled, the number that didn't toggle, and a percentage that toggled. The toggle report command is intended to be used as follows: 1 Enable statistics collection with the toggle add command (CR-216). 2 Run the simulation with the run command (CR-197).
3 Produce the report with the toggle report command..
You can produce this same information using the coverage report command (CR-106).
Port collapsing and toggle coverage

The simulator collapses certain ports that are connected to the same signal in order to improve performance, and those collapsed signals will not appear in the report. If you want to ensure that you are reporting all signals in the design, use the -nocollapse argument to vsim when you load your design. The -nocollapse argument degrades simulator performance, so it should be used only when it is absolutely necessary to see all signals in a toggle report.
Excluding nodes from toggle coverage

You can disable toggle coverage with the toggle disable command (CR-218). This command disables toggle statistics collection on the specified nodes and provides a method of implementing coverage exclusions for toggle coverage. It is intended to be used as follows: 1 Enable toggle statistics collection for all signals using the -cover t/x argument to vcom or vlog. 2 Exclude certain signals by disabling them with the toggle disable command. The toggle enable command (CR-219) re-enables toggle statistics collection on nodes whose toggle coverage has previously been disabled via the toggle disable command. (See the Command Reference for correct syntax.)
Toggle coverage UM-303
Excluding bus bits from toggle coverage

You can exclude bus bits from toggle coverage using special source code pragmas.
Verilog syntax
// coverage toggle_ignore <simple_signal_name> "<list>"
VHDL syntax
-- coverage toggle_ignore <simple_signal_name> "<list>"
The following rules apply to using these pragmas: <list> is a space-separated list of bit indices or ranges, where a range is two integers separated by ':' or '-'. If using a range, the range must be in the same ascending or descending order as the signal declaration. Quotes are required around the list. The pragma must be placed within the declarative region of the module or architecture in which <simple_signal_name> is declared.
Excluding VHDL enum signals from toggle statistics

You can exclude individual enums or ranges of enums from toggle coverage and reporting by specifying enum exclusions in source code pragmas or by using the -exclude argument to the toggle add command (CR-216).
Setting a coverage threshold

You can specify a percentage above or below which you dont want to see coverage statistics. For example, you might set a threshhold of 85% such that only objects with coverage below that percentage are displayed. Anything above that percentage is filtered. You can set a filter using either a dialog or toolbar icons (see below). To access the dialog, right-click any object in the Instance Coverage pane and select Set filter.
See "Filter instance list dialog" (GR-93) for details on this dialog.
Excluding objects from coverage UM-305
Excluding objects from coverage

You can exclude any number of lines or files, or condition and expression UDP truth table rows, from coverage statistics collection. The line exclusions can be instance-specific or they can apply to all instances in the enclosing design unit. Truth table row exclusions cannot be instance specific at this time. You can also exclude nodes from toggle statistics collection using the toggle disable command (CR-218). The following methods can be used for excluding objects: Exclude lines/files via the GUI (UM-305) Exclude lines/files with pragmas (UM-305) Exclude lines/files with a filter file (UM-306) Exclude lines/rows from UDP truth tables (UM-307) Exclude lines/rows with the coverage exclude command (UM-307) Exclude nodes from toggle statistics (UM-308) Exclude VHDL enum signals from toggle statistics (UM-308)
Exclude lines/files via the GUI

There are several locations in the GUI where you can access commands to exclude lines or files: Right-click a file in Files tab of the Workspace pane and select Code Coverage > Exclude Selected File from the popup menu. Right-click an entry in the Main window Missed Coverage pane and select Exclude Selection or Exclude Selection For Instance <inst_name> from the popup menu. Right-click a line in the Hits column of the Source window and select Exclude Coverage Line xxx, Exclude Coverage Line xxx For Instance <inst_name>, or Exclude Entire File.
Exclude lines/files with pragmas

ModelSim also supports the use of source code pragmas to selectively turn coverage off and on. In Verilog, the pragmas are:
// coverage off // coverage on
In VHDL, the pragmas are:

-- coverage off -- coverage on
Bracket the line or lines you want to exclude with these pragmas. Here are some points to keep in mind about using these pragmas: Pragmas are enforced at the design unit level only. For example, if you put "-- coverage off" before an architecture declaration, all statements in that architecture will be excluded from coverage; however, statements in all following design units will be included in statement coverage (until the next "--coverage off").
Pragmas cannot be used to exclude specific subconditions or subexpressions within lines, although they can be used for individual case statement alternatives.
Exclude lines/files with a filter file

Exclusion filter files specify files and line numbers or condition and expression UDP truth table rows (see below for details) that you wish to exclude from coverage statistics. You can create the filter file in any text editor or save the current filter in the Current Exclusions pane by selecting the pane and then choosing File > Save. To load the filter during a future analysis, select the Current Exclusions pane and select File > Open.
Syntax
# Comment <filename>... [[<range> ...] [<line#> ...]] | all
or
begin instance <instance_name>... <inst_filename>... [[<range> ...] [<line#> ...]] | all end instance
Arguments
#
Comment character.
<filename>
The name of the file you want to exclude. Required if you are not specifying an instance. The filter file may include an unlimited number of filename entries, each on its own line. You may use environment variables in the pathname.
begin instance <instance_name>
The name of an instance for which you want to exclude lines. Required if you dont specify <filename>. The filter file may include an unlimited number of instances.
<inst_filename>
The name of the file(s) that compose the instance from which you are excluding lines. Required, unless all is specified.
<range> ...
A range of line numbers you want to exclude. Optional. Enter the range in "#-#" format. For example, 32-35. You can specify multiple ranges separated by spaces.
<line#> ...
A line number that you want to exclude. Optional. You can specify multiple line numbers separated by spaces.
all
When used with <filename>, specifies that all lines in the file should be excluded. When used with <instance_name>, specifies that all lines in the instance and all instances contained within the specified instance should be excluded. Required if a range or line number is not specified.
Excluding objects from coverage UM-307
Example
control.vhd 72-76 84 93 testring.vhd all begin instance /test_delta/chip/bid01_inst src/delta/buffers.vhd 45-46 end instance
Default filter file

The Tcl preference variable PrefCoverage(pref_InitFilterFrom) specifies a default filter file to read when a design is loaded with the -coverage switch. By default this variable is not set. See "Simulator GUI preferences" (GR-251) for details on changing this variable. A file named workingExclude.cov appears in the design directory when you specify exclusions in the GUI. This file remains after quitting simulation.
Exclude lines/rows from UDP truth tables

You can also use exclusion filter files to exclude lines and rows from condition and expression UDP truth tables.
Syntax
-c | -e {<ln> <rn|rn1-rn2>...}
Arguments
-c | -e
Determines whether to exclude condition (-c) or expression (-e) UDP truth table rows.
<ln> ...
The line number containing the condition or expression. The line number and list of row numbers are surrounded by curly braces.
<rn | rn1 - rn2>
A space separated list of row numbers or ranges of row numbers referring to the UDP truth table rows that you want excluded.
Example
control.vhd 72-76 -c {78 1 3-6}
In this example, lines 72 through 76 will be excluded from code coverage in control.vhd file. Rows 1 and 3 through 6 in the condition truth table on line 78 will also be excluded.
Exclude lines/rows with the coverage exclude command

You can use the coverage exclude command (CR-103) for direct exclusion of specific lines in a source file or rows within a condition or expression truth table.
Exclude nodes from toggle statistics

To exclude nodes from toggle statistics collection, use the toggle disable command (CR218).
Exclude VHDL enum signals from toggle statistics

You can exclude individual enums or ranges of enums from toggle coverage and reporting by specifying enum exclusions in source code pragmas or by using the -exclude argument to the toggle add command (CR-216).
Reporting coverage data UM-309
Reporting coverage data

You have several options for creating reports on coverage data. To create reports when a simulation is loaded, use either the coverage report command (CR-106), the toggle report command (CR-220) (see Toggle coverage reporting (UM-301) in this chapter), or the Coverage Report dialog. To create reports when a simulation isnt loaded, use the vcover report command (CR-258).
Command example
Here is a sample command sequence that outputs a code coverage report and saves the coverage data:
vlog vlog vlog vlog -cover -cover -cover -cover bcesx bcesx bcesx bcesx ../rtl/host/top.v ../rtl/host/a.v ../rtl/host/b.v ../rtl/host/c.v
vsim -c -coverage top run 1 ms coverage report -file d:\\sample\\coverage_rep coverage save d:\\sample\\coverage
GUI example
To access the Coverage Report dialog, right-click any object in the Files tab of the Workspace pane and select Code Coverage > Code Coverage Reports; or, select Tools > Code Coverage > Reports.
See "Coverage Report dialog" (GR-90) for details on this dialog.
Setting a default coverage mode

You can specify a default coverage mode that persists from one ModelSim session to the next through the preference variable PrefCoverage(DefaultCoverageMode). The modes available allow you to specify that lists and data given in each report are listed by: file, design unit, or instance. By default, the report is listed by file. See "Simulator GUI preferences" (GR-251) for details on changing this variable.
XML output
You can output coverage reports in XML format by checking Write XML Format in the Coverage Report dialog or by using the -xml argument to the coverage report command (CR-106). The following example is an abbreviated "By Instance" report that includes line details:
<?xml version="1.0"?> <report lines="1" byInstance="1"> <instance path="/test_delta/chip/control_126k_inst" du="mode_two_control"> <source_table files="1"> <file fn="0" path="C:/modelsim_examples/coverage/Modetwo.v"></file> </source_table> <statements active="30" hits="17" percent="56.7"> </statements> <statement_data> <stmt fn="0" ln="39" st="1" hits="82"> </stmt> <stmt fn="0" ln="42" st="1" hits="82"> </stmt> <stmt fn="0" ln="44" st="1" hits="82"> </stmt>
"fn" stands for filename, "ln" stands for line number, and "st" stands for statement. There is also an XSL stylesheet named covreport.xsl located in <install_dir>/modeltech/ examples/misc. Use it as a foundation for building your own customized report translators.
Sample reports
Below are abbreviated coverage reports with descriptions of select fields.
Zero counts report by file
The "%" field shows the percentage of statements in the file that had zero coverage.
Instance report with line details
The "Stmt" field identifies the number of statements with zero coverage on that line.
Branch report
If an IF Branch ends in an "else" clause, the "else" count will be shown. Otherwise, an "All False" count will be given, which indicates how many times none of the conditions evaluated "true." If "INF" appears in the Count column, it indicates that the coverage count has exceeded ~4 billion (232 -1). ***0*** indicates a zero count for that branch.
Saving and reloading coverage data

Raw coverage data can be saved and then reloaded later. Saved data can also be merged with coverage statistics from the current simulation or previously loaded data. You can perform these operations via the command line, the graphic interface, or the $coverage_save Verilog system task (see "System tasks and functions specific to ModelSim" (UM-128)).
From the command line

The coverage save command (CR-110) saves current coverage statistics to a file that can be reloaded later, preserving instance-specific information. The coverage reload command (CR-105) seeds the coverage statistics of the current simulation with the output of a previous coverage save command. This allows you, for example, to gather statistics from multiple simulation runs into a single report. The -incremental option adds the data to the data currently in memory.
From the graphic interface

To save raw coverage data, select Tools > Code Coverage > Save.
The "Save on exit" field causes coverage data to be saved automatically when the simulator exits.
Saving and reloading coverage data UM-315
To reload previously saved coverage data, select Tools > Coverage > Load.
See "Load Coverage Data dialog" (GR-89) for details on this dialog.
With the vcover utility

The vcover code coverage utility supports instance-specific toggles and summing the instances of each design unit. It also supports source code annotation and printing of condition and expression truth tables. The merge utility, vcover merge, allows you to merge sets of coverage data without first loading a design. It is a standard ModelSim utility that can be invoked from within the GUI or from the command line. See the vcover merge command (CR-255) in the ModelSim Command Reference for further details.
Coverage statistics details

This section describes how condition and expression coverage statistics are calculated.
Condition coverage
Condition coverage analyzes the decision made in "if" and ternary statements and is an extension to branch coverage. A truth table is constructed for the condition expression and counts are kept for each row of the truth table that occurs. For example, the following IF statement:
Line 180: IF (a or b) THEN x := 0; else x := 1; endif;
reflects this truth table. Truth table for line 180 counts Row 1 Row 2 Row 3 unknown 5 0 8 0 a 1 0 b 1 0 (a or b) 1 1 0
Row 1 indicates that (a or b) is true if a is true, no matter what b is. The "counts" column indicates that this combination has executed 5 times. The '-' character means "don't care." Likewise, row 2 indicates that the result is true if b is true no matter what a is, and this combination has executed zero times. Finally, row 3 indicates that the result is always zero when a is zero and b is zero, and that this combination has executed 8 times. The unknown row indicates how many times the line was executed when one of the variables had an unknown state. If more than one row matches the input, each matching row will be counted. If that is not the behavior you want and you would prefer no counts to be incremented on multiple matches, set "CoverCountAll=0" in your modelsim.ini file. Values that are vectors are treated as subexpressions external to the table until they resolve to a boolean result. For example, take the IF statement:
Line 38:IF ((e = '1') AND (bus = "0111")) ...
Coverage statistics details UM-317
A truth table will be generated in which bus = "0111" is evaluated as a subexpression and the result, which is boolean, becomes an input to the truth table. The truth table looks as follows: Truth table for line 38 counts Row 1 Row 2 Row 3 unknown 0 10 1 0 e 0 1 (bus="0111") 0 1 (e=1) AND ( bus = "0111") 0 0 1
Index expressions also serve as inputs to the table. Conditions containing function calls cannot be handled and will be ignored for condition coverage. If a line contains a condition that is uncovered - some part of its truth table was not encountered - that line will appear in the Missed Coverage pane under the Conditions tab. When that line is selected, the condition truth table will appear in the Details pane and the line will be highlighted in the Source window. Condition coverage truth tables are printed in coverage reports when the Condition Coverage type is selected in the Coverage Reports dialog (see "Reporting coverage data" (UM-309)), or when the -lines argument is specified in the coverage report command (CR106) and one or more of the rows has a zero hit count. To force the table to be printed even when it is 100% covered, use the -dump argument to the coverage report command.
Expression coverage
Expression coverage analyzes the expressions on the right hand side of assignment statements and counts when these expressions are executed. For expressions that involve logical operators, a truth table is constructed and counts are tabulated for conditions matching rows in the truth table. For example, take the statement:
Line 236: x <= a xor (not b(0));
This statement results in the following truth table, with associated counts. Truth table for line 236 counts Row 1 Row 2 Row 3 Row 4 unknown 1 0 2 0 0 a 0 0 1 1 b(0) 0 1 0 1 (a xor (not b(0))) 1 0 0 1
If a line contains an expression that is uncovered (some part of its truth table was not encountered) that line will appear in the Missed Coverage pane under the Expressions tab. When that line is selected, the expression truth table will appear in the Details pane and the line will be highlighted in the Source window. As with condition coverage, expression coverage truth tables are printed in coverage reports when the Expression Coverage type is selected in the Coverage Reports dialog (see "Reporting coverage data" (UM-309)) or when the -lines argument is specified in the coverage report command (CR-106) and one or more of the rows has a zero hit count. To force the table to be printed even when it is 100% covered, use the -dump argument for the coverage report command (CR-106).
UM-319
12 - C Debug
Chapter contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-320 UM-321 UM-321 UM-322 UM-322 UM-323 UM-325 UM-325 UM-326 UM-327 UM-327 UM-328 UM-329 UM-330 UM-331 UM-331 UM-333 UM-333 UM-334 UM-335 Supported platforms and gdb versions . . . Running C Debug on Windows platforms Setting up C Debug . . . . . Running C Debug from a DO file . Setting breakpoints. . . . . . . . . . . .
Stepping in C Debug . . . . . . . . Known problems with stepping in C Debug . Finding function entry points with Auto find bp . Identifying all registered function calls Enabling Auto step mode . . . Example . . . . . . . Auto find bp versus Auto step mode Debugging functions during elaboration FLI functions in initialization mode PLI functions in initialization mode VPI functions in initialization mode Completing design load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Debugging functions when quitting simulation C Debug command reference . . . . .
Note: The functionality described in this chapter requires a cdebug license feature in your ModelSim license file. Please contact your Mentor Graphics sales representative if you currently do not have such a feature.
UM-320 12 - C Debug
Introduction
C Debug allows you to interactively debug FLI/PLI/VPI/SystemC C/C++ source code with the open-source gdb debugger. Even though C Debug doesnt provide access to all gdb features, you may wish to read gdb documentation for additional information. Please be aware of the following caveats before using C Debug: C Debug is an interface to the open-source gdb debugger. We have not customized gdb source code, and C Debug doesnt remove any of the limitations or bugs of gdb. We assume that you are competent with C or C++ coding and C debugging in general. Recommended usage is that you invoke C Debug once for a given simulation and then quit both C Debug and ModelSim. Starting and stopping C Debug more than once during a single simulation session may cause problems for gdb. The gdb debugger has a known bug that makes it impossible to set breakpoints reliably in constructors or destructors. Be careful while stepping through code which may end up calling constructors of SystemC objects; it may crash the debugger. Generally you should not have an existing .gdbinit file. If you do, make certain you havent done any of the following: defined your own commands or renamed existing commands; used 'set annotate...', 'set height...', 'set width...', or 'set print...'; set breakpoints or watchpoints. To use C Debug on Windows platforms, you must compile your source code with gcc/ g++. See "Running C Debug on Windows platforms" (UM-321) below.
Supported platforms and gdb versions UM-321
Supported platforms and gdb versions

ModelSim ships with the gdb 6.0 debugger. Testing has shown this version to be the most reliable for SystemC applications. However, for FLI/PLI applications, you can also use a current installation of gdb if you prefer. C Debug has been tested on these platforms with these versions of gdb: Platform 32-bit Solaris 2.6, 7, 8, 9 32- and 64-bit HP-UX 11.0a, 11.11b 64-bit HP-UX B.11.22 on Itanium 2 32-bit AIX 4.2, 4.3 32-bit Redhat Linux 7.2 or later 32-bit Windows NT and NT-based platforms (XP, win2K, etc.) Opteron / SuSE Linux 9.0 or Redhat EWS 3.0 (32-bit mode only) x86 / Redhat Linux 6.0 to 7.1 Opteron & Athlon 64 / Redhat EWS 3.0 Required gdb version gdb-5.0-sol-2.6 wdb version 3.3 or later wdb version 4.2 gdb-5.1-aix-4.2 /usr/bin/gdb 5.2 or later gdb 6.0 from MinGW-32 gdb 6.0 or later /usr/bin/gdb 5.2 or later gdb 5.3.92 or 6.1.1
a.You must install kernel patch PHKL_22568 (or a later patch that supersedes PHKL_22568) on HP-UX 11.0. If you do not, you will see the following error message when trying to enable C Debug:
# Unable to find dynamic library list. # error from C debugger
b.You must install B.11.11.0306 Gold Base Patches for HP-UX 11i, June 2003. To invoke C Debug, you must have the following: A cdebug license feature; contact Model Technology sales for more information. The correct gdb debugger version for your platform.
Running C Debug on Windows platforms

To use C Debug on Windows, you must compile your C/C++ source code using the gcc/g++ compiler supplied with ModelSim. Source compiled with Microsoft Visual C++ is not debuggable using C Debug. The g++ compiler is installed in the following location:
../modeltech/gcc-3.2.3-mingw32/
UM-322 12 - C Debug
Setting up C Debug
Before viewing your SystemC/C/C++ source code, you must set up the C Debug path and options. To set up C Debug, follow these steps: 1 Compile and link your C code with the -g switch (to create debug symbols) and without -O (or any other optimization switches you normally use). See Chapter 6 - SystemC simulation for information on compiling and linking SystemC code. See the FLI Reference Manual or Appendix C - Verilog PLI / VPI / DPI for information on compiling and linking C code. 2 Specify the path to the gdb debugger by selecting Tools > C Debug > C Debug Setup.
Select "default" to point at the Model Technology supplied version of gdb or "custom" to point at a separate installation. 3 Start the debugger by selecting Tools > C Debug > Start C Debug. ModelSim will start the debugger automatically if you set a breakpoint in a SystemC file. 4 If you are not using gcc, or otherwise havent specified a source directory, specify a source directory for your C code with the following command:
ModelSim> gdb dir <srcdirpath1>[:<srcdirpath2>[...]]
Running C Debug from a DO file

You can run C Debug from a DO file but there is a configuration issue of which you should be aware. It takes C Debug a few moments to start-up. If you try to execute a run command before C Debug is fully loaded, you may see an error like the following:
# # # # # ** Error: Stopped in C debugger, unable to real_run mti_run 10us Error in macro ./do_file line 8 Stopped in C debugger, unable to real_run mti_run 10us while executing "run 10us
In your DO file, add the command cdbg_wait_for_starting to alleviate this problem. For example:
cdbg enable_auto_step on cdbg set_debugger /modelsim/5.8c_32/common/linux cdbg debug_on cdbg_wait_for_starting run 10us
Setting breakpoints UM-323
Setting breakpoints
Breakpoints in C Debug work much like normal HDL breakpoints. You can create and edit them with ModelSim commands (bp (CR-61), bd (CR-56), enablebp (CR-130), disablebp (CR-124)) or via a Source window in the ModelSim GUI (see "File-line breakpoints" (GR240)). Some differences do exist: The Breakpoints dialog in the ModelSim GUI doesnt list C breakpoints. C breakpoint id numbers require a "c." prefix when referenced in a command. When using the bp command (CR-61) to set a breakpoint in a C file, you must use the -c argument. Here are some example commands:
bp -c *0x400188d4
Sets a C breakpoint at the hex address 400188d4. Note the * prefix for the hex address.
bp -c or_checktf
Sets a C breakpoint at the entry to function or_checktf.

bp -c or.c 91
Sets a C breakpoint at line 91 of or.c.

enablebp c.1
Enables C breakpoint number 1. The graphic below shows a C file with one enabled breakpoint (on line 44) and one disabled breakpoint (on line 48).
UM-324 12 - C Debug
Clicking the red diamonds with your right (third) mouse button pops up a menu with commands for removing or enabling/disabling the breakpoints
Note: The gdb debugger has a known bug that makes it impossible to set breakpoints reliably in constructors or destructors. Do not set breakpoints in constructors of SystemC objects; it may crash the debugger.
Stepping in C Debug UM-325
Stepping in C Debug
Stepping in C Debug works much like you would expect. You use the same buttons and commands that you use when working with an HDL-only design. Button Step steps the current simulation to the next statement; if the next statement is a call to a C function that was compiled with debug info, ModelSim will step into the function Step Over statements are executed but treated as simple statements instead of entered and traced line-by-line; C functions are not stepped into unless you have an enabled breakpoint in the C file Continue Run continue the current simulation run until the end of the specified run length or until it hits a breakpoint or specified break event Menu equivalent Tools > C Debug > Run > Step Other equivalents use the step command at the CDBG> prompt see: step (CR-213) command
Tools > C Debug > Run > Step -Over
use the step -over command at the CDBG> prompt see: step (CR-213) command
Tools > C Debug > Run > Continue
use the run -continue command at the CDBG> prompt see: run (CR-197)
Known problems with stepping in C Debug

The following are known limitations which relate to problems with gdb: The gdb debugger has a known bug that makes it impossible to set breakpoints reliably in constructors or destructors. Be careful while stepping through code which may end up calling constructors of SystemC objects; it may crash the debugger. With some platform and compiler versions, step may actually behave like run -continue when in a C file. This is a gdb quirk that results from not having any debugging information when in an internal function to VSIM (i.e., any FLI or VPI function). In these situations, use step -over to move line-by-line.
UM-326 12 - C Debug
Finding function entry points with Auto find bp

ModelSim can automatically locate and set breakpoints at all currently known function entry points (i.e., PLI/VPI system tasks and functions and callbacks; and FLI subprograms and callbacks and processes created with mti_CreateProcess). Select Tools > C Debug > Auto find bp to invoke this feature. The Auto find bp command provides a "snapshot" of your design when you invoke the command. If additional callbacks get registered later in the simulation, ModelSim will not identify these new function entry points unless you re-execute the Auto find bp command. If you want functions to be identified regardless of when they are registered, use "Identifying all registered function calls" (UM-327) instead. The Auto find bp command sets breakpoints in an enabled state and doesnt toggle that state to account for step -over or run -continue commands. This may result in unexpected behavior. For example, say you have invoked the Auto find bp command and you are currently stopped on a line of code that calls a C function. If you execute a step -over or run -continue command, ModelSim will stop on the breakpoint set in the called C file.
Identifying all registered function calls UM-327
Identifying all registered function calls

Auto step mode automatically identifies and sets breakpoints at registered function calls (i.e., PLI/VPI system tasks and functions and callbacks; and FLI subprograms and callbacks and processes created with mti_CreateProcess). Auto step mode is helpful when you are not entirely familiar with a design and its associated C routines. As you step through the design, ModelSim steps into and displays the associated C file when you hit a C function call in your HDL code. If you execute a step -over or run -continue command, ModelSim does not step into the C code. When you first enable Auto step mode, ModelSim scans your design and sets enabled breakpoints at all currently known function entry points. As you step through the simulation, Auto step continues looking for newly registered callbacks and sets enabled breakpoints at any new entry points it identifies. Once you execute a step -over or run -continue command, Auto step disables the breakpoints it set, and the simulation continues running. The next time you execute a step command, the automatic breakpoints are re-enabled and Auto step sets breakpoints on any new entry points it identifies. Note that Auto step does not disable user-set breakpoints.
Enabling Auto step mode

To enable Auto step mode, follow these steps: 1 Configure C Debug as described in "Setting up C Debug" (UM-322). 2 Select Tools > C Debug > Enable auto step. 3 Load and run your design.
UM-328 12 - C Debug
Example
The graphic below shows a simulation that has stopped at a user-set breakpoint on a PLI system task.
Because Auto step mode is enabled, ModelSim automatically sets a breakpoint in the underlying xor_gate.c file. If you click the step button at this point, ModelSim will step into that file.
Identifying all registered function calls UM-329
Auto find bp versus Auto step mode

As noted in "Finding function entry points with Auto find bp" (UM-326), the Auto find bp command also locates and sets breakpoints at function entry points. Note the following differences between Auto find bp and Auto step mode: Auto find bp provides a "snapshot" of currently known function entry points at the time you invoke the command. Auto step mode continues to locate and set automatic breakpoints in newly registered function calls as the simulation continues. In other words, Auto find bp is static while Auto step mode is dynamic. Auto find bp sets automatic breakpoints in an enabled state and doesnt change that state to account for step-over or run-continue commands. Auto step mode enables and disables automatic breakpoints depending on how you step through the design. In cases where you invoke both features, Auto step mode takes precedence over Auto find bp. In other words, even if Auto find bp has set enabled breakpoints, if you then invoke Auto step mode, it will toggle those breakpoints to account for step-over and run-continue commands.
UM-330 12 - C Debug
Debugging functions during elaboration

Initialization mode (not supported for Windows platforms) allows you to examine and debug functions that are called during elaboration (i.e., while your design is in the process of loading). When you select this mode, ModelSim sets special breakpoints for foreign architectures and PLI/VPI modules that allow you to set breakpoints in the initialization functions. When the design finishes loading, the special breakpoints are automatically deleted, and any breakpoints that you set are disabled (unless you specify Keep user init bps in the C debug setup dialog). To run C Debug in initialization mode, follow these steps: 1 Start C Debug by selecting Tools > C Debug > Start C Debug before loading your design. 2 Select Tools > C Debug > Init mode. 3 Load your design. As the design loads, ModelSim prints to the Transcript the names and/or hex addresses of called functions. For example the Transcript below shows a function pointer to a foreign architecture:
To set a breakpoint on that function, you would type:

bp -c *0x4001b571
or
bp -c and_gate_init
Debugging functions during elaboration UM-331
ModelSim in turn reports that it has set a breakpoint at line 37 of the and_gate.c file. As you continue through the design load using run -continue, ModelSim hits that breakpoint and displays the file and associated line in a Source window.
FLI functions in initialization mode

There are two kinds of FLI functions that you may encounter in initialization mode. The first is a foreign architecture which was shown above. The second is a foreign function. ModelSim produces a Transcript message like the following when it encounters a foreign function during initialization:
# Shared object file './all.sl' # Function name 'in_params' # Function ptr '0x4001a950'. Foreign function. # C breakpoint c.1 # 0x0814fc96 in mti_cdbg_shared_objects_loaded ()
You can set a breakpoint on the function using either the function name (i.e., bp -c in_params) or the function pointer (i.e., bp -c *0x4001a950). Note, however, that foreign functions arent called during initialization. You would hit the breakpoint only during runtime and then only if you enabled the breakpoint after initialization was complete or had specified Keep user init bps in the C debug setup dialog.
PLI functions in initialization mode

There are two methods for registering callback functions in the PLI: 1) using a veriusertfs array to define all usertf entries; and 2) adding an init_usertfs function to explicitly register each usertfs entry (see "Registering DPI applications" (UM-480) for more details). The messages ModelSim produces in initialization mode vary depending on which method you use.
UM-332 12 - C Debug
ModelSim produces a Transcript message like the following when it encounters a veriusertfs array during initialization:
# vsim -pli ./veriuser.sl mux_tb # Loading ./veriuser.sl # Shared object file './veriuser.sl' # veriusertfs array - registering calltf # Function ptr '0x40019518'. $or_c. # C breakpoint c.1 # 0x0814fc96 in mti_cdbg_shared_objects_loaded cont # Shared object file './veriuser.sl' # veriusertfs array - registering checktf # Function ptr '0x40019570'. $or_c. # C breakpoint c.1 # 0x0814fc96 in mti_cdbg_shared_objects_loaded cont # Shared object file './veriuser.sl' # veriusertfs array - registering sizetf # Function ptr '0x0'. $or_c. # C breakpoint c.1 # 0x0814fc96 in mti_cdbg_shared_objects_loaded cont # Shared object file './veriuser.sl' # veriusertfs array - registering misctf # Function ptr '0x0'. $or_c. # C breakpoint c.1 # 0x0814fc96 in mti_cdbg_shared_objects_loaded
()
()
()
()
You can set breakpoints on non-null callbacks using the function pointer (e.g., bp -c *0x40019570). You cannot set breakpoints on null functions. The sizetf and misctf entries in the example above are null (the function pointer is '0x0'). ModelSim reports the entries in multiples of four with at least one entry each for calltf, checktf, sizetf, and misctf. Checktf and sizetf functions are called during initialization but calltf and misctf are not called until runtime. The second registration method uses init_usertfs functions for each usertfs entry. ModelSim produces a Transcript message like the following when it encounters an init_usertfs function during initialization:
# Shared object file './veriuser.sl' # Function name 'init_usertfs' # Function ptr '0x40019bec'. Before first call of init_usertfs. # C breakpoint c.1 # 0x0814fc96 in mti_cdbg_shared_objects_loaded ()
You can set a breakpoint on the function using either the function name (i.e., bp -c init_usertfs) or the function pointer (i.e., bp -c *0x40019bec). ModelSim will hit this breakpoint as you continue through initialization.
Debugging functions during elaboration UM-333
VPI functions in initialization mode

VPI functions are registered via routines placed in a table named vlog_startup_routines (see "Registering VPI applications" (UM-478) for more details). ModelSim produces a Transcript message like the following when it encounters a vlog_startup_routines table during initialization:
# Shared object file './vpi_test.sl' # vlog_startup_routines array # Function ptr '0x4001d310'. Before first call using function pointer. # C breakpoint c.1 # 0x0814fc96 in mti_cdbg_shared_objects_loaded ()
You can set a breakpoint on the function using the function pointer (i.e., bp -c *0x4001d310). ModelSim will hit this breakpoint as you continue through initialization.
Completing design load

If you are through looking at the initialization code you can select Tools > C Debug > Complete load at any time, and ModelSim will continue loading the design without stopping. The one exception to this is if you have set a breakpoint in a LoadDone callback and also specified Keep user init bps in the "C Debug setup dialog" (GR-100).
UM-334 12 - C Debug
Debugging functions when quitting simulation

Stop on quit mode allows you to debug functions that are called when the simulator exits. Such functions include those referenced by an mti_AddQuitCB function in FLI code, misctf functions called by a quit or $finish in PLI code, or cbEndofSimulation functions called by a quit or $finish in VPI code. To enable Stop on quit mode, follow these steps: 1 Start C Debug by selecting Tools > C Debug > Start C Debug. 2 Select Tools > C Debug > C Debug Setup. 3 Select Stop on quit in the C Debug setup dialog.
With this mode enabled, if you have set a breakpoint in a quit callback function, C Debug will stop at the breakpoint after you issue the quit command in ModelSim. This allows you to step and examine the code in the quit callback function. Invoke run -continue when you are done looking at the C code. Note that whether or not a C breakpoint was hit, when you return to the VSIM> prompt, youll need to quit C Debug by selecting Tools > C Debug > Quit C Debug before finally quitting the simulation.
C Debug command reference UM-335
C Debug command reference

The table below provides a brief description of the commands that can be invoked when C Debug is running. Follow the links to the ModelSim SE Command Reference for complete command syntax. Command bd (CR-56) bp (CR-61) -c change (CR-67) describe (CR-123) Description deletes a previously set C breakpoint sets a C breakpoint changes the value of a C variable prints the type information of a C variable disables a previously set C breakpoint enables a previously disabled C breakpoint prints the value of a C variable Corresponding menu command right click breakpoint in Source window and select Remove Breakpoint click the desired line number in the Source window none select the C variable name in the Source window and select Tools > Describe or right click and select Describe. right click breakpoint in Source window and select Disable Breakpoint right click breakpoint in Source window and select Enable Breakpoint select the C variable name in the Source window and select Tools > Examine or right click and select Examine none none none click the run -continue button on the Main or Source window toolbar Tools > C Debug > Run > Finish Tools > C Debug > Show
disablebp (CR-124) enablebp (CR-130) examine (CR-132)
gdb dir (CR-144) pop (CR-174) push (CR-186) run (CR-197) -continue run (CR-197) -finish show (CR-209)
sets the source directory search path for the C debugger moves the specified number of call frames up the C callstack moves the specified number of call frames down the C callstack continues running the simulation after stopping continues running the simulation until control returns to the calling function displays the names and types of the local variables and arguments of the current C function single step in the C debugger to the next executable line of C code; step goes into function calls, whereas step -over does not displays a stack trace of the C call stack
step (CR-213)
click the step or step -over button on the Main or Source window toolbar
tb (CR-215)
Tools > C Debug > Traceback
UM-336 12 - C Debug
UM-337
13 - Profiling performance and memory use

Chapter contents
Introducing performance and memory profiling . A statistical sampling profiler . . . . . A memory allocation profiler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-338 UM-338 UM-338 UM-339 UM-339 UM-341 UM-341 UM-342 UM-343 UM-344 UM-344 UM-345 UM-347 UM-348 UM-351 UM-352
Getting started . . . . . . . . . . . . . Enabling the memory allocation profiler . . . . . Enabling the statistical sampling profiler . . . . . Collecting memory allocation and performance data . . Running the profiler on Windows with FLI/PLI/VPI code Interpreting profiler data . Interpreting profiler data The Ranked View . The Call Tree view The Structural View Viewing profile details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Analyzing C code performance Reporting profiler results . .
The ModelSim profiler combines a statistical sampling profiler with a memory allocation profiler to provide instance specific execution and memory allocation data. It allows you to quickly determine how your memory is being allocated and easily identify areas in your simulation where performance can be improved. The profiler can be used at all levels of design simulation Functional, RTL, and Gate Level and has the potential to save hours of regression test time. In addition, ASIC and FPGA design flows benefit from the use of this tool. Note: The functionality described in this chapter requires a profiler license feature in your ModelSim license file. Please contact your Mentor Graphics sales representative if you currently do not have such a feature.
Platform information
Profiling is not supported on Opteron / Athlon 64 platforms.
UM-338 13 - Profiling performance and memory use
Introducing performance and memory profiling

The profiler provides an interactive graphical representation of both memory and CPU usage on a per instance basis. It shows you what part of your design is consuming resources (CPU cycles or memory), allowing you to more quickly find problem areas in your code. The profiler enables those familiar with the design and validation environment to find firstlevel improvements in a matter of minutes. For example, the statistical sampling profiler might show the following: non-accelerated VITAL library cells that are impacting simulation run time objects in the sensitivity list that are not required, resulting in a process that consumes more simulation time than necessary a testbench process that is active even though it is not needed an inefficient C module random number processes that are consuming simulation resources in a testbench running in non-random mode With this information, you can make changes to the VHDL or Verilog source code that will speed up the simulation. The memory allocation profiler provides insight into how much memory different parts of the design are consuming. The two major areas of concern are typically: 1) memory usage during elaboration, and 2) during simulation. If memory is exhausted during elaboration, for example, memory profiling may provide insights into what part(s) of the design are memory intensive. Or, if your HDL or PLI/FLI code is allocating memory and not freeing it when appropriate, the memory profiler will indicate excessive memory use in particular portions of the design.
A statistical sampling profiler

The profilers statistical sampling profiler samples the current simulation at a userdetermined rate (every <n> milliseconds of real or "wall-clock" time, not simulation time) and records what is executing at each sample point. The advantage of statistical sampling is that an entire simulation need not be run to get good information about what parts of your design are using the most simulation time. A few thousand samples, for example, can be accumulated before pausing the simulation to see where simulation time is being spent. The statistical profiler reports only on the samples that it can attribute to user code. For example, if you use the -nodebug argument to vcom (CR-246) or vlog (CR-293), it cannot report sample results.
A memory allocation profiler

The profilers memory allocation profiler records every memory allocation and deallocation that takes place in the context of elaborating and simulating the design. It makes a record of the design element that is active at the time of allocation so memory resources can be attributed to appropriate parts of the design. This provides insights into memory usage that can help you re-code designs to, for example, minimize memory use, correct memory leaks, and change optimization parameters used at compile time.
Getting started UM-339
Getting started
Memory allocation profiling and statistical sampling are enabled separately.
Enabling the memory allocation profiler

To record memory usage during elaboration and simulation, enable memory allocation profiling when the design is loaded with the -memprof argument to the vsim command.
vsim -memprof <design_unit>
Note that profile-data collection for the call tree is off by default. See "The Call Tree view" (UM-345) for additional information on collecting call-stack data. You can use the graphic user interface as follows to perform the same task. 1 Select Simulate > Start Simulation or the Simulate icon, to open the Start Simulation dialog box. 2 Select the Others tab. 3 Click the Enable memory profiling checkbox to select it.
4 Click OK to load the design with memory allocation profiling enabled.
If memory allocation during elaboration is not a concern, the memory allocation profiler can be enabled at any time after the design is loaded by doing any one of the following: select Tools > Profile > Memory use the -m argument with the profile on command (CR-180)
profile on -m
click the Memory Profiling icon
Handling large files

To allow memory allocation profiling of large designs, where the design itself plus the data required to keep track of memory allocation exceeds the memory available on the machine, the memory profiler allows you to route raw memory allocation data to an external file. This allows you to save the memory profile with minimal memory impact on the simulator, regardless of the size of your design. The external data file is created during elaboration by using either the -memprof+file=<filename> or the -memprof+fileonly=<filename> argument with the vsim command (CR-305). The -memprof+file=<filename> option will collect memory profile data during both elaboration and simulation and save it to the named external file and makes the data available for viewing and reporting during the current simulation. The -memprof+fileonly=<filename> option will collect memory profile data during both elaboration and simulation and save it to only the named external file. No data is saved for viewing and reporting during the current simulation, which reduces the overall amount of memory required by memory allocation profiling. Alternatively, you can save memory profile data from the simulation only by using either the -m -file <filename> or the -m -fileonly <filename> argument with the profile on command (CR-180). The -m -file <filename> option saves memory profile data from simulation to the designated external file and makes the data available for viewing and reporting during the current simulation. The -m -fileonly <filename> option saves memory profile data from simulation to only the designated external file. No data is saved for viewing and reporting during the current simulation, which reduces the overall amount of memory required by memory allocation profiling. After elaboration and/or simulation is complete, a separate session can be invoked and the profile data can be read in with the profile reload command (CR-182) for analysis. It should be noted, however, that this command will clear all performance and memory profiling data collected to that point (implicit profile clear). Any currently loaded design will be unloaded (implicit quit -sim), and run-time profiling will be turned off (implicit profile off -m -p). If a new design is loaded after you have read the raw profile data, then all internal profile data is cleared (implicit profile clear), but run-time profiling is not turned back on.
Getting started UM-341
Enabling the statistical sampling profiler

To enable the profilers statistical sampling profiler prior to a simulation run, do any one of the following: select Tools > Profile > Performance use the profile on command (CR-180) click the Performance Profiling icon
Collecting memory allocation and performance data

Both memory allocation profiling and statistical sampling occur during the execution of a ModelSim run command. With profiling enabled, all subsequent run commands will collect memory allocation data and performance statistics. Profiling results are cumulative each run command performed with profiling enabled will add new information to the data already gathered. To clear this data, select Tools > Profile > Clear Profile Data or use the profile clear command (CR-177). With the profiler enabled and a run command initiated, the simulator will provide a "Profiling" message in the transcript to indicate that profiling has started. If the statistical sampling profiler and the memory allocation profiler are on, the status bar will display the number of Profile Samples collected and the amount of memory allocated, as shown below. Each profile sample will become a data point in the simulations performance profile.
Turning profiling off

You can turn off the statistical sampling profiler or the memory allocation profiler by doing any one of the following: deselect the Performance and/or Memory options in the Tools > Profile menu deselect the Performance Profiling and Memory Profiling icons in the toolbar use the profile off command (CR-179) with the -p or -m arguments. Any ModelSim run commands that follow will not be profiled.
Running the profiler on Windows with FLI/PLI/VPI code

If you need to run the profiler under Windows on a design that contains FLI/PLI/VPI code, add these two switches to the compiling/linking command:
/DEBUG /DEBUGTYPE:COFF
These switches add symbols to the .dll file that the profiler can use in its report.
Interpreting profiler data UM-343
Interpreting profiler data

The utility of the data supplied by the profiler depends in large part on how your code is written. In cases where a single model or instance consumes a high percentage of simulation time or requires a high percentage of memory, the statistical sampling profiler or the memory allocation profiler quickly identifies that object, allowing you to implement a change that runs faster or requires less memory. More commonly, simulation time or memory allocation will be spread among a handful of modules or entities for example, 30% of simulation time split between models X, Y, and Z; or 20% of memory allocation going to models A, B, C and D. In such situations, careful examination and improvement of each model may result in overall speed improvement or more efficient memory allocation. There are times, however, when the statistical sampling and memory allocation profilers tell you nothing more than that simulation time or memory allocation is fairly equally distributed throughout your design. In such situations, the profiler provides little helpful information and improvement must come from a higher level examination of how the design can be changed or optimized.
Viewing profiler results

The profiler provides three views of the collected data Ranked, Call Tree and Structural. All three views are enabled by selecting View > Profile > View or by typing view profilemain at the VSIM prompt. This opens the Profile pane. The Profile pane includes selection tabs for the Ranked, Call Tree, and Structural views.
The Ranked View

The Ranked view displays the results of the statistical performance profiler and the memory allocation profiler for each function or instance. By default, ranked profiler results are sorted by values in the In% column, which shows the percentage of the total samples collected for each function or instance. You can sort ranked results by any other column by simply clicking the column heading. Click the down arrow to the left of the Name column to open a Configure Columns dialog, which allows you to select which columns are to be hidden or displayed. The use of colors in the display provides an immediate visual indication of where your design is spending most of its simulation time. By default, red text indicates functions or instances that are consuming 5% or more of simulation time.
Click here to hide or display columns
The Ranked view does not provide hierarchical, function-call information.
Viewing profiler results UM-345
The Call Tree view

Data collection for the call tree is off by default. Collection can be turned on from the VSIM command prompt with profile option collect_calltrees on and off with profile option collect_calltrees off. Call stack data collection can also be turned on with the memprof+call argument to the vsim command (CR-305). By default, profiler results in the Call Tree view are sorted according to the Under(%) column, which shows the percentage of the total samples collected for each function or instance and all supporting routines or instances. Sort results by any other column by clicking the column heading. As in the Ranked view, red object names indicate functions or instances that, by default, are consuming 5% or more of simulation time. The Call Tree view differs from the Ranked view in two important respects. Entries in the Name column of the Call Tree view are indented in hierarchical order to indicate which functions or routines call which others. A %Parent column in the Call Tree view allows you to see what percentage of a parent routines simulation time is used in which subroutines. The Call Tree view presents data in a call-stack format that provides more context than does the ranked view about where simulation time is spent. For example, your models may contain several instances of a utility function that computes the maximum of 3-delay values. A Ranked view might reveal that the simulation spent 60% of its time in this utility function, but would not tell you which routine or routines were making the most use of it. The Call Tree view will reveal which line is calling the function most frequently. Using this information, you might decide that instead of calling the function every time to compute the maximum of the 3-delays, this spot in your VHDL code can be used to compute it just once. You can then store the maximum delay value in a local variable. The two %Parent columns in the Call Tree view show the percent of simulation time or allocated memory a given function or instance is using of its parents total simulation time or available memory. From these columns, you can calculate the percentage of total simulation time or memory taken up by any function. For example, if a particular parent entry used 10% of the total simulation time or allocated memory, and it called a routine that used 80% of its simulation time or memory, then the percentage of total simulation time spent in, or memory allocated to, that routine would be 80% of 10%, or 8%. In addition to these differences, the Ranked view displays any particular function only once, regardless of where it was used. In the Call Tree view, the function can appear multiple times each time in the context of where it was used.
Viewing profiler results UM-347
The Structural View

The Structural view displays instance-specific performance and memory profile information in a hierarchical structure format identical to the structural view in the Workspace. It contains the same information found in the Call Tree view but adds an additional dimension with which to categorize performance samples and memory allocation. It shows how call stacks are associated with different instances in the design. For example, in the illustration that follows, TCL_Flush and TCL_Close appear under both test_sm and sm_0.
In the Call Tree and Structural views, you can expand and collapse the various levels to hide data that is not useful to the current analysis and/or is cluttering the display. Click on the '+' box next to an object name to expand the hierarchy and show supporting functions and/or instances beneath it. Click the '-' box to collapse all levels beneath the entry.
(UM-345)
Note that profile-data collection for the call tree is off by default. See "The Call Tree view" for additional information on collecting call-stack data. You can also right click any function or instance in the Call Tree and Structural views to obtain popup menu selections for rooting the display to the currently selected item, to ascend the displayed root one level, or to expand and collapse the hierarchy. See Profiler popup menu commands (GR-178).
Toggling display of call stack entries

By default call stack entries do not display in the Structural tab. To display call stack entries, right-click in the pane and select Show Calls.
Viewing profile details

The Profiler increases visibility into simulation performance and memory usage with dynamic links to the Source window and the Profile Details pane. The Profile Details pane is enabled by selecting View > Profile > View Details or by entering the view profiledetails command at the VSIM prompt. You can also right-click any function or instance in the Ranked, Call Tree, or Structural views to open a popup menu that includes options for viewing profile details. The following options are available:
View Source
When View Source is selected the Source window opens to the location of the selected function in the source code.
Function Usage
When Function Usage is selected, the Profile Details pane opens and displays all instances using the selected function. In the Profile Details pane shown below, all the instances using function Tcl_Close are displayed. The statistical performance and memory allocation data shows how much simulation time and memory is used by Tcl_Close in each instance.
Instance Usage
When Instance Usage is selected all instances with the same definition as the selected instance will be displayed in the Profile Details pane.
View Instantiation
When View Instantiation is selected the Source window opens to the point in the source code where the selected instance is instantiated.
Viewing profile details UM-349
Callers & Callees

When Callers & Callees is selected, callers and callees for the selected function are displayed in the Profile Details window. Items above the selected function are callers; items below are callees. The selected function is distinguished with an arrow on the left and in 'hotForeground' color as shown below.
Display in Call Tree

When Display in Call Tree is selected the Call Tree view of the Profile window expands to display all occurrences of the selected function and puts the selected function into a search buffer so you can easily cycle across all occurrences of that function. Note that profile-data collection for the call tree is off by default. See "The Call Tree view" (UM-345) for additional information on collecting call-stack data.
Display in Structural
When Display in Structural is selected the Structural view of the Profile window expands to display all occurrences of the selected function and puts the selected function into a search buffer so you can easily cycle across all occurrences of that function.
Integration with Source windows

The Ranked, Call Tree and Structural profile views are all dynamically linked to Source window. You can double-click any function or instance in the Ranked, Call Tree and Structural views to bring up that object in a Source window with the selected line highlighted.
You can perform the same task by right-clicking any function or instance in any one of the three Profile views and selecting View Source from the popup menu that opens. When you right-click an instance in the Structural profile view, the View Instantiation selection will become active in the popup menu. Selecting this option opens the instantiation in a Source window and highlights it. The right-click popup menu also allows you to change the root instance of the display, ascend to the next highest root instance, or reset the root instance to the top level instance. The selection of a context in the structure tab of the Workspace pane will cause the root display to be set in the Structural view.
Analyzing C code performance UM-351
Analyzing C code performance

You can include C code in your design via SystemC, the Verilog PLI/VPI, or the ModelSim FLI. The profiler can be used to determine the impact of these C modules on simulator performance. Factors that can affect simulator performance when a design includes C code are as follows: PLI/FLI applications with large sensitivity lists Calling operating system functions from C code Calling the simulators command interpreter from C code Inefficient C code In addition, the Verilog PLI/VPI requires maintenance of the simulators internal data structures as well as the PLI/VPI data structures for portability. (VHDL does not have this problem in ModelSim because the FLI gets information directly from the simulator.)
Reporting profiler results

You can create performance and memory profile reports using the Profile Report dialog or the profile report command (CR-183). For example, the command
profile report -calltree -file calltree.rpt -cutoff 2
will produce a Call Tree profile report in a text file called calltree.rpt, as shown here.
See the profile report command (CR-183) in the Command Reference for complete details on profiler reporting options.
Reporting profiler results UM-353
Select Tools > Profile > Profile Report to open the Profile Report dialog. From the dialog below, a Structural profile report will be created from the root instance pathname, /test_sm/ sm_seq0. The report will include function call hierarchy and three structure levels. Both performance and memory data will be displayed with a cutoff of 3% - meaning, the report will not contain any functions or instances that do not use 3% or more of simulation time or memory. The report will be written to a file called profile.out and, since the "View file" box is selected, it will be generated and displayed in Notepad when the OK button is clicked.
See Profile Report dialog (GR-94) for details on dialog options.
UM-355
14 - Signal Spy
Chapter contents
Introduction . . . . . Designed for testbenches . disable_signal_spy . enable_signal_spy . init_signal_driver . init_signal_spy . signal_force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-356 UM-356 UM-357 UM-358 UM-359 UM-362 UM-365 UM-367 UM-369 UM-370 UM-371 UM-374 UM-377 UM-379
signal_release .
$disable_signal_spy $enable_signal_spy $init_signal_driver . $init_signal_spy $signal_force . $signal_release . . . .
This chapter describes the Signal SpyTM procedures and system tasks. These allow you to monitor, drive, force, and release hierarchical objects in VHDL or mixed designs.
UM-356 14 - Signal Spy
Introduction
The Verilog language allows access to any signal from any other hierarchical block without having to route it via the interface. This means you can use hierarchical notation to either assign or determine the value of a signal in the design hierarchy from a testbench. This capability fails when a Verilog testbench attempts to reference a signal in a VHDL block or reference a signal in a Verilog block through a VHDL level of hierarchy. This limitation exists because VHDL does not allow hierarchical notation. In order to reference internal hierarchical signals, you have to resort to defining signals in a global package and then utilize those signals in the hierarchical blocks in question. But, this requires that you keep making changes depending on the signals that you want to reference. The Signal Spy procedures and system tasks overcome the aforementioned limitations. They allow you to monitor (spy), drive, force, or release hierarchical objects in a VHDL or mixed design. The VHDL procedures are provided via the "Util package" (UM-87) within the modelsim_lib library. To access the procedures you would add lines like the following to your VHDL code:
library modelsim_lib; use modelsim_lib.util.all;
The Verilog tasks are available as built-in "System tasks and functions" (UM-125). The table below shows the VHDL procedures and their corresponding Verilog system tasks. VHDL procedures disable_signal_spy (UM-357) enable_signal_spy (UM-358) init_signal_driver (UM-359) init_signal_spy (UM-362) signal_force (UM-365) signal_release (UM-367) Verilog system tasks $disable_signal_spy (UM-369) $enable_signal_spy (UM-370) $init_signal_driver (UM-371) $init_signal_spy (UM-374) $signal_force (UM-377) $signal_release (UM-379)
Designed for testbenches

Signal Spy limits the portability of your code. HDL code with Signal Spy procedures or tasks works only in ModelSim, not other simulators. We therefore recommend using Signal Spy only in testbenches, where portability is less of a concern, and the need for such a tool is more applicable.
disable_signal_spy UM-357
disable_signal_spy
The disable_signal_spy() procedure disables the associated init_signal_spy. The association between the disable_signal_spy call and the init_signal_spy call is based on specifying the same src_object and dest_object arguments to both functions. The disable_signal_spy call can only affect init_signal_spy calls that had their control_state argument set to "0" or "1".
Syntax
disable_signal_spy(src_object, dest_object, verbose)
Returns
Nothing
Arguments
Name src_object
Type string
Description Required. A full hierarchical path (or relative path with reference to the calling block) to a VHDL signal or Verilog register/net. This path should match the path that was specified in the init_signal_spy call that you wish to disable. Required. A full hierarchical path (or relative path with reference to the calling block) to a VHDL signal or Verilog register/net. This path should match the path that was specified in the init_signal_spy call that you wish to disable. Optional. Possible values are 0 or 1. Specifies whether you want a message reported in the transcript stating that a disable occurred and the simulation time that it occurred. Default is 0, no message
dest_object
string
verbose
integer
Related procedures
init_signal_spy (UM-362), enable_signal_spy (UM-358)
Example
See init_signal_spy Example (UM-364)
enable_signal_spy
The enable_signal_spy() procedure enables the associated init_signal_spy. The association between the enable_signal_spy call and the init_signal_spy call is based on specifying the same src_object and dest_object arguments to both functions. The enable_signal_spy call can only affect init_signal_spy calls that had their control_state argument set to "0" or "1".
Syntax
enable_signal_spy(src_object, dest_object, verbose)
Returns
Nothing
Arguments
Name src_object
Type string
Description Required. A full hierarchical path (or relative path with reference to the calling block) to a VHDL signal or Verilog register/net. This path should match the path that was specified in the init_signal_spy call that you wish to enable. Required. A full hierarchical path (or relative path with reference to the calling block) to a VHDL signal or Verilog register/net. This path should match the path that was specified in the init_signal_spy call that you wish to enable. Optional. Possible values are 0 or 1. Specifies whether you want a message reported in the transcript stating that an enable occurred and the simulation time that it occurred. Default is 0, no message
dest_object
string
verbose
integer
Related procedures
init_signal_spy (UM-362), disable_signal_spy (UM-357)
Example
See init_signal_spy Example (UM-364)
init_signal_driver UM-359
init_signal_driver
The init_signal_driver() procedure drives the value of a VHDL signal or Verilog net (called the src_object) onto an existing VHDL signal or Verilog net (called the dest_object). This allows you to drive signals or nets at any level of the design hierarchy from within a VHDL architecture (e.g., a testbench). The init_signal_driver procedure drives the value onto the destination signal just as if the signals were directly connected in the HDL code. Any existing or subsequent drive or force of the destination signal, by some other means, will be considered with the init_signal_driver value in the resolution of the signal.
Call only once

The init_signal_driver procedure creates a persistent relationship between the source and destination signals. Hence, you need to call init_signal_driver only once for a particular pair of signals. Once init_signal_driver is called, any change on the source signal will be driven on the destination signal until the end of the simulation. Thus, we recommend that you place all init_signal_driver calls in a VHDL process. You need to code the VHDL process correctly so that it is executed only once. The VHDL process should not be sensitive to any signals and should contain only init_signal_driver calls and a simple wait statement. The process will execute once and then wait forever. See the example below.
Syntax
init_signal_driver(src_object, dest_object, delay, delay_type, verbose)
Returns
Nothing
Arguments
Name src_object
Type string
Description Required. A full hierarchical path (or relative path with reference to the calling block) to a VHDL signal or Verilog net. Use the path separator to which your simulation is set (i.e., "/" or "."). A full hierarchical path must begin with a "/" or ".". The path must be contained within double quotes. Required. A full hierarchical path (or relative path with reference to the calling block) to an existing VHDL signal or Verilog net. Use the path separator to which your simulation is set (i.e., "/" or "."). A full hierarchical path must begin with a "/" or ".". The path must be contained within double quotes. Optional. Specifies a delay relative to the time at which the src_object changes. The delay can be an inertial or transport delay. If no delay is specified, then a delay of zero is assumed. Optional. Specifies the type of delay that will be applied. The value must be either mti_inertial or mti_transport. The default is mti_inertial. Optional. Possible values are 0 or 1. Specifies whether you want a message reported in the Transcript stating that the src_object is driving the dest_object. Default is 0, no message.
dest_object
string
delay
time
delay_type
del_mode
verbose
integer
Related procedures
init_signal_spy (UM-362), signal_force (UM-365), signal_release (UM-367)
Limitations
When driving a Verilog net, the only delay_type allowed is inertial. If you set the delay type to mti_transport, the setting will be ignored and the delay type will be mti_inertial. Any delays that are set to a value less than the simulator resolution will be rounded to the nearest resolution unit; no special warning will be issued.
init_signal_driver UM-361
Example
library IEEE, modelsim_lib; use IEEE.std_logic_1164.all; use modelsim_lib.util.all; entity testbench is end; architecture only of testbench is signal clk0 : std_logic; begin gen_clk0 : process begin clk0 <= '1' after 0 ps, '0' after 20 ps; wait for 40 ps; end process gen_clk0; drive_sig_process : process begin init_signal_driver("clk0", "/testbench/uut/blk1/clk", open, open, 1); init_signal_driver("clk0", "/testbench/uut/blk2/clk", 100 ps, mti_transport); wait; end process drive_sig_process; ... end;
The above example creates a local clock (clk0) and connects it to two clocks within the design hierarchy. The .../blk1/clk will match local clk0 and a message will be displayed. The open entries allow the default delay and delay_type while setting the verbose parameter to a 1. The .../blk2/clk will match the local clk0 but be delayed by 100 ps.
init_signal_spy
The init_signal_spy() procedure mirrors the value of a VHDL signal or Verilog register/net (called the src_object) onto an existing VHDL signal or Verilog register (called the dest_object). This allows you to reference signals, registers, or nets at any level of hierarchy from within a VHDL architecture (e.g., a testbench). The init_signal_spy procedure only sets the value onto the destination signal and does not drive or force the value. Any existing or subsequent drive or force of the destination signal, by some other means, will override the value that was set by init_signal_spy.
Call only once

The init_signal_spy procedure creates a persistent relationship between the source and destination signals. Hence, you need to call init_signal_spy once for a particular pair of signals. Once init_signal_spy is called, any change on the source signal will mirror on the destination signal until the end of the simulation unless the control_state is set. The control_state determines whether the mirroring of values can be enabled/disabled and what the initial state is. Subsequent control of whether the mirroring of values is enabled/ disabled is handled by the enable_signal_spy and disable_signal_spy calls. We recommend that you place all init_signal_spy calls in a VHDL process. You need to code the VHDL process correctly so that it is executed only once. The VHDL process should not be sensitive to any signals and should contain only init_signal_spy calls and a simple wait statement. The process will execute once and then wait forever, which is the desired behavior. See the example below.
Syntax
init_signal_spy(src_object, dest_object, verbose, control_state)
Returns
Nothing
init_signal_spy UM-363
Arguments
Name src_object
Type string
Description Required. A full hierarchical path (or relative path with reference to the calling block) to a VHDL signal or Verilog register/net. Use the path separator to which your simulation is set (i.e., "/" or "."). A full hierarchical path must begin with a "/" or ".". The path must be contained within double quotes. Required. A full hierarchical path (or relative path with reference to the calling block) to an existing VHDL signal or Verilog register. Use the path separator to which your simulation is set (i.e., "/" or "."). A full hierarchical path must begin with a "/" or ".". The path must be contained within double quotes. Optional. Possible values are 0 or 1. Specifies whether you want a message reported in the Transcript stating that the src_objects value is mirrored onto the dest_object. Default is 0, no message. Optional. Possible values are -1, 0, or 1. Specifies whether or not you want the ability to enable/disable mirroring of values and, if so, specifies the initial state. The default is -1, no ability to enable/disable and mirroring is enabled. "0" turns on the ability to enable/ disable and initially disables mirroring. "1" turns on the ability to enable/disable and initially enables mirroring.
dest_object
string
verbose
integer
control_state
integer
Related procedures
init_signal_driver (UM-359), signal_force (UM-365), signal_release (UM-367), enable_signal_spy (UM-358), disable_signal_spy (UM-357)
Limitations
When mirroring the value of a Verilog register/net onto a VHDL signal, the VHDL signal must be of type bit, bit_vector, std_logic, or std_logic_vector. Verilog memories (arrays of registers) are not supported.
Example
library ieee; library modelsim_lib; use ieee.std_logic_1164.all; use modelsim_lib.util.all; entity top is end; architecture only of top is signal top_sig1 : std_logic; begin ... spy_process : process begin init_signal_spy("/top/uut/inst1/sig1","/top/top_sig1",1,1); wait; end process spy_process; ... spy_enable_disable : process(enable_sig) begin if (enable_sig = '1') then enable_signal_spy("/top/uut/inst1/sig1","/top/top_sig1",0); elseif (enable_sig = '0') disable_signal_spy("/top/uut/inst1/sig1","/top/top_sig1",0); end if; end process spy_enable_disable; ... end;
In this example, the value of /top/uut/inst1/sig1 is mirrored onto /top/top_sig1. A message is issued to the transcript. The ability to control the mirroring of values is turned on and the init_signal_spy is initially enabled. The mirroring of values will be disabled when enable_sig transitions to a 0 and enable when enable_sig transitions to a 1.
signal_force UM-365
signal_force
The signal_force() procedure forces the value specified onto an existing VHDL signal or Verilog register or net (called the dest_object). This allows you to force signals, registers, or nets at any level of the design hierarchy from within a VHDL architecture (e.g., a testbench). A signal_force works the same as the force command (CR-141) with the exception that you cannot issue a repeating force. The force will remain on the signal until a signal_release, a force or release command, or a subsequent signal_force is issued. Signal_force can be called concurrently or sequentially in a process.
Syntax
signal_force( dest_object, value, rel_time, force_type, cancel_period, verbose )
Returns
Nothing
Arguments
Name dest_object
Type string
Description Required. A full hierarchical path (or relative path with reference to the calling block) to an existing VHDL signal or Verilog register/net. Use the path separator to which your simulation is set (i.e., "/" or "."). A full hierarchical path must begin with a "/" or ".". The path must be contained within double quotes. Required. Specifies the value to which the dest_object is to be forced. The specified value must be appropriate for the type. Optional. Specifies a time relative to the current simulation time for the force to occur. The default is 0. Optional. Specifies the type of force that will be applied. The value must be one of the following; default, deposit, drive, or freeze. The default is "default" (which is "freeze" for unresolved objects or "drive" for resolved objects). See the force command (CR-141) for further details on force type.
value
string
rel_time
time
force_type
forcetype
Name cancel_period
Type time
Description Optional. Cancels the signal_force command after the specified period of time units. Cancellation occurs at the last simulation delta cycle of a time unit. A value of zero cancels the force at the end of the current time period. Default is -1 ms. A negative value means that the force will not be cancelled. Optional. Possible values are 0 or 1. Specifies whether you want a message reported in the Transcript stating that the value is being forced on the dest_object at the specified time. Default is 0, no message.
verbose
integer
Related procedures
init_signal_driver (UM-359), init_signal_spy (UM-362), signal_release (UM-367)
Limitations
You cannot force bits or slices of a register; you can force only the entire register.
Example
library IEEE, modelsim_lib; use IEEE.std_logic_1164.all; use modelsim_lib.util.all; entity testbench is end; architecture only of testbench is begin force_process : process begin signal_force("/testbench/uut/blk1/reset", "1", 0 ns, freeze, open, 1); signal_force("/testbench/uut/blk1/reset", "0", 40 ns, freeze, 2 ms, 1); wait; end process force_process; ... end;
The above example forces reset to a "1" from time 0 ns to 40 ns. At 40 ns, reset is forced to a "0", 2 ms after the second signal_force call was executed. If you want to skip parameters so that you can specify subsequent parameters, you need to use the keyword "open" as a placeholder for the skipped parameter(s). The first signal_force procedure illustrates this, where an "open" for the cancel_period parameter means that the default value of -1 ms is used.
signal_release UM-367
signal_release
The signal_release() procedure releases any force that was applied to an existing VHDL signal or Verilog register/net (called the dest_object). This allows you to release signals, registers or nets at any level of the design hierarchy from within a VHDL architecture (e.g., a testbench). A signal_release works the same as the noforce command (CR-164). Signal_release can be called concurrently or sequentially in a process.
Syntax
signal_release( dest_object, verbose )
Returns
Nothing
Arguments
Name dest_object
Type string
Description Required. A full hierarchical path (or relative path with reference to the calling block) to an existing VHDL signal or Verilog register/net. Use the path separator to which your simulation is set (i.e., "/" or "."). A full hierarchical path must begin with a "/" or ".". The path must be contained within double quotes. Optional. Possible values are 0 or 1. Specifies whether you want a message reported in the Transcript stating that the signal is being released and the time of the release. Default is 0, no message.
verbose
integer
Related procedures
init_signal_driver (UM-359), init_signal_spy (UM-362), signal_force (UM-365)
Limitations
You cannot release a bit or slice of a register; you can release only the entire register.
Example
library IEEE, modelsim_lib; use IEEE.std_logic_1164.all; use modelsim_lib.util.all; entity testbench is end; architecture only of testbench is signal release_flag : std_logic; begin stim_design : process begin ... wait until release_flag = '1'; signal_release("/testbench/dut/blk1/data", 1); signal_release("/testbench/dut/blk1/clk", 1); ... end process stim_design; ... end;
The above example releases any forces on the signals data and clk when the signal release_flag is a "1". Both calls will send a message to the transcript stating which signal was released and when.
$disable_signal_spy UM-369
$disable_signal_spy
The $disable_signal_spy() system task disables the associated $init_signal_spy task. The association between the $disable_signal_spy task and the $init_signal_spy task is based on specifying the same src_object and dest_object arguments to both tasks. The $disable_signal_spy task can only affect $init_signal_spy tasks that had their control_state argument set to "0" or "1".
Syntax
$disable_signal_spy(src_object, dest_object, verbose)
Returns
Nothing
Arguments
Name src_object
Type string
Description Required. A full hierarchical path (or relative path with reference to the calling block) to a VHDL signal or Verilog register/net. This path should match the path that was specified in the init_signal_spy call that you wish to disable. Required. A full hierarchical path (or relative path with reference to the calling block) to a VHDL signal or Verilog register/net. This path should match the path that was specified in the init_signal_spy call that you wish to disable. Optional. Possible values are 0 or 1. Specifies whether you want a message reported in the transcript stating that a disable occurred and the simulation time that it occurred. Default is 0, no message
dest_object
string
verbose
integer
Related tasks
$init_signal_spy (UM-374), $enable_signal_spy (UM-370)
Example
See $init_signal_spy Example (UM-375)
$enable_signal_spy
The $enable_signal_spy() system task enables the associated $init_signal_spy task. The association between the $enable_signal_spy task and the $init_signal_spy task is based on specifying the same src_object and dest_object arguments to both tasks. The $enable_signal_spy task can only affect $init_signal_spys tasks that had their control_state argument set to "0" or "1".
Syntax
$enable_signal_spy(src_object, dest_object, verbose)
Returns
Nothing
Arguments
Name src_object
Type string
Description Required. A full hierarchical path (or relative path with reference to the calling block) to a VHDL signal or Verilog register/net. This path should match the path that was specified in the init_signal_spy call that you wish to enable. Required. A full hierarchical path (or relative path with reference to the calling block) to a VHDL signal or Verilog register/net. This path should match the path that was specified in the init_signal_spy call that you wish to enable. Optional. Possible values are 0 or 1. Specifies whether you want a message reported in the transcript stating that an enable occurred and the simulation time that it occurred. Default is 0, no message
dest_object
string
verbose
integer
Related tasks
$init_signal_spy (UM-374), $disable_signal_spy (UM-369)
Example
See $init_signal_spy Example (UM-375)
$init_signal_driver UM-371
$init_signal_driver
The $init_signal_driver() system task drives the value of a VHDL signal or Verilog net (called the src_object) onto an existing VHDL signal or Verilog register/net (called the dest_object). This allows you to drive signals or nets at any level of the design hierarchy from within a Verilog module (e.g., a testbench). The $init_signal_driver system task drives the value onto the destination signal just as if the signals were directly connected in the HDL code. Any existing or subsequent drive or force of the destination signal, by some other means, will be considered with the $init_signal_driver value in the resolution of the signal.
Call only once

The $init_signal_driver system task creates a persistent relationship between the source and destination signals. Hence, you need to call $init_signal_driver only once for a particular pair of signals. Once $init_signal_driver is called, any change on the source signal will be driven on the destination signal until the end of the simulation. Thus, we recommend that you place all $init_signal_driver calls in a Verilog initial block. See the example below.
Syntax
$init_signal_driver(src_object, dest_object, delay, delay_type, verbose)
Returns
Nothing
Arguments
Name src_object
Type string
Description Required. A full hierarchical path (or relative path with reference to the calling block) to a VHDL signal or Verilog net. Use the path separator to which your simulation is set (i.e., "/" or "."). A full hierarchical path must begin with a "/" or ".". The path must be contained within double quotes. Required. A full hierarchical path (or relative path with reference to the calling block) to an existing VHDL signal or Verilog net. Use the path separator to which your simulation is set (i.e., "/" or "."). A full hierarchical path must begin with a "/" or ".". The path must be contained within double quotes.
dest_object
string
Name delay
Type integer, real, or time
Description Optional. Specifies a delay relative to the time at which the src_object changes. The delay can be an inertial or transport delay. If no delay is specified, then a delay of zero is assumed. Optional. Specifies the type of delay that will be applied. The value must be either 0 (inertial) or 1 (transport). The default is 0. Optional. Possible values are 0 or 1. Specifies whether you want a message reported in the Transcript stating that the src_object is driving the dest_object. Default is 0, no message.
delay_type
integer
verbose
integer
Related tasks
$init_signal_spy (UM-374), $signal_force (UM-377), $signal_release (UM-379)
Limitations
When driving a Verilog net, the only delay_type allowed is inertial. If you set the delay type to 1 (transport), the setting will be ignored, and the delay type will be inertial. Any delays that are set to a value less than the simulator resolution will be rounded to the nearest resolution unit; no special warning will be issued. Verilog memories (arrays of registers) are not supported.
$init_signal_driver UM-373
Example
`timescale 1 ps / 1 ps module testbench; reg clk0; initial begin clk0 = 1; forever begin #20 clk0 = ~clk0; end end initial begin $init_signal_driver("clk0", "/testbench/uut/blk1/clk", , , 1); $init_signal_driver("clk0", "/testbench/uut/blk2/clk", 100, 1); end ... endmodule
The above example creates a local clock (clk0) and connects it to two clocks within the design hierarchy. The .../blk1/clk will match local clk0 and a message will be displayed. The .../blk2/clk will match the local clk0 but be delayed by 100 ps. For the second call to work, the .../blk2/clk must be a VHDL based signal, because if it were a Verilog net a 100 ps inertial delay would consume the 40 ps clock period. Verilog nets are limited to only inertial delays and thus the setting of 1 (transport delay) would be ignored.
$init_signal_spy
The $init_signal_spy() system task mirrors the value of a VHDL signal or Verilog register/ net (called the src_object) onto an existing VHDL signal or Verilog register (called the dest_object). This allows you to reference signals, registers, or nets at any level of hierarchy from within a Verilog module (e.g., a testbench). The $init_signal_spy system task only sets the value onto the destination signal and does not drive or force the value. Any existing or subsequent drive or force of the destination signal, by some other means, will override the value set by $init_signal_spy.
Call only once

The $init_signal_spy system task creates a persistent relationship between the source and the destination signal. Hence, you need to call $init_signal_spy only once for a particular pair of signals. Once $init_signal_spy is called, any change on the source signal will mirror on the destination signal until the end of the simulation unless the control_state is set. The control_state determines whether the mirroring of values can be enabled/disabled and and what the initial state is. Subsequent control of whether the mirroring of values is enabled/disabled is handled by the $enable_signal_spy and $disable_signal_spy tasks. We recommend that you place all $init_signal_spy tasks in a Verilog initial block. See the example below.
Syntax
$init_signal_spy(src_object, dest_object, verbose, control_state)
Returns
Nothing
Arguments
Name src_object
Type string
Description Required. A full hierarchical path (or relative path with reference to the calling block) to a VHDL signal or Verilog register/net. Use the path separator to which your simulation is set (i.e., "/" or "."). A full hierarchical path must begin with a "/" or ".". The path must be contained within double quotes. Required. A full hierarchical path (or relative path with reference to the calling block) to a Verilog register or VHDL signal. Use the path separator to which your simulation is set (i.e., "/" or "."). A full hierarchical path must begin with a "/" or ".". The path must be contained within double quotes.
dest_object
string
$init_signal_spy UM-375
Name verbose
Type integer
Description Optional. Possible values are 0 or 1. Specifies whether you want a message reported in the Transcript stating that the src_objects value is mirrored onto the dest_object. Default is 0, no message. Optional. Possible values are -1, 0, or 1. Specifies whether or not you want the ability to enable/disable mirroring of values and, if so, specifies the initial state. The default is -1, no ability to enable/disable and mirroring is enabled. "0" turns on the ability to enable/ disable and initially disables mirroring. "1" turns on the ability to enable/disable and initially enables mirroring.
control_state
integer
Related tasks
$init_signal_driver (UM-371), $signal_force (UM-377), $signal_release (UM-379), $disable_signal_spy (UM-369)
Limitations
When mirroring the value of a VHDL signal onto a Verilog register, the VHDL signal must be of type bit, bit_vector, std_logic, or std_logic_vector. Verilog memories (arrays of registers) are not supported.
Example
module top; ... reg top_sig1; reg enable_reg; ... initial begin $init_signal_spy(".top.uut.inst1.sig1",".top.top_sig1",1,1); end always @ (posedge enable_reg) begin $enable_signal_spy(".top.uut.inst1.sig1",".top.top_sig1",0); end always @ (negedge enable_reg) begin $disable_signal_spy(".top.uut.inst1.sig1",".top.top_sig1",0); end ... endmodule
In this example, the value of .top.uut.inst1.sig1 is mirrored onto .top.top_sig1. A message is issued to the transcript. The ability to control the mirroring of values is turned on and the init_signal_spy is initially enabled. The mirroring of values will be disabled when enable_reg transitions to a 0 and enabled when enable_reg transitions to a 1.
$signal_force UM-377
$signal_force
The $signal_force() system task forces the value specified onto an existing VHDL signal or Verilog register/net (called the dest_object). This allows you to force signals, registers, or nets at any level of the design hierarchy from within a Verilog module (e.g., a testbench). A $signal_force works the same as the force command (CR-141) with the exception that you cannot issue a repeating force. The force will remain on the signal until a $signal_release, a force or release command, or a subsequent $signal_force is issued. $signal_force can be called concurrently or sequentially in a process.
Syntax
$signal_force( dest_object, value, rel_time, force_type, cancel_period, verbose )
Returns
Nothing
Arguments
Name dest_object
Type string
Description Required. A full hierarchical path (or relative path with reference to the calling block) to an existing VHDL signal or Verilog register/net. Use the path separator to which your simulation is set (i.e., "/" or "."). A full hierarchical path must begin with a "/" or ".". The path must be contained within double quotes. Required. Specifies the value to which the dest_object is to be forced. The specified value must be appropriate for the type. Optional. Specifies a time relative to the current simulation time for the force to occur. The default is 0. Optional. Specifies the type of force that will be applied. The value must be one of the following; 0 (default), 1 (deposit), 2 (drive), or 3 (freeze). The default is "default" (which is "freeze" for unresolved objects or "drive" for resolved objects). See the force command (CR-141) for further details on force type.
value
string
rel_time
integer, real, or time integer
force_type
Name cancel_period
Type integer, real, time
Description Optional. Cancels the $signal_force command after the specified period of time units. Cancellation occurs at the last simulation delta cycle of a time unit. A value of zero cancels the force at the end of the current time period. Default is -1. A negative value means that the force will not be cancelled. Optional. Possible values are 0 or 1. Specifies whether you want a message reported in the Transcript stating that the value is being forced on the dest_object at the specified time. Default is 0, no message.
verbose
integer
Related tasks
$init_signal_driver (UM-371), $init_signal_spy (UM-374), $signal_release (UM-379)
Limitations
You cannot force bits or slices of a register; you can force only the entire register. Verilog memories (arrays of registers) are not supported.
Example
`timescale 1 ns / 1 ns module testbench; initial begin $signal_force("/testbench/uut/blk1/reset", "1", 0, 3, , 1); $signal_force("/testbench/uut/blk1/reset", "0", 40, 3, 200000, 1); end ... endmodule
The above example forces reset to a "1" from time 0 ns to 40 ns. At 40 ns, reset is forced to a "0", 200000 ns after the second $signal_force call was executed.
$signal_release UM-379
$signal_release
The $signal_release() system task releases any force that was applied to an existing VHDL signal or Verilog register/net (called the dest_object). This allows you to release signals, registers, or nets at any level of the design hierarchy from within a Verilog module (e.g., a testbench). A $signal_release works the same as the noforce command (CR-164). $signal_release can be called concurrently or sequentially in a process.
Syntax
$signal_release( dest_object, verbose )
Returns
Nothing
Arguments
Name dest_object
Type string
Description Required. A full hierarchical path (or relative path with reference to the calling block) to an existing VHDL signal or Verilog register/net. Use the path separator to which your simulation is set (i.e., "/" or "."). A full hierarchical path must begin with a "/" or ".". The path must be contained within double quotes. Optional. Possible values are 0 or 1. Specifies whether you want a message reported in the Transcript stating that the signal is being released and the time of the release. Default is 0, no message.
verbose
integer
Related tasks
$init_signal_driver (UM-371), $init_signal_spy (UM-374), $signal_force (UM-377)
Limitations
You cannot release a bit or slice of a register; you can release only the entire register.
Example
module testbench; reg release_flag; always @(posedge release_flag) begin $signal_release("/testbench/dut/blk1/data", 1); $signal_release("/testbench/dut/blk1/clk", 1); end ... endmodule
The above example releases any forces on the signals data and clk when the register release_flag transitions to a "1". Both calls will send a message to the transcript stating which signal was released and when.
UM-381
15 - Standard Delay Format (SDF) Timing Annotation

Chapter contents
Specifying SDF files for simulation Instance specification . . . SDF specification with the GUI Errors and warnings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-382 UM-382 UM-383 UM-383 UM-384 UM-384 UM-385 UM-386 UM-386 UM-387 UM-390 UM-391 UM-391 UM-392 UM-393 UM-393 UM-394 UM-394 UM-395 UM-395
VHDL VITAL SDF . . . . . SDF to VHDL generic matching . Resolving errors . . . . . Verilog SDF . . . . . . . The $sdf_annotate system task . SDF to Verilog construct matching Optional edge specifications . . Optional conditions . . . . Rounded timing values . . .
SDF for mixed VHDL and Verilog designs Interconnect delays. . . . . . . . . . . .
Troubleshooting . . . . . . . . . . . . . Specifying the wrong instance . . . . . . . . Mistaking a component or module name for an instance label Forgetting to specify the instance . . . . . . . .
This chapter discusses ModelSims implementation of SDF (Standard Delay Format) timing annotation. Included are sections on VITAL SDF and Verilog SDF, plus troubleshooting. Verilog and VHDL VITAL timing data can be annotated from SDF files by using the simulators built-in SDF annotator. ASIC and FPGA vendors usually provide tools that create SDF files for use with their cell libraries. Refer to your vendors documentation for details on creating SDF files for your library. Many vendors also provide instructions on using their SDF files and libraries with ModelSim. The SDF specification was originally created for Verilog designs, but it has also been adopted for VHDL VITAL designs. In general, the designer does not need to be familiar with the details of the SDF specification because the cell library provider has already supplied tools that create SDF files that match their libraries. Note: ModelSim will read SDF files that were compressed using gzip. Other compression formats (e.g., Unix zip) are not supported.
UM-382 15 - Standard Delay Format (SDF) Timing Annotation
Specifying SDF files for simulation

ModelSim supports SDF versions 1.0 through 4.0 (except the NETDELAY statement). The simulators built-in SDF annotator automatically adjusts to the version of the file. Use the following vsim (CR-305) command-line options to specify the SDF files, the desired timing values, and their associated design instances:
-sdfmin [<instance>=]<filename> -sdftyp [<instance>=]<filename> -sdfmax [<instance>=]<filename>
Any number of SDF files can be applied to any instance in the design by specifying one of the above options for each file. Use -sdfmin to select minimum, -sdftyp to select typical, and -sdfmax to select maximum timing values from the SDF file.
Instance specification
The instance paths in the SDF file are relative to the instance to which the SDF is applied. Usually, this instance is an ASIC or FPGA model instantiated under a testbench. For example, to annotate maximum timing values from the SDF file myasic.sdf to an instance u1 under a top-level named testbench, invoke the simulator as follows:
vsim -sdfmax /testbench/u1=myasic.sdf testbench
If the instance name is omitted then the SDF file is applied to the top-level. This is usually incorrect because in most cases the model is instantiated under a testbench or within a larger system level simulation. In fact, the design can have several models, each having its own SDF file. In this case, specify an SDF file for each instance. For example,
vsim -sdfmax /system/u1=asic1.sdf -sdfmax /system/u2=asic2.sdf system
Specifying SDF files for simulation UM-383
SDF specification with the GUI

As an alternative to the command-line options, you can specify SDF files in the Start Simulation dialog box under the SDF tab.
You can access this dialog by invoking the simulator without any arguments or by selecting Simulate > Start Simulation. See the GUI chapter for a description of this dialog. For Verilog designs, you can also specify SDF files by using the $sdf_annotate system task. See "The $sdf_annotate system task" (UM-386) for more details.
Errors and warnings

Errors issued by the SDF annotator while loading the design prevent the simulation from continuing, whereas warnings do not. Use the -sdfnoerror option with vsim (CR-305) to change SDF errors to warnings so that the simulation can continue. Warning messages can be suppressed by using vsim with either the -sdfnowarn or +nosdfwarn options. Another option is to use the SDF tab from the Start Simulation dialog box (shown above). Select Disable SDF warnings (-sdfnowarn +nosdfwarn) to disable warnings, or select Reduce SDF errors to warnings (-sdfnoerror) to change errors to warnings. See "Troubleshooting" (UM-394) for more information on errors and warnings and how to avoid them.
VHDL VITAL SDF

VHDL SDF annotation works on VITAL cells only. The IEEE 1076.4 VITAL ASIC Modeling Specification describes how cells must be written to support SDF annotation. Once again, the designer does not need to know the details of this specification because the library provider has already written the VITAL cells and tools that create compatible SDF files. However, the following summary may help you understand simulator error messages. For additional VITAL specification information, see "VITAL specification and source code" (UM-84).
SDF to VHDL generic matching

An SDF file contains delay and timing constraint data for cell instances in the design. The annotator must locate the cell instances and the placeholders (VHDL generics) for the timing data. Each type of SDF timing construct is mapped to the name of a generic as specified by the VITAL modeling specification. The annotator locates the generic and updates it with the timing value from the SDF file. It is an error if the annotator fails to find the cell instance or the named generic. The following are examples of SDF constructs and their associated generic names: SDF construct (IOPATH a y (3)) (IOPATH (posedge clk) q (1) (2)) (INTERCONNECT u1/y u2/a (5)) (SETUP d (posedge clk) (5)) (HOLD (negedge d) (posedge clk) (5)) (SETUPHOLD d clk (5) (5)) (WIDTH (COND (reset==1b0) clk) (5)) Matching VHDL generic name tpd_a_y tpd_clk_q_posedge tipd_a tsetup_d_clk_noedge_posedge thold_d_clk_negedge_posedge tsetup_d_clk & thold_d_clk tpw_clk_reset_eq_0
VHDL VITAL SDF UM-385
Resolving errors
If the simulator finds the cell instance but not the generic then an error message is issued. For example,
** Error (vsim-SDF-3240) myasic.sdf(18): Instance /testbench/dut/u1 does not have a generic named tpd_a_y
In this case, make sure that the design is using the appropriate VITAL library cells. If it is, then there is probably a mismatch between the SDF and the VITAL cells. You need to find the cell instance and compare its generic names to those expected by the annotator. Look in the VHDL source files provided by the cell library vendor. If none of the generic names look like VITAL timing generic names, then perhaps the VITAL library cells are not being used. If the generic names do look like VITAL timing generic names but dont match the names expected by the annotator, then there are several possibilities: The vendors tools are not conforming to the VITAL specification. The SDF file was accidentally applied to the wrong instance. In this case, the simulator also issues other error messages indicating that cell instances in the SDF could not be located in the design. The vendors library and SDF were developed for the older VITAL 2.2b specification. This version uses different name mapping rules. In this case, invoke vsim (CR-305) with the -vital2.2b option:
vsim -vital2.2b -sdfmax /testbench/u1=myasic.sdf testbench
For more information on resolving errors see "Troubleshooting" (UM-394).
Verilog SDF
Verilog designs can be annotated using either the simulator command-line options or the $sdf_annotate system task (also commonly used in other Verilog simulators). The command-line options annotate the design immediately after it is loaded, but before any simulation events take place. The $sdf_annotate task annotates the design at the time it is called in the Verilog source code. This provides more flexibility than the command-line options.
The $sdf_annotate system task

The syntax for $sdf_annotate is:
Syntax
$sdf_annotate (["<sdffile>"], [<instance>], ["<config_file>"], ["<log_file>"], ["<mtm_spec>"], ["<scale_factor>"], ["<scale_type>"]);
Arguments
"<sdffile>"
String that specifies the SDF file. Required.

<instance>
Hierarchical name of the instance to be annotated. Optional. Defaults to the instance where the $sdf_annotate call is made.
"<config_file>"
String that specifies the configuration file. Optional. Currently not supported, this argument is ignored.
"<log_file>"
String that specifies the logfile. Optional. Currently not supported, this argument is ignored.
"<mtm_spec>"
String that specifies the delay selection. Optional. The allowed strings are "minimum", "typical", "maximum", and "tool_control". Case is ignored and the default is "tool_control". The "tool_control" argument means to use the delay specified on the command line by +mindelays, +typdelays, or +maxdelays (defaults to +typdelays).
"<scale_factor>"
String that specifies delay scaling factors. Optional. The format is "<min_mult>:<typ_mult>:<max_mult>". Each multiplier is a real number that is used to scale the corresponding delay in the SDF file.
"<scale_type>"
String that overrides the <mtm_spec> delay selection. Optional. The <mtm_spec> delay selection is always used to select the delay scaling factor, but if a <scale_type> is specified, then it will determine the min/typ/max selection from the SDF file. The allowed strings are "from_min", "from_minimum", "from_typ", "from_typical", "from_max", "from_maximum", and "from_mtm". Case is ignored, and the default is "from_mtm", which means to use the <mtm_spec> value.
Verilog SDF UM-387
Examples
Optional arguments can be omitted by using commas or by leaving them out if they are at the end of the argument list. For example, to specify only the SDF file and the instance to which it applies:
$sdf_annotate("myasic.sdf", testbench.u1);
To also specify maximum delay values:

$sdf_annotate("myasic.sdf", testbench.u1, , , "maximum");
SDF to Verilog construct matching

The annotator matches SDF constructs to corresponding Verilog constructs in the cells. Usually, the cells contain path delays and timing checks within specify blocks. For each SDF construct, the annotator locates the cell instance and updates each specify path delay or timing check that matches. An SDF construct can have multiple matches, in which case each matching specify statement is updated with the SDF timing value. SDF constructs are matched to Verilog constructs as follows: IOPATH is matched to specify path delays or primitives: SDF (IOPATH (posedge clk) q (3) (4)) (IOPATH a y (3) (4)) Verilog (posedge clk => q) = 0; buf u1 (y, a);
The IOPATH construct usually annotates path delays. If ModelSim cant locate a corresponding specify path delay, it returns an error unless you use the +sdf_iopath_to_prim_ok argument to vsim (CR-305). If you specify that argument and the module contains no path delays, then all primitives that drive the specified output port are annotated. INTERCONNECT and PORT are matched to input ports: SDF (INTERCONNECT u1.y u2.a (5)) (PORT u2.a (5)) Verilog input a; inout a;
Both of these constructs identify a module input or inout port and create an internal net that is a delayed version of the port. This is called a Module Input Port Delay (MIPD). All primitives, specify path delays, and specify timing checks connected to the original port are reconnected to the new MIPD net. PATHPULSE and GLOBALPATHPULSE are matched to specify path delays: SDF (PATHPULSE a y (5) (10)) (GLOBALPATHPULSE a y (30) (60)) Verilog (a => y) = 0; (a => y) = 0;
If the input and output ports are omitted in the SDF, then all path delays are matched in the cell. DEVICE is matched to primitives or specify path delays: SDF (DEVICE y (5)) (DEVICE y (5)) Verilog and u1(y, a, b); (a => y) = 0; (b => y) = 0;
If the SDF cell instance is a primitive instance, then that primitives delay is annotated. If it is a module instance, then all specify path delays are annotated that drive the output port specified in the DEVICE construct (all path delays are annotated if the output port is omitted). If the module contains no path delays, then all primitives that drive the specified output port are annotated (or all primitives that drive any output port if the output port is omitted). SETUP is matched to $setup and $setuphold: SDF (SETUP d (posedge clk) (5)) (SETUP d (posedge clk) (5)) Verilog $setup(d, posedge clk, 0); $setuphold(posedge clk, d, 0, 0);
HOLD is matched to $hold and $setuphold: SDF (HOLD d (posedge clk) (5)) (HOLD d (posedge clk) (5)) Verilog $hold(posedge clk, d, 0); $setuphold(posedge clk, d, 0, 0);
SETUPHOLD is matched to $setup, $hold, and $setuphold: SDF (SETUPHOLD d (posedge clk) (5) (5)) (SETUPHOLD d (posedge clk) (5) (5)) (SETUPHOLD d (posedge clk) (5) (5)) RECOVERY is matched to $recovery: SDF (RECOVERY (negedge reset) (posedge clk) (5)) Verilog $recovery(negedge reset, posedge clk, 0); Verilog $setup(d, posedge clk, 0); $hold(posedge clk, d, 0); $setuphold(posedge clk, d, 0, 0);
Verilog SDF UM-389
REMOVAL is matched to $removal: SDF (REMOVAL (negedge reset) (posedge clk) (5)) Verilog $removal(negedge reset, posedge clk, 0);
RECREM is matched to $recovery, $removal, and $recrem: SDF (RECREM (negedge reset) (posedge clk) (5) (5)) (RECREM (negedge reset) (posedge clk) (5) (5)) (RECREM (negedge reset) (posedge clk) (5) (5)) SKEW is matched to $skew: SDF (SKEW (posedge clk1) (posedge clk2) (5)) WIDTH is matched to $width: SDF (WIDTH (posedge clk) (5)) PERIOD is matched to $period: SDF (PERIOD (posedge clk) (5)) NOCHANGE is matched to $nochange: SDF (NOCHANGE (negedge write) addr (5) (5)) Verilog $nochange(negedge write, addr, 0, 0); Verilog $period(posedge clk, 0); Verilog $width(posedge clk, 0); Verilog $skew(posedge clk1, posedge clk2, 0); Verilog $recovery(negedge reset, posedge clk, 0); $removal(negedge reset, posedge clk, 0); $recrem(negedge reset, posedge clk, 0);
Optional edge specifications

Timing check ports and path delay input ports can have optional edge specifications. The annotator uses the following rules to match edges: A match occurs if the SDF port does not have an edge. A match occurs if the specify port does not have an edge. A match occurs if the SDF port edge is identical to the specify port edge. A match occurs if explicit edge transitions in the specify port edge overlap with the SDF port edge. These rules allow SDF annotation to take place even if there is a difference between the number of edge-specific constructs in the SDF file and the Verilog specify block. For example, the Verilog specify block may contain separate setup timing checks for a falling and rising edge on data with respect to clock, while the SDF file may contain only a single setup check for both edges: SDF (SETUP data (posedge clock) (5)) (SETUP data (posedge clock) (5)) Verilog $setup(posedge data, posedge clk, 0); $setup(negedge data, posedge clk, 0);
In this case, the cell accommodates more accurate data than can be supplied by the tool that created the SDF file, and both timing checks correctly receive the same value. Likewise, the SDF file may contain more accurate data than the model can accommodate. SDF (SETUP (posedge data) (posedge clock) (4)) (SETUP (negedge data) (posedge clock) (6)) Verilog $setup(data, posedge clk, 0); $setup(data, posedge clk, 0);
In this case, both SDF constructs are matched and the timing check receives the value from the last one encountered. Timing check edge specifiers can also use explicit edge transitions instead of posedge and negedge. However, the SDF file is limited to posedge and negedge. For example, SDF (SETUP data (posedge clock) (5)) Verilog $setup(data, edge[01, 0x] clk, 0);
The explicit edge specifiers are 01, 0x, 10, 1x, x0, and x1. The set of [01, 0x, x1] is equivalent to posedge, while the set of [10, 1x, x0] is equivalent to negedge. A match occurs if any of the explicit edges in the specify port match any of the explicit edges implied by the SDF port.
Verilog SDF UM-391
Optional conditions
Timing check ports and path delays can have optional conditions. The annotator uses the following rules to match conditions: A match occurs if the SDF does not have a condition. A match occurs for a timing check if the SDF port condition is semantically equivalent to the specify port condition. A match occurs for a path delay if the SDF condition is lexically identical to the specify condition. Timing check conditions are limited to very simple conditions, therefore the annotator can match the expressions based on semantics. For example, SDF (SETUP data (COND (reset!=1) (posedge clock)) (5)) Verilog $setup(data, posedge clk &&& (reset==0), 0);
The conditions are semantically equivalent and a match occurs. In contrast, path delay conditions may be complicated and semantically equivalent conditions may not match. For example, SDF (COND (r1 || r2) (IOPATH clk q (5))) (COND (r1 || r2) (IOPATH clk q (5))) Verilog if (r1 || r2) (clk => q) = 5; // matches if (r2 || r1) (clk => q) = 5; // does not match
The annotator does not match the second condition above because the order of r1 and r2 are reversed.
Rounded timing values

The SDF TIMESCALE construct specifies time units of values in the SDF file. The annotator rounds timing values from the SDF file to the time precision of the module that is annotated. For example, if the SDF TIMESCALE is 1ns and a value of .016 is annotated to a path delay in a module having a time precision of 10ps (from the timescale directive), then the path delay receives a value of 20ps. The SDF value of 16ps is rounded to 20ps. Interconnect delays are rounded to the time precision of the module that contains the annotated MIPD.
SDF for mixed VHDL and Verilog designs

Annotation of a mixed VHDL and Verilog design is very flexible. VHDL VITAL cells and Verilog cells can be annotated from the same SDF file. This flexibility is available only by using the simulators SDF command-line options. The Verilog $sdf_annotate system task can annotate Verilog cells only. See the vsim command (CR-305) for more information on SDF command-line options.
Interconnect delays UM-393
Interconnect delays
An interconnect delay represents the delay from the output of one device to the input of another. ModelSim can model single interconnect delays or multisource interconnect delays for Verilog, VHDL/VITAL, or mixed designs. See the vsim command for more information on the relevant command-line arguments. Timing checks are performed on the interconnect delayed versions of input ports. This may result in misleading timing constraint violations, because the ports may satisfy the constraint while the delayed versions may not. If the simulator seems to report incorrect violations, be sure to account for the effect of interconnect delays.

ModelSim offers a number of options for disabling timing checks on a "global" or individual basis. The table below provides a summary of those options. See the command and argument descriptions in the ModelSim Command Reference for more details. Command and argument vlog +notimingchecks vlog +nospecify vsim +no_neg_tchk vsim +no_notifier vsim +no_tchk_msg vsim +notimingchecks vsim +nospecify Effect disables timing check system tasks for all instances in the specified Verilog design disables specify path delays and timing checks for all instances in the specified Verilog design disables negative timing check limits by setting them to zero for all instances in the specified design disables the toggling of the notifier register argument of the timing check system tasks for all instances in the specified design disables error messages issued by timing check system tasks when timing check violations occur for all instances in the specified design disables Verilog and VITAL timing checks for all instances in the specified design disables specify path delays and timing checks for all instances in the specified design
Troubleshooting
Specifying the wrong instance
By far, the most common mistake in SDF annotation is to specify the wrong instance to the simulators SDF options. The most common case is to leave off the instance altogether, which is the same as selecting the top-level design unit. This is generally wrong because the instance paths in the SDF are relative to the ASIC or FPGA model, which is usually instantiated under a top-level testbench. See "Instance specification" (UM-382) for an example. A common example for both VHDL and Verilog testbenches is provided below. For simplicity, the test benches do nothing more than instantiate a model that has no ports.
VHDL testbench
entity testbench is end; architecture only of testbench is component myasic end component; begin dut : myasic; end;
Verilog testbench
module testbench; myasic dut(); endmodule
The name of the model is myasic and the instance label is dut. For either testbench, an appropriate simulator invocation might be:
vsim -sdfmax /testbench/dut=myasic.sdf testbench
Optionally, you can leave off the name of the top-level:

vsim -sdfmax /dut=myasic.sdf testbench
The important thing is to select the instance for which the SDF is intended. If the model is deep within the design hierarchy, an easy way to find the instance name is to first invoke the simulator without SDF options, view the structure pane, navigate to the model instance, select it, and enter the environment command (CR-131). This command displays the instance name that should be used in the SDF command-line option.
Troubleshooting UM-395
Mistaking a component or module name for an instance label

Another common error is to specify the component or module name rather than the instance label. For example, the following invocation is wrong for the above testbenches:
vsim -sdfmax /testbench/myasic=myasic.sdf testbench
This results in the following error message:

** Error (vsim-SDF-3250) myasic.sdf(0): Failed to find INSTANCE /testbench/myasic.
Forgetting to specify the instance

If you leave off the instance altogether, then the simulator issues a message for each instance path in the SDF that is not found in the design. For example,
vsim -sdfmax myasic.sdf testbench
Results in:
** Error (vsim-SDF-3250) myasic.sdf(0): Failed to find INSTANCE /testbench/u1 ** Error (vsim-SDF-3250) myasic.sdf(0): Failed to find INSTANCE /testbench/u2 ** Error (vsim-SDF-3250) myasic.sdf(0): Failed to find INSTANCE /testbench/u3 ** Error (vsim-SDF-3250) myasic.sdf(0): Failed to find INSTANCE /testbench/u4 ** Error (vsim-SDF-3250) myasic.sdf(0): Failed to find INSTANCE /testbench/u5 ** Warning (vsim-SDF-3432) myasic.sdf: This file is probably applied to the wrong instance. ** Warning (vsim-SDF-3432) myasic.sdf: Ignoring subsequent missing instances from this file.
After annotation is done, the simulator issues a summary of how many instances were not found and possibly a suggestion for a qualifying instance:
** Warning (vsim-SDF-3440) myasic.sdf: Failed to find any of the 358 instances from this file. ** Warning (vsim-SDF-3442) myasic.sdf: Try instance /testbench/dut. It contains all instance paths from this file.
The simulator recommends an instance only if the file was applied to the top-level and a qualifying instance is found one level down. Also see "Resolving errors" (UM-385) for specific VHDL VITAL SDF troubleshooting.
UM-397
16 - Value Change Dump (VCD) Files

Chapter contents
Creating a VCD file . . . Flow for four-state VCD file Flow for extended VCD file Case sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-398 UM-398 UM-398 UM-398 UM-399 UM-399 UM-400 UM-402 UM-403 UM-404 UM-404 UM-404 UM-405 UM-408 UM-408 UM-409 UM-409 UM-409 UM-411
Using extended VCD as stimulus . . . . . . . . Simulating with input values from a VCD file . . . Replacing instances with output values from a VCD file . ModelSim VCD commands and VCD tasks . Compressing files with VCD tasks . . A VCD file from source to output . VHDL source code . . . VCD simulator commands . VCD output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Capturing port driver data . . . . . . Driver states . . . . . . . . Driver strength . . . . . . . Identifier code . . . . . . . . Resolving values . . . . . . . Example VCD output from vcd dumpports
This chapter describes how to use VCD files in ModelSim. The VCD file format is specified in the IEEE 1364 standard. It is an ASCII file containing header information, variable definitions, and variable value changes. VCD is in common use for Verilog designs, and is controlled by VCD system task calls in the Verilog source code. ModelSim provides command equivalents for these system tasks and extends VCD support to VHDL designs. The ModelSim commands can be used on VHDL, Verilog, or mixed designs. If you need vendor-specific ASIC design-flow documentation that incorporates VCD, please contact your ASIC vendor.
UM-398 16 - Value Change Dump (VCD) Files
Creating a VCD file

There are two flows in ModelSim for creating a VCD file. One flow produces a four-state VCD file with variable changes in 0, 1, x, and z with no strength information; the other produces an extended VCD file with variable changes in all states and strength information and port driver data. Both flows will also capture port driver changes unless filtered out with optional command-line arguments.
Flow for four-state VCD file

First, compile and load the design:
% % % % cd ~/modeltech/examples/misc vlib work vlog counter.v tcounter.v vsim test_counter
Next, with the design loaded, specify the VCD file name with the vcd file command (CR237) and add objects to the file with the vcd add command (CR-227):
VSIM VSIM VSIM VSIM 1> 2> 3> 4> vcd file myvcdfile.vcd vcd add /test_counter/dut/* run quit -f
There will now be a VCD file in the working directory.
Flow for extended VCD file

First, compile and load the design:
% % % % cd ~/modeltech/examples/misc vlib work vlog counter.v tcounter.v vsim test_counter
Next, with the design loaded, specify the VCD file name and objects to add with the vcd dumpports command (CR-230):
VSIM 1> vcd dumpports -file myvcdfile.vcd /test_counter/dut/* VSIM 3> run VSIM 4> quit -f
There will now be an extended VCD file in the working directory.
Ignoring strength ranges

By default ModelSim uses strength ranges for resolving conflicts as specified by IEEE 1364-2005. You can ignore strength ranges using the -no_strength_range argument to the vcd dumpports command (CR-230). See "Resolving values" (UM-409) for more details.
Case sensitivity
VHDL is not case sensitive so ModelSim converts all signal names to lower case when it produces a VCD file. Conversely, Verilog designs are case sensitive so ModelSim maintains case when it produces a VCD file.
Using extended VCD as stimulus UM-399
Using extended VCD as stimulus

You can use an extended VCD file as stimulus to re-simulate your design. There are two ways to do this: 1) simulate the top level of a design unit with the input values from an extended VCD file; and 2) specify one or more instances in a design to be replaced with the output values from the associated VCD file.
Simulating with input values from a VCD file

When simulating with inputs from an extended VCD file, you can simulate only one design unit at a time. In other words, you can apply the VCD file inputs only to the top level of the design unit for which you captured port data. The general procedure includes two steps: 1 Create a VCD file for a single design unit using the vcd dumpports command (CR-230). 2 Resimulate the single design unit using the -vcdstim argument to vsim (CR-305). Note that -vcdstim works only with VCD files that were created by a ModelSim simulation.
Example 1 Verilog counter

First, create the VCD file for the single instance using vcd dumpports:
% cd ~/modeltech/examples/misc % vlib work % vlog counter.v tcounter.v % vsim test_counter VSIM 1> vcd dumpports -file counter.vcd /test_counter/dut/* VSIM 2> run VSIM 3> quit -f
Next, rerun the counter without the testbench, using the -vcdstim argument:
% vsim -vcdstim counter.vcd counter VSIM 1> add wave /* VSIM 2> run 200
Example 2 VHDL adder

First, create the VCD file using vcd dumpports:
% cd ~/modeltech/examples/misc % vlib work % vcom gates.vhd adder.vhd stimulus.vhd % vsim testbench2 VSIM 1> vcd dumpports -file addern.vcd /testbench2/uut/* VSIM 2> run 1000 VSIM 3> quit -f
Next, rerun the adder without the testbench, using the -vcdstim argument:
% vsim -vcdstim addern.vcd addern -gn=8 -do "add wave /*; run 1000"
Example 3 Mixed-HDL design

First, create three VCD files, one for each module:
% cd ~/modeltech/examples/tutorials/mixed/projects % vlib work % vlog cache.v memory.v proc.v % vcom util.vhd set.vhd top.vhd % vsim top VSIM 1> vcd dumpports -file proc.vcd /top/p/* VSIM 2> vcd dumpports -file cache.vcd /top/c/* VSIM 3> vcd dumpports -file memory.vcd /top/m/* VSIM 4> run 1000 VSIM 5> quit -f
Next, rerun each module separately, using the captured VCD stimulus:
% vsim -vcdstim proc.vcd proc -do "add wave /*; run 1000" VSIM 1> quit -f % vsim -vcdstim cache.vcd cache -do "add wave /*; run 1000" VSIM 1> quit -f % vsim -vcdstim memory.vcd memory -do "add wave /*; run 1000" VSIM 1> quit -f
Replacing instances with output values from a VCD file

Replacing instances with output values from a VCD file lets you simulate without the instances source or even the compiled object. The general procedure includes two steps: 1 Create VCD files for one or more instances in your design using the vcd dumpports command (CR-230). If necessary, use the -vcdstim switch to handle port order problems (see below). 2 Re-simulate your design using the -vcdstim <instance>=<filename> argument to vsim (CR-305). Note that this works only with VCD files that were created by a ModelSim simulation.
Example
In the following example, the three instances /top/p, /top/c, and /top/m are replaced in simulation by the output values found in the corresponding VCD files. First, create VCD files for all instances you want to replace:
vcd vcd vcd run dumpports -vcdstim -file proc.vcd /top/p/* dumpports -vcdstim -file cache.vcd /top/c/* dumpports -vcdstim -file memory.vcd /top/m/* 1000
Next, simulate your design and map the instances to the VCD files you created:
vsim top -vcdstim /top/p=proc.vcd -vcdstim /top/c=cache.vcd -vcdstim /top/m=memory.vcd
Using extended VCD as stimulus UM-401
Port order issues

The -vcdstim argument to the vcd dumpports command ensures the order that port names appear in the VCD file matches the order that they are declared in the instances module or entity declaration. Consider the following module declaration:
module proc(clk, addr, data, rw, strb, rdy); input clk, rdy; output addr, rw, strb; inout data;
The order of the ports in the module line (clk, addr, data, ...) does not match the order of those ports in the input, output, and inout lines (clk, rdy, addr, ...). In this case the -vcdstim argument to the vcd dumpports command needs to be used. In cases where the order is the same, you do not need to use the -vcdstim argument to vcd dumpports. Also, module declarations of the form:
module proc(input clk, output addr, inout data, ...)
do not require use of the argument.
ModelSim VCD commands and VCD tasks

ModelSim VCD commands map to IEEE Std 1364 VCD system tasks and appear in the VCD file along with the results of those commands. The table below maps the VCD commands to their associated tasks. VCD commands vcd add (CR-227) vcd checkpoint (CR-228) vcd file (CR-237) vcd flush (CR-241) vcd limit (CR-242) vcd off (CR-243) vcd on (CR-244) VCD system tasks $dumpvars $dumpall $dumpfile $dumpflush $dumplimit $dumpoff $dumpon
ModelSim also supports extended VCD (dumpports system tasks). The table below maps the VCD dumpports commands to their associated tasks. VCD dumpports commands vcd dumpports (CR-230) vcd dumpportsall (CR-232) vcd dumpportsflush (CR-233) vcd dumpportslimit (CR-234) vcd dumpportsoff (CR-235) vcd dumpportson (CR-236) VCD system tasks $dumpports $dumpportsall $dumpportsflush $dumpportslimit $dumpportsoff $dumpportson
ModelSim supports multiple VCD files. This functionality is an extension of the IEEE Std 1364 specification. The tasks behave the same as the IEEE equivalent tasks such as $dumpfile, $dumpvar, etc. The difference is that $fdumpfile can be called multiple times to create more than one VCD file, and the remaining tasks require a filename argument to associate their actions with a specific file. VCD commands vcd add (CR-227)
-file <filename>
VCD system tasks $fdumpvars $fdumpall $fdumpfile $fdumpflush
vcd checkpoint (CR-228) <filename> vcd files (CR-239) <filename> vcd flush (CR-241) <filename>
ModelSim VCD commands and VCD tasks UM-403
VCD commands vcd limit (CR-242) <filename> vcd off (CR-243) <filename> vcd on (CR-244) <filename>
VCD system tasks $fdumplimit $fdumpoff $fdumpon
Compressing files with VCD tasks

ModelSim can produce compressed VCD files using the gzip compression algorithm. Since we cannot change the syntax of the system tasks, we act on the extension of the output file name. If you specify a .gz extension on the filename, ModelSim will compress the output.
A VCD file from source to output

The following example shows the VHDL source, a set of simulator commands, and the resulting VCD output.
VHDL source code

The design is a simple shifter device represented by the following VHDL source code:
library IEEE; use IEEE.STD_LOGIC_1164.all; entity SHIFTER_MOD is port (CLK, RESET, data_in : IN STD_LOGIC; Q : INOUT STD_LOGIC_VECTOR(8 downto 0)); END SHIFTER_MOD ; architecture RTL of SHIFTER_MOD is begin process (CLK,RESET) begin if (RESET = '1') then Q <= (others => '0') ; elsif (CLK'event and CLK = '1') then Q <= Q(Q'left - 1 downto 0) & data_in ; end if ; end process ; end ;
VCD simulator commands

At simulator time zero, the designer executes the following commands:
vcd file output.vcd vcd add -r * force reset 1 0 force data_in 0 0 force clk 0 0 run 100 force clk 1 0, 0 50 -repeat 100 run 100 vcd off force reset 0 0 force data_in 1 0 run 100 vcd on run 850 force reset 1 0 run 50 vcd checkpoint quit -sim
A VCD file from source to output UM-405
VCD output
The VCD file created as a result of the preceding scenario would be called output.vcd. The following pages show how it would look.
VCD output
$date Thu Sep 18 11:07:43 2003 $end $version ModelSim Version 6.1 $end $timescale 1ns $end $scope module shifter_mod $end $var wire 1 ! clk $end $var wire 1 " reset $end $var wire 1 # data_in $end $var wire 1 $ q [8] $end $var wire 1 % q [7] $end $var wire 1 & q [6] $end $var wire 1 ' q [5] $end $var wire 1 ( q [4] $end $var wire 1 ) q [3] $end $var wire 1 * q [2] $end $var wire 1 + q [1] $end $var wire 1 , q [0] $end $upscope $end $enddefinitions $end #0 $dumpvars 0! 1" 0# 0$ 0% 0& 0' 0( 0) 0* 0+ 0, $end #100 1! #150 0! #200 1! $dumpoff x! x" x# x$ x% x& x' x(
x) x* x+ x, $end #300 $dumpon 1! 0" 1# 0$ 0% 0& 0' 0( 0) 0* 0+ 1, $end #350 0! #400 1! 1+ #450 0! #500 1! 1* #550 0! #600 1! 1) #650 0! #700 1! 1( #750 0! #800 1! 1' #850 0! #900 1! 1& #950 0! #1000 1! 1% #1050 0! #1100 1! 1$ #1150
A VCD file from source to output UM-407
0! 1" 0, 0+ 0* 0) 0( 0' 0& 0% 0$ #1200 1! $dumpall 1! 1" 1# 0$ 0% 0& 0' 0( 0) 0* 0+ 0, $end
Capturing port driver data

Some ASIC vendors toolkits read a VCD file format that provides details on port drivers. This information can be used, for example, to drive a tester. See the ASIC vendors documentation for toolkit specific information. In ModelSim use the vcd dumpports command (CR-230) to create a VCD file that captures port driver data. Each time an external or internal port driver changes values, a new value change is recorded in the VCD file with the following format:
p<state> <0 strength> <1 strength> <identifier_code>
Driver states
The driver states are recorded as TSSI states if the direction is known, as detailed in this table: Input (testfixture) D low U high N unknown Z tri-state d low (two or more drivers active) u high (two or more drivers active) Output (dut) L low H high X unknown T tri-state l low (two or more drivers active) h high (two or more drivers active)
If the direction is unknown, the state will be recorded as one of the following: Unknown direction 0 low (both input and output are driving low) 1 high (both input and output are driving high) ? unknown (both input and output are driving unknown) F three-state (input and output unconnected) A unknown (input driving low and output driving high) a unknown (input driving low and output driving unknown) B unknown (input driving high and output driving low) b unknown (input driving high and output driving unknown) C unknown (input driving unknown and output driving low) c unknown (input driving unknown and output driving high)
Capturing port driver data UM-409
Unknown direction f unknown (input and output three-stated)
Driver strength
The recorded 0 and 1 strength values are based on Verilog strengths: Strength 0 highz 1 small 2 medium 3 weak 4 large 5 pull 6 strong 7 supply W,H,L U,X,0,1,- VHDL std_logic mappings Z
Identifier code
The <identifier_code> is an integer preceded by < that starts at zero and is incremented for each port in the order the ports are specified. Also, the variable type recorded in the VCD header is "port".
Resolving values
The resolved values written to the VCD file depend on which options you specify when creating the file.
Default behavior
By default ModelSim generates output according to IEEE 1364-2005. The standard states that the values 0 (both input and output are active with value 0) and 1 (both input and output are active with value 1) are conflict states. The standard then defines two strength ranges: Strong: strengths 7, 6, and 5 Weak: strengths 4, 3, 2, 1 The rules for resolving values are as follows: If the input and output are driving the same value with the same range of strength, the resolved value is 0 or 1, and the strength is the stronger of the two. If the input is driving a strong strength and the output is driving a weak strength, the resolved value is D, d, U or u, and the strength is the strength of the input. If the input is driving a weak strength and the output is driving a strong strength, the resolved value is L, l, H or h, and the strength is the strength of the output.
Ignoring strength ranges

You may wish to ignore strength ranges and have ModelSim handle each strength separately. Any of the following options will produce this behavior: Use the -no_strength_range argument to the vcd dumpports command (CR-230) Use an optional argument to $dumpports (see "Extended $dumpports syntax" (UM-410) below) Use the +dumpports+no_strength_range argument to vsim command (CR-305) In this situation, ModelSim reports strengths for both the zero and one components of the value if the strengths are the same. If the strengths are different, ModelSim reports only the winning strength. In other words, the two strength values either match (e.g., pA 5 5 !) or the winning strength is shown and the other is zero (e.g., pH 0 5 !).
Extended $dumpports syntax

ModelSim extends the $dumpports system task in order to support exclusion of strength ranges. The extended syntax is as follows:
$dumpports (scope_list, file_pathname, ncsim_file_index, file_format)
The nc_sim_index argument is required yet ignored by ModelSim. It is required only to be compatible with NCSims argument list. The file_format argument accepts the following values or an ORed combination thereof (see examples below): File_format value 0 2 4 8 Here are some examples:
// ignore strength range $dumpports(top, "filename", 0, 0) // compress and ignore strength range $dumpports(top, "filename", 0, 4) // print direction and ignore strength range $dumpports(top, "filename", 0, 8) // compress, print direction, and ignore strength range $dumpports(top, "filename", 0, 12)
Meaning Ignore strength range Use strength ranges; produces IEEE 1364-compliant behavior Compress the EVCD output Include port direction information in the EVCD file header; same as using -direction argument to vcd dumpports
Capturing port driver data UM-411
Example VCD output from vcd dumpports

This section demonstrates how vcd dumpports resolves values based on certain combinations of driver values and strengths and whether or not you use strength ranges.
Sample driver data

time 0 100 200 300 900 27400 27500 27600 in value 0 0 0 0 1 1 1 1 out value 0 0 0 0 0 1 1 1 in strength value (range) 7 (strong) 6 (strong) 5 (strong) 4 (weak) 6 (strong) 5 (strong) 4 (weak) 3 (weak) out strength value (range) 7 (strong) 7 (strong) 7 (strong) 7 (strong) 7 (strong) 4 (weak) 4 (weak) 4 (weak)
Vcd output with strength ranges

Given the driver data above and use of 1364 strength ranges, here is what the VCD file output would look like:
#0 p0 7 0 #100 p0 7 0 #200 p0 7 0 #300 pL 7 0 #900 pB 7 0 #27400 pU 0 5 #27500 p1 0 4 #27600 p1 0 4 <0 <0 <0 <0 <0 <0 <0 <0
Vcd output ignoring strength ranges

Here is what the output would look like if you ignore strength ranges:
#0 p0 7 #100 pL 7 #200 pL 7 #300 pL 7 0 <0 0 <0 0 <0 0 <0
#900 pL 7 0 #27400 pU 0 5 #27500 p1 0 4 #27600 pH 0 4
<0 <0 <0 <0
UM-413
17 - Tcl and macros (DO files)

Chapter contents
Introduction . . . . . . Tcl features within ModelSim. Tcl References. . . . . Tcl commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-414 UM-414 UM-414 UM-415 UM-416 UM-418 UM-420 UM-420 UM-420 UM-420 UM-420 UM-421 UM-421 UM-422 UM-422 UM-423 UM-425 UM-429 UM-429 UM-429 UM-429 UM-430 UM-431 UM-431
Tcl command syntax . . . . . if command syntax . . . . Command substitution . . . Command separator . . . . Multiple-line commands . . . Evaluation order . . . . . Tcl relational expression evaluation Variable substitution . . . . System commands. . . . . List processing . . . . . . . . . . . . . .
ModelSim Tcl commands .
ModelSim Tcl time commands Tcl examples . . . . .
Macros (DO files) . . . . . . . . . . . Creating DO files . . . . . . . . . . Using Parameters with DO files . . . . . . Deleting a file from a .do script . . . . . . Making macro parameters optional . . . . . Useful commands for handling breakpoints and errors Error action in DO files . . . . . . . .
UM-414 17 - Tcl and macros (DO files)
Introduction
This chapter provides an overview of Tcl (tool command language) as used with ModelSim. Macros in ModelSim are simply Tcl scripts that contain ModelSim and, optionally, Tcl commands. Tcl is a scripting language for controlling and extending ModelSim. Within ModelSim you can develop implementations from Tcl scripts without the use of C code. Because Tcl is interpreted, development is rapid; you can generate and execute Tcl scripts on the fly without stopping to recompile or restart ModelSim. In addition, if ModelSim does not provide the command you need, you can use Tcl to create your own commands.
Tcl features within ModelSim

Using Tcl with ModelSim gives you these features: command history (like that in C shells) full expression evaluation and support for all C-language operators a full range of math and trig functions support of lists and arrays regular expression pattern matching procedures the ability to define your own commands command substitution (that is, commands may be nested) robust scripting language for macros Note: You must be using ModelSim SE to customize the interface.
Tcl References
Two books about Tcl are Tcl and the Tk Toolkit by John K. Ousterhout, published by Addison-Wesley Publishing Company, Inc., and Practical Programming in Tcl and Tk by Brent Welch published by Prentice Hall. You can also consult the following online references: Select Help > Tcl Man Pages.
Tcl commands UM-415
Tcl commands
For complete information on Tcl commands, select Help > Tcl Man Pages. Also see "Simulator GUI preferences" (GR-251) for information on Tcl preference variables. ModelSim command names that conflict with Tcl commands have been renamed or have been replaced by Tcl commands. See the list below:
Previous ModelSim command continue format list | wave if list nolist | nowave set source wave
Command changed to (or replaced by) run (CR-197) with the -continue option write format (CR-341) with either list or wave specified
418)
replaced by the Tcl if command, see "if command syntax" (UMfor more information
add list (CR-44) delete (CR-122) with either list or wave specified replaced by the Tcl set command vsource (CR-323) add wave (CR-49)
Tcl command syntax

The following eleven rules define the syntax and semantics of the Tcl language. Additional details on if command syntax (UM-418). 1 A Tcl script is a string containing one or more commands. Semi-colons and newlines are command separators unless quoted as described below. Close brackets ("]") are command terminators during command substitution (see below) unless quoted. 2 A command is evaluated in two steps. First, the Tcl interpreter breaks the command into words and performs substitutions as described below. These substitutions are performed in the same way for all commands. The first word is used to locate a command procedure to carry out the command, then all of the words of the command are passed to the command procedure. The command procedure is free to interpret each of its words in any way it likes, such as an integer, variable name, list, or Tcl script. Different commands interpret their words differently. 3 Words of a command are separated by white space (except for newlines, which are command separators). 4 If the first character of a word is a double-quote (""") then the word is terminated by the next double-quote character. If semi-colons, close brackets, or white space characters (including newlines) appear between the quotes then they are treated as ordinary characters and included in the word. Command substitution, variable substitution, and backslash substitution are performed on the characters between the quotes as described below. The double-quotes are not retained as part of the word. 5 If the first character of a word is an open brace ("{") then the word is terminated by the matching close brace ("}"). Braces nest within the word: for each additional open brace there must be an additional close brace (however, if an open brace or close brace within the word is quoted with a backslash then it is not counted in locating the matching close brace). No substitutions are performed on the characters between the braces except for backslash-newline substitutions described below, nor do semi-colons, newlines, close brackets, or white space receive any special interpretation. The word will consist of exactly the characters between the outer braces, not including the braces themselves. 6 If a word contains an open bracket ("[") then Tcl performs command substitution. To do this it invokes the Tcl interpreter recursively to process the characters following the open bracket as a Tcl script. The script may contain any number of commands and must be terminated by a close bracket ("]"). The result of the script (i.e. the result of its last command) is substituted into the word in place of the brackets and all of the characters between them. There may be any number of command substitutions in a single word. Command substitution is not performed on words enclosed in braces.
Tcl command syntax UM-417
7 If a word contains a dollar-sign ("$") then Tcl performs variable substitution: the dollarsign and the following characters are replaced in the word by the value of a variable. Variable substitution may take any of the following forms:
$name
Name is the name of a scalar variable; the name is terminated by any character that isn't a letter, digit, or underscore.
$name(index)
Name gives the name of an array variable and index gives the name of an element within that array. Name must contain only letters, digits, and underscores. Command substitutions, variable substitutions, and backslash substitutions are performed on the characters of index.
${name}
Name is the name of a scalar variable. It may contain any characters whatsoever except for close braces. There may be any number of variable substitutions in a single word. Variable substitution is not performed on words enclosed in braces. 8 If a backslash ("\") appears within a word then backslash substitution occurs. In all cases but those described below the backslash is dropped and the following character is treated as an ordinary character and included in the word. This allows characters such as double quotes, close brackets, and dollar signs to be included in words without triggering special processing. The following table lists the backslash sequences that are handled specially, along with the value that replaces each sequence.
\a \b \f \n \r \t \v \<newline>whiteSpace
Audible alert (bell) (0x7). Backspace (0x8). Form feed (0xc). Newline (0xa). Carriage-return (0xd). Tab (0x9). Vertical tab (0xb). A single space character replaces the backslash, newline, and all spaces and tabs after the newline. This backslash sequence is unique in that it is replaced in a separate pre-pass before the command is actually parsed. This means that it will be replaced even when it occurs between braces, and the resulting space will be treated as a word separator if it isn't in braces or quotes. Backslash ("\"). The digits ooo (one, two, or three of them) give the octal value of the character.
\\ \ooo
\xhh
The hexadecimal digits hh give the hexadecimal value of the character. Any number of digits may be present.
Backslash substitution is not performed on words enclosed in braces, except for backslash-newline as described above. 9 If a hash character ("#") appears at a point where Tcl is expecting the first character of the first word of a command, then the hash character and the characters that follow it, up through the next newline, are treated as a comment and ignored. The comment character only has significance when it appears at the beginning of a command. 10 Each character is processed exactly once by the Tcl interpreter as part of creating the words of a command. For example, if variable substitution occurs then no further substitutions are performed on the value of the variable; the value is inserted into the word verbatim. If command substitution occurs then the nested command is processed entirely by the recursive call to the Tcl interpreter; no substitutions are performed before making the recursive call and no additional substitutions are performed on the result of the nested script. 11 Substitutions do not affect the word boundaries of a command. For example, during variable substitution the entire value of the variable becomes part of a single word, even if the variable's value contains spaces.
if command syntax
The Tcl if command executes scripts conditionally. Note that in the syntax below the "?" indicates an optional argument.
Syntax
if expr1 ?then? body1 elseif expr2 ?then? body2 elseif ... ?else? ?bodyN?
Description
The if command evaluates expr1 as an expression. The value of the expression must be a boolean (a numeric value, where 0 is false and anything else is true, or a string value such as true or yes for true and false or no for false); if it is true then body1 is executed by passing it to the Tcl interpreter. Otherwise expr2 is evaluated as an expression and if it is true then body2 is executed, and so on. If none of the expressions evaluates to true then bodyN is executed. The then and else arguments are optional "noise words" to make the command easier to read. There may be any number of elseif clauses, including zero. BodyN may also be omitted as long as else is omitted too. The return value from the command is the result of the body script that was executed, or an empty string if none of the expressions was non-zero and there was no bodyN.
Command substitution
Placing a command in square brackets [ ] will cause that command to be evaluated first and its results returned in place of the command. An example is:
set a 25 set b 11 set c 3 echo "the result is [expr ($a + $b)/$c]"
will output:
"the result is 12"
This feature allows VHDL variables and signals, and Verilog nets and registers to be accessed using:
[examine -<radix> name]
The %name substitution is no longer supported. Everywhere %name could be used, you now can use [examine -value -<radix> name] which allows the flexibility of specifying command options. The radix specification is optional.
Command separator
A semicolon character (;) works as a separator for multiple commands on the same line. It is not required at the end of a line in a command sequence.
Multiple-line commands
With Tcl, multiple-line commands can be used within macros and on the command line. The command line prompt will change (as in a C shell) until the multiple-line command is complete. In the example below, note the way the opening brace { is at the end of the if and else lines. This is important because otherwise the Tcl scanner won't know that there is more coming in the command and will try to execute what it has up to that point, which won't be what you intend.
if { [exa sig_a] == "0011ZZ"} { echo "Signal value matches" do macro_1.do } else { echo "Signal value fails" do macro_2.do }
Evaluation order
An important thing to remember when using Tcl is that anything put in curly brackets {} is not evaluated immediately. This is important for if-then-else statements, procedures, loops, and so forth.
Tcl relational expression evaluation

When you are comparing values, the following hints may be useful: Tcl stores all values as strings, and will convert certain strings to numeric values when appropriate. If you want a literal to be treated as a numeric value, don't quote it.
if {[exa var_1] == 345}...
The following will also work:

if {[exa var_1] == "345"}...
However, if a literal cannot be represented as a number, you must quote it, or Tcl will give you an error. For instance:
if {[exa var_2] == 001Z}...
will give an error.

if {[exa var_2] == "001Z"}...
will work okay. Don't quote single characters in single quotes:

if {[exa var_3] == 'X'}...
will give an error

if {[exa var_3] == "X"}...
will work okay.
For the equal operator, you must use the C operator "==". For not-equal, you must use the C operator "!=".
Variable substitution
When a $<var_name> is encountered, the Tcl parser will look for variables that have been defined either by ModelSim or by you, and substitute the value of the variable. Note: Tcl is case sensitive for variable names.
To access environment variables, use the construct:

$env(<var_name>) echo My user name is $env(USER)
Environment variables can also be set using the env array:

set env(SHELL) /bin/csh
See "Simulator state variables" (UM-455) for more information about ModelSim-defined variables.
System commands
To pass commands to the UNIX shell or DOS window, use the Tcl exec command:
echo The date is [exec date]
List processing
In Tcl a "list" is a set of strings in curly braces separated by spaces. Several Tcl commands are available for creating lists, indexing into lists, appending to lists, getting the length of lists and shifting lists. These commands are: Command syntax lappend var_name val1 val2 ... lindex list_name index linsert list_name index val1 val2 ... list val1, val2 ... llength list_name lrange list_name first last lreplace list_name first last val1, val2, ... Description appends val1, val2, etc. to list var_name returns the index-th element of list_name; the first element is 0 inserts val1, val2, etc. just before the index-th element of list_name returns a Tcl list consisting of val1, val2, etc. returns the number of elements in list_name returns a sublist of list_name, from index first to index last; first or last may be "end", which refers to the last element in the list replaces elements first through last with val1, val2, etc.
Two other commands, lsearch and lsort, are also available for list manipulation. See the Tcl man pages (Help > Tcl Man Pages) for more information on these commands.
ModelSim Tcl commands

These additional commands enhance the interface between Tcl and ModelSim. Only brief descriptions are provided here; for more information and command syntax see the ModelSim Command Reference. Command alias (CR-54) find (CR-137) lshift (CR-150) lsublist (CR-151) printenv (CR-176) Description creates a new Tcl procedure that evaluates the specified commands; used to create a user-defined alias locates incrTcl classes and objects takes a Tcl list as argument and shifts it in-place one place to the left, eliminating the 0th element returns a sublist of the specified Tcl list that matches the specified Tcl glob pattern echoes to the Transcript pane the current names and values of all environment variables
ModelSim Tcl time commands UM-423
ModelSim Tcl time commands

ModelSim Tcl time commands make simulator-time-based values available for use within other Tcl procedures. Time values may optionally contain a units specifier where the intervening space is also optional. If the space is present, the value must be quoted (e.g. 10ns, "10 ns"). Time values without units are taken to be in the UserTimeScale. Return values are always in the current Time Scale Units. All time values are converted to a 64-bit integer value in the current Time Scale. This means that values smaller than the current Time Scale will be truncated to 0.
Conversions
Command intToTime <intHi32> <intLo32>
Description converts two 32-bit pieces (high and low order) into a 64-bit quantity (Time in ModelSim is a 64-bit integer) converts a <real> number to a 64-bit integer in the current Time Scale returns the value of <time> multiplied by the <scaleFactor> integer
RealToTime <real> scaleTime <time> <scaleFactor>
Relations
Command eqTime <time> <time> neqTime <time> <time> gtTime <time> <time> gteTime <time> <time> ltTime <time> <time> lteTime <time> <time>
Description evaluates for equal evaluates for not equal evaluates for greater than evaluates for greater than or equal evaluates for less than evaluates for less than or equal
All relation operations return 1 or 0 for true or false respectively and are suitable return values for TCL conditional expressions. For example,
if {[eqTime $Now 1750ns]} { ... }
Arithmetic
Command addTime <time> <time> divTime <time> <time> mulTime <time> <time> subTime <time> <time> Description add time 64-bit integer divide 64-bit integer multiply subtract time
Tcl examples UM-425
Tcl examples
This is an example of using the Tcl while loop to copy a list from variable a to variable b, reversing the order of the elements along the way:
set b [list] set i [expr {[llength $a] - 1}] while {$i >= 0} { lappend b [lindex $a $i] incr i -1 }
This example uses the Tcl for command to copy a list from variable a to variable b, reversing the order of the elements along the way:
set b [list] for {set i [expr {[llength $a] - 1}]} {$i >= 0} {incr i -1} { lappend b [lindex $a $i] }
This example uses the Tcl foreach command to copy a list from variable a to variable b, reversing the order of the elements along the way (the foreach command iterates over all of the elements of a list):
set b [list] foreach i $a { set b [linsert $b 0 $i] }
This example shows a list reversal as above, this time aborting on a particular element using the Tcl break command:
set b [list] foreach i $a { if {$i = "ZZZ"} break set b [linsert $b 0 $i] }
This example is a list reversal that skips a particular element by using the Tcl continue command:
set b [list] foreach i $a { if {$i = "ZZZ"} continue set b [linsert $b 0 $i] }
The next example works in UNIX only. In a Windows environment, the Tcl exec command will execute compiled files only, not system commands.) The example shows how you can access system information and transfer it into VHDL variables or signals and Verilog nets or registers. When a particular HDL source breakpoint occurs, a Tcl function is called that gets the date and time and deposits it into a VHDL signal of type STRING. If a particular environment variable (DO_ECHO) is set, the function also echoes the new date and time to the transcript file by examining the VHDL variable. (in VHDL source):
signal datime : string(1 to 28) := " ";# 28 spaces
(on VSIM command line or in macro):

proc set_date {} { global env set do_the_echo [set env(DO_ECHO)] set s [clock format [clock seconds]] force -deposit datime $s if {do_the_echo} { echo "New time is [examine -value datime]" } } bp src/waveadd.vhd 133 {set_date; continue} --sets the breakpoint to call set_date
This next example shows a complete Tcl script that restores multiple Wave windows to their state in a previous simulation, including signals listed, geometry, and screen position. It also adds buttons to the Main window toolbar to ease management of the wave files.
## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## This file contains procedures to manage multiple wave files. Source this file from the command line or as a startup script. source <path>/wave_mgr.tcl add_wave_buttons Add wave management buttons to the main toolbar (new, save and load) new_wave Dialog box creates a new wave window with the user provided name named_wave <name> Creates a new wave window with the specified title save_wave <file-root> Saves name, window location and contents for all open windows wave windows Creates <file-root><n>.do file for each window where <n> is 1 to the number of windows. Default file-root is "wave". Also creates windowSet.do file that contains title and geometry info. load_wave Opens where Also <file-root> and loads wave windows for all files matching <file-root><n>.do <n> are the numbers from 1-9. Default <file-root> is "wave". runs windowSet.do file if it exists.
## Add wave management buttons to the main toolbar proc add_wave_buttons {} { _add_menu main controls right SystemMenu SystemWindowFrame {Load Waves} \ load_wave _add_menu main controls right SystemMenu SystemWindowFrame {Save Waves} \ save_wave _add_menu main controls right SystemMenu SystemWindowFrame {New Wave} \ new_wave } ## Simple Dialog requests name of new wave window. Defaults to Wave<n> proc new_wave {} { global vsimPriv set defaultName "Wave[llength $vsimPriv(WaveWindows)]" set windowName [GetValue . "Create Named Wave Window:" $defaultName ]
Tcl examples UM-427
if {$windowName == ""} { # Dialog canceled # abort operation return } ## Debug puts "Window name: $windowName\n" if {$windowName == "{}"} { set windowName "" } if {$windowName != ""} { named_wave $windowName } else { named_wave $defaultName } } ## Creates a new wave window with the provided name (defaults to "Wave") proc named_wave {{name "Wave"}} { set newWave [view -new wave] if {[string length $name] > 0} { wm title $newWave $name } } ## Writes out format of all wave windows, stores geometry and title info in ## windowSet.do file. Removes any extra files with the same fileroot. ## Default file name is wave<n> starting from 1. proc save_wave {{fileroot "wave"}} { global vsimPriv set n 1 if {[catch {open windowSet_$fileroot.do w 755} fileId]} { error "Open failure for $fileroot ($fileId)" } foreach w $vsimPriv(WaveWindows) { echo "Saving: [wm title $w]" set filename $fileroot$n.do if {[file exists $filename]} { # Use different file set n2 0 while {[file exists ${fileroot}${n}${n2}.do]} { incr n2 } set filename ${fileroot}${n}${n2}.do } write format wave -window $w $filename puts $fileId "wm title $w \"[wm title $w]\"" puts $fileId "wm geometry $w [wm geometry $w]" puts $fileId "mtiGrid_colconfig $w.grid name -width \ [mtiGrid_colcget $w.grid name -width]" puts $fileId "mtiGrid_colconfig $w.grid value -width \ [mtiGrid_colcget $w.grid value -width]" flush $fileId incr n } foreach f [lsort [glob -nocomplain $fileroot\[$n-9\].do]] { echo "Removing: $f" exec rm $f
} } } ## Provide file root argument and load_wave restores all saved windows. ## Default file root is "wave". proc load_wave {{fileroot "wave"}} { foreach f [lsort [glob -nocomplain $fileroot\[1-9\].do]] { echo "Loading: $f" view -new wave do $f } if {[file exists windowSet_$fileroot.do]} { do windowSet_$fileroot.do } } ...
This next example specifies the compiler arguments and lets you compile any number of files.
set Files [list] set nbrArgs $argc for {set x 1} {$x <= $nbrArgs} {incr x} { set lappend Files $1 shift } eval vcom -93 -explicit -noaccel $Files
This example is an enhanced version of the last one. The additional code determines whether the files are VHDL or Verilog and uses the appropriate compiler and arguments depending on the file type. Note that the macro assumes your VHDL files have a .vhd file extension.
set set set for vhdFiles [list] vFiles [list] nbrArgs $argc {set x 1} {$x <= $nbrArgs} {incr x} { if {[string match *.vhd $1]} { lappend vhdFiles $1 } else { lappend vFiles $1 } shift
} if {[llength $vhdFiles] > 0} { eval vcom -93 -explicit -noaccel $vhdFiles } if {[llength $vFiles] > 0} { eval vlog $vFiles }
Macros (DO files) UM-429
Macros (DO files)

ModelSim macros (also called DO files) are simply scripts that contain ModelSim and, optionally, Tcl commands. You invoke these scripts with the Tools > Execute Macro menu selection or the do command (CR-125).
Creating DO files
You can create DO files, like any other Tcl script, by typing the required commands in any editor and saving the file. Alternatively, you can save the transcript as a DO file (see "Saving the transcript file" (GR-19)). All "event watching" commands (e.g. onbreak (CR-170), onerror (CR-172), etc.) must be placed before run (CR-197) commands within the macros in order to take effect. The following is a simple DO file that was saved from the transcript. It is used in the dataset exercise in the ModelSim Tutorial. This DO file adds several signals to the Wave window, provides stimulus to those signals, and then advances the simulation.
add wave ld add wave rst add wave clk add wave d add wave q force -freeze clk 0 0, 1 {50 ns} -r 100 force rst 1 force rst 0 10 force ld 0 force d 1010 onerror {cont} run 1700 force ld 1 run 100 force ld 0 run 400 force rst 1 run 200 force rst 0 10 run 1500
Using Parameters with DO files

You can increase the flexibility of DO files by using parameters. Parameters specify values that are passed to the corresponding parameters $1 through $9 in the macro file. For example say the macro "testfile" contains the line bp $1 $2. The command below would place a breakpoint in the source file named design.vhd at line 127:
do testfile design.vhd 127
There is no limit on the number of parameters that can be passed to macros, but only nine values are visible at one time. You can use the shift command (CR-208) to see the other parameters.
Deleting a file from a .do script

To delete a file from a .do script, use the Tcl file command as follows:
file delete myfile.log
This will delete the file "myfile.log." You can also use the transcript file command to perform a deletion:
transcript file () transcript file my file.log
The first line will close the current log file. The second will open a new log file. If it has the same name as an existing file, it will replace the previous one.
Making macro parameters optional

If you want to make macro parameters optional (i.e., be able to specify fewer parameter values with the do command than the number of parameters referenced in the macro), you must use the argc (UM-455) simulator state variable. The argc simulator state variable returns the number of parameters passed. The examples below show several ways of using argc.
Example 1
This macro specifies the files to compile and handles 0-2 compiler arguments as parameters. If you supply more arguments, ModelSim generates a message.
switch $argc { 0 {vcom file1.vhd file2.vhd file3.vhd } 1 {vcom $1 file1.vhd file2.vhd file3.vhd } 2 {vcom $1 $2 file1.vhd file2.vhd file3.vhd } default {echo Too many arguments. The macro accepts 0-2 args. }
Example 2
This macro specifies the compiler arguments and lets you compile any number of files.
variable Files "" set nbrArgs $argc for {set x 1} {$x <= $nbrArgs} {incr x} { set Files [concat $Files $1] shift } eval vcom -93 -explicit -noaccel $Files
Example 3
This macro is an enhanced version of the one shown in example 2. The additional code determines whether the files are VHDL or Verilog and uses the appropriate compiler and arguments depending on the file type. Note that the macro assumes your VHDL files have a .vhd file extension.
variable vhdFiles "" variable vFiles "" set nbrArgs $argc set vhdFilesExist 0 set vFilesExist 0 for {set x 1} {$x <= $nbrArgs} {incr x} { if {[string match *.vhd $1]} { set vhdFiles [concat $vhdFiles $1] set vhdFilesExist 1 } else { set vFiles [concat $vFiles $1]
Macros (DO files) UM-431
set vFilesExist 1 } shift } if {$vhdFilesExist == 1} { eval vcom -93 -explicit -noaccel $vhdFiles } if {$vFilesExist == 1} { eval vlog $vFiles }
Useful commands for handling breakpoints and errors

If you are executing a macro when your simulation hits a breakpoint or causes a run-time error, ModelSim interrupts the macro and returns control to the command line. The following commands may be useful for handling such events. (Any other legal command may be executed as well.) command run (CR-197) -continue onbreak (CR-170) onElabError (CR-171) onerror (CR-172) status (CR-212) abort (CR-42) pause (CR-173) result continue as if the breakpoint had not been executed, completes the run (CR-197) that was interrupted specify a command to run when you hit a breakpoint within a macro specify a command to run when an error is encountered during elaboration specify a command to run when an error is encountered within a macro get a traceback of nested macro calls when a macro is interrupted terminate a macro once the macro has been interrupted or paused cause the macro to be interrupted; the macro can be resumed by entering a resume command (CR-196) via the command line
You can also set the OnErrorDefaultAction Tcl variable to determine what action ModelSim takes when an error occurs. To set the variable on a permanent basis, you must define the variable in a modelsim.tcl file (see "The modelsim.tcl file" (GR-253) for details).
Error action in DO files

If a command in a macro returns an error, ModelSim does the following: 1 If an onerror (CR-172) command has been set in the macro script, ModelSim executes that command. The onerror (CR-172) command must be placed prior to the run command in the DO file to take effect. 2 If no onerror command has been specified in the script, ModelSim checks the OnErrorDefaultAction variable. If the variable is defined, its action will be invoked.
3 If neither 1 or 2 is true, the macro aborts.
Using the Tcl source command with DO files

Either the do command or Tcl source command can execute a DO file, but they behave differently. With the source command, the DO file is executed exactly as if the commands in it were typed in by hand at the prompt. Each time a breakpoint is hit, the Source window is updated to show the breakpoint. This behavior could be inconvenient with a large DO file containing many breakpoints. When a do command is interrupted by an error or breakpoint, it does not update any windows, and keeps the DO file "locked". This keeps the Source window from flashing, scrolling, and moving the arrow when a complex DO file is executed. Typically an onbreak resume command is used to keep the macro running as it hits breakpoints. Add an onbreak abort command to the DO file if you want to exit the macro and update the Source window.
UM-433
A - Simulator variables
Appendix contents
Variable settings report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-434 UM-434 UM-435 UM-436 UM-437 UM-438 UM-439 UM-439 UM-441 UM-442 UM-444 UM-449 UM-451 UM-454 UM-455 UM-455 UM-456 Environment variables . . . . . . . . . . Creating environment variables in Windows . . . Referencing environment variables within ModelSim Removing temp files (VSOUT) . . . . . . Control variables located in INI files . . . . [Library] library path variables . . . . [vlog] Verilog compiler control variables. . [vcom] VHDL compiler control variables . [sccom] SystemC compiler control variables . [vsim] simulator control variables . . . [msg_system] message system variables . . Commonly used INI variables . . . . Variable precedence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Simulator state variables . . . . . . . Referencing simulator state variables . . . Special considerations for the now variable .
This appendix documents the following types of variables: environment variables Variables referenced and set according to operating system conventions. Environment variables prepare the ModelSim environment prior to simulation. simulator control variables Variables used to control compiler, simulator , and various other functions. simulator state variables Variables that provide feedback on the state of the current simulation.
UM-434 A - Simulator variables
Variable settings report

The report command (CR-192) returns a list of current settings for either the simulator state or simulator control variables. Use the following commands at either the ModelSim or VSIM prompt:
report simulator state report simulator control
Environment variables
Before compiling or simulating, several environment variables may be set to provide the functions described in the table below. The variables are in the autoexec.bat file on Windows 98/Me machines, and set through the System control panel on NT/2000/XP machines. For UNIX, the variables are typically found in the .login script. The LM_LICENSE_FILE variable is required; all others are optional. Environment variable DOPATH Description used by ModelSim to search for DO files (macros); consists of a colon-separated (semi-colon for Windows) list of paths to directories; this environment variable can be overridden by the DOPATH Tcl preference variable The DOPATH environment variable isnt accessible when you invoke vsim from a Unix shell or from a Windows command prompt. It is accessible once ModelSim or vsim is invoked. If you need to invoke from a shell or command line and use the DOPATH environment variable, use the following syntax:
vsim -do "do <dofile_name>" <design_unit>
EDITOR HOME HOME_0IN LM_LICENSE_FILE
specifies the editor to invoke with the edit command (CR-129) used by ModelSim to look for an optional graphical preference file and optional location map file; see: "Control variables located in INI files" (UM-438) identifies the location of the 0-In executables directory; see 0-In documentation for more information used by the ModelSim license file manager to find the location of the license file; may be a colon-separated (semi-colon for Windows) set of paths, including paths to other vendor license files; REQUIRED set by all ModelSim tools to the directory in which the binary executable resides; DO NOT SET THIS VARIABLE! used by ModelSim to find Tcl libraries for Tcl/Tk 8.3 and vsim; may also be used to specify a startup DO file; defaults to /modeltech/../tcl; may be set to an alternate path used by ModelSim tools to find source files based on easily reallocated "soft" paths; optional
MODEL_TECH MODEL_TECH_TCL
MGC_LOCATION_MAP
Environment variables UM-435
Environment variable MODELSIM
Description used by all ModelSim tools to find the modelsim.ini file; consists of a path including the file name. An alternative use of this variable is to set it to the path of a project file (<Project_Root_Dir>/<Project_Name>.mpf). This allows you to use project settings with command line tools. However, if you do this, the .mpf file will replace modelsim.ini as the initialization file for all ModelSim tools. specifies the location where user interface preferences will be stored. Setting this variable with the path of a file will cause ModelSim to use this file instead of the default location (the users HOME directory in UNIX, and in the registry in Windows). The file does not need to exist beforehand, ModelSim will initialize it. Also, if this file is read-only, ModelSim will not update or otherwise modify the file. This variable may contain a relative pathname in which case the file will be relative to the working directory at the time the tool is started. used by ModelSim to look for an optional graphical preference file; can be a colon-separated (UNIX) or semi-colon separated (Windows) list of file paths creates an mti_trace_cosim file containing debugging information about FLI/PLI/ VPI function calls; set to any value before invoking the simulator. limits the size of the VSOUT temp file (generated by the ModelSim kernel); the value of the variable is the size of k-bytes; TMPDIR (below) controls the location of this file, STDOUT controls the name; default = 10, 0 = no limit; does not control the size of the transcript file specifies the directory into which object libraries are compiled when using the -compile_uselibs argument to the vlog command (CR-293) if set to 1, disables memory mapping in ModelSim; this should be used only when running on Linux 7.1; it will decrease the speed with which ModelSim reads files used by ModelSim to search for PLI object files for loading; consists of a space-separated list of file or path names the VSOUT temp file (generated by the simulator kernel) is deleted when the simulator exits; the file is not deleted if you specify a filename for VSOUT with STDOUT; specifying a name and location (use TMPDIR) for the VSOUT file will also help you locate and delete the file in event of a crash (an unnamed VSOUT file is not deleted after a crash either) specifies the path to a tempnam() generated file (VSOUT) containing all stdout from the simulation kernel
MODELSIM_PREFERE NCES
MODELSIM_TCL MTI_COSIM_TRACE MTI_TF_LIMIT
MTI_USELIB_DIR NOMMAP PLIOBJS STDOUT
TMPDIR (Unix) TMP (Windows)
Creating environment variables in Windows

In addition to the predefined variables shown above, you can define your own environment variables. This example shows a user-defined library path variable that can be referenced by the vmap command to add library mapping to the modelsim.ini file.
Using Windows 98/Me

Open and edit the autoexec.bat file by adding this line:
set MY_PATH=\temp\work
Restart Windows to initialize the new variable.
Using Windows NT/2000/XP

Right-click the My Computer icon and select Properties, then select the Environment tab (in Windows 2000/XP select the Advanced tab and then Environment Variables). Add the new variable with this dataVariable:MY_PATH and Value:\temp\work. Click Set and Apply to initialize the variable.
Library mapping with environment variables

Once the MY_PATH variable is set, you can use it with the vmap command (CR-304) to add library mappings to the current modelsim.ini file. If youre using the vmap command from a DOS prompt type:
vmap MY_VITAL %MY_PATH%
If youre using vmap from the ModelSim/VSIM prompt type:

vmap MY_VITAL \$MY_PATH
If you used DOS vmap, this line will be added to the modelsim.ini:
MY_VITAL = c:\temp\work
If vmap is used from the ModelSim/VSIM prompt, the modelsim.ini file will be modified with this line:
MY_VITAL = $MY_PATH
You can easily add additional hierarchy to the path. For example,
vmap MORE_VITAL %MY_PATH%\more_path\and_more_path vmap MORE_VITAL \$MY_PATH\more_path\and_more_path
The "$" character in the examples above is Tcl syntax that precedes a variable. The "\" character is an escape character that keeps the variable from being evaluated during the execution of vmap.
Referencing environment variables within ModelSim

There are two ways to reference environment variables within ModelSim. Environment variables are allowed in a FILE variable being opened in VHDL. For example,
use std.textio.all; entity test is end; architecture only of test is begin process FILE in_file : text is in "$ENV_VAR_NAME"; begin wait; end process; end;
Environment variables may also be referenced from the ModelSim command line or in macros using the Tcl env array mechanism:
Environment variables UM-437
echo "$env(ENV_VAR_NAME)"
Note: Environment variable expansion does not occur in files that are referenced via the -f argument to vcom, vlog, or vsim.
Removing temp files (VSOUT)

The VSOUT temp file is the communication mechanism between the simulator kernel and the ModelSim GUI. In normal circumstances the file is deleted when the simulator exits. If ModelSim crashes, however, the temp file must be deleted manually. Specifying the location of the temp file with TMPDIR (above) will help you locate and remove the file.
Control variables located in INI files

Initialization (INI) files contain control variables that specify reference library paths and compiler and simulator settings. The default initialization file is modelsim.ini and is located in your install directory. To set these variables, edit the initialization file directly with any text editor. The syntax for variables in the file is:
<variable> = <value>
Comments within the file are preceded with a semicolon ( ; ). The following tables list the variables by section, and in order of their appearance within the INI file: INI file sections [Library] library path variables (UM-439) [vlog] Verilog compiler control variables (UM-439) [vcom] VHDL compiler control variables (UM-441) [sccom] SystemC compiler control variables (UM-442) [vsim] simulator control variables (UM-444) [msg_system] message system variables (UM-449)
Control variables located in INI files UM-439
[Library] library path variables

Variable name ieee Value range any valid path; may include environment variables any valid path; may include environment variables any valid path; may include environment variables any valid path; may include environment variables any valid path; may include environment variables any valid path; may include environment variables any valid path; may include environment variables any valid path; may include environment variables Purpose sets the path to the library containing IEEE and Synopsys arithmetic packages; the default is $MODEL_TECH/../ieee sets the path to the library containing Model Technology VHDL utilities such as Signal Spy; the default is $MODEL_TECH/../modelsim_lib sets the path to the VHDL STD library; the default is $MODEL_TECH/../std sets the path to the libraries for MGC standard developers kit; the default is $MODEL_TECH/../std_developerskit sets the path to the accelerated arithmetic packages; the default is $MODEL_TECH/../ synopsys sets the path to the library containing VHDL/ Verilog type mappings; the default is $MODEL_TECH/../verilog sets the path to the VITAL 2000 library; the default is $MODEL_TECH/../vital2000 points to another modelsim.ini file whose library path variables will also be read; the pathname must include "modelsim.ini"; only one others variable can be specified in any modelsim.ini file.
modelsim_lib
std std_developerskit
synopsys
verilog
vital2000 others
[vlog] Verilog compiler control variables

Variable name DisableOpt Hazard GenerateLoopIterationMax Value range 0, 1 0, 1 natural integer (>=0) natural integer (>=0) Purpose if 1, disables all optimizations enacted by the compiler; same as the -O0 argument to vlog if 1, turns on Verilog hazard checking (orderdependent accessing of global variables) the maximum number of iterations permitted for a generate loop; restricting this permits the implementation to recognize infinite generate loops the maximum depth permitted for a recursive generate instantiation; restricting this permits the implementation to recognize infinite recursions Default off (0) off (0) 100000
GenerateRecursionDepthMax
200
Variable name Incremental MultiFileCompilationUnit
Value range 0, 1 0, 1
Purpose if 1, turns on incremental compilation of modules if 1, treats all files within the compilation command line as a single compilation unit; default behavior in versions prior to 6.1was to treat all files as separate compilation units if 1, turns off inclusion of debugging info within design units if 1, enables `protect directive processing; see "Compiler directives specific to ModelSim" (UM-133) for details if 1, turns off "loading..." messages if 1, generates a warning whenever an unknown plus argument is encountered if 1, displays lint warning messages if 1, displays warning when the simulator encounters constructs which code coverage cant handle if 1, displays warning messages about non-LRM compliance in order to match Cadence behavior if 1, shows source line containing error if 1, disables SystemVerilog and Verilog 2001 support and makes compiler compatible with IEEE Std 1364-1995
Default off (0) off (0)
NoDebug Protect
0, 1 0, 1
off (0) off (0)
Quiet Show_BadOptionWarning Show_Lint Show_WarnCantDoCoverage Show_ WarnMatchCadence Show_source vlog95compat
0, 1 0, 1 0, 1 0,1 0, 1 0, 1 0, 1
off (0) off (0) off (0) on (1) on (1) off (0) off (0)
[vcom] VHDL compiler control variables

Variable name BindAtCompile Value range 0, 1 Purpose if 1, instructs ModelSim to perform VHDL default binding at compile time rather than load time; see "Default binding" (UM-76) for further details if 1, turns on limited synthesis rule compliance checking; checks only signals used (read) by a process; also, understands only combinational logic, not clocked logic if 1, disables all optimizations enacted by the compiler; same as the -O0 argument to vcom if 1, turns on resolving of ambiguous function overloading in favor of the "explicit" function declaration (not the one automatically created by the compiler for each type declaration) if 1, ignores VITAL compliance checking errors if 1, changes case statement static errors to warnings if 1, turns off inclusion of debugging info within design units if 1, run time index checks are disabled if 1, disables errors caused by aggregates that are not locally static if 1, disables run time range checking if 1, turns off acceleration of the VITAL packages if 1, turns off VITAL compliance checking if 0, turns off optimization for the IEEE std_logic_1164 package if 1, overrides NoCaseStaticError and NoOthersStaticError if 1, turns off "loading..." messages if 1, instructs the compiler not to generate a default binding during compilation if 1, turns on lint-style checking if 1, shows source line containing error if 0, turns off VITAL optimization warnings Default off (0)
CheckSynthesis
0, 1
off (0)
DisableOpt Explicit
0, 1 0, 1
off (0) on (1)
IgnoreVitalErrors NoCaseStaticError NoDebug NoIndexCheck NoOthersStaticError NoRangeCheck NoVital NoVitalCheck Optimize_1164 PedanticErrors Quiet RequireConfigForAllDefault Binding Show_Lint Show_source Show_VitalChecksOpt
0, 1 0, 1 0, 1 0, 1 0, 1 0, 1 0, 1 0, 1 0, 1 0, 1 0, 1 0, 1 0, 1 0, 1 0, 1
off (0) off (0) off (0) off (0) off (0) off (0) off (0) off (0) on (1) off(0) off (0) off (0) off (0) off (0) on (1)
Variable name Show_VitalChecksWarnings Show_WarnCantDoCoverage
Value range 0, 1 0, 1
Purpose if 0, turns off VITAL compliance-check warnings if 0, turns off warnings when the simulator encounters constructs which code coverage cant handle if 0, turns off unbound-component warnings if 0, turns off process-without-a-wait-statement warnings if 0, turns off null-range warnings if 0, turns off no-space-in-time-literal warnings if 0, turns off multiple-drivers-on-unresolved-signal warnings if 0, turns off warnings about signal value dependency at elaboration if 0, turns off warnings about VHDL-1993 constructs in VHDL-1987 code if 0, turns off warnings about locally static errors deferred until run time if 0, enables support for VHDL-1987; if 1, enables support for VHDL-1993; if 2, enables support for VHDL-2002
Default on (1) on (1)
Show_Warning1 Show_Warning2 Show_Warning3 Show_Warning4 Show_Warning5 Show_Warning9 Show_Warning10 Show_WarnLocallyStaticErr or VHDL93
0, 1 0, 1 0, 1 0, 1 0, 1 0, 1 0, 1 0, 1 0, 1, 2
on (1) on (1) on (1) on (1) on (1) on (1) on (1) on (1) 2
[sccom] SystemC compiler control variables

Variable name CppOptions Value range any valid C+++ compiler options C++ compiler path Purpose adds any specified C++ compiler options to the sccom command line at the time of invocation Default none
CppPath
If used, variables should point directly to the location of the g++ executable, such as:
% CppPath /usr/bin/g++
none
This variable is not required when running SystemC designs. By default, you should install and use the built-in g++ compiler that comes with ModelSim SccomLogfile SccomVerbose 0, 1 0, 1 if 1, creates a logfile for sccom if 1, turns on verbose messages from sccom (CR-199): see "-verbose" (CR-201) for details off (0) off (0)
Variable name UseScv
Value range 0, 1
Purpose if 1, turns on use of SCV include files and library; see"-scv" (CR-201) for details
Default off (0)
[vsim] simulator control variables

Variable name AssertFile AssertionFormat Value range any valid filename see next column Purpose alternative file for storing VHDL assertion messages defines format of VHDL assertion messages; fields include: %S - severity level %R - report message %T - time of assertion %D - delta %I - instance or region pathname (if available) %i - instance pathname with process %O - process name %K - kind of object path points to; returns Instance, Signal, Process, or Unknown %P - instance or region path without leaf process %F - file %L - line number of assertion, or if from subprogram, line from which call is made %% - print % character defines format of messages for VHDL assertions that trigger a breakpoint; see AssertionFormat for options Default transcript "** %S: %R\n Time: %T Iteration: %D%I\n"
AssertionFormatBreak
see AssertionFormat above
"** %S: %R\n Time: %T Iteration: %D %K: %i File: %F\n" "** %S: %R\n Time: %T Iteration: %D %K: %i File: %F\n" "** %S: %R\n Time: %T Iteration: %D %K: %i File: %F\n"
AssertionFormatError
defines format of messages for VHDL Error assertions; see AssertionFormat for options; if undefined, AssertionFormat is used unless assertion causes a breakpoint in which case AssertionFormatBreak is used defines format of messages for VHDL Fail assertions; see AssertionFormat for options; if undefined, AssertionFormat is used unless assertion causes a breakpoint in which case AssertionFormatBreak is used
AssertionFormatFail
Variable name AssertionFormatFatal
Value range see AssertionFormat above
Purpose defines format of messages for VHDL Fatal assertions; see AssertionFormat for options; if undefined, AssertionFormat is used unless assertion causes a breakpoint in which case AssertionFormatBreak is used defines format of messages for VHDL Note assertions; see AssertionFormat for options; if undefined, AssertionFormat is used unless assertion causes a breakpoint in which case AssertionFormatBreak is used defines format of messages for VHDL Warning assertions; see AssertionFormat for options; if undefined, AssertionFormat is used unless assertion causes a breakpoint in which case AssertionFormatBreak is used defines severity of VHDL assertions that cause a simulation break (0 = note, 1 = warning, 2 = error, 3 = failure, 4 = fatal) if 0, vsim ignores unrecognized plusargs; if 1, vsim produces warnings for unrecognized plusargs, but will simulate ignoring the unrecognized plusargs; if 2, vsim produces errors for unrecognized plusargs and exits if 1, checkpoint files are written in compressed format sets the name of a file in which to store the Main window command history controls the number of VHDL files open concurrently; this number should be less than the current limit setting for max file descriptors; 0 = unlimited the dataset separator for fully-rooted contexts, for example sim:/top; must not be the same character as PathSeparator
Default "** %S: %R\n Time: %T Iteration: %D %K: %i File: %F\n" "** %S: %R\n Time: %T Iteration: %D%I\n" "** %S: %R\n Time: %T Iteration: %D%I\n" 3
AssertionFormatNote
AssertionFormatWarning
BreakOnAssertion
0-4
CheckPlusargs
0, 1, 2
off (0)
CheckpointCompressMode CommandHistory ConcurrentFileLimit
0, 1 any valid filename any positive integer
on (1) commented out (;) 40
DatasetSeparator
any character except those with special meaning (i.e., \, {, }, etc.)
Variable name DefaultForceKind
Value range freeze, drive, or deposit
Purpose defines the kind of force used when not otherwise specified
Default drive for resolved signals; freeze for unresolved signals symbolic
DefaultRadix
symbolic, binary, octal, decimal, unsigned, hexadecimal, ascii one or more of: -force, -noassertions, -nobreakpoint, -nofcovers, -nolist, -nolog, -nowave 0, 1 Any non-quoted string containing at a minimum a %s followed by a %d comma seperated list of filenames 0,1 0,1 0,1 0,1 positive integer
a numeric radix may be specified as a name or number (i.e., binary can be specified as binary or 2; octal as octal or 8; etc.)
DefaultRestartOptions
sets default behavior for the restart command
commented out (;)
DelayFileOpen GenerateFormat
if 1, open VHDL87 files on first read or write, else open files when elaborated controls the format of a generate statement label (don't quote it)
off (0) %s__%d
GlobalSharedObjectsList IgnoreError IgnoreFailure IgnoreNote IgnoreWarning IterationLimit
loads the specified PLI/FLI shared objects with global symbol visibility if 1, ignore VHDL assertion errors if 1, ignore VHDL assertion failures if 1, ignore VHDL assertion notes if 1, ignore VHDL assertion warnings limit on simulation kernel iterations allowed without advancing time
commented out (;) off (0) off (0) off (0) off (0) 5000
Variable name License
Value range one ore more <license_option> described in the next column; separate by spaces if using multiple entries
Purpose if set, controls ModelSim license file search; license options include: lnlonly - only use msimhdlsim and hdlsim mixedonly - exclude single language licenses nomgc - exclude MGC licenses nolnl - exclude language neutral licenses nomix - exclude msimhdlmix and hdlmix nomti - exclude MTI licenses noqueue - do not wait in license queue if no licenses are available noslvhdl - exclude qhsimvh and vsim noslvlog - exclude qhsimvl and vsimvlog plus - only use PLUS license vlog - only use VLOG license vhdl - only use VHDL license viewsim - accepts a simulation license rather than being queued for a viewer license see also the vsim command (CR-305) <license_option>
Default search all licenses
NumericStdNoWarnings
0, 1
if 1, warnings generated within the accelerated numeric_std and numeric_bit packages are suppressed used for hierarchical pathnames; must not be the same character as DatasetSeparator; you must specify either a slash (/) or a period (.) when creating a virtual bus simulator resolution; no space between value and units (i.e., 10fs, not 10 fs); overridden by the -t argument to vsim (CR305); if your delays get truncated, set the resolution smaller; this value must be less than or equal to the UserTimeUnit (described below) default simulation length in units specified by the UserTimeUnit variable sets the default time unit for SystemC simulations
off (0)
PathSeparator
any character except those with special meaning (i.e., \, {, }, etc.) fs, ps, ns, us, ms, or sec with optional prefix of 1, 10, or 100
Resolution
ns
RunLength ScTimeUnit
positive integer fs, ps, ns, us, ms, or sec with optional prefix of 1, 10, or 100 0, 1
100 1 ns
ShowUnassociatedScNameWa rning
if 1, displays unassociated SystemC name warnings
off (1)
Variable name ShowUndebuggableScTypeWa rning Startup
Value range 0, 1 = do <DO filename>; any valid macro (do) file 0, 1
Purpose if 1, displays undebuggable SystemC type warnings specifies the ModelSim startup macro; see the do command (CR-125)
Default on (1) commented out (;)
StdArithNoWarnings
if 1, warnings generated within the accelerated Synopsys std_arith packages are suppressed sets the maximum number of VHDL integer values to record with toggle coverage file for saving command transcript; environment variables may be included in the pathname controls VHDL and Verilog files open for write; 0 = Buffered, 1 = Unbuffered instructs vsim to use /usr/lib/libCsup_v2.sl for shared object loading; for use only on HP-UX 11.00 when you have compiled FLI/PLI/VPI C++ code with aCC's -AA option specifies scaling for the Wave window and the default time units to use for commands such as force (CR-141) and run (CR-197); should generally be set to default, in which case it takes the value of the Resolution variable list of dynamically loadable objects for Verilog PLI/VPI applications; see Appendix C - Verilog PLI / VPI / DPI controls whether a warning is issued when the change command changes the value of a VHDL constant or generic controls the number of visible hierarchical regions of a signal name shown in the "Wave window" (GR-194); the default value of zero displays the full name, a setting of one or above displays the corresponding level(s) of hierarchy
off (0)
ToggleMaxIntValues
positive integer
100
TranscriptFile
any valid filename 0, 1 0, 1
transcript
UnbufferedOutput UseCsupV2
0 off (0)
UserTimeUnit
fs, ps, ns, us, ms, sec, or default
default
Veriuser
one or more valid shared object names 0, 1
commented out (;) on (1)
WarnConstantChange
WaveSignalNameWidth
0, positive integer
Variable name WLFCacheSize
Value range positive integer
Purpose sets the number of megabytes for the WLF reader cache; WLF reader caching caches blocks of the WLF file to reduce redundant file I/O if 0, WLF file records values at every change of the logged objects; if 1, WLF file records values only at the end of each delta step; if 2, WLF file records values only at the end of a simulator time step turns WLF file compression on (1) or off (0) specifies whether a WLF file should be deleted when the simulation ends; if set to 0, the file is not deleted; if set to 1, the file is deleted specifies the default WLF file name specifies whether the viewing of waveforms is optimized; default is enabled specifies whether to save all design hierarchy in the WLF file (1) or only regions containing logged signals (0) WLF file size limit; limits WLF file by size (as closely as possible) to the specified number of megabytes; if both size and time limits are specified the most restrictive is used; setting to 0 results in no limit WLF file time limit; limits WLF file by time (as closely as possible) to the specified amount of time. If both time and size limits are specified the most restrictive is used; setting to 0 results in no limit
Default 0
WLFCollapseMode
0, 1, 2
WLFCompress WLFDeleteOnQuit
0, 1 0, 1
1 0
WLFFilename WLFOptimize WLFSaveAllRegions
0, 1 0, 1 0, 1
vsim.wlf 1 0
WLFSizeLimit
0 - positive integer of MB
WLFTimeLimit
0 - positive integer, time unit is optional
[msg_system] message system variables

The message system variables help you identify and troubleshoot problems while using the application. See also Message system (UM-458). Variable name error Value range list of message numbers Purpose changes the severity of the listed message numbers to "error"; see "Changing message severity level" (UM458) for more information Default none
Variable name note
Value range list of message numbers list of message numbers list of message numbers
Purpose changes the severity of the listed message numbers to "note"; see "Changing message severity level" (UM458) for more information suppresses the listed message numbers; see "Changing message severity level" (UM-458) for more information changes the severity of the listed message numbers to "warning"; see "Changing message severity level" (UM-458) for more information
Default none
suppress
none
warning
none
Commonly used INI variables

Several of the more commonly used modelsim.ini variables are further explained below.
Environment variables
You can use environment variables in your initialization files. Use a dollar sign ($) before the environment variable name. For example:
[Library] work = $HOME/work_lib test_lib = ./$TESTNUM/work ... [vsim] IgnoreNote = $IGNORE_ASSERTS IgnoreWarning = $IGNORE_ASSERTS IgnoreError = 0 IgnoreFailure = 0
There is one environment variable, MODEL_TECH, that you cannot and should not set. MODEL_TECH is a special variable set by Model Technology software. Its value is the name of the directory from which the VCOM or VLOG compilers or VSIM simulator was invoked. MODEL_TECH is used by the other Model Technology tools to find the libraries.
Hierarchical library mapping

By adding an "others" clause to your modelsim.ini file, you can have a hierarchy of library mappings. If the ModelSim tools dont find a mapping in the modelsim.ini file, then they will search only the library section of the initialization file specified by the "others" clause. For example:
[Library] asic_lib = /cae/asic_lib work = my_work others = /install_dir/modeltech/modelsim.ini
Since the file referred to by the "others" clause may itself contain an "others" clause, you can use this feature to chain a set of hierarchical INI files for library mappings.
Creating a transcript file

A feature in the system initialization file allows you to keep a record of everything that occurs in the transcript: error messages, assertions, commands, command outputs, etc. To do this, set the value for the TranscriptFile line in the modelsim.ini file to the name of the file in which you would like to record the ModelSim history.
; Save the command window contents to this file TranscriptFile = trnscrpt
You can disable the creation of the transcript file by using the following ModelSim command immediately after ModelSim starts:
transcript file ""
Using a startup file

The system initialization file allows you to specify a command or a do file that is to be executed after the design is loaded. For example:
; VSIM Startup command Startup = do mystartup.do
The line shown above instructs ModelSim to execute the commands in the macro file named mystartup.do.
; VSIM Startup command Startup = run -all
The line shown above instructs VSIM to run until there are no events scheduled. See the do command (CR-125) for additional information on creating do files.
Turning off assertion messages

You can turn off assertion messages from your VHDL code by setting a switch in the modelsim.ini file. This option was added because some utility packages print a huge number of warnings.
[vsim] IgnoreNote = 1 IgnoreWarning = 1 IgnoreError = 1 IgnoreFailure = 1
Turning off warnings from arithmetic packages

You can disable warnings from the Synopsys and numeric standard packages by adding the following lines to the [vsim] section of the modelsim.ini file.
[vsim] NumericStdNoWarnings = 1 StdArithNoWarnings = 1
Force command defaults

The force command has -freeze, -drive, and -deposit options. When none of these is specified, then -freeze is assumed for unresolved signals and -drive is assumed for resolved signals. But if you prefer -freeze as the default for both resolved and unresolved signals, you can change the defaults in the modelsim.ini file.
[vsim] ; Default Force Kind ; The choices are freeze, drive, or deposit DefaultForceKind = freeze
Restart command defaults

The restart command has -force, -nobreakpoint, -nofcovers, -nolist, -nolog, and -nowave options. You can set any of these as defaults by entering the following line in the modelsim.ini file:
DefaultRestartOptions = <options>
where <options> can be one or more of -force, -nobreakpoint, -nofcovers, -nolist, -nolog, and -nowave. Example:
DefaultRestartOptions = -nolog -force
VHDL standard
You can specify which version of the 1076 Std ModelSim follows by default using the VHDL93 variable:
[vcom] ; VHDL93 variable selects language version as the default. ; Default is VHDL-2002. ; Value of 0 or 1987 for VHDL-1987. ; Value of 1 or 1993 for VHDL-1993. ; Default or value of 2 or 2002 for VHDL-2002. VHDL93 = 2002
Opening VHDL files

You can delay the opening of VHDL files with an entry in the INI file if you wish. Normally VHDL files are opened when the file declaration is elaborated. If the DelayFileOpen option is enabled, then the file is not opened until the first read or write to that file.
[vsim] DelayFileOpen = 1
Variable precedence
Note that some variables can be set in a .modelsim file (Registry in Windows) or a .ini file. A variable set in the .modelsim file takes precedence over the same variable set in a .ini file. For example, assume you have the following line in your modelsim.ini file:
TranscriptFile = transcript
And assume you have the following line in your .modelsim file:
set PrefMain(file) {}
In this case the setting in the .modelsim file overrides that in the modelsim.ini file, and a transcript file will not be produced.
Simulator state variables UM-455
Simulator state variables

Unlike other variables that must be explicitly set, simulator state variables return a value relative to the current simulation. Simulator state variables can be useful in commands, especially when used within ModelSim DO files (macros). The variables are referenced in commands by prefixing the name with a dollar sign ($). Variable argc architecture Result returns the total number of parameters passed to the current macro returns the name of the top-level architecture currently being simulated; for a configuration or Verilog module, this variable returns an empty string returns the name of the top-level configuration currently being simulated; returns an empty string if no configuration returns the number of the current simulator iteration returns the name of the top-level VHDL entity or Verilog module currently being simulated returns the library name for the current region returns the current depth of macro call nesting represents a macro parameter, where n can be an integer in the range 1-9 always returns the current simulation time with time units (e.g., 110,000 ns) Note: will return a comma between thousands when time resolution is a unary unit (i.e., 1ns, 1ps, 1fs): returns the current simulation time without time units (e.g., 100000) when time resolution is a multiple of the unary unit (i.e., 10ns, 100ps, 10fs): returns the current simulation time with time units (e.g. 110000 ns) Note: will not return comma between thousands returns the current simulation time resolution
configuration delta entity library MacroNestingLevel n Now now
resolution
Referencing simulator state variables

Variable values may be referenced in simulator commands by preceding the variable name with a dollar sign ($). For example, to use the now and resolution variables in an echo command type:
echo "The time is $now $resolution."
Depending on the current simulator state, this command could result in:
The time is 12390 ps 10ps.
If you do not want the dollar sign to denote a simulator variable, precede it with a "\". For example, \$now will not be interpreted as the current simulator time.
Special considerations for the now variable

For the when command (CR-327), special processing is performed on comparisons involving the now variable. If you specify "when {$now=100}...", the simulator will stop at time 100 regardless of the multiplier applied to the time resolution. You must use 64-bit time operators if the time value of now will exceed 2147483647 (the limit of 32-bit numbers). For example:
if { [gtTime $now 2us] } { . . .
See "ModelSim Tcl time commands" (UM-423) for details on 64-bit time operators.
UM-457
B - Error and warning messages

Appendix contents
Message system . . . . . Message format . . . . Getting more information . . Changing message severity level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-458 UM-458 UM-458 UM-458 UM-460 UM-460 UM-460 UM-460 UM-461 UM-463 UM-463 UM-463 UM-464 UM-464 UM-464 UM-466 UM-467 UM-468 UM-468 UM-469 UM-470
Suppressing warning messages . . . Suppressing VCOM warning messages Suppressing VLOG warning messages Suppressing VSIM warning messages Exit codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Miscellaneous messages . . Empty port name warning. Lock message . . . . Metavalue detected warning Sensitivity list warning . Tcl Initialization error 2 . Too few port connections . VSIM license lost . . .
sccom error messages . . . . . . . . Failed to load sc lib error: undefined symbol . Multiply defined symbols . . . . . .
Enforcing strict 1076 compliance with -pedanticerrors
This appendix documents various status and warning messages that are produced by ModelSim.
UM-458 B - Error and warning messages
Message system
The ModelSim message system helps you identify and troubleshoot problems while using the application. The messages display in a standard format in the Transcript pane. Accordingly, you can also access them from a saved transcript file (see "Saving the transcript file" (GR-19) for more details).
Message format
The format for the messages is:
** <SEVERITY LEVEL>: ([<Tool>[-<Group>]]-<MsgNum>) <Message>
SEVERITY LEVEL may be one of the following: severity level Note Warning Error Fatal INTERNAL ERROR meaning This is an informational message. There may be a problem that will affect the accuracy of your results. The tool cannot complete the operation. The tool cannot complete execution. This is an unexpected error that should be reported to your support representative.
Tool indicates which ModelSim tool was being executed when the message was generated. For example tool could be vcom, vdel, vsim, etc. Group indicates the topic to which the problem is related. For example group could be FLI, PLI, VCD, etc.
Example
# ** Error: (vsim-PLI-3071) ./src/19/testfile(77): $fdumplimit : Too few arguments.
Getting more information

Each message is identified by a unique MsgNum id. You can access additional information about a message using the unique id and the verror (CR-265) command. For example:
% verror 3071 Message # 3071: Not enough arguments are being passed to the specified system task or function.
Changing message severity level

You can change the severity of or suppress notes, warnings, and errors that come from vcom, vlog, and vsim. You cannot change the severity of or suppress Fatal or Internal messages.
Message system UM-459
There are two ways to modify the severity of or suppress notes, warnings, and errors: Use the -error, -note, -suppress, and -warning arguments to sccom (CR-199), vcom (CR246), vlog (CR-293), or vsim (CR-305). See the command descriptions in the ModelSim Command Reference for details on those arguments. Set a permanent default in the [msg_system] section of the modelsim.ini file. See "Control variables located in INI files" (UM-438) for more information.
Suppressing warning messages

You can suppress some warning messages. For example, you may receive warning messages about unbound components about which you are not concerned.
Suppressing VCOM warning messages

Use the -nowarn <number> argument to vcom (CR-246) to suppress a specific warning message. For example:
vcom -nowarn 1
Suppresses unbound component warning messages. Alternatively, warnings may be disabled for all compiles via the modelsim.ini file (see "[vcom] VHDL compiler control variables" (UM-441)). The warning message numbers are:
1 = unbound component 2 = process without a wait statement 3 = null range 4 = no space in time literal 5 = multiple drivers on unresolved signal 6 = compliance checks 7 = optimization messages 8 = lint checks 9 = signal value dependency at elaboration 10 = VHDL93 constructs in VHDL87 code 14 = locally static error deferred until simulation run
These numbers are category-of-warning message numbers. They are unrelated to vcom arguments that are specified by numbers, such as vcom -87 which disables support for VHDL-1993 and 2002.
Suppressing VLOG warning messages

Use the +nowarn<CODE> argument to vlog (CR-293) to suppress a specific warning message. Warnings that can be disabled include the <CODE> name in square brackets in the warning message. For example:
vlog +nowarnDECAY
Suppresses decay warning messages.
Suppressing VSIM warning messages

Use the +nowarn<CODE> argument to vsim (CR-305) to suppress a specific warning message. Warnings that can be disabled include the <CODE> name in square brackets in the warning message. For example:
vsim +nowarnTFMPC
Suppresses warning messages about too few port connections.
Exit codes UM-461
Exit codes
The table below describes exit codes used by ModelSim tools. Exit code 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 19 22 42 43 44 45 90 99 100 Description Normal (non-error) return Incorrect invocation of tool Previous errors prevent continuing Cannot create a system process (execv, fork, spawn, etc.) Licensing problem Cannot create/open/find/read/write a design library Cannot create/open/find/read/write a design unit Cannot open/read/write/dup a file (open, lseek, write, mmap, munmap, fopen, fdopen, fread, dup2, etc.) File is corrupted or incorrect type, version, or format of file Memory allocation error General language semantics error General language syntax error Problem during load or elaboration Problem during restore Problem during refresh Communication problem (Cannot create/read/write/close pipe/socket) Version incompatibility License manager not found/unreadable/unexecutable (vlm/mgvlm) SystemC link error Lost license License read/write failure Modeltech daemon license checkout failure #44 Modeltech daemon license checkout failure #45 Assertion failure (SEVERITY_QUIT) Unexpected error in tool GUI Tcl initialization failure
Exit code 101 102 111 202 204 205 206 208 210 211 213 214 215 216 217 218 230 231
Description GUI Tk initialization failure GUI IncrTk initialization failure X11 display error Interrupt (SIGINT) Illegal instruction (SIGILL) Trace trap (SIGTRAP) Abort (SIGABRT) Floating point exception (SIGFPE) Bus error (SIGBUS) Segmentation violation (SIGSEGV) Write on a pipe with no reader (SIGPIPE) Alarm clock (SIGALRM) Software termination signal from kill (SIGTERM) User-defined signal 1 (SIGUSR1) User-defined signal 2 (SIGUSR2) Child status change (SIGCHLD) Exceeded CPU limit (SIGXCPU) Exceeded file size limit (SIGXFSZ)
Miscellaneous messages UM-463
Miscellaneous messages
This section describes miscellaneous messages which may be associated with ModelSim.
Compilation of DPI export TFs error

Message text
# ** Fatal: (vsim-3740) Can't locate a C compiler for compilation of DPI export tasks/functions.
Meaning
ModelSim was unable to locate a C compiler to compile the DPI exported tasks or functions in your design.
Suggested action
Make sure that a C compiler is visible from where you are running the simulation.
Empty port name warning

Message text
# ** WARNING: [8] <path/file_name>: empty port name in port list.
Meaning
ModelSim reports these warnings if you use the -lint argument to vlog (CR-293). It reports the warning for any NULL module ports.
Suggested action
If you wish to ignore this warning, do not use the -lint argument.
Lock message
Message text
waiting for lock by user@user. Lockfile is <library_path>/_lock
Meaning
The _lock file is created in a library when you begin a compilation into that library, and it is removed when the compilation completes. This prevents simultaneous updates to the library. If a previous compile did not terminate properly, ModelSim may fail to remove the _lock file.
Suggested action
Manually remove the _lock file after making sure that no one else is actually using that library.
Metavalue detected warning

Message text
Warning: NUMERIC_STD.">": metavalue detected, returning FALSE
Meaning
This warning is an assertion being issued by the IEEE numeric_std package. It indicates that there is an 'X' in the comparison.
Suggested action
The message does not indicate which comparison is reporting the problem since the assertion is coming from a standard package. To track the problem, note the time the warning occurs, restart the simulation, and run to one time unit before the noted time. At this point, start stepping the simulator until the warning appears. The location of the blue arrow in a Source window will be pointing at the line following the line with the comparison. These messages can be turned off by setting the NumericStdNoWarnings variable to 1 from the command line or in the modelsim.ini file.
Sensitivity list warning

Message text
signal is read by the process but is not in the sensitivity list
Meaning
ModelSim outputs this message when you use the -check_synthesis argument to vcom (CR-246). It reports the warning for any signal that is read by the process but is not in the sensitivity list.
Suggested action
There are cases where you may purposely omit signals from the sensitivity list even though they are read by the process. For example, in a strictly sequential process, you may prefer to include only the clock and reset in the sensitivity list because it would be a design error if any other signal triggered the process. In such cases, your only option is to not use the -check_synthesis argument.
Tcl Initialization error 2

Message text
Tcl_Init Error 2 : Can't find a usable Init.tcl in the following directories : ./../tcl/tcl8.3 .
Meaning
This message typically occurs when the base file was not included in a Unix installation. When you install ModelSim, you need to download and install 3 files from the ftp site. These files are:
modeltech-base.tar.gz modeltech-docs.tar.gz modeltech-<platform>.exe.gz If you install only the <platform> file, you will not get the Tcl files that are located in the base file. This message could also occur if the file or directory was deleted or corrupted.
Suggested action
Reinstall ModelSim with all three files.
Too few port connections

Message text
# ** Warning (vsim-3017): foo.v(1422): [TFMPC] - Too few port connections. Expected 2, found 1. # Region: /foo/tb
Meaning
This warning occurs when an instantiation has fewer port connections than the corresponding module definition. The warning doesnt necessarily mean anything is wrong; it is legal in Verilog to have an instantiation that doesnt connect all of the pins. However, someone that expects all pins to be connected would like to see such a warning. Here are some examples of legal instantiations that will and will not cause the warning message. Module definition:
module foo (a, b, c, d);
Instantiation that does not connect all pins but will not produce the warning:
foo inst1(e, f, g, );
positional association named association
foo inst1(.a(e), .b(f), .c(g), .d());
Instantiation that does not connect all pins but will produce the warning:
foo inst1(e, f, g);
positional association named association
foo inst1(.a(e), .b(f), .c(g));
Any instantiation above will leave pin d unconnected but the first example has a placeholder for the connection. Heres another example:
foo inst1(e, , g, h); foo inst1(.a(e), .b(), .c(g), .d(h));
Suggested actions
Check that there is not an extra comma at the end of the port list. (e.g., model(a,b,) ). The extra comma is legal Verilog and implies that there is a third port connection that is unnamed. If you are purposefully leaving pins unconnected, you can disable these messages using the +nowarnTFMPC argument to vsim.
VSIM license lost

Message text
Console output: Signal 0 caught... Closing vsim vlm child. vsim is exiting with code 4 FATAL ERROR in license manager transcript/vsim output: # ** Error: VSIM license lost; # Time: 5027 ns Iteration: # ** Fatal: Unable to kill and # Time: 5027 ns Iteration:
attempting to re-establish. 2 restart license process. 2
Meaning
ModelSim queries the license server for a license at regular intervals. Usually these "License Lost" error messages indicate that network traffic is high, and communication with the license server times out.
Suggested action
Anything you can do to improve network communication with the license server will probably solve or decrease the frequency of this problem.
Failed to find libswift entry

Message text
** Error: Failed to find LMC Smartmodel libswift entry in project file. # Fatal: Foreign module requested halt
Meaning
ModelSim could not locate the libswift entry and therefore could not link to the Logic Modeling library.
Suggested action
Uncomment the appropriate libswift entry in the [lmc] section of the modelsim.ini or project .mpf file. See "VHDL SmartModel interface" (UM-532) for more information.
sccom error messages

This section describes sccom (CR-199) error messages which may be associated with ModelSim.
Failed to load sc lib error: undefined symbol

Message text
# ** Error: (vsim-3197) Load of "/home/cmg/newport2_systemc/chip/vhdl/work/ systemc.so" failed: ld.so.1: /home/icds_nut/modelsim/5.8a/sunos5/vsimk: fatal: relocation error: file /home/cmg/newport2_systemc/chip/vhdl/work/systemc.so: symbol _Z28host_respond_to_vhdl_requestPm: referenced symbol not found. # ** Error: (vsim-3676) Could not load shared library /home/cmg/ newport2_systemc/chip/vhdl/work/systemc.so for SystemC module 'host_xtor'.
Meaning
The causes for such an error could be: missing symbol definition bad link order specified in sccom -link multiply defined symbols (see "Multiple symbol definitions" (UM-168)
sccom error messages UM-469
Suggested action
If the undefined symbol is a C function in your code or a library you are linking with, be sure that you declared it as an extern "C" function:
extern "C" void myFunc();
The order in which you place the -link option within the sccom -link command is critical. Make sure you have used it appropriately. See sccom (CR-199) for syntax and usage information. See "Misplaced "-link" option" (UM-168) for further explanation of error and correction.
Multiply defined symbols

Message text
work/sc/gensrc/test_ringbuf.o: In function `test_ringbuf::clock_generator(void)': work/sc/gensrc/test_ringbuf.o(.text+0x4): multiple definition of `test_ringbuf::clock_generator(void)' work/sc/test_ringbuf.o(.text+0x4): first defined here
Meaning
The most common type of error found during sccom -link operation is the multiple symbol definition error. This typically arises when the same global symbol is present in more than one .o file. Several causes are likely: A common cause of multiple symbol definitions involves incorrect definition of symbols in header files. If you have an out-of-line function (one that isnt preceded by the "inline" keyword) or a variable defined (i.e., not just referenced or prototyped, but truly defined) in a .h file, you can't include that .h file in more than one .cpp file. Another cause of errors is due to ModelSims name association feature. The name association feature automatically generates .cpp files in the work library. These files "include" your header files. Thus, while it might appear as though you have included your header file in only one .cpp file, from the linkers point of view, it is included in multiple .cpp files.
Suggested action
Make sure you dont have any out-of-line functions. Use the "inline" keyword. See "Multiple symbol definitions" (UM-168).
Enforcing strict 1076 compliance with -pedanticerrors

The optional -pedanticerrors argument to vcom (CR-246) enforces strict compliance to the IEEE 1076 LRM in the cases listed below. The default behavior for these cases is to issue an unsuppressable warning message. If you compile with -pedanticerrors, the warnings change to an error, unless otherwise noted. Descriptions in quotes are actual warning/error messages emitted by vcom. As noted, in some cases you can suppress the warning using -nowarn [level]. Type conversion between array types, where the element subtypes of the arrays do not have identical constraints. "Extended identifier terminates at newline character (0xa)." "Extended identifier contains non-graphic character 0x%x." "Extended identifier \"%s\" contains no graphic characters." "Extended identifier \"%s\" did not terminate with backslash character." "An abstract literal and an identifier must have a separator between them." This is for forming physical literals, which comprise an optional numeric literal, followed by a separator, followed by an identifier (the unit name). Warning is level 4, which means "-nowarn 4" will suppress it. In VHDL 1993 or 2002, a subprogram parameter was declared using VHDL 1987 syntax (which means that it was a class VARIABLE parameter of a file type, which is the only way to do it in VHDL 1987 and is illegal in later VHDLs). Warning is level 10. "Shared variables must be of a protected type." Applies to VHDL 2002 only. Expressions evaluated during elaboration cannot depend on signal values. Warning is level 9. "Non-standard use of output port '%s' in PSL expression." Warning is level 11. "Non-standard use of linkage port '%s' in PSL expression." Warning is level 11. Type mark of type conversion expression must be a named type or subtype, it can't have a constraint on it. When the actual in a PORT MAP association is an expression, it must be a (globally) static expression. The port must also be of mode IN. The expression in the CASE and selected signal assignment statements must follow the rules given in 8.8 of the LRM. In certain cases we can relax these rules, but -pedanticerrors forces strict compliance. A CASE choice expression must be a locally static expression. We allow it to be only globally static, but -pedanticerrors will check that it is locally static. Same rule for selected signal assignment statement choices. Warning level is 8. When making a default binding for a component instantiation, ModelSim's non-standard search rules found a matching entity. VHDL 2002 LRM Section 5.2.2 spells out the standard search rules. Warning level is 1. Both FOR GENERATE and IF GENERATE expressions must be globally static. We allow non-static expressions unless -pedanticerrors is present.
Enforcing strict 1076 compliance with -pedanticerrors UM-471
When the actual part of an association element is in the form of a conversion function call [or a type conversion], and the formal is of an unconstrained array type, the return type of the conversion function [type mark of the type conversion] must be of a constrained array subtype. We relax this (with a warning) unless -pedanticerrors is present when it becomes an error. OTHERS choice in a record aggregate must refer to at least one record element. In an array aggregate of an array type whose element subtype is itself an array, all expressions in the array aggregate must have the same index constraint, which is the element's index constraint. No warning is issued; the presence of -pedanticerrors will produce an error. Non-static choice in an array aggregate must be the only choice in the only element association of the aggregate. The range constraint of a scalar subtype indication must have bounds both of the same type as the type mark of the subtype indication. The index constraint of an array subtype indication must have index ranges each of whose both bounds must be of the same type as the corresponding index subtype. When compiling VHDL 1987, various VHDL 1993 and 2002 syntax is allowed. Use -pedanticerrors to force strict compliance. Warnings are all level 10.
UM-473
C - Verilog PLI / VPI / DPI

Chapter contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-475 UM-480 UM-478 UM-478 UM-480 UM-481 UM-484 UM-484 UM-497 UM-497 UM-497 UM-498 UM-499 UM-500 UM-501 UM-502 UM-504 UM-505 UM-506 UM-507 UM-508 UM-510 UM-512 UM-512 UM-513 UM-513 UM-514 UM-514 UM-514 UM-514 UM-515 UM-516 Registering DPI applications . Registering VPI applications . Example . . . . . Registering DPI applications . DPI use flow . . . . .
Compiling and linking C applications for PLI/VPI/DPI . Compiling and linking C++ applications for PLI/VPI/DPI Specifying application files to load . . . . . . PLI/VPI file loading . . . . . . . . . DPI file loading . . . . . . . . . . Loading shared objects with global symbol visibility PLI example VPI example DPI example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The PLI callback reason argument. The sizetf callback function PLI object handles . . . . . . . . . . .
Third party PLI applications . Support for VHDL objects . .
IEEE Std 1364 ACC routines . IEEE Std 1364 TF routines .
SystemVerilog DPI access routines Verilog-XL compatible routines .
64-bit support for PLI . . . . . . . . . Using 64-bit ModelSim with 32-bit Applications. PLI/VPI tracing . . . . The purpose of tracing files Invoking a trace . . . Syntax . . . . . . Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Debugging PLI/VPI/DPI application code.
UM-474 C - Verilog PLI / VPI / DPI
HP-UX specific warnings .
UM-516
Introduction UM-475
Introduction
This appendix describes the ModelSim implementation of the Verilog PLI (Programming Language Interface), VPI (Verilog Procedural Interface) and SystemVerilog DPI (Direct Programming Interface). These three interfaces provide a mechanism for defining tasks and functions that communicate with the simulator through a C procedural interface. There are many third party applications available that interface to Verilog simulators through the PLI (see "Third party PLI applications" (UM-506)). In addition, you may write your own PLI/ VPI/DPI applications. ModelSim Verilog implements the PLI as defined in the IEEE Std 1364, with the exception of the acc_handle_datapath() routine. We did not implement the acc_handle_datapath() routine because the information it returns is more appropriate for a static timing analysis tool. The VPI is partially implemented as defined in the IEEE Std 1364-2005. The list of currently supported functionality can be found in the following file:
<install_dir>/modeltech/docs/technotes/Verilog_VPI.note
ModelSim SystemVerilog implements DPI as defined in IEEE Std P1800-2005. The IEEE Std 1364 is the reference that defines the usage of the PLI/VPI routines, and the IEEE Std P1800-2005 Language Reference Manual (LRM) defines the usage of DPI routines. This manual describes only the details of using the PLI/VPI/DPI with ModelSim Verilog and SystemVerilog.
Registering PLI applications

Each PLI application must register its system tasks and functions with the simulator, providing the name of each system task and function and the associated callback routines. Since many PLI applications already interface to Verilog-XL, ModelSim Verilog PLI applications make use of the same mechanism to register information about each system task and function in an array of s_tfcell structures. This structure is declared in the veriuser.h include file as follows:
typedef int (*p_tffn)(); typedef struct t_tfcell { short type;/* USERTASK, USERFUNCTION, or USERREALFUNCTION */ short data;/* passed as data argument of callback function */ p_tffn checktf; /* argument checking callback function */ p_tffn sizetf; /* function return size callback function */ p_tffn calltf; /* task or function call callback function */ p_tffn misctf; /* miscellaneous reason callback function */ char *tfname;/* name of system task or function */ /* The following fields are ignored by ModelSim Verilog */ int forwref; char *tfveritool; char *tferrmessage; int hash; struct t_tfcell *left_p; struct t_tfcell *right_p; char *namecell_p; int warning_printed; } s_tfcell, *p_tfcell;
The various callback functions (checktf, sizetf, calltf, and misctf) are described in detail in the IEEE Std 1364. The simulator calls these functions for various reasons. All callback functions are optional, but most applications contain at least the calltf function, which is called when the system task or function is executed in the Verilog code. The first argument to the callback functions is the value supplied in the data field (many PLI applications don't use this field). The type field defines the entry as either a system task (USERTASK) or a system function that returns either a register (USERFUNCTION) or a real (USERREALFUNCTION). The tfname field is the system task or function name (it must begin with $). The remaining fields are not used by ModelSim Verilog. On loading of a PLI application, the simulator first looks for an init_usertfs function, and then a veriusertfs array. If init_usertfs is found, the simulator calls that function so that it can call mti_RegisterUserTF() for each system task or function defined. The mti_RegisterUserTF() function is declared in veriuser.h as follows:
void mti_RegisterUserTF(p_tfcell usertf);
Registering PLI applications UM-477
The storage for each usertf entry passed to the simulator must persist throughout the simulation because the simulator de-references the usertf pointer to call the callback functions. We recommend that you define your entries in an array, with the last entry set to 0. If the array is named veriusertfs (as is the case for linking to Verilog-XL), then you don't have to provide an init_usertfs function, and the simulator will automatically register the entries directly from the array (the last entry must be 0). For example,
s_tfcell veriusertfs[] = { {usertask, 0, 0, 0, abc_calltf, 0, "$abc"}, {usertask, 0, 0, 0, xyz_calltf, 0, "$xyz"}, {0} /* last entry must be 0 */ };
Alternatively, you can add an init_usertfs function to explicitly register each entry from the array:
void init_usertfs() { p_tfcell usertf = veriusertfs; while (usertf->type) mti_RegisterUserTF(usertf++); }
It is an error if a PLI shared library does not contain a veriusertfs array or an init_usertfs function. Since PLI applications are dynamically loaded by the simulator, you must specify which applications to load (each application must be a dynamically loadable library, see "Compiling and linking C applications for PLI/VPI/DPI" (UM-484)). The PLI applications are specified as follows (note that on a Windows platform the file extension would be .dll): As a list in the Veriuser entry in the modelsim.ini file:
Veriuser = pliapp1.so pliapp2.so pliappn.so
As a list in the PLIOBJS environment variable:

% setenv PLIOBJS "pliapp1.so pliapp2.so pliappn.so"
As a -pli argument to the simulator (multiple arguments are allowed):

-pli pliapp1.so -pli pliapp2.so -pli pliappn.so
The various methods of specifying PLI applications can be used simultaneously. The libraries are loaded in the order listed above. Environment variable references can be used in the paths to the libraries in all cases.
Registering VPI applications

Each VPI application must register its system tasks and functions and its callbacks with the simulator. To accomplish this, one or more user-created registration routines must be called at simulation startup. Each registration routine should make one or more calls to vpi_register_systf() to register user-defined system tasks and functions and vpi_register_cb() to register callbacks. The registration routines must be placed in a table named vlog_startup_routines so that the simulator can find them. The table must be terminated with a 0 entry.
Example
PLI_INT32 MyFuncCalltf( PLI_BYTE8 *user_data ) { ... } PLI_INT32 MyFuncCompiletf( PLI_BYTE8 *user_data ) { ... } PLI_INT32 MyFuncSizetf( PLI_BYTE8 *user_data ) { ... } PLI_INT32 MyEndOfCompCB( p_cb_data cb_data_p ) { ... } PLI_INT32 MyStartOfSimCB( p_cb_data cb_data_p ) { ... } void RegisterMySystfs( void ) { vpiHandle tmpH; s_cb_data callback; s_vpi_systf_data systf_data; systf_data.type = vpiSysFunc; systf_data.sysfunctype = vpiSizedFunc; systf_data.tfname = "$myfunc"; systf_data.calltf = MyFuncCalltf; systf_data.compiletf = MyFuncCompiletf; systf_data.sizetf = MyFuncSizetf; systf_data.user_data = 0; tmpH = vpi_register_systf( &systf_data ); vpi_free_object(tmpH); callback.reason = cbEndOfCompile; callback.cb_rtn = MyEndOfCompCB; callback.user_data = 0; tmpH = vpi_register_cb( &callback ); vpi_free_object(tmpH); callback.reason = cbStartOfSimulation; callback.cb_rtn = MyStartOfSimCB; callback.user_data = 0; tmpH = vpi_register_cb( &callback ); vpi_free_object(tmpH); } void (*vlog_startup_routines[ ] ) () = { RegisterMySystfs, 0 /* last entry must be 0 */ };
Registering VPI applications UM-479
Loading VPI applications into the simulator is the same as described in "Registering DPI applications" (UM-480).
Using PLI and VPI together

PLI and VPI applications can co-exist in the same application object file. In such cases, the applications are loaded at startup as follows: If an init_usertfs() function exists, then it is executed and only those system tasks and functions registered by calls to mti_RegisterUserTF() will be defined. If an init_usertfs() function does not exist but a veriusertfs table does exist, then only those system tasks and functions listed in the veriusertfs table will be defined. If an init_usertfs() function does not exist and a veriusertfs table does not exist, but a vlog_startup_routines table does exist, then only those system tasks and functions and callbacks registered by functions in the vlog_startup_routines table will be defined. As a result, when PLI and VPI applications exist in the same application object file, they must be registered in the same manner. VPI registration functions that would normally be listed in a vlog_startup_routines table can be called from an init_usertfs() function instead.
Registering DPI applications

DPI applications do not need to be registered. However, each DPI imported or exported task or function must be identified using SystemVerilog import DPI-C or export DPICsyntax. Examples of the syntax follow:
export DPI-C task t1; task t1(input int i, output int o); . . . end task import DPI-C function void f1(input int i, output int o);
Your code must provide imported functions or tasks, compiled with an external compiler. An imported task must return an int value, "1" indicating that it is returning due to a disable, or "0" indicating otherwise. These imported functions or objects may then be loaded as a shared library into the simulator with either the command line option -sv_lib <lib> or -sv_liblist <bootstrap_file>. For example,
vlog dut.v gcc -shared -o imports.so imports.c vsim -sv_lib imports top -do <do_file>
The -sv_lib option specifies the shared library name, without an extension. A file extension is added by the tool, as appropriate to your platform. For a list of file extensions accepted by platform, see "DPI file loading" (UM-497). You can also use the command line options -sv_root and -sv_liblist to control the process for loading imported functions and tasks. These options are defined in the IEEE std P18002005 LRM.
DPI use flow UM-481
DPI use flow

Correct use of ModelSim DPI depends on the flow presented in this section.
Step 1 Create header

.v
vlog -dpiheader dpiheader.h
vlog
dpiheader.h
Step 2 Include header

#include "dpiheader.h"
vsim
Step 1.5 Required for Windows only
.c
vsim -dpiexportobj <exportobj>
gcc
<exportobj> C compiler
mtipli.lib
.o
compiled user code
ld/link
loader/linker
Step 3 Compile and load/link C code
<test>.so
shared object
vsim
Step 4 Simulate
vsim -sv_lib <test>
Steps in flow
1 Run vlog (CR-293) to generate a dpiheader.h file. This file defines the interface between C and ModelSim for exported and imported tasks and functions. Though the dpiheader.h is a user convenience file rather than requirement, including dpiheader.h in your C code can immediately solve problems caused by an improperly defined interface. An example command for creating the header file would be:
vlog -dpiheader <dpiheader>.h
Required for Windows only; Run a preliminary invocation of vsim (CR-305) with the -dpiexportobj argument. Because of limitations with the linker/loader provided on Windows, this additional step is required. You must create the exported task/function compiled object file (exportobj) by running a preliminary vsim command, such as:
vsim -dpiexportobj exportobj
2 Include the dpiheader.h file in your C code. ModelSim recommends that any user DPI C code that accesses exported tasks/functions, or defines imported tasks/functions, will include the dpiheader.h file. This allows the C compiler to verify the interface between C and ModelSim. 3 Compile the C code into a shared object. Compile your code, providing any .a or other .o files required. For Windows users: In this step, the object file is bound into the .dll that you created using the -dpiexportobj argument. 4 Simulate the design. When simulating, specify the name of the imported DPI C shared object (according to the SystemVerilog LRM).
Use model for read-only work libraries

You may want to create the work library as a read-only entity, which enables multiple users to simultaneously share the design library at runtime. The steps are as follows:
Windows and RS6000/RS64

On these platorms, simply change the permissions on the design library to read only by issuing a command such as "chmod -R a-w <libname>". Do this after you have finished compiling with vlog (CR-293)/vcom (CR-246).
All other platforms

If a design contains no DPI export tasks or functions, the work library can be changed by simply changing the permissions, as shown for win32 and rs6000/rs64 above. For designs that contain DPI export tasks and functions, and are not run on Windows or RS6000/RS64, by default vsim (CR-305) creates a shared object in directory <libname>/ _dpi. This shared object is called exportwrapper.so (Linux and Solaris) or exportwrapper.sl (hp700, hppa64, and hpux_ia64). If you are using a read-only library, vsim must not create any objects in the library. To prevent vsim from creating objects in the library at runtime, the vsim -dpiexportobj flow is available on all platforms. Use this flow after compilation, but before you start simulation using the design library. An example command sequence on Linux would be:
vlib work vlog -dpiheader dpiheader.h test.sv
DPI use flow UM-483
gcc -shared -o test.so test.c vsim -c -dpiexportobj work/_dpi/exportwrapper top chmod -R a-w work
The library is now ready for simulation by multiple simultaneous users, as follows:
vsim top -sv_lib test
The work/_dpi/exportwrapper argument provides a basename for the shared object. At runtime, vsim automatically checks to see if the file work/_dpi/exportwrapper.so is upto-date with respect to its C source code. If it is out of date, an error message is issued and elaboration stops.
Compiling and linking C applications for PLI/VPI/DPI

The following platform-specific instructions show you how to compile and link your PLI/VPI/DPI C applications so that they can be loaded by ModelSim. Various native C/ C++ compilers are supported on different platforms. The gcc compiler is supported on all platforms. The following PLI/VPI/DPI routines are declared in the include files located in the ModelSim <install_dir>/modeltech/include directory: acc_user.h veriuser.h vpi_user.h svdpi.h declares the ACC routines declares the TF routines declares the VPI routines declares DPI routines
The following instructions assume that the PLI, VPI, or DPI application is in a single source file. For multiple source files, compile each file as specified in the instructions and link all of the resulting object files together with the specified link instructions. Although compilation and simulation switches are platform-specific, loading shared libraries is the same for all platforms. For information on loading libraries for PLI/VPI see "PLI/VPI file loading" (UM-497). For DPI loading instructions, see "DPI file loading" (UM497).
For all UNIX platforms

The information in this section applies to all UNIX platforms.
app.so
If app.so is not in your current directory, you must tell the OS where to search for the shared object. You can do this one of two ways: Add a path before app.so in the command line option or control variable (The path may include environment variables.) Put the path in a UNIX shell environment variable: LD_LIBRARY_PATH= <library path without filename> (for Solaris/Linux) or SHLIB_PATH= <library path without filename> (for HP-UX)
Windows platforms
Microsoft Visual C 4.1 or later
cl -c -I<install_dir>\modeltech\include app.c link -dll -export:<init_function> app.obj \ <install_dir>\modeltech\win32\mtipli.lib -out:app.dll
For the Verilog PLI, the <init_function> should be "init_usertfs". Alternatively, if there is no init_usertfs function, the <init_function> specified on the command line should be "veriusertfs". For the Verilog VPI, the <init_function> should be "vlog_startup_routines".
Compiling and linking C applications for PLI/VPI/DPI UM-485
These requirements ensure that the appropriate symbol is exported, and thus ModelSim can find the symbol when it dynamically loads the DLL. When executing cl commands in a DO file, use the /NOLOGO switch to prevent the Microsoft C compiler from writing the logo banner to stderr. Writing the logo causes Tcl to think an error occurred. If you need to run the profiler (see Chapter 13 - Profiling performance and memory use) on a design that contains PLI/VPI code, add these two switches to the link commands shown above:
These switches add symbols to the .dll that the profiler can use in its report.
MinGW gcc 3.2.3

gcc -c -I<install_dir>\modeltech\include app.c gcc -shared -o app.dll app.o -L<install_dir>\modeltech\win32 -lmtipli
ModelSim requires the use of MinGW gcc compiler rather than the Cygwin gcc compiler. MinGW gcc is available on the ModelSim FTP site. Remember to add the path to your gcc executable in the Windows environment variables.
For DPI imports

When linking the shared objects, be sure to specify one export option for each DPI imported task or function in your linking command line. You can use the -isymfile argument from the vlog (CR-293) command to obtain a complete list of all imported tasks/functions expected by ModelSim. As an alternative to specifying one -export option for each imported task or function, you can make use of the __declspec (dllexport) macro supported by Visual C. You can place this macro before every DPI import task or function declaration in your C source. All the marked functions will be available for use by vsim as DPI import tasks and functions.
DPI special flow for exported tasks and functions

Since the Windows platform lacks the necessary runtime linking capabilities, you must perform an additional manual step in order to prepare shared objects containing calls to exported SystemVerilog tasks or functions. You need to invoke a special run of vsim (CR305). The command is as follows:
vsim <top du list> -dpiexportobj <objname> <other args>
The -dpiexportobj generates an object file <objname>.obj that contains "glue" code for exported tasks and functions. You must add that object file to the link line for your .dll, listed after the other object files. For example, a link line for MinGW would be:
gcc -shared -o app.dll app.obj <objname>.obj -L<install_dir>\modeltech\win32 -lmtipli
and a link line for Visual C would be:

link -dll -export:<init_function> app.obj <objname>.obj\ <install_dir>\modeltech\win32\mtipli.lib -out:app.dll
32-bit Linux platform
If your PLI/VPI/DPI application uses anything from a system library, you will need to specify that library when you link your PLI/VPI/DPI application. For example, to use the standard C library, specify -lc to the ld command.
gcc compiler
gcc -c -I/<install_dir>/modeltech/include app.c ld -shared -E -Bsymbolic -o app.so app.o -lc
When using -Bsymbolic with ld, all symbols are first resolved within the shared library at link time. This will result in a list of undefined symbols. This is only a warning for shared libraries and can be ignored. If you are using ModelSim RedHat version 6.0 through 7.1, you also need to add the -noinhibit-exec switch when you specify -Bsymbolic. The compiler switch -freg-struct-return must be used when compiling any FLI application code that contains foreign functions that return real or time values.
64-bit Linux for IA64 platform

64-bit Linux is supported on RedHat Linux Advanced Workstation 2.1 for Itanium 2.
gcc compiler (gcc 3.2 or later)

gcc -c -fPIC -I/<install_dir>/modeltech/include app.c ld -shared -Bsymbolic -E --allow-shlib-undefined -o app.so app.o
If your PLI/VPI/DPI application requires a user or vendor-supplied C library, or an additional system library, you will need to specify that library when you link your PLI/VPI/ DPI application. For example, to use the system math library libm, specify -lm to the ld command:
gcc -c -fPIC -I/<install_dir>/modeltech/include math_app.c ld -shared -Bsymbolic -E --allow-shlib-undefined -o math_app.so math_app.o -lm
64-bit Linux for Opteron/Athlon 64 and EM64T platforms

64-bit Linux is supported on RedHat Linux EWS 3.0 for Opteron/Athlon 64 and EM64T.
gcc compiler (gcc 3.2 or later)

gcc -c -fPIC -I/<install_dir>/modeltech/include app.c ld -shared -Bsymbolic -E --allow-shlib-undefined -o app.so app.o
To compile for 32-bit operation, specify the -m32 argument on the gcc command line. If your PLI/VPI/DPI application requires a user or vendor-supplied C library, or an additional system library, you will need to specify that library when you link your PLI/VPI/ DPI application. For example, to use the system math library libm, specify -lm to the ld command:
gcc -c -fPIC -I/<install_dir>/modeltech/include math_app.c ld -shared -Bsymbolic -E --allow-shlib-undefined \ -o math_app.so math_app.o -lm
32-bit Solaris platform

If your PLI/VPI/DPI application uses anything from a system library, you will need to specify that library when you link your PLI/VPI/DPI application. For example, to use the standard C library, specify -lc to the ld command.
gcc compiler
gcc -c -I/<install_dir>/modeltech/include app.c ld -G -Bsymbolic -o app.so app.o -lc
cc compiler
cc -c -I/<install_dir>/modeltech/include app.c ld -G -Bsymbolic -o app.so app.o -lc
When using -Bsymbolic with ld, all symbols are first resolved within the shared library at link time. This will result in a list of undefined symbols. This is only a warning for shared libraries and can be ignored.

gcc compiler
gcc -c -I<install_dir>/modeltech/include -m64 -fPIC app.c gcc -shared -o app.so -m64 app.o
This was tested with gcc 3.2.2. You may need to add the location of libgcc_s.so.1 to the LD_LIBRARY_PATH environment variable.
cc compiler
cc -v -xarch=v9 -O -I<install_dir>/modeltech/include -c app.c ld -G -Bsymbolic app.o -o app.so
32-bit HP700 platform

A shared library is created by creating object files that contain position-independent code (use the +z or -fPIC compiler argument) and by linking as a shared library (use the -b linker argument). If your PLI/VPI/DPI application uses anything from a system library, youll need to specify that library when you link your PLI/VPI/DPI application. For example, to use the standard C library, specify -lc to the ld command.
gcc compiler
gcc -c -fPIC -I/<install_dir>/modeltech/include app.c ld -b -o app.sl app.o -lc
cc compiler
cc -c +z +DD32 -I/<install_dir>/modeltech/include app.c
ld -b -o app.sl app.o -lc
Note that -fPIC may not work with all versions of gcc.
64-bit HP platform
cc compiler
cc -v +DD64 -O -I<install_dir>/modeltech/include -c app.c ld -b -o app.sl app.o -lc
64-bit HP for IA64 platform

cc compiler (/opt/ansic/bin/cc, /usr/ccs/bin/ld)
cc -c +DD64 -I/<install_dir>/modeltech/include app.c ld -b -o app.sl app.o
If your PLI/VPI/DPI application requires a user or vendor-supplied C library, or an additional system library, you will need to specify that library when you link your PLI/VPI/ DPI application. For example, to use the system math library, specify '-lm' to the 'ld' command:
cc -c +DD64 -I/<install_dir>/modeltech/include math_app.c ld -b -o math_app.sl math_app.o -lm
32-bit IBM RS/6000 platform

ModelSim loads shared libraries on the IBM RS/6000 workstation. The shared library must import ModelSim's PLI/VPI/DPI symbols, and it must export the PLI or VPI applications initialization function or table. ModelSim's export file is located in the ModelSim installation directory in rs6000/mti_exports. If your PLI/VPI/DPI application uses anything from a system library, youll need to specify that library when you link your PLI/VPI/DPI application. For example, to use the standard C library, specify -lc to the ld command. The resulting object must be marked as shared reentrant using these gcc or cc compiler commands for AIX 4.x:
gcc compiler
gcc -c -I/<install_dir>/modeltech/include app.c ld -o app.sl app.o -bE:app.exp \ -bI:/<install_dir>/modeltech/rs6000/mti_exports -bM:SRE -bnoentry -lc
cc compiler
cc -c -I/<install_dir>/modeltech/include app.c ld -o app.sl app.o -bE:app.exp \ -bI:/<install_dir>/modeltech/rs6000/mti_exports -bM:SRE -bnoentry -lc
The app.exp file must export the PLI/VPI initialization function or table. For the PLI, the exported symbol should be "init_usertfs". Alternatively, if there is no init_usertfs function, then the exported symbol should be "veriusertfs". For the VPI, the exported symbol should be "vlog_startup_routines". These requirements ensure that the appropriate symbol is exported, and thus ModelSim can find the symbol when it dynamically loads the shared object.
When using AIX 4.3 in 32-bit mode, you must add the -DUSE_INTTYPES switch to the compile command lines. This switch prevents a name conflict that occurs between inttypes.h and mti.h.
For DPI imports

When linking the shared objects, be sure to specify -bE:<isymfile> option on the link command line. <isymfile> is the name of the file generated by the-isymfile argument to the vlog (CR-293) command. Once you have created the <isymfile>, it contains a complete list of all imported tasks and functions expected by ModelSim.

Since the RS6000 platform lacks the necessary runtime linking capabilities, you must perform an additional manual step in order to prepare shared objects containing calls to exported SystemVerilog tasks or functions shared object file. You need to invoke a special run of vsim (CR-305). The command is as follows:
The -dpiexportobj generates the object file <objname>.o that contains "glue" code for exported tasks and functions. You must add that object file to the link line, listed after the other object files. For example, a link line would be:
ld -o app.so app.o <objname>.o -bE:<isymfile> -bI:/<install_dir>/modeltech/rs6000/mti_exports -bM:SRE -bnoentry -lc

Only versions 4.3 and later of AIX support the 64-bit platform. A gcc 64-bit compiler is not available at this time.
VisualAge cc compiler
cc -c -q64 -I/<install_dir>/modeltech/include app.c ld -o app.s1 app.o -b64 -bE:app.exports \ -bI:/<install_dir>/modeltech/rs64/mti_exports -bM:SRE -bnoentry -lc
For DPI imports


ld -o app.dll app.o <objname>.o -bE:<isymfile> -bI:/<install_dir>/modeltech/rs6000/mti_exports -bM:SRE -bnoentry -lc
Compiling and linking C++ applications for PLI/VPI/DPI UM-491
Compiling and linking C++ applications for PLI/VPI/DPI

ModelSim does not have direct support for any language other than standard C; however, C++ code can be loaded and executed under certain conditions. Since ModelSim's PLI/VPI/DPI functions have a standard C prototype, you must prevent the C++ compiler from mangling the PLI/VPI/DPI function names. This can be accomplished by using the following type of extern:
extern "C" { <PLI/VPI/DPI application function prototypes> }
The header files veriuser.h, acc_user.h, and vpi_user.h, svdpi.h already include this type of extern. You must also put the PLI/VPI/DPI shared library entry point (veriusertfs, init_usertfs, or vlog_startup_routines) inside of this type of extern. The following platform-specific instructions show you how to compile and link your PLI/VPI/DPI C++ applications so that they can be loaded by ModelSim. Although compilation and simulation switches are platform-specific, loading shared libraries is the same for all platforms. For information on loading libraries, see "DPI file loading" (UM-497).
For PLI/VPI only

If app.so is not in your current directory you must tell Solaris where to search for the shared object. You can do this one of two ways: Add a path before app.so in the foreign attribute specification. (The path may include environment variables.) Put the path in a UNIX shell environment variable: LD_LIBRARY_PATH= <library path without filename>
Windows platforms
Microsoft Visual C++ 4.1 or later
cl -c [-GX] -I<install_dir>\modeltech\include app.cxx link -dll -export:<init_function> app.obj \ <install_dir>\modeltech\win32\mtipli.lib /out:app.dll
The -GX argument enables exception handling. For the Verilog PLI, the <init_function> should be "init_usertfs". Alternatively, if there is no init_usertfs function, the <init_function> specified on the command line should be "veriusertfs". For the Verilog VPI, the <init_function> should be "vlog_startup_routines". These requirements ensure that the appropriate symbol is exported, and thus ModelSim can find the symbol when it dynamically loads the DLL. When executing cl commands in a DO file, use the /NOLOGO switch to prevent the Microsoft C compiler from writing the logo banner to stderr. Writing the logo causes Tcl to think an error occurred.
If you need to run the profiler (see Chapter 13 - Profiling performance and memory use) on a design that contains PLI/VPI code, add these two switches to the link command shown above:
These switches add symbols to the .dll that the profiler can use in its report.
MinGW C++ version 3.2.3

g++ -c -I<install_dir>\modeltech\include app.cpp g++ -shared -o app.dll app.o -L<install_dir>\modeltech\win32 -lmtipli
ModelSim requires the use of MinGW gcc compiler rather than the Cygwin gcc compiler. MinGW gcc is available on the ModelSim FTP site.
For DPI imports

When linking the shared objects, be sure to specify one -export option for each DPI imported task or function in your linking command line. You can use Verilogs -isymfile option to obtain a complete list of all imported tasks and functions expected by ModelSim.

Since the Windows platform lacks the necessary runtime linking complexity, you must perform an additional manual step in order to compile the HDL source files into the shared object file. You need to invoke a special run of vsim. The command is as follows:
The -dpiexportobj generates the object file <objname>.obj that contains "glue" code for exported tasks and functions. You must add that object file to the link line, listed after the other object files. For example, if the object name was dpi1, the link line for MinGW would be:
g++ -shared -o app.dll app.obj <objname>.obj -L<install_dir>\modeltech\win32 -lmtipli
32-bit Linux platform

GNU C++ version 2.95.3 or later
g++ -c -fPIC -I<install_dir>/modeltech/include app.cpp g++ -shared -fPIC -o app.so app.o
64-bit Linux for IA64 platform

64-bit Linux is supported on RedHat Linux Advanced Workstation 2.1 for Itanium 2.
GNU C++ compiler version gcc 3.2 or later

g++ -c -fPIC -I/<install_dir>/modeltech/include app.cpp ld -shared -Bsymbolic -E --allow-shlib-undefined -o app.so app.o
If your PLI/VPI application requires a user or vendor-supplied C library, or an additional system library, you will need to specify that library when you link your PLI/VPI application. For example, to use the system math library libm, specify '-lm' to the 'ld' command:
g++ -c -fPIC -I/<install_dir>/modeltech/include math_app.cpp ld -shared -Bsymbolic -E --allow-shlib-undefined -o math_app.so math_app.o -lm
64-bit Linux for Opteron/Athlon 64 and EM64T platforms

64-bit Linux is supported on RedHat Linux EWS 3.0 for Opteron/Athlon 64 and EM64T.

g++ -c -fPIC -I/<install_dir>/modeltech/include app.cpp ld -shared -Bsymbolic -E --allow-shlib-undefined -o app.so app.o
To compile for 32-bit operation, specify the -m32 argument on the gcc command line. If your PLI/VPI/DPI application requires a user or vendor-supplied C library, or an additional system library, you will need to specify that library when you link your PLI/VPI/ DPI application. For example, to use the system math library libm, specify -lm to the ld command:
g++ -c -fPIC -I/<install_dir>/modeltech/include math_app.cpp ld -shared -Bsymbolic -E --allow-shlib-undefined -o math_app.so math_app.o -lm

If your PLI/VPI application uses anything from a system library, you will need to specify that library when you link your PLI/VPI application. For example, to use the standard C library, specify -lc to the ld command.

g++ -c -I/<install_dir>/modeltech/include app.cpp ld -G -Bsymbolic -o app.so app.o -lc
Sun Forte C++ compiler

cc -c -I/<install_dir>/modeltech/include app.cpp ld -G -Bsymbolic -o app.so app.o -lc

g++ -c -I<install_dir>/modeltech/include -m64 -fPIC app.cpp g++ -shared -o app.so -m64 app.o
This was tested with gcc 3.2.2. You may need to add the location of libgcc_s.so.1 to the LD_LIBRARY_PATH environment variable.
cc compiler
cc -v -xarch=v9 -O -I<install_dir>/modeltech/include -c app.cpp ld -G -Bsymbolic app.o -o app.so
32-bit HP700 platform

A shared library is created by creating object files that contain position-independent code (use the +z or -fPIC compiler argument) and by linking as a shared library (use the -b linker argument). If your PLI/VPI application uses anything from a system library, youll need to specify that library when you link your PLI/VPI application. For example, to use the standard C library, specify -lc to the ld command.
GNU C++ compiler

g++ -c -fPIC -I/<install_dir>/modeltech/include app.cpp ld -b -o app.sl app.o -lc
cc compiler
cc -c +z +DD32 -I/<install_dir>/modeltech/include app.cpp ld -b -o app.sl app.o -lc
Note that -fPIC may not work with all versions of gcc.
64-bit HP platform
cc compiler
cc -v +DD64 -O -I<install_dir>/modeltech/include -c app.cpp ld -b -o app.sl app.o -lc
64-bit HP for IA64 platform

HP ANSI C++ compiler (/opt/ansic/bin/cc, /usr/ccs/bin/ld)
cc -c +DD64 -I/<install_dir>/modeltech/include app.cpp ld -b -o app.sl app.o
If your PLI/VPI application requires a user or vendor-supplied C library, or an additional system library, you will need to specify that library when you link your PLI/VPI application. For example, to use the system math library, specify '-lm' to the 'ld' command:
cc -c +DD64 -I/<install_dir>/modeltech/include math_app.c ld -b -o math_app.sl math_app.o -lm

ModelSim loads shared libraries on the IBM RS/6000 workstation. The shared library must import ModelSim's PLI/VPI symbols, and it must export the PLI or VPI applications initialization function or table. ModelSim's export file is located in the ModelSim installation directory in rs6000/mti_exports. If your PLI/VPI application uses anything from a system library, youll need to specify that library when you link your PLI/VPI application. For example, to use the standard C library,
specify -lc to the ld command. The resulting object must be marked as shared reentrant using these gcc or cc compiler commands for AIX 4.x:

g++ -c -I/<install_dir>/modeltech/include app.cpp ld -o app.sl app.o -bE:app.exp \ -bI:/<install_dir>/modeltech/rs6000/mti_exports -bM:SRE -bnoentry -lc
VisualAge C++ compiler

cc -c -I/<install_dir>/modeltech/include app.cpp ld -o app.sl app.o -bE:app.exp \ -bI:/<install_dir>/modeltech/rs6000/mti_exports -bM:SRE -bnoentry -lc
The app.exp file must export the PLI/VPI initialization function or table. For the PLI, the exported symbol should be "init_usertfs". Alternatively, if there is no init_usertfs function, then the exported symbol should be "veriusertfs". For the VPI, the exported symbol should be "vlog_startup_routines". These requirements ensure that the appropriate symbol is exported, and thus ModelSim can find the symbol when it dynamically loads the shared object. When using AIX 4.3 in 32-bit mode, you must add the -DUSE_INTTYPES switch to the compile command lines. This switch prevents a name conflict that occurs between inttypes.h and mti.h.
For DPI imports


ld -o app.dll app.o <objname>.o -bE:<isymfile> -bI:/<install_dir>/modeltech/rs6000/mti_exports -bM:SRE -bnoentry -lc

Only version 4.3 and later of AIX supports the 64-bit platform. A gcc 64-bit compiler is not available at this time.
VisualAge C++ compiler

cc -c -q64 -I/<install_dir>/modeltech/include app.cpp
ld -o app.s1 app.o -b64 -bE:app.exports \ -bI:/<install_dir>/modeltech/rs64/mti_exports -bM:SRE -bnoentry -lc
For DPI imports


ld -o app.so app.o <objname>.o -bE:<isymfile> -bI:/<install_dir>/modeltech/rs6000/mti_exports -bM:SRE -bnoentry -lc
Specifying application files to load UM-497
Specifying application files to load

PLI and VPI file loading is identical. DPI file loading uses switches to the vsim command.
PLI/VPI file loading

The PLI/VPI applications are specified as follows: As a list in the Veriuser entry in the modelsim.ini file:
Veriuser = pliapp1.so pliapp2.so pliappn.so
As a list in the PLIOBJS environment variable:

% setenv PLIOBJS "pliapp1.so pliapp2.so pliappn.so"
As a -pli argument to the simulator (multiple arguments are allowed):

-pli pliapp1.so -pli pliapp2.so -pli pliappn.so
Note: On Windows platforms, the file names shown above should end with .dll rather than .so. The various methods of specifying PLI/VPI applications can be used simultaneously. The libraries are loaded in the order listed above. Environment variable references can be used in the paths to the libraries in all cases. See also Appendix A - Simulator variables for more information on the modelsim.ini file.
DPI file loading

DPI applications are specified to vsim (CR-305) using the following SystemVerilog arguments: -sv_lib <name> specifies a library name to be searched and used. No filename extensions must be specified. (The extensions ModelSim expects are: .sl for HP, .dll for Win32, .so for all other platforms.) specifies a new prefix for shared objects as specified by -sv_lib specifies a bootstrap file to use
-sv_root <name> -sv_liblist
When the simulator finds an imported task or function, it searches for the symbol in the collection of shared objects specified using these arguments. For example, you can specify the DPI application as follows:
vsim -sv_lib dpiapp1 -sv_lib dpiapp2 -sv_lib dpiappn
It is a mistake to specify DPI import tasks and functions (tf) inside PLI/VPI shared objects. However, a DPI import tf can make calls to PLI/VPI C code, providing that vsim -gblso was used to mark the PLI/VPI shared object with global symbol visibility. See "Loading shared objects with global symbol visibility" (UM-498).
Loading shared objects with global symbol visibility

On Unix platforms you can load shared objects such that all symbols in the object have global visibility. To do this, use the -gblso argument to vsim when you load your PLI/VPI application. For example:
vsim -pli obj1.so -pli obj2.so -gblso obj1.so
The -gblso argument works in conjunction with the GlobalSharedObjectList variable in the modelsim.ini file. This variable allows user C code in other shared objects to refer to symbols in a shared object that has been marked as global. All shared objects marked as global are loaded by the simulator earlier than any non-global shared objects.
PLI example UM-499
PLI example
The following example is a trivial, but complete PLI application.
hello.c: #include "veriuser.h" static PLI_INT32 hello() { io_printf("Hi there\n"); return 0; } s_tfcell veriusertfs[] = { {usertask, 0, 0, 0, hello, 0, "$hello"}, {0} /* last entry must be 0 */ }; hello.v: module hello; initial $hello; endmodule Compile the PLI code for the Solaris operating system: % cc -c -I<install_dir>/modeltech/include hello.c % ld -G -o hello.sl hello.o Compile the Verilog code: % vlib work % vlog hello.v Simulate the design: % vsim -c -pli hello.sl hello # Loading work.hello # Loading ./hello.sl VSIM 1> run -all # Hi there VSIM 2> quit
VPI example
The following example is a trivial, but complete VPI application. A general VPI example can be found in <install_dir>/modeltech/examples/verilog/vpi.
hello.c: #include "vpi_user.h" static PLI_INT32 hello(PLI_BYTE8 * param) { vpi_printf( "Hello world!\n" ); return 0; } void RegisterMyTfs( void ) { s_vpi_systf_data systf_data; vpiHandle systf_handle; systf_data.type = vpiSysTask; systf_data.sysfunctype = vpiSysTask; systf_data.tfname = "$hello"; systf_data.calltf = hello; systf_data.compiletf = 0; systf_data.sizetf = 0; systf_data.user_data = 0; systf_handle = vpi_register_systf( &systf_data ); vpi_free_object( systf_handle ); } void (*vlog_startup_routines[])() = { RegisterMyTfs, 0 }; hello.v: module hello; initial $hello; endmodule Compile the VPI code for the Solaris operating system: % gcc -c -I<install_dir>/include hello.c % ld -G -o hello.sl hello.o Compile the Verilog code: % vlib work % vlog hello.v Simulate the design: % vsim -c -pli hello.sl hello # Loading work.hello # Loading ./hello.sl VSIM 1> run -all # Hello world! VSIM 2> quit
DPI example UM-501
DPI example
The following example is a trivial but complete DPI application. For win32 and RS6000 platforms, an additional step is required. For additional examples, see the <install_dir>/ modeltech/examples/systemverilog/dpi directory.
hello_c.c: #include "svdpi.h" #include "dpiheader.h" int c_task(int i, int *o) { printf("Hello from c_task()\n"); verilog_task(i, o); /* Call back into Verilog */ *o = i; return(0); /* Return success (required by tasks) */
}
hello.v: module hello_top; int ret; export "DPI-C" task verilog_task; task verilog_task(input int i, output int o); #10; $display("Hello from verilog_task()"); endtask import "DPI-C" context task c_task(input int i, output int o); initial begin c_task(1, ret); end endmodule
// Call the c task named 'c_task()'
Compile the Verilog code: % vlib work % vlog -sv -dpiheader dpiheader.h hello.v Compile the DPI code for the Solaris operating system: % gcc -c -g -I<install_dir>/modeltech/include hello_c.c % ld -G -o hello_c.so hello_c.o Simulate the design: % vsim -c -sv_lib hello_c hello_top # Loading work.hello_c # Loading ./hello_c.so VSIM 1> run -all # Hello from c_task() # Hello from verilog_task() VSIM 2> quit
The PLI callback reason argument

The second argument to a PLI callback function is the reason argument. The values of the various reason constants are defined in the veriuser.h include file. See IEEE Std 1364 for a description of the reason constants. The following details relate to ModelSim Verilog, and may not be obvious in the IEEE Std 1364. Specifically, the simulator passes the reason values to the misctf callback functions under the following circumstances:
reason_endofcompile
For the completion of loading the design.

reason_finish
For the execution of the $finish system task or the quit command.
reason_startofsave
For the start of execution of the checkpoint command, but before any of the simulation state has been saved. This allows the PLI application to prepare for the save, but it shouldn't save its data with calls to tf_write_save() until it is called with reason_save.
reason_save
For the execution of the checkpoint command. This is when the PLI application must save its state with calls to tf_write_save().
reason_startofrestart
For the start of execution of the restore command, but before any of the simulation state has been restored. This allows the PLI application to prepare for the restore, but it shouldn't restore its state with calls to tf_read_restart() until it is called with reason_restart. The reason_startofrestart value is passed only for a restore command, and not in the case that the simulator is invoked with -restore.
reason_restart
For the execution of the restore command. This is when the PLI application must restore its state with calls to tf_read_restart().
reason_reset
For the execution of the restart command. This is when the PLI application should free its memory and reset its state. We recommend that all PLI applications reset their internal state during a restart as the shared library containing the PLI code might not be reloaded. (See the -keeploaded and -keeploadedrestart arguments to vsim for related information.)
reason_endofreset
For the completion of the restart command, after the simulation state has been reset but before the design has been reloaded.
reason_interactive
For the execution of the $stop system task or any other time the simulation is interrupted and waiting for user input.
reason_scope
For the execution of the environment command or selecting a scope in the structure window. Also for the call to acc_set_interactive_scope() if the callback_flag argument is non-zero.
reason_paramvc
For the change of value on the system task or function argument.
The PLI callback reason argument UM-503
reason_synch
For the end of time step event scheduled by tf_synchronize().

reason_rosynch
For the end of time step event scheduled by tf_rosynchronize().

reason_reactivate
For the simulation event scheduled by tf_setdelay().

reason_paramdrc
Not supported in ModelSim Verilog.

reason_force

reason_release

reason_disable
The sizetf callback function

A user-defined system function specifies the width of its return value with the sizetf callback function, and the simulator calls this function while loading the design. The following details on the sizetf callback function are not found in the IEEE Std 1364: If you omit the sizetf function, then a return width of 32 is assumed. The sizetf function should return 0 if the system function return value is of Verilog type "real". The sizetf function should return -32 if the system function return value is of Verilog type "integer".
PLI object handles UM-505
PLI object handles

Many of the object handles returned by the PLI ACC routines are pointers to objects that naturally exist in the simulation data structures, and the handles to these objects are valid throughout the simulation, even after the acc_close() routine is called. However, some of the objects are created on demand, and the handles to these objects become invalid after acc_close() is called. The following object types are created on demand in ModelSim Verilog:
accOperator (acc_handle_condition) accWirePath (acc_handle_path) accTerminal (acc_handle_terminal, acc_next_cell_load, acc_next_driver, and acc_next_load) accPathTerminal (acc_next_input and acc_next_output) accTchkTerminal (acc_handle_tchkarg1 and acc_handle_tchkarg2) accPartSelect (acc_handle_conn, acc_handle_pathin, and acc_handle_pathout)
If your PLI application uses these types of objects, then it is important to call acc_close() to free the memory allocated for these objects when the application is done using them. If your PLI application places value change callbacks on accRegBit or accTerminal objects, do not call acc_close() while these callbacks are in effect.
Third party PLI applications

Many third party PLI applications come with instructions on using them with ModelSim Verilog. Even without the instructions, it is still likely that you can get it to work with ModelSim Verilog as long as the application uses standard PLI routines. The following guidelines are for preparing a Verilog-XL PLI application to work with ModelSim Verilog. Generally, a Verilog-XL PLI application comes with a collection of object files and a veriuser.c file. The veriuser.c file contains the registration information as described above in "Registering DPI applications" (UM-480). To prepare the application for ModelSim Verilog, you must compile the veriuser.c file and link it to the object files to create a dynamically loadable object (see "Compiling and linking C applications for PLI/VPI/DPI" (UM-484)). For example, if you have a veriuser.c file and a library archive libapp.a file that contains the application's object files, then the following commands should be used to create a dynamically loadable object for the Solaris operating system:
% cc -c -I<install_dir>/modeltech/include veriuser.c % ld -G -o app.sl veriuser.o libapp.a
The PLI application is now ready to be run with ModelSim Verilog. All that's left is to specify the resulting object file to the simulator for loading using the Veriuser entry in the modesim.ini file, the -pli simulator argument, or the PLIOBJS environment variable (see "Registering DPI applications" (UM-480)). Note: On the HP700 platform, the object files must be compiled as position-independent code by using the +z compiler argument. Since, the object files supplied for Verilog-XL may be compiled for static linking, you may not be able to use the object files to create a dynamically loadable object for ModelSim Verilog. In this case, you must get the third party application vendor to supply the object files compiled as position-independent code.
Support for VHDL objects UM-507
Support for VHDL objects

The PLI ACC routines also provide limited support for VHDL objects in either an all VHDL design or a mixed VHDL/Verilog design. The following table lists the VHDL objects for which handles may be obtained and their type and fulltype constants: Type accArchitecture accArchitecture accArchitecture accArchitecture accArchitecture Fulltype accArchitecture accEntityVitalLevel0 accArchVitalLevel0 accArchVitalLevel1 accForeignArch Description instantiation of an architecture instantiation of an architecture whose entity is marked with the attribute VITAL_Level0 instantiation of an architecture which is marked with the attribute VITAL_Level0 instantiation of an architecture which is marked with the attribute VITAL_Level1 instantiation of an architecture which is marked with the attribute FOREIGN and which does not contain any VHDL statements or objects other than ports and generics instantiation of an architecture which is marked with the attribute FOREIGN and which contains some VHDL statements or objects besides ports and generics block statement for loop statement foreign scope created by mti_CreateRegion() generate statement package declaration signal declaration
accArchitecture
accForeignArchMixed
accBlock accForLoop accForeign accGenerate accPackage accSignal
accBlock accForLoop accShadow accGenerate accPackage accSignal
The type and fulltype constants for VHDL objects are defined in the acc_vhdl.h include file. All of these objects (except signals) are scope objects that define levels of hierarchy in the structure window. Currently, the PLI ACC interface has no provision for obtaining handles to generics, types, constants, variables, attributes, subprograms, and processes.
IEEE Std 1364 ACC routines

ModelSim Verilog supports the following ACC routines: cc_append_delays cc_append_pulsere cc_close cc_collect cc_compare_handles cc_configure cc_count cc_fetch_argc cc_fetch_argv cc_fetch_attribute cc_fetch_attribute_int cc_fetch_attribute_str cc_fetch_defname cc_fetch_delay_mode cc_fetch_delays cc_fetch_direction cc_fetch_edge cc_fetch_fullname cc_fetch_fulltype cc_fetch_index cc_fetch_location acc_fetch_name acc_fetch_paramtype acc_fetch_paramval acc_fetch_polarity acc_fetch_precision acc_handle_loconn acc_fetch_pulsere acc_handle_modpath acc_fetch_range acc_handle_notifier acc_fetch_size acc_handle_object acc_fetch_tfarg acc_handle_parent acc_fetch_itfarg acc_handle_path acc_fetch_tfarg_int acc_handle_pathin acc_fetch_itfarg_int acc_handle_pathout acc_fetch_tfarg_str acc_handle_port acc_fetch_itfarg_str acc_handle_scope acc_fetch_timescale_info acc_handle_simulated_net acc_fetch_type acc_handle_tchk acc_fetch_type_str acc_handle_tchkarg1 acc_fetch_value acc_handle_tchkarg2 acc_free acc_handle_terminal acc_handle_by_name acc_handle_tfarg acc_handle_calling_mod_m acc_handle_itfarg acc_handle_condition acc_handle_conn acc_handle_hiconn acc_handle_interactive_scop e
IEEE Std 1364 ACC routines UM-509
acc_handle_tfinst acc_initialize acc_next acc_next_bit acc_next_cell acc_next_cell_load acc_next_child acc_next_driver acc_next_hiconn acc_next_input acc_next_load acc_next_loconn acc_next_modpath
acc_next_net acc_next_output acc_next_parameter acc_next_port acc_next_portout acc_next_primitive acc_next_scope acc_next_specparam acc_next_tchk acc_next_terminal acc_next_topmod acc_object_in_typelist acc_object_of_type
acc_product_type acc_product_version acc_release_object acc_replace_delays acc_replace_pulsere acc_reset_buffer acc_set_interactive_scope acc_set_pulsere acc_set_scope acc_set_value acc_vcl_add acc_vcl_delete acc_version
acc_fetch_paramval() cannot be used on 64-bit platforms to fetch a string value of a parameter. Because of this, the function acc_fetch_paramval_str() has been added to the PLI for this use. acc_fetch_paramval_str() is declared in acc_user.h. It functions in a manner similar to acc_fetch_paramval() except that it returns a char *. acc_fetch_paramval_str() can be used on all platforms.
IEEE Std 1364 TF routines

ModelSim Verilog supports the following TF (task and function) routines; io_mcdprintf io_printf mc_scan_plusargs tf_add_long tf_asynchoff tf_iasynchoff tf_asynchon tf_iasynchon tf_clearalldelays tf_iclearalldelays tf_compare_long tf_copypvc_flag tf_icopypvc_flag tf_divide_long tf_dofinish tf_dostop tf_error tf_evaluatep tf_ievaluatep tf_exprinfo tf_iexprinfo tf_getcstringp tf_igetcstringp tf_getinstance tf_getlongp tf_igetlongp tf_getlongtime tf_igetlongtime tf_getnextlongtime tf_getp tf_igetp tf_getpchange tf_igetpchange tf_getrealp tf_igetrealp tf_getrealtime tf_igetrealtime tf_gettime tf_igettime tf_gettimeprecision tf_igettimeprecision tf_gettimeunit tf_igettimeunit tf_getworkarea tf_igetworkarea tf_long_to_real tf_longtime_tostr tf_message tf_mipname tf_imipname tf_movepvc_flag tf_imovepvc_flag tf_multiply_long tf_nodeinfo tf_inodeinfo tf_nump tf_inump tf_propagatep tf_ipropagatep tf_putlongp tf_iputlongp tf_putp tf_iputp tf_putrealp tf_iputrealp tf_read_restart tf_real_to_long tf_rosynchronize tf_irosynchronize
IEEE Std 1364 TF routines UM-511
tf_scale_longdelay tf_scale_realdelay tf_setdelay tf_isetdelay tf_setlongdelay tf_isetlongdelay tf_setrealdelay tf_isetrealdelay tf_setworkarea tf_isetworkarea tf_sizep tf_isizep
tf_spname tf_ispname tf_strdelputp tf_istrdelputp tf_strgetp tf_istrgetp tf_strgettime tf_strlongdelputp tf_istrlongdelputp tf_strrealdelputp tf_istrrealdelputp tf_subtract_long
tf_synchronize tf_isynchronize tf_testpvc_flag tf_itestpvc_flag tf_text tf_typep tf_itypep tf_unscale_longdelay tf_unscale_realdelay tf_warning tf_write_save
SystemVerilog DPI access routines

ModelSim SystemVerilog supports all routines defined in the "svdpi.h" file defined in P1800-2005, with the exception of functions related to open array processing.
Verilog-XL compatible routines

The following PLI routines are not defined in IEEE Std 1364, but ModelSim Verilog provides them for compatibility with Verilog-XL.
char *acc_decompile_exp(handle condition)
This routine provides similar functionality to the Verilog-XL acc_decompile_expr routine. The condition argument must be a handle obtained from the acc_handle_condition routine. The value returned by acc_decompile_exp is the string representation of the condition expression.
char *tf_dumpfilename(void)
This routine returns the name of the VCD file.

void tf_dumpflush(void)
A call to this routine flushes the VCD file buffer (same effect as calling $dumpflush in the Verilog code).
int tf_getlongsimtime(int *aof_hightime)
This routine gets the current simulation time as a 64-bit integer. The low-order bits are returned by the routine, while the high-order bits are stored in the aof_hightime argument.
64-bit support for PLI UM-513
64-bit support for PLI

The PLI function acc_fetch_paramval() cannot be used on 64-bit platforms to fetch a string value of a parameter. Because of this, the function acc_fetch_paramval_str() has been added to the PLI for this use. acc_fetch_paramval_str() is declared in acc_user.h. It functions in a manner similar to acc_fetch_paramval() except that it returns a char *. acc_fetch_paramval_str() can be used on all platforms.
Using 64-bit ModelSim with 32-bit Applications

If you have 32-bit PLI/VPI/DPI applications and wish to use 64-bit ModelSim, you will need to port your code to 64 bits by moving from the ILP32 data model to the LP64 data model. We strongly recommend that you consult the 64-bit porting guides for Sun and HP.
PLI/VPI tracing
The foreign interface tracing feature is available for tracing PLI and VPI function calls. Foreign interface tracing creates two kinds of traces: a human-readable log of what functions were called, the value of the arguments, and the results returned; and a set of C-language files that can be used to replay what the foreign interface code did.
The purpose of tracing files

The purpose of the logfile is to aid you in debugging PLI or VPI code. The primary purpose of the replay facility is to send the replay files to MTI support for debugging co-simulation problems, or debugging PLI/VPI problems for which it is impractical to send the PLI/VPI code. We still need you to send the VHDL/Verilog part of the design to actually execute a replay, but many problems can be resolved with the trace only.
Invoking a trace
To invoke the trace, call vsim (CR-305) with the -trace_foreign argument:
Syntax
vsim -trace_foreign <action> [-tag <name>]
Arguments
<action>
Specifies one of the following actions:
Value 1 2
Action create log only create replay only
Result writes a local file called "mti_trace_<tag>" writes local files called "mti_data_<tag>.c", "mti_init_<tag>.c", "mti_replay_<tag>.c" and "mti_top_<tag>.c"
3
-tag <name>
create both log and replay
Used to give distinct file names for multiple traces. Optional.
PLI/VPI tracing UM-515
Examples
vsim -trace_foreign 1 mydesign
Creates a logfile.
vsim -trace_foreign 3 mydesign
Creates both a logfile and a set of replay files.

vsim -trace_foreign 1 -tag 2 mydesign
Creates a logfile with a tag of "2". The tracing operations will provide tracing during all user foreign code-calls, including PLI/VPI user tasks and functions (calltf, checktf, sizetf and misctf routines), and Verilog VCL callbacks.
Debugging PLI/VPI/DPI application code

ModelSim offers the optional C Debug feature. This tool allows you to interactively debug SystemC/C/C++ source code with the open-source gdb debugger. See Chapter 12 - C Debug for details. If you dont have access to C Debug, continue reading for instructions on how to attach to an external C debugger. In order to debug your PLI/VPI/DPI application code in a debugger, you must first: 1 Compile the application code with debugging information (using the -g option) and without optimizations (for example, dont use the -O option). 2 Load vsim into a debugger. Even though vsim is stripped, most debuggers will still execute it. You can invoke the debugger directly on vsimk, the simulation kernal where your application code is loaded (for example, "ddd `which vsimk`"), or you can attach the debugger to an already running vsim process. In the second case, you must attach to the PID for vsimk, and you must specify the full path to the vsimk executable (for example, "gdb $MTI_HOME/ sunos5/vsimk 1234"). On Solaris, AIX, and Linux systems you can use either gdb or ddd. On HP-UX systems you can use the wdb debugger from HP. You will need version 1.2 or later. 3 Set an entry point using breakpoint. Since initially the debugger recognizes only vsim's PLI/VPI/DPI function symbols, when invoking the debugger directly on vsim you need to place a breakpoint in the first PLI/VPI/DPI function that is called by your application code. An easy way to set an entry point is to put a call to acc_product_version() as the first executable statement in your application code. Then, after vsim has been loaded into the debugger, set a breakpoint in this function. Once you have set the breakpoint, run vsim with the usual arguments. When the breakpoint is reached, the shared library containing your application code has been loaded. 4 In some debuggers, you must use the share command to load the application's symbols. At this point all of the application's symbols should be visible. You can now set breakpoints in and single step through your application code.
HP-UX specific warnings

On HP-UX you might see some warning messages that vsim does not have debugging information available. This is normal. If you are using Exceed to access an HP machine from Windows NT, it is recommended that you run vsim in command line or batch mode because your NT machine may hang if you run vsim in GUI mode. Click on the "go" button, or use F5 or the go command to execute vsim in wdb. You might also see a warning about not finding "__dld_flags" in the object file. This warning can be ignored. You should see a list of libraries loaded into the debugger. It should include the library for your PLI/VPI/DPI application. Alternatively, you can use share to load only a single library.
UM-517
D - ModelSim shortcuts
Appendix contents
Command shortcuts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-517 UM-517 UM-519 UM-522 UM-523 Command history shortcuts
Main and Source window mouse and keyboard shortcuts . List window keyboard shortcuts . . . . . . . . . .
Wave window mouse and keyboard shortcuts .
This appendix is a collection of the keyboard and command shortcuts available in the ModelSim GUI.
Command shortcuts
You may abbreviate command syntax, but theres a catch the minimum number of characters required to execute a command are those that make it unique. Remember, as we add new commands some of the old shortcuts may not work. For this reason ModelSim does not allow command name abbreviations in macro files. This minimizes your need to update macro files as new commands are added. Multiple commands may be entered on one line if they are separated by semi-colons (;). For example:
ModelSim> vlog -nodebug=ports level3.v level2.v ; vlog -nodebug top.v
The return value of the last function executed is the only one printed to the transcript. This may cause some unexpected behavior in certain circumstances. Consider this example:
vsim -c -do "run 20 ; simstats ; quit -f" top
You probably expect the simstats results to display in the Transcript window, but they will not, because the last command is quit -f. To see the return values of intermediate commands, you must explicitly print the results. For example:
vsim -do "run 20 ; echo [simstats]; quit -f" -c top
Command history shortcuts

The simulator command history may be reviewed, or commands may be reused, with these shortcuts at the ModelSim/VSIM prompt: Shortcut
!! !n
Description repeats the last command repeats command number n; n is the VSIM prompt number (e.g., for this prompt: VSIM 12>, n =12)
UM-518 D - ModelSim shortcuts
Shortcut
!abc ^xyzâb^
Description repeats the most recent command starting with "abc" replaces "xyz" in the last command with "ab" scrolls through the command history with the keyboard arrows left-click once on a previous ModelSim or VSIM prompt in the transcript to copy the command typed at that prompt to the active cursor shows the last few commands (up to 50 are kept)
up and down arrows click on prompt
his or history
Main and Source window mouse and keyboard shortcuts UM-519
Main and Source window mouse and keyboard shortcuts

The following mouse actions and special keystrokes can be used to edit commands in the entry region of the Main window. They can also be used in editing the file displayed in the Source window and all Notepad windows (enter the notepad command within ModelSim to open the Notepad editor). Mouse - UNIX < left-button - click > < left-button - press > + drag < shift - left-button - press > < left-button - double-click > < left-button - double-click > + drag < control - left-button - click > < left-button - click > on previous ModelSim or VSIM prompt < middle-button - click > < middle-button - press > + drag none none Mouse - Windows Result move the insertion cursor select extend selection select word select word + word move insertion cursor without changing the selection copy and paste previous command string to current prompt paste clipboard scroll the window
Keystrokes - UNIX < left | right arrow > < control > < left | right arrow >
Keystrokes - Windows
Result move cursor left | right one character move cursor left | right one word extend selection of text extend selection of text by word scroll through command history (in Source window, moves cursor one line up | down) moves cursor up | down one paragraph move cursor to the beginning of the text move cursor to the end of the text
< shift > < left | right | up | down arrow > < control > < shift > < left | right arrow > < up | down arrow > < control > < up | down > < control > < home > < control > < end > < backspace >, < control-h > < delete >, < control-d > none < backspace > < delete > esc
delete character to the left delete character to the right cancel
Keystrokes - UNIX < alt > < alt > < F4 > < control - a >, < home > < control - b > < control - d > < control - e >, < end > < control - f > < control - k > < control - n > < control - o > < control - p > < control - s > < F3 > < control - t > < control - u > < control - v >, PageDn < control - w > < control - x >, < control - s> < control - y >, F18 none < control - \ > < control - ->, < control - / > < meta - "<" > < meta - ">" > < meta - v >, PageUp < Meta - w>
Result activate or inactivate menu bar mode close active window
< home >
move cursor to the beginning of the line move cursor left delete character to the right
< end > <right arrow>
move cursor to the end of the line move cursor right one character delete to the end of line move cursor one line down (Source window only under Windows)
none
insert a new line character at the cursor move cursor one line up (Source window only under Windows)
< control - f >
find find next reverse the order of the two characters on either side of the cursor delete line
PageDn < control - x > < control - s > < control - v > < control - a >
move cursor down one screen cut the selection save paste the selection select the entire contents of the widget clear any selection in the widget
< control - Z > none none PageUp < control - c >
undoes previous edits in the Source window move cursor to the beginning of the file move cursor to the end of the file move cursor up one screen copy selection
Main and Source window mouse and keyboard shortcuts UM-521
Keystrokes - UNIX < F8 >
Result search for the most recent command that matches the characters typed (Main window only) run simulation continue simulation
< F9> < F10 > < F11 > < F12>
single-step step-over
The Main window allows insertions or pastes only after the prompt; therefore, you dont need to set the cursor when copying strings to the command line.
List window keyboard shortcuts

Using the following keys when the mouse cursor is within the List window will cause the indicated actions:
Key <left arrow> <right arrow> <up arrow> <down arrow> <page up> <control-up arrow> <page down> <control-down arrow> <tab> <shift-tab> <shift-left arrow> <shift-right arrow> <control-f> Windows <control-s> UNIX
Action scroll listing left (selects and highlights the item to the left of the currently selected item) scroll listing right (selects and highlights the item to the right of the currently selected item) scroll listing up scroll listing down scroll listing up by page scroll listing down by page
searches forward (down) to the next transition on the selected signal searches backward (up) to the previous transition on the selected signal (does not function on HP workstations) extends selection left/right opens the Find dialog box to find the specified item label within the list display
Wave window mouse and keyboard shortcuts UM-523
Wave window mouse and keyboard shortcuts

The following mouse actions and keystrokes can be used in the Wave window.
Mouse action < control - left-button - drag down and right>a < control - left-button - drag up and right> < control - left-button - drag up and left> <left-button - drag> (Select mode) < middle-button - drag> (Zoom mode) < control - left-button - click on a scroll arrow >
Result zoom area (in) zoom out zoom fit moves closest cursor scrolls window to very top or bottom (vertical scroll) or far left or right (horizontal scroll) scrolls window to position of click
< middle mouse-button - click in scroll bar trough> (UNIX) only
a. If you enter zoom mode by selecting View > Mouse Mode > Zoom Mode, you do not need to hold down the <Ctrl> key.
Keystroke s i I or + o O or f or F l or L r or R <up arrow>/ <down arrow> <left arrow> <right arrow>
Action bring into view and center the currently active cursor zoom in (mouse pointer must be over the cursor or waveform panes) zoom out (mouse pointer must be over the cursor or waveform panes) zoom full (mouse pointer must be over the cursor or waveform panes) zoom last (mouse pointer must be over the cursor or waveform panes) zoom range (mouse pointer must be over the cursor or waveform panes) with mouse over waveform pane, scrolls entire window up/ down one line; with mouse over pathname or values pane, scrolls highlight up/down one line scroll pathname, values, or waveform pane left scroll pathname, values, or waveform pane right
Keystroke <page up> <page down> <tab> <shift-tab> <control-f> Windows <control-s> UNIX <control-left arrow> <control-right arrow>
Action scroll waveform pane up by a page scroll waveform pane down by a page search forward (right) to the next transition on the selected signal - finds the next edge search backward (left) to the previous transition on the selected signal - finds the previous edge open the find dialog box; searches within the specified field in the pathname pane for text strings scroll pathname, values, or waveform pane left by a page scroll pathname, values, or waveform pane right by a page
UM-525
E - System initialization
Appendix contents
Files accessed during startup . . . . . . . . . . . . . . . . . . . . . . . UM-526 UM-527 UM-529 Environment variables accessed during startup Initialization sequence . . . . . . .
ModelSim goes through numerous steps as it initializes the system during startup. It accesses various files and environment variables to determine library mappings, configure the GUI, check licensing, and so forth.
UM-526 E - System initialization
Files accessed during startup

The table below describes the files that are read during startup. They are listed in the order in which they are accessed.
File modelsim.ini location map file pref.tcl .modelsim (UNIX) or Windows registry modelsim.tcl
Purpose contains initial tool settings; see "Control variables located in INI files" (UM-438) for specific details on the modelsim.ini file used by ModelSim tools to find source files based on easily reallocated "soft" paths; default file name is mgc_location_map contains defaults for fonts, colors, prompts, window positions, and other simulator window characteristics contains last working directory, project file, printer defaults, and other user-customized GUI characteristics contains user-customized settings for fonts, colors, prompts, other GUI characteristics; maintained for backwards compatibility with older versions (see "The modelsim.tcl file" (GR-253)) if available, loads last project file which is specified in the registry (Windows) or $(HOME)/.modelsim (UNIX); see "What are projects?" (UM-36) for details on project settings
<project_name>.mpf
Environment variables accessed during startup UM-527
Environment variables accessed during startup

The table below describes the environment variables that are read during startup. They are listed in the order in which they are accessed. For more information on environment variables, see "Environment variables" (UM-434). Environment variable MODEL_TECH MODEL_TECH_OVERRIDE MODELSIM MGC_WD MGC_LOCATION_MAP MODEL_TECH_TCL HOME MGC_HOME TCL_LIBRARY Purpose set by ModelSim to the directory in which the binary executables reside (e.g., ../modeltech/<platform>/) provides an alternative directory for the binary executables; MODEL_TECH is set to this path identifies the pathname of the modelsim.ini file identifies the Mentor Graphics working directory identifies the pathname of the location map file; set by ModelSim if not defined identifies the pathname of all Tcl libraries installed with ModelSim identifies your login directory (UNIX only) identifies the pathname of the MGC tool suite identifies the pathname of the Tcl library; set by ModelSim to the same pathname as MODEL_TECH_TCL; must point to libraries supplied by Model Technology identifies the pathname of the Tk library; set by ModelSim to the same pathname as MODEL_TECH_TCL; must point to libraries supplied by Model Technology identifies the pathname of the [incr]Tcl library; set by ModelSim to the same path as MODEL_TECH_TCL; must point to libraries supplied by Model Technology identifies the pathname of the [incr]Tk library; set by ModelSim to the same pathname as MODEL_TECH_TCL; must point to libraries supplied by Model Technology identifies the pathname of the Tcl files that are used by ModelSim; set by ModelSim to the same pathname as MODEL_TECH_TCL; must point to libraries supplied by Model Technology creates an mti_trace_cosim file containing debugging information about FLI/PLI/VPI function calls; set to any value before invoking the simulator identifies the path to all Tcl libraries installed with ModelSim determines which version of ModelSim to use on platforms that support both 32- and 64-bit versions when ModelSim executables are invoked from the modeltech/bin directory by a Unix shell command (using full path specification or PATH search)
TK_LIBRARY
ITCL_LIBRARY
ITK_LIBRARY
VSIM_LIBRARY
MTI_COSIM_TRACE MTI_LIB_DIR MTI_VCO_MODE
Environment variable MODELSIM_TCL
Purpose identifies the pathname to a user preference file (e.g., C:\modeltech\modelsim.tcl); can be a list of file pathnames, separated by semicolons (Windows) or colons (UNIX); note that user preferences are now stored in the .modelsim file (Unix) or registry (Windows); ModelSim will still read this environment variable but it will then save all the settings to the .modelsim file when you exit the tool
Initialization sequence UM-529
Initialization sequence
The following list describes in detail ModelSims initialization sequence. The sequence includes a number of conditional structures, the results of which are determined by the existence of certain files and the current settings of environment variables. In the steps below, names in uppercase denote environment variables (except MTI_LIB_DIR which is a Tcl variable). Instances of $(NAME) denote paths that are determined by an environment variable (except $(MTI_LIB_DIR) which is determined by a Tcl variable). 1 Determines the path to the executable directory (../modeltech/<platform>). Sets MODEL_TECH to this path, unless MODEL_TECH_OVERRIDE exists, in which case MODEL_TECH is set to the same value as MODEL_TECH_OVERRIDE. 2 Finds the modelsim.ini file by evaluating the following conditions: use MODELSIM if it exists; else use $(MGC_WD)/modelsim.ini; else use ./modelsim.ini; else use $(MODEL_TECH)/modelsim.ini; else use $(MODEL_TECH)/../modelsim.ini; else use $(MGC_HOME)/lib/modelsim.ini; else set path to ./modelsim.ini even though the file doesnt exist 3 Finds the location map file by evaluating the following conditions: use MGC_LOCATION_MAP if it exists (if this variable is set to "no_map", ModelSim skips initialization of the location map); else use mgc_location_map if it exists; else use $(HOME)/mgc/mgc_location_map; else use $(HOME)/mgc_location_map; else use $(MGC_HOME)/etc/mgc_location_map; else use $(MGC_HOME)/shared/etc/mgc_location_map; else use $(MODEL_TECH)/mgc_location_map; else use $(MODEL_TECH)/../mgc_location_map; else use no map 4 Reads various variables from the [vsim] section of the modelsim.ini file. See "[vsim] simulator control variables" (UM-444) for more details. 5 Parses any command line arguments that were included when you started ModelSim and reports any problems. 6 Defines the following environment variables: use MODEL_TECH_TCL if it exists; else
set MODEL_TECH_TCL=$(MODEL_TECH)/../tcl set TCL_LIBRARY=$(MODEL_TECH_TCL)/tcl8.3 set TK_LIBRARY=$(MODEL_TECH_TCL)/tk8.3 set ITCL_LIBRARY=$(MODEL_TECH_TCL)/itcl3.0 set ITK_LIBRARY=$(MODEL_TECH_TCL)/itk3.0 set VSIM_LIBRARY=$(MODEL_TECH_TCL)/vsim 7 Initializes the simulators Tcl interpreter. 8 Checks for a valid license (a license is not checked out unless specified by a modelsim.ini setting or command line option). The next four steps relate to initializing the graphical user interface. 9 Sets Tcl variable MTI_LIB_DIR=$(MODEL_TECH_TCL) 10 Loads $(MTI_LIB_DIR)/vsim/pref.tcl. 11 Loads gui preferences, project file, etc. from the registry (Windows) or $(HOME)/ .modelsim (UNIX). 12 Searches for the modelsim.tcl file by evaluating the following conditions: use MODELSIM_TCL environment variable if it exists (if MODELSIM_TCL is a list of files, each file is loaded in the order that it appears in the list); else use ./modelsim.tcl; else use $(HOME)/modelsim.tcl if it exists That completes the initialization sequence. Also note the following about the modelsim.ini file: When you change the working directory within ModelSim, the tool reads the [library], [vcom], and [vlog] sections of the local modelsim.ini file. When you make changes in the compiler or simulator options dialog or use the vmap command, the tool updates the appropriate sections of the file. The pref.tcl file references the default .ini file via the [GetPrivateProfileString] Tcl command. The .ini file that is read will be the default file defined at the time pref.tcl is loaded.
UM-531
F - Logic Modeling SmartModels

Appendix contents
VHDL SmartModel interface . . . . . . Creating foreign architectures with sm_entity Vector ports . . . . . . . . . Command channel. . . . . . . . SmartModel Windows . . . . . . Memory arrays . . . . . . . . Verilog SmartModel interface . . . . . . Linking the LMTV interface to the simulator. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-532 UM-533 UM-535 UM-536 UM-537 UM-538 UM-539 UM-539
The Logic Modeling SWIFT-based SmartModel library can be used with ModelSim VHDL and Verilog. The SmartModel library is a collection of behavioral models supplied in binary form with a procedural interface that is accessed by the simulator. This appendix describes how to use the SmartModel library with ModelSim. The SmartModel library must be obtained from Logic Modeling along with the documentation that describes how to use it. This appendix only describes the specifics of using the library with ModelSim. A 32-bit SmartModel will not run with a 64-bit version of SE. When trying to load the operating system specific 32-bit library into the 64-bit executable, the pointer sizes will be incorrect. Note: The VHDL SmartModel interface is available as an add-on to the PE and Designer versions. Contact your Mentor Graphics sales representative for more information.
UM-532 F - Logic Modeling SmartModels
VHDL SmartModel interface

ModelSim VHDL interfaces to a SmartModel through a foreign architecture. The foreign architecture contains a foreign attribute string that associates a specific SmartModel with the architecture. On elaboration of the foreign architecture, the simulator automatically loads the SmartModel library software and establishes communication with the specific SmartModel.
Enabling the interface

To enable the SmartModel interface you must do the following: Set the LMC_HOME environment variable to the root of the SmartModel library installation directory. Consult Logic Modeling's documentation for details. Uncomment the appropriate libswift entry in the modelsim.ini file for your operating system. If you are running the Windows operating system, you must also comment out the default libsm entry (precede the line with the ";" character) and uncomment the libsm entry for the Windows operating system. The libswift and libsm entries are found under the [lmc] section of the default modelsim.ini file located in the ModelSim installation directory. The default settings are as follows:
[lmc] ; ModelSim's interface to Logic Modeling's SmartModel SWIFT software libsm = $MODEL_TECH/libsm.sl ; ModelSim's interface to Logic Modeling's SmartModel SWIFT software (Windows NT) ; libsm = $MODEL_TECH/libsm.dll ; Logic Modeling's SmartModel SWIFT software (HP 9000 Series 700) ; libswift = $LMC_HOME/lib/hp700.lib/libswift.sl ; Logic Modeling's SmartModel SWIFT software (IBM RISC System/6000) ; libswift = $LMC_HOME/lib/ibmrs.lib/swift.o ; Logic Modeling's SmartModel SWIFT software (Sun4 Solaris) ; libswift = $LMC_HOME/lib/sun4Solaris.lib/libswift.so ; Logic Modeling's SmartModel SWIFT software (Windows NT) ; libswift = $LMC_HOME/lib/pcnt.lib/libswift.dll ; Logic Modeling's SmartModel SWIFT software (Linux) ; libswift = $LMC_HOME/lib/x86_linux.lib/libswift.so
The libsm entry points to the ModelSim dynamic link library that interfaces the foreign architecture to the SmartModel software. The libswift entry points to the Logic Modeling dynamic link library software that accesses the SmartModels. The simulator automatically loads both the libsm and libswift libraries when it elaborates a SmartModel foreign architecture. By default, the libsm entry points to the libsm.sl supplied in the ModelSim installation directory indicated by the MODEL_TECH environment variable. ModelSim automatically sets the MODEL_TECH environment variable to the appropriate directory containing the executables and binaries for the current operating system.
VHDL SmartModel interface UM-533
Creating foreign architectures with sm_entity

The ModelSim sm_entity tool automatically creates entities and foreign architectures for SmartModels. Its usage is as follows: Syntax
sm_entity [-] [-xe] [-xa] [-c] [-all] [-v] [-93] [<SmartModelName>...]
Arguments
-
Read SmartModel names from standard input.

-xe
Do not generate entity declarations.

-xa
Do not generate architecture bodies.

-c
Generate component declarations.

-all
Select all models installed in the SmartModel library.

-v
Display progress messages.

-93
Use extended identifiers where needed.

<SmartModelName>
Name of a SmartModel (see the SmartModel library documentation for details on SmartModel names). By default, the sm_entity tool writes an entity and foreign architecture to stdout for each SmartModel name listed on the command line. Optionally, you can include the component declaration (-c), exclude the entity (-xe), and exclude the architecture (-xa). The simplest way to prepare SmartModels for use with ModelSim VHDL is to generate the entities and foreign architectures for all installed SmartModels, and compile them into a library named lmc. This is easily accomplished with the following commands:
% sm_entity -all > sml.vhd % vlib lmc % vcom -work lmc sml.vhd
To instantiate the SmartModels in your VHDL design, you also need to generate component declarations for the SmartModels. Add these component declarations to a package named sml (for example), and compile the package into the lmc library:
% sm_entity -all -c -xe -xa > smlcomp.vhd
Edit the resulting smlcomp.vhd file to turn it into a package of SmartModel component declarations as follows:
library ieee; use ieee.std_logic_1164.all; package sml is <component declarations go here> end sml;
Compile the package into the lmc library:

% vcom -work lmc smlcomp.vhd
The SmartModels can now be referenced in your design by adding the following library and use clauses to your code:
library lmc; use lmc.sml.all;
The following is an example of an entity and foreign architecture created by sm_entity for the cy7c285 SmartModel.
library ieee; use ieee.std_logic_1164.all; entity cy7c285 is generic (TimingVersion : STRING := "CY7C285-65"; DelayRange : STRING := "Max"; MemoryFile : STRING := "memory" ); port ( A0 : in std_logic; A1 : in std_logic; A2 : in std_logic; A3 : in std_logic; A4 : in std_logic; A5 : in std_logic; A6 : in std_logic; A7 : in std_logic; A8 : in std_logic; A9 : in std_logic; A10 : in std_logic; A11 : in std_logic; A12 : in std_logic; A13 : in std_logic; A14 : in std_logic; A15 : in std_logic; CS : in std_logic; O0 : out std_logic; O1 : out std_logic; O2 : out std_logic; O3 : out std_logic; O4 : out std_logic; O5 : out std_logic; O6 : out std_logic; O7 : out std_logic; WAIT_PORT : inout std_logic ); end; architecture SmartModel of cy7c285 is attribute FOREIGN : STRING; attribute FOREIGN of SmartModel : architecture is "sm_init $MODEL_TECH/libsm.sl ; cy7c285"; begin end SmartModel;
Entity details
The entity name is the SmartModel name (you can manually change this name if you like). The port names are the same as the SmartModel port names (these names must not be changed). If the SmartModel port name is not a valid VHDL identifier, then sm_entity automatically converts it to a valid name. If sm_entity is invoked with the -93 option, then the identifier is converted to an extended identifier, and the resulting entity must also be compiled with the -93 option. If the -93 option had been specified in the example above, then WAIT would have been converted to \WAIT\. Note that in this example the port WAIT was converted to WAIT_PORT because wait is a VHDL reserved word. The port types are std_logic. This data type supports the full range of SmartModel logic states. The DelayRange, TimingVersion, and MemoryFile generics represent the SmartModel attributes of the same name. Consult your SmartModel library documentation for a description of these attributes (and others). Sm_entity creates a generic for each attribute of the particular SmartModel. The default generic value is the default attribute value that the SmartModel has supplied to sm_entity.
Architecture details
The first part of the foreign attribute string (sm_init) is the same for all SmartModels. The second part ($MODEL_TECH/libsm.sl) is taken from the libsm entry in the initialization file, modelsim.ini. The third part (cy7c285) is the SmartModel name. This name correlates the architecture with the SmartModel at elaboration.
Vector ports
The entities generated by sm_entity only contain single-bit ports, never vectored ports. This is necessary because ModelSim correlates entity ports with the SmartModel SWIFT interface by name. However, for ease of use in component instantiations, you may want to create a custom component declaration and component specification that groups ports into vectors. You can also rename and reorder the ports in the component declaration. You can also reorder the ports in the entity declaration, but you can't rename them! The following is an example component declaration and specification that groups the address and data ports of the CY7C285 SmartModel:
component cy7c285 generic ( TimingVersion : STRING := "CY7C285-65"; DelayRange : STRING := "Max"; MemoryFile : STRING := "memory" ); port ( A : in std_logic_vector (15 downto 0); CS : in std_logic; O : out std_logic_vector (7 downto 0); WAIT_PORT : inout std_logic ); end component;
for all: cy7c285 use entity work.cy7c285 port map (A0 => A(0), A1 => A(1),
A2 => A(2), A3 => A(3), A4 => A(4), A5 => A(5), A6 => A(6), A7 => A(7), A8 => A(8), A9 => A(9), A10 => A(10), A11 => A(11), A12 => A(12), A13 => A(13), A14 => A(14), A15 => A(15), CS => CS, O0 => O(0), O1 => O(1), O2 => O(2), O3 => O(3), O4 => O(4), O5 => O(5), O6 => O(6), O7 => O(7), WAIT_PORT => WAIT_PORT );
Command channel
The command channel is a SmartModel feature that lets you invoke SmartModel specific commands. These commands are documented in the SmartModel library documentation from Synopsys. ModelSim provides access to the Command Channel from the command line. The form of a SmartModel command is:
lmc <instance_name>|-all "<SmartModel command>"
The instance_name argument is either a full hierarchical name or a relative name of a SmartModel instance. A relative name is relative to the current environment setting (see environment command (CR-131)). For example, to turn timing checks off for SmartModel /top/u1:
lmc /top/u1 "SetConstraints Off"
Use -all to apply the command to all SmartModel instances. For example, to turn timing checks off for all SmartModel instances:
lmc -all "SetConstraints Off"
There are also some SmartModel commands that apply globally to the current simulation session rather than to models. The form of a SmartModel session command is:
lmcsession "<SmartModel session command>"
SmartModel Windows
Some models in the SmartModel library provide access to internal registers with a feature called SmartModel Windows. Refer to Logic Modelings SmartModel library documentation (available on Synopsys web site) for details on this feature. The simulator interface to this feature is described below. Window names that are not valid VHDL or Verilog identifiers are converted to VHDL extended identifiers. For example, with a window named z1I10.GSR.OR, ModelSim will treat the name as \z1I10.GSR.OR\ (for all commands including lmcwin, add wave, and examine). You must then use that name in all commands. For example,
add wave /top/swift_model/\z1I10.GSR.OR\
Extended identifiers are case sensitive.
ReportStatus
The ReportStatus command displays model information, including the names of window registers. For example,
lmc /top/u1 ReportStatus
SmartModel Windows description:

WA "Read-Only (Read Only)" WB "1-bit" WC "64-bit"
This model contains window registers named wa, wb, and wc. These names can be used in subsequent window (lmcwin) commands.
SmartModel lmcwin commands

The following window commands are supported: lmcwin read <window_instance> [-<radix>] lmcwin write <window_instance> <value> lmcwin enable <window_instance> lmcwin disable <window_instance> lmcwin release <window_instance> Each command requires a window instance argument that identifies a specific model instance and window name. For example, /top/u1/wa refers to window wa in model instance /top/u1.
lmcwin read
The lmcwin read command displays the current value of a window. The optional radix argument is -binary, -decimal, or -hexadecimal (these names can be abbreviated). The default is to display the value using the std_logic characters. For example, the following command displays the 64-bit window wc in hexadecimal:
lmcwin read /top/u1/wc -h
lmcwin write
The lmcwin write command writes a value into a window. The format of the value argument is the same as used in other simulator commands that take value arguments. For example, to write 1 to window wb, and all 1s to window wc:
lmcwin write /top/u1/wb 1 lmcwin write /top/u1/wc X"FFFFFFFFFFFFFFFF"
lmcwin enable
The lmcwin enable command enables continuous monitoring of a window. The specified window is added to the model instance as a signal (with the same name as the window) of type std_logic or std_logic_vector. This signal's values can then be referenced in simulator commands that read signal values, such as the add list command (CR-44) shown below. The window signal is continuously updated to reflect the value in the model. For example, to list window wa:
lmcwin enable /top/u1/wa add list /top/u1/wa
lmcwin disable
The lmcwin disable command disables continuous monitoring of a window. The window signal is not deleted, but it no longer is updated when the models window register changes value. For example, to disable continuous monitoring of window wa:
lmcwin disable /top/u1/wa
lmcwin release
Some windows are actually nets, and the lmcwin write command behaves more like a continuous force on the net. The lmcwin release command disables the effect of a previous lmcwin write command on a window net.
Memory arrays
A memory model usually makes the entire register array available as a window. In this case, the window commands operate only on a single element at a time. The element is selected as an array reference in the window instance specification. For example, to read element 5 from the window memory mem:
lmcwin read /top/u2/mem(5)
Omitting the element specification defaults to element 0. Also, continuous monitoring is limited to a single array element. The associated window signal is updated with the most recently enabled element for continuous monitoring.
Verilog SmartModel interface UM-539
Verilog SmartModel interface

The SWIFT SmartModel library, beginning with release r40b, provides an optional library of Verilog modules and a PLI application that communicates between a simulator's PLI and the SWIFT simulator interface. The Logic Modeling documentation refers to this as the Logic Models to Verilog (LMTV) interface. To install this option, you must select the simulator type "Verilog" when you run Logic Modelings SmartInstall program.
Linking the LMTV interface to the simulator

Synopsys provides a dynamically loadable library that links ModelSim to the LMTV interface. See chapter 5, "Using MTI Verilog with Synopsys Models," in the "Simulator Configuration Guide for Synopsys Models" (available on Synopsys web site) for directions on how to link to this library.
UM-541
License Agreement
End-User License Agreement

IMPORTANT - USE OF THIS SOFTWARE IS SUBJECT TO LICENSE RESTRICTIONS. CAREFULLY READ THIS LICENSE AGREEMENT BEFORE USING THE SOFTWARE. This license is a legal Agreement concerning the use of Software between you, the end user, either individually or as an authorized representative of the company acquiring the license, and Mentor Graphics Corporation and Mentor Graphics (Ireland) Limited acting directly or through their subsidiaries or authorized distributors (collectively Mentor Graphics). USE OF SOFTWARE INDICATES YOUR COMPLETE AND UNCONDITIONAL ACCEPTANCE OF THE TERMS AND CONDITIONS SET FORTH IN THIS AGREEMENT. If you do not agree to these terms and conditions, promptly return, or, if received electronically, certify destruction of Software and all accompanying items within five days after receipt of Software and receive a full refund of any license fee paid. END-USER LICENSE AGREEMENT 1. GRANT OF LICENSE. The software programs you are installing, downloading, or have acquired with this Agreement, including any updates, modifications, revisions, copies, documentation and design data (Software) are copyrighted, trade secret and confidential information of Mentor Graphics or its licensors who maintain exclusive title to all Software and retain all rights not expressly granted by this Agreement. Mentor Graphics grants to you, subject to payment of appropriate license fees, a nontransferable, nonexclusive license to use Software solely: (a) in machine-readable, object-code form; (b) for your internal business purposes; and (c) on the computer hardware or at the site for which an applicable license fee is paid, or as authorized by Mentor Graphics. A site is restricted to a one-half mile (800 meter) radius. Mentor Graphics standard policies and programs, which vary depending on Software, license fees paid or service plan purchased, apply to the following and are subject to change: (a) relocation of Software; (b) use of Software, which may be limited, for example, to execution of a single session by a single user on the authorized hardware or for a restricted period of time (such limitations may be communicated and technically implemented through the use of authorization codes or similar devices); (c) support services provided, including eligibility to receive telephone support, updates, modifications, and revisions. Current standard policies and programs are available upon request. ESD SOFTWARE. If you purchased a license to use embedded software development (ESD) Software, Mentor Graphics grants to you a nontransferable, nonexclusive license to reproduce and distribute executable files created using ESD compilers, including the ESD run-time libraries distributed with ESD C and C++ compiler Software that are linked into a composite program as an integral part of your compiled computer program, provided that you distribute these files only in conjunction with your compiled computer program. Mentor Graphics does NOT grant you any right to duplicate or incorporate copies of Mentor Graphics' real-time operating systems or other ESD Software, except those explicitly granted in this section, into your products without first signing a separate agreement with Mentor Graphics for such purpose. BETA CODE. Portions or all of certain Software may contain code for experimental testing and evaluation (Beta Code), which may not be used without Mentor Graphics explicit authorization. Upon Mentor Graphics authorization, Mentor Graphics grants to you a temporary, nontransferable, nonexclusive license for experimental use to test and evaluate the Beta Code without charge for a limited period of time specified by Mentor
2.
3.
UM-542
License Agreement
Graphics. This grant and your use of the Beta Code shall not be construed as marketing or offering to sell a license to the Beta Code, which Mentor Graphics may choose not to release commercially in any form. If Mentor Graphics authorizes you to use the Beta Code, you agree to evaluate and test the Beta Code under normal conditions as directed by Mentor Graphics. You will contact Mentor Graphics periodically during your use of the Beta Code to discuss any malfunctions or suggested improvements. Upon completion of your evaluation and testing, you will send to Mentor Graphics a written evaluation of the Beta Code, including its strengths, weaknesses and recommended improvements. You agree that any written evaluations and all inventions, product improvements, modifications or developments that Mentor Graphics conceived or made during or subsequent to this Agreement, including those based partly or wholly on your feedback, will be the exclusive property of Mentor Graphics. Mentor Graphics will have exclusive rights, title and interest in all such property. The provisions of this subsection shall survive termination or expiration of this Agreement. 4. RESTRICTIONS ON USE. You may copy Software only as reasonably necessary to support the authorized use. Each copy must include all notices and legends embedded in Software and affixed to its medium and container as received from Mentor Graphics. All copies shall remain the property of Mentor Graphics or its licensors. You shall maintain a record of the number and primary location of all copies of Software, including copies merged with other software, and shall make those records available to Mentor Graphics upon request. You shall not make Software available in any form to any person other than employees and contractors, excluding Mentor Graphics' competitors, whose job performance requires access. You shall take appropriate action to protect the confidentiality of Software and ensure that any person permitted access to Software does not disclose it or use it except as permitted by this Agreement. Except as otherwise permitted for purposes of interoperability as specified by applicable and mandatory local law, you shall not reverse-assemble, reverse-compile, reverse-engineer or in any way derive from Software any source code. You may not sublicense, assign or otherwise transfer Software, this Agreement or the rights under it, whether by operation of law or otherwise (attempted transfer), without Mentor Graphics prior written consent and payment of Mentor Graphics then-current applicable transfer charges. Any attempted transfer without Mentor Graphics' prior written consent shall be a material breach of this Agreement and may, at Mentor Graphics' option, result in the immediate termination of the Agreement and licenses granted under this Agreement. The terms of this Agreement, including without limitation, the licensing and assignment provisions shall be binding upon your heirs, successors in interest and assigns. The provisions of this section 4 shall survive the termination or expiration of this Agreement. 5. LIMITED WARRANTY. 5.1. Mentor Graphics warrants that during the warranty period Software, when properly installed, will substantially conform to the functional specifications set forth in the applicable user manual. Mentor Graphics does not warrant that Software will meet your requirements or that operation of Software will be uninterrupted or error free. The warranty period is 90 days starting on the 15th day after delivery or upon installation, whichever first occurs. You must notify Mentor Graphics in writing of any nonconformity within the warranty period. This warranty shall not be valid if Software has been subject to misuse, unauthorized modification or installation. MENTOR GRAPHICS' ENTIRE LIABILITY AND YOUR EXCLUSIVE REMEDY SHALL BE, AT MENTOR GRAPHICS' OPTION, EITHER (A) REFUND OF THE PRICE PAID UPON RETURN OF SOFTWARE TO MENTOR GRAPHICS OR (B) MODIFICATION OR REPLACEMENT OF SOFTWARE THAT DOES NOT MEET THIS LIMITED WARRANTY, PROVIDED YOU HAVE OTHERWISE
UM-543
License Agreement
COMPLIED WITH THIS AGREEMENT. MENTOR GRAPHICS MAKES NO WARRANTIES WITH RESPECT TO: (A) SERVICES; (B) SOFTWARE WHICH IS LICENSED TO YOU FOR A LIMITED TERM OR LICENSED AT NO COST; OR (C) EXPERIMENTAL BETA CODE; ALL OF WHICH ARE PROVIDED AS IS. 5.2. THE WARRANTIES SET FORTH IN THIS SECTION 5 ARE EXCLUSIVE. NEITHER MENTOR GRAPHICS NOR ITS LICENSORS MAKE ANY OTHER WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, WITH RESPECT TO SOFTWARE OR OTHER MATERIAL PROVIDED UNDER THIS AGREEMENT. MENTOR GRAPHICS AND ITS LICENSORS SPECIFICALLY DISCLAIM ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT OF INTELLECTUAL PROPERTY. 6. LIMITATION OF LIABILITY. EXCEPT WHERE THIS EXCLUSION OR RESTRICTION OF LIABILITY WOULD BE VOID OR INEFFECTIVE UNDER APPLICABLE LAW, IN NO EVENT SHALL MENTOR GRAPHICS OR ITS LICENSORS BE LIABLE FOR I NDIRECT, SPECIAL, I NCIDENTAL, OR CONSEQUENTIAL DAMAGES (INCLUDING LOST PROFITS OR SAVINGS) WHETHER BASED ON CONTRACT, TORT OR ANY OTHER LEGAL THEORY, EVEN IF MENTOR GRAPHICS OR ITS LICENSORS HAVE BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. IN NO EVENT SHALL MENTOR GRAPHICS' OR ITS LICENSORS' LIABILITY UNDER THIS AGREEMENT EXCEED THE AMOUNT PAID BY YOU FOR THE SOFTWARE OR SERVICE GIVING RISE TO THE CLAIM. IN THE CASE WHERE NO AMOUNT WAS PAID, MENTOR GRAPHICS AND ITS LICENSORS SHALL HAVE NO LIABILITY FOR ANY DAMAGES WHATSOEVER. LIFE ENDANGERING ACTIVITIES. NEITHER MENTOR GRAPHICS NOR ITS LICENSORS SHALL BE LIABLE FOR ANY DAMAGES RESULTING FROM OR IN CONNECTION WITH THE USE OF SOFTWARE IN ANY APPLICATION WHERE THE FAILURE OR INACCURACY OF THE SOFTWARE MIGHT RESULT IN DEATH OR PERSONAL INJURY. INDEMNIFICATION. YOU AGREE TO INDEMNIFY AND HOLD HARMLESS MENTOR GRAPHICS AND ITS LICENSORS FROM ANY CLAIMS, LOSS, COST, DAMAGE, EXPENSE, OR LIABILITY, INCLUDING ATTORNEYS' FEES, ARISING OUT OF OR IN CONNECTION WITH YOUR USE OF SOFTWARE AS DESCRIBED IN SECTION 7. INFRINGEMENT. 9.1. Mentor Graphics will defend or settle, at its option and expense, any action brought against you alleging that Software infringes a patent or copyright or misappropriates a trade secret in the United States, Canada, Japan, or member state of the European Patent Office. Mentor Graphics will pay any costs and damages finally awarded against you that are attributable to the infringement action. You understand and agree that as conditions to Mentor Graphics' obligations under this section you must: (a) notify Mentor Graphics promptly in writing of the action; (b) provide Mentor Graphics all reasonable information and assistance to defend or settle the action; and (c) grant Mentor Graphics sole authority and control of the defense or settlement of the action. 9.2. If an infringement claim is made, Mentor Graphics may, at its option and expense: (a) replace or modify Software so that it becomes noninfringing; (b) procure for you
7.
8.
9.
UM-544
License Agreement
the right to continue using Software; or (c) require the return of Software and refund to you any license fee paid, less a reasonable allowance for use. 9.3. Mentor Graphics has no liability to you if infringement is based upon: (a) the combination of Software with any product not furnished by Mentor Graphics; (b) the modification of Software other than by Mentor Graphics; (c) the use of other than a current unaltered release of Software; (d) the use of Software as part of an infringing process; (e) a product that you make, use or sell; (f) any Beta Code contained in Software; (g) any Software provided by Mentor Graphics licensors who do not provide such indemnification to Mentor Graphics customers; or (h) infringement by you that is deemed willful. In the case of (h) you shall reimburse Mentor Graphics for its attorney fees and other costs related to the action upon a final judgment. 9.4. THIS SECTION 9 STATES THE ENTIRE LIABILITY OF MENTOR GRAPHICS AND ITS LICENSORS AND YOUR SOLE AND EXCLUSIVE REMEDY WITH RESPECT TO ANY ALLEGED PATENT OR COPYRIGHT INFRINGEMENT OR TRADE SECRET MISAPPROPRIATION BY ANY SOFTWARE LICENSED UNDER THIS AGREEMENT. 10. TERM. This Agreement remains effective until expiration or termination. This Agreement will immediately terminate upon notice if you exceed the scope of license granted or otherwise fail to comply with the provisions of Sections 1, 2, or 4. For any other material breach under this Agreement, Mentor Graphics may terminate this Agreement upon 30 days written notice if you are in material breach and fail to cure such breach within the 30-day notice period. If Software was provided for limited term use, this Agreement will automatically expire at the end of the authorized term. Upon any termination or expiration, you agree to cease all use of Software and return it to Mentor Graphics or certify deletion and destruction of Software, including all copies, to Mentor Graphics reasonable satisfaction. 11. EXPORT. Software is subject to regulation by local laws and United States government agencies, which prohibit export or diversion of certain products, information about the products, and direct products of the products to certain countries and certain persons. You agree that you will not export any Software or direct product of Software in any manner without first obtaining all necessary approval from appropriate local and United States government agencies. 12. RESTRICTED RIGHTS NOTICE. Software was developed entirely at private expense and is commercial computer software provided with RESTRICTED RIGHTS. Use, duplication or disclosure by the U.S. Government or a U.S. Government subcontractor is subject to the restrictions set forth in the license agreement under which Software was obtained pursuant to DFARS 227.7202-3(a) or as set forth in subparagraphs (c)(1) and (2) of the Commercial Computer Software - Restricted Rights clause at FAR 52.227-19, as applicable. Contractor/manufacturer is Mentor Graphics Corporation, 8005 SW Boeckman Road, Wilsonville, Oregon 97070-7777 USA. 13. THIRD PARTY BENEFICIARY. For any Software under this Agreement licensed by Mentor Graphics from Microsoft or other licensors, Microsoft or the applicable licensor is a third party beneficiary of this Agreement with the right to enforce the obligations set forth herein. 14. AUDIT RIGHTS. With reasonable prior notice, Mentor Graphics shall have the right to audit during your normal business hours all records and accounts as may contain information regarding your compliance with the terms of this Agreement. Mentor Graphics shall keep in confidence all information gained as a result of any audit. Mentor
UM-545
License Agreement
Graphics shall only use or disclose such information as necessary to enforce its rights under this Agreement. 15. CONTROLLING LAW AND JURISDICTION. THIS AGREEMENT SHALL BE GOVERNED BY AND CONSTRUED UNDER THE LAWS OF THE STATE OF OREGON, USA, IF YOU ARE LOCATED IN NORTH OR SOUTH AMERICA, AND THE LAWS OF IRELAND IF YOU ARE LOCATED OUTSIDE OF NORTH AND SOUTH AMERICA. All disputes arising out of or in relation to this Agreement shall be submitted to the exclusive jurisdiction of Dublin, Ireland when the laws of Ireland apply, or Wilsonville, Oregon when the laws of Oregon apply. This section shall not restrict Mentor Graphics right to bring an action against you in the jurisdiction where your place of business is located. The United Nations Convention on Contracts for the International Sale of Goods does not apply to this Agreement. 16. SEVERABILITY. If any provision of this Agreement is held by a court of competent jurisdiction to be void, invalid, unenforceable or illegal, such provision shall be severed from this Agreement and the remaining provisions will remain in full force and effect. 17. PAYMENT TERMS AND MISCELLANEOUS. You will pay amounts invoiced, in the currency specified on the applicable invoice, within 30 days from the date of such invoice. This Agreement contains the parties' entire understanding relating to its subject matter and supersedes all prior or contemporaneous agreements, including but not limited to any purchase order terms and conditions, except valid license agreements related to the subject matter of this Agreement (which are physically signed by you and an authorized agent of Mentor Graphics) either referenced in the purchase order or otherwise governing this subject matter. This Agreement may only be modified in writing by authorized representatives of the parties. Waiver of terms or excuse of breach must be in writing and shall not constitute subsequent consent, waiver or excuse. The prevailing party in any legal action regarding the subject matter of this Agreement shall be entitled to recover, in addition to other relief, reasonable attorneys' fees and expenses. Rev. 040401, Part Number 221417
UM-546
Index
CR = Command Reference, UM = Users Manual, GR = GUI Reference add wave command CR-49 add_cmdhelp command CR-53 aggregates, SystemC UM-159 alias command CR-54 analog signal formatting CR-50, GR-221 supported signal types GR-221 annotating differences, wave compare UM-268 annotating interconnect delays, v2k_int_delays CR-319 api_version in error message UM-167 architecture simulator state variable UM-455 archives described UM-55 archives, library CR-292 argc simulator state variable UM-455 argument CR-296 arguments passing to a DO file UM-429 arguments, accessing commandl-line UM-163 arithmetic package warnings, disabling UM-452 array of sc_signal<T> UM-159 arrays indexes CR-12 slices CR-13, CR-16 AssertFile .ini file variable UM-444 AssertionFormat .ini file variable UM-444 AssertionFormatBreak .ini file variable UM-444 AssertionFormatError .ini file variable UM-444 AssertionFormatFail .ini file variable UM-444 AssertionFormatFatal .ini file variable UM-445 AssertionFormatNote .ini file variable UM-445 AssertionFormatWarning .ini file variable UM-445 assertions configuring from the GUI GR-86 file and line number UM-444 message display GR-86 messages turning off UM-452 setting format of messages UM-444 testing for with onbreak command CR-170 warnings, locating UM-444 attributes, of signals, using in expressions CR-25 auto find bp command UM-326 auto step mode, C Debug UM-327
Symbols
#, comment character UM-418 $disable_signal_spy UM-369 $enable_signal_spy UM-370 +typdelays CR-299 .so, shared object file loading PLI/VPI C applications UM-484 loading PLI/VPI C++ applications UM-491 {} CR-16 hasX, hasX CR-25
Numerics
0-In tools setting environment variable UM-434 1076, IEEE Std UM-27 differences between versions UM-72 1364, IEEE Std UM-27, UM-102 2001, keywords, disabling CR-299 64-bit libraries UM-62 64-bit time now variable UM-456 Tcl time commands UM-423 64-bit vsim, using with 32-bit FLI apps UM-513
A
abort command CR-42 absolute time, using @ CR-19 ACC routines UM-508 accelerated packages UM-61 access hierarchical objects UM-355 limitations in mixed designs UM-173 Active Processes pane GR-107 see also windows, Active Processes pane Add file to Project dialog GR-51 Add Folder dialog GR-53 add list command CR-44 add log command CR-148 add memory command CR-47 add watch command CR-48
Index
B
bad magic number error message UM-213 balloon dialog, toggling on/off GR-238 balloon popup C Debug GR-100 base (radix) List window UM-247 Memory window GR-166 Wave window UM-243 batch_mode command CR-55 batch-mode simulations UM-26 halting CR-330 bd (breakpoint delete) command CR-56 binary radix, mapping to std_logic values CR-30 BindAtCompile .ini file variable UM-441 binding, VHDL, default UM-76 bitwise format UM-267 blocking assignments UM-119 bookmark add wave command CR-57 bookmark delete wave command CR-58 bookmark goto wave command CR-59 bookmark list wave command CR-60 bookmarks Source window GR-187 Wave window UM-238 bp (breakpoint) command CR-61 brackets, escaping CR-16 break on signal value CR-327 stop simulation run GR-41 BreakOnAssertion .ini file variable UM-445 breakpoints C code UM-323 conditional CR-327 continuing simulation after CR-197 deleting CR-56, GR-186, GR-240 listing CR-61 setting CR-61, GR-186 setting automatically in C code UM-327 signal breakpoints (when statements) CR-327 Source window, viewing in GR-182 time-based in when statements CR-331 .bsm file UM-287 buffered/unbuffered output UM-448 busses escape characters in CR-16 RTL-level, reconstructing UM-223
user-defined CR-50, UM-252 buswise format UM-267
C
C applications compiling and linking UM-484 debugging UM-319 C callstack moving down CR-186 moving up CR-174 C Debug UM-319 auto find bp UM-326 auto step mode UM-327 debugging functions during elaboration UM-330 debugging functions when exiting UM-334 function entry points, finding UM-326 initialization mode UM-330 menu reference GR-36, GR-37 registered function calls, identifying UM-327 running from a DO file UM-322 Stop on quit mode UM-334 C Debug setup dialog GR-100 C debugging CR-65 C++ applications compiling and linking UM-491 cancelling scheduled events, performance UM-98 case choice, must be locally static CR-249 case sensitivity named port associations UM-192 VHDL vs. Verilog CR-16 causality, tracing in Dataflow window UM-280 cd (change directory) command CR-64 cdbg command CR-65 cdbg_wait_for_starting command UM-322 cell libraries UM-123 cells hiding in Dataflow window GR-133, GR-134 change command CR-67 change directory, disabled GR-24 Change Memory dialog GR-162 Change Selected Variable dialog GR-149 chasing X UM-281 -check_synthesis argument CR-247 warning message UM-464 CheckPlusargs .ini file variable (VLOG) UM-445 CheckpointCompressMode .ini file variable UM-445 CheckSynthesis .ini file variable UM-441 class member selection, syntax CR-13
Index
class of sc_signal<T> UM-159 cleanup SystemC state-based code UM-154 clean-up of SystemC state-based code UM-154 clock change, sampling signals at UM-256 clocked comparison UM-264 Code Coverage $coverage_save system function UM-128 by instance UM-292 columns in workspace GR-109 condition coverage UM-292, UM-316 coverage clear command CR-102 coverage exclude command CR-103 coverage reload command CR-105 coverage report command CR-106 coverage save command CR-110 Current Exclusions pane GR-114 data types supported UM-293 Details pane GR-116 display filter toolbar GR-120 enabling with vcom or vlog UM-295 enabling with vsim UM-296 excluding lines/files UM-305 exclusion filter files UM-306 expression coverage UM-292, UM-317 important notes UM-294 Instance Coverage pane GR-115 Main window coverage data UM-297 merge utility UM-315 merging report files CR-105 merging reports CR-255 missed branches GR-113 missed coverage GR-113 pragma exclusions UM-305 reports UM-309 Source window data UM-298 source window details GR-117 statistics in Main window UM-297 toggle coverage UM-292 excluding signals CR-218 toggle coverage in Signals window UM-300 toggle details GR-116 vcover report command CR-258 vcover utility UM-315 Workspace pane GR-109 code profiling UM-337 collapsing ports, and coverage reporting UM-302 collapsing time and delta steps UM-221 colorization, in Source window GR-188 columns
hide/showing in GUI GR-249 moving GR-249 sorting by GR-249 Combine Selected Signals dialog GR-144 combining signals, busses CR-50, UM-252 command history GR-33 command line args, accessing vsim sc_arg command CR-320 CommandHistory .ini file variable UM-445 command-line arguments, accessing UM-163 command-line mode UM-25 commands CR-33CR-350 abort CR-42 add list CR-44 add memory CR-47 add watch CR-48 add wave CR-49 alias CR-54 batch_mode CR-55 bd (breakpoint delete) CR-56 bookmark add wave CR-57 bookmark delete wave CR-58 bookmark goto wave CR-59 bookmark list wave CR-60 bp (breakpoint) CR-61 cd (change directory) CR-64 cdbg CR-65 change CR-67 compare add CR-69 compare annotate CR-73, CR-76 compare clock CR-74 compare close CR-80 compare delete CR-79 compare info CR-81 compare list CR-82 compare open CR-94 compare options CR-83 compare reload CR-87 compare savediffs CR-90 compare saverules CR-91 compare see CR-92 compare start CR-89 configure CR-98 coverage clear CR-102 coverage exclude CR-103 coverage reload CR-105 coverage report CR-106 coverage save CR-110 dataset alias CR-111 dataset clear CR-112
Index
dataset close CR-113 dataset config CR-114 dataset info CR-115 dataset list CR-116 dataset open CR-117 dataset rename CR-118, CR-119 dataset snapshot CR-120 delete CR-122 describe CR-123 disablebp CR-124 do CR-125 drivers CR-126 dumplog64 CR-127 echo CR-128 edit CR-129 enablebp CR-130 environment CR-131 event watching in DO file UM-429 examine CR-132 exit CR-136 find CR-137 force CR-141 gdb dir CR-144 help CR-145 history CR-146 layout CR-147 log CR-148 lshift CR-150 lsublist CR-151 mem compare CR-152 mem display CR-153 mem list CR-155 mem load CR-156 mem save CR-159 mem search CR-161 noforce CR-164 nolog CR-165 notation conventions CR-10 notepad CR-167 noview CR-168 nowhen CR-169 onbreak CR-170 onElabError CR-171 onerror CR-172 pause CR-173 pop CR-174 printenv CR-175, CR-176 profile clear CR-177 profile interval CR-178 profile off CR-179
profile on CR-180 profile option CR-181 profile reload CR-182 profile report CR-183 push CR-186 pwd CR-187 quietly CR-188 quit CR-189 radix CR-190 readers CR-191 report CR-192 restart CR-194 resume CR-196 run CR-197 sccom CR-199 scgenmod CR-203 searchlog CR-205 setenv CR-207 shift CR-208 show CR-209 status CR-212 step CR-213 stop CR-214 system UM-421 tb (traceback) CR-215 toggle add CR-216 toggle disable CR-218 toggle enable CR-219 toggle report CR-220 toggle reset CR-222 transcript CR-223 transcript file CR-224 TreeUpdate CR-342 tssi2mti CR-225 unsetenv CR-226 variables referenced in CR-18 vcd add CR-227 vcd checkpoint CR-228 vcd comment CR-229 vcd dumpports CR-230 vcd dumpportsall CR-232 vcd dumpportsflush CR-233 vcd dumpportslimit CR-234 vcd dumpportsoff CR-235 vcd dumpportson CR-236 vcd file CR-237 vcd files CR-239 vcd flush CR-241 vcd limit CR-242 vcd off CR-243
Index
vcd on CR-244 vcom CR-246 vcover convert CR-254 vcover merge CR-255 vcover rank CR-257 vcover report CR-258 vdel CR-263 vdir CR-264 verror CR-265 vgencomp CR-266 view CR-268 virtual count CR-270 virtual define CR-271 virtual delete CR-272 virtual describe CR-273 virtual expand CR-274 virtual function CR-275 virtual hide CR-278 virtual log CR-279 virtual nohide CR-281 virtual nolog CR-282 virtual region CR-284 virtual save CR-285 virtual show CR-286 virtual signal CR-287 virtual type CR-290 vlib CR-292 vlog CR-293 vmake CR-302 vmap CR-304 vsim CR-305 VSIM Tcl commands UM-422 vsimDate CR-322 vsimId CR-322 vsimVersion CR-322 wave CR-324 WaveActivateNextPane CR-342 WaveRestoreCursors CR-342 WaveRestoreZoom CR-342 when CR-327 where CR-332 wlf2log CR-333 wlf2vcd CR-335 wlfman CR-336 wlfrecover CR-340 write format CR-341 write list CR-343 write preferences CR-344 write report CR-345 write timing CR-346
write transcript CR-347 write tssi CR-348 write wave CR-350 comment character Tcl and DO files UM-418 comment characters in VSIM commands CR-10 compare add region UM-263 add signals UM-262 by signal UM-262 clocked UM-264 difference markers UM-267 displayed in List window UM-269 icons UM-269 method UM-264 options UM-266 pathnames UM-267 reference dataset UM-260 reference region UM-263 tab UM-261 test dataset UM-261 timing differences UM-267 tolerance UM-264 values UM-268 wave window display UM-267 compare add command CR-69 compare annotate command CR-73, CR-76 compare by region UM-263 compare clock command CR-74 compare close command CR-80 compare delete command CR-79 compare info command CR-81 compare list command CR-82 Compare Memory dialog GR-164 compare open command CR-94 compare options command CR-83 compare reload command CR-87 compare savediffs command CR-90 compare saverules command CR-91 compare see command CR-92 compare signal, virtual restrictions UM-252 compare simulations UM-211 compare start command CR-89 compatibility, of vendor libraries CR-264 compile order auto generate UM-44 changing UM-44 SystemVerilog packages UM-105 Compile Order dialog GR-75
Index
Compile Source Files dialog dialogs Compile Source Files GR-65 compiler directives UM-132 IEEE Std 1364-2000 UM-132 XL compatible compiler directives UM-134 Compiler Options dialog GR-66 compiling overview UM-23 changing order in the GUI UM-44 gensrc errors during UM-168 graphic interface to GR-65 grouping files UM-45 order, changing in projects UM-44 properties, in projects UM-50 range checking in VHDL CR-251, UM-71 SystemC CR-199, CR-203, UM-143 converting sc_main() UM-143 exporting top level module UM-144 for source level debug UM-146 invoking sccom UM-146 linking the compiled source UM-151 modifying source code UM-143 replacing sc_start() UM-143 using sccom vs. raw C++ compiler UM-149 Verilog CR-293, UM-103 incremental compilation UM-105 XL uselib compiler directive UM-110 XL compatible options UM-109 VHDL CR-246, UM-70 at a specified line number CR-248 selected design units (-just eapbc) CR-248 standard package (-s) CR-251 VITAL packages UM-86 compiling C code, gcc UM-485 component declaration generating SystemC from Verilog or VHDL UM208 generating VHDL from Verilog UM-189 vgencomp for SystemC UM-208 vgencomp for VHDL UM-189 component, default binding rules UM-76 Compressing files VCD tasks UM-403 compressing files VCD files CR-230, CR-239 concatenation directives CR-29 of signals CR-28, CR-287 ConcurrentFileLimit .ini file variable UM-445
conditional breakpoints CR-327 configuration simulator state variable UM-455 configurations instantiation in mixed designs UM-188 Verilog UM-112 configurations, simulating CR-305 configure command CR-98 connectivity, exploring UM-277 constants in case statements CR-249 values of, displaying CR-123, CR-132 construction parameters, SystemC UM-164 context menu List window GR-138 context menus Library tab UM-57 context sensitivity UM-257 control function, SystemC UM-175 control_foreign_signal() function UM-173 conversion, radix CR-190 convert real to time UM-90 convert time to real UM-89 coverage merging data UM-314 saving raw data UM-314 see also Code Coverage setting default mode UM-310 coverage clear command CR-102 coverage exclude command CR-103 coverage reload command CR-105 coverage report command CR-106 Coverage Report dialog GR-90 coverage reports UM-309 reporting all signals UM-302 sample reports UM-311 xml format UM-310 coverage save command CR-110 $coverage_save system function UM-128 covreport.xsl UM-310 CppOptions .ini file variable (sccom) UM-442 CppPath .ini file variable (sccom) UM-442 Create a New Library dialog GR-45 Create Project dialog GR-44 Create Project File dialog GR-50 current exclusions pragmas UM-305 Current Exclusions pane GR-114 cursors adding, deleting, locking, naming UM-233 link to Dataflow window UM-276
Index
measuring time with UM-233 trace events with UM-280 Wave window UM-233 customizing via preference variables GR-251
D
deltas explained UM-77 data types Code Coverage UM-293 Dataflow Options dialog GR-133 Dataflow Page Setup dialog GR-131 Dataflow window UM-274, GR-121 automatic cell hiding GR-133, GR-134 menu bar GR-122 options GR-133, GR-134 pan UM-279 zoom UM-279 see also windows, Dataflow window dataflow.bsm file UM-287 dataset alias command CR-111 Dataset Browser UM-218, GR-55 dialog GR-55 dataset clear command CR-112 dataset close command CR-113 dataset config command CR-114 dataset info command CR-115 dataset list command CR-116 dataset open command CR-117 dataset rename command CR-118, CR-119 Dataset Snapshot UM-220 dataset snapshot command CR-120 datasets UM-211 environment command, specifying with CR-131 managing UM-218 openingdialogs Open File GR-46 reference UM-260 restrict dataset prefix display UM-219 test UM-261 DatasetSeparator .ini file variable UM-445 debuggable SystemC objects UM-155 debugging C code UM-319 debugging the design, overview UM-24 declarations, hiding implicit with explicit CR-252 default binding
BindAtCompile .ini file variable UM-441 disabling UM-76 default binding rules UM-76 default coverage mode, setting UM-310 Default editor, changing UM-434 default SystemC parameter values, overriding UM-164 DefaultForceKind .ini file variable UM-446 DefaultRadix .ini file variable UM-446 DefaultRestartOptions variable UM-446, UM-453 +define+ CR-294 Define Clock dialog GR-171 delay delta delays UM-77 interconnect CR-309 modes for Verilog models UM-123 SDF files UM-381 stimulus delay, specifying GR-170 +delay_mode_distributed CR-294 +delay_mode_path CR-294 +delay_mode_unit CR-294 +delay_mode_zero CR-294 delayed CR-25 DelayFileOpen .ini file variable UM-446 delaying test signal, Waveform Comparison GR-226 delete command CR-122 deleting library contents UM-57 delta collapsing UM-221 delta simulator state variable UM-455 deltas collapsing in the List window GR-146 collapsing in WLF files CR-313 hiding in the List window CR-99, GR-146 in List window UM-253 referencing simulator iteration as a simulator state variable UM-455 dependencies, checking CR-264 dependent design units UM-70 describe command CR-123 descriptions of HDL items GR-186 design library creating UM-56 logical name, assigning UM-58 mapping search rules UM-59 resource type UM-54 VHDL design units UM-70 working type UM-54 design loading, interrupting CR-305 design object icons, described GR-14 design portability and SystemC UM-147 design units UM-54
Index
report of units simulated CR-345 Verilog adding to a library CR-293 details code coverage GR-116 dialogs GR-55 Add file to Project GR-51 Add Folder GR-53 C Debug setup GR-100 Change Memory GR-162 Change Selected Variable GR-149 Combine Selected Signals GR-144 Compare Memory GR-164 Compile Order GR-75 Compiler Options GR-66 Coverage Report GR-90 Create a New Library GR-45 Create Project GR-44 Create Project File GR-50 Dataflow Options GR-133 Dataflow Page Setup GR-131 Define Clock GR-171 Export Memory GR-160 File Breakpoint GR-99 Filter instance list GR-93 Find in dataflow GR-132 Find in List GR-139 Find in Locals GR-150 Find in memory GR-165 Find in Process GR-108 Force Selected Signal GR-169 List Signal Properties GR-142 List Signal Search GR-140 Load Coverage Data GR-89 Modify Breakpoints GR-96 Modify Display Properties GR-145 Preferences GR-103 Print GR-128 Print Postscript GR-130 Profile Report GR-94, GR-180 Project Compiler Settings GR-56 Project Settings GR-63 Properties (memory) GR-166 Restart GR-88 Runtime Options GR-85 Signal Breakpoints GR-98 Simulation Configuration GR-52 Start Simulation GR-76 SystemC Link dialog GR-74 directories
mapping libraries CR-304 moving libraries UM-59 directory, changing, disabled GR-24 disable_signal_spy UM-357 disablebp command CR-124 DisableOpt .ini file variable UM-439, UM-441 distributed delay mode UM-124 dividers adding from command line CR-49 Wave window UM-244 DLL files, loading UM-484, UM-491 do command CR-125 DO files (macros) CR-125 error handling UM-431 executing at startup UM-434, UM-448 parameters, passing to UM-429 Tcl source command UM-432 docking window panes GR-245 documentation UM-31 DOPATH environment variable UM-434 DPI export TFs UM-463 DPI export TFs UM-463 DPI use flow UM-481 drag & drop preferences GR-102 drivers Dataflow Window UM-277 show in Dataflow window UM-257 Wave window UM-257 drivers command CR-126 drivers, multiple on unresolved signal GR-59, GR-68 dump files, viewing in the simulator CR-245 dumplog64 command CR-127 dumpports tasks, VCD files UM-402
E
echo command CR-128 edit command CR-129 Editing in notepad windows UM-519 in the Main window UM-519 in the Source window UM-519 EDITOR environment variable UM-434 editor, default, changing UM-434 embedded wave viewer UM-278 empty port name warning UM-463 enable_signal_spy UM-358
Index
enablebp command CR-130 encryption +protect argument CR-298 protect compiler directive UM-133 -nodebug argument (vcom) CR-249 -nodebug argument (vlog) CR-296 securing pre-compiled libraries UM-66 end_of_construction() function UM-163 end_of_simulation() function UM-163 ENDFILE function UM-82 ENDLINE function UM-82 endprotect compiler directive UM-133 entities default binding rules UM-76 entities, specifying for simulation CR-320 entity simulator state variable UM-455 enumerated types user defined CR-290 environment command CR-131 environment variables UM-434 accessed during startup UM-527 reading into Verilog code CR-294 referencing from command line UM-436 referencing with VHDL FILE variable UM-436 setting in Windows UM-435 specifying library locations in modelsim.ini file UM-439 specifying UNIX editor CR-129 state of CR-176 TranscriptFile, specifying location of UM-448 used in Solaris linking for FLI UM-484, UM-491 using in pathnames CR-16 using with location mapping UM-63 variable substitution using Tcl UM-421 environment, displaying or changing pathname CR-131 error cant locate C compiler UM-463 Error .ini file variable UM-449 errors "api_version" in UM-167 bad magic number UM-213 getting details about messages CR-265 getting more information UM-458 libswift entry not found UM-467 multiple definition UM-168 onerror command CR-172 out-of-line function UM-168 SDF, disabling CR-311 severity level, changing UM-458 SystemC loading UM-166
SystemVerilog, missing declaration UM-440 Tcl_init error UM-464 void function UM-168 VSIM license lost UM-467 escape character CR-16 event order changing in Verilog CR-293 in Verilog simulation UM-117 event queues UM-117 event watching commands, placement of UM-429 events, tracing UM-280 examine command CR-132 examine tooltip toggling on/off GR-238 exclusion filter files UM-306 excluding udp truth table rows UM-307 exclusions lines and files UM-305 exit codes UM-461 exit command CR-136 expand net UM-277 Explicit .ini file variable UM-441 Export Memory dialog GR-160 export TFs, in DPI UM-463 Exporting SystemC modules to Verilog UM-199 exporting SystemC modules to VHDL UM-209 exporting top SystemC module UM-144 Expression Builder UM-241 configuring a List trigger with UM-254 saving expressions to Tcl variable UM-241 extended identifiers CR-17 in mixed designs UM-188, UM-208
F
-f CR-295 F8 function key UM-521 features, new UM-255 field descriptions coverage reports UM-311 FIFOs, viewing SystemC UM-161 File Breakpoint dialog GR-99 File compression VCD tasks UM-403 file compression SDF files UM-381 VCD files CR-230, CR-239
Index
file format MTI memory data GR-161 file I/O TextIO package UM-79 VCD files UM-397 file-line breakpoints GR-186 files opening in GUI GR-46 files, grouping for compile UM-45 filter processes GR-107 Filter instance list dialog GR-93 filtering signals in Objects window GR-168 filters for Code Coverage UM-306 find command CR-137 Find in dataflow dialog GR-132 Find in List dialog GR-139 Find in Locals dialog GR-150 Find in memory dialog GR-165 Find in Process dialog GR-108 Find in Transcript dialog dialogs Find in Transcript GR-54 fixed-point types, in SystemC compiling for support UM-162 construction parameters for UM-164 FLI debugging UM-319 folders, in projects UM-48 font scaling for dual monitors GR-33 fonts controlling in X-sessions GR-15 scaling GR-15 force command CR-141 defaults UM-453 Force Selected Signal dialog GR-169 foreign model loading SmartModels UM-532 foreign module declaration Verilog example CR-204, UM-195 VHDL example UM-204 foreign module declaration, SystemC UM-194 format file UM-249 List window CR-341 Wave window CR-341, UM-249 FPGA libraries, importing UM-65 function calls, identifying with C Debug UM-327
Functional coverage merging databases offline CR-255 functions SystemC control UM-175 observe UM-175 unsupported UM-162
G
-g C++ compiler option UM-157 g++, alternate installations UM-147 gdb setting source directory CR-144 gdb debugger UM-320 gdb dir command CR-144 generate statements, Veilog UM-113 GenerateFormat .ini file variable UM-446 GenerateLoopIterationMax .ini file variable UM-439 GenerateRecursionDepthMax .ini variable UM-439 generics assigning or overriding values with -g and -G CR307 examining generic values CR-132 limitation on assigning composite types CR-307 SystemC instantiating VHDL UM-204 VHDL UM-178 get_resolution() VHDL function UM-87 glitches disabling generation from command line CR-315 from GUI GR-78 global visibility PLI/FLI shared objects CR-308, UM-498 GlobalSharedObjectsList .ini file variable UM-446 graphic interface UM-225, UM-273, GR-11 grayed-out menu options UM-257 grouping files for compile UM-45 grouping objects, Monitor window GR-192 GUI preferences, saving GR-252 GUI_expression_format CR-23 GUI expression builder UM-241 syntax CR-24
Index
H
hasX CR-25 Hazard .ini file variable (VLOG) UM-439 hazards -hazards argument to vlog CR-295 -hazards argument to vsim CR-316 limitations on detection UM-120 help command CR-145 hierarchical references SystemC/HDL designs UM-175 hierarchical references, mixed-language UM-173 hierarchy driving signals in UM-359, UM-371 forcing signals in UM-88, UM-365, UM-377 referencing signals in UM-88, UM-362, UM-374 releasing signals in UM-88, UM-367, UM-379 viewing signal names without GR-237 highlighting, in Source window GR-188 history of commands shortcuts for reuse CR-20, UM-517 history command CR-146 HOME environment variable UM-434 HOME_0IN environment variable UM-434 HP aCC, restrictions on compiling with UM-148
I
I/O TextIO package UM-79 VCD files UM-397 icons shapes and meanings GR-14 ieee .ini file variable UM-439 IEEE libraries UM-61 IEEE Std 1076 UM-27 differences between versions UM-72 IEEE Std 1364 UM-27, UM-102 IgnoreError .ini file variable UM-446 IgnoreFailure .ini file variable UM-446 IgnoreNote .ini file variable UM-446 IgnoreVitalErrors .ini file variable UM-441 IgnoreWarning .ini file variable UM-446 implicit operator, hiding with vcom -explicit CR-252 importing FPGA libraries UM-65 importing memory patterns GR-157 +incdir+ CR-295 Incremental .ini file variable UM-440
incremental compilation automatic UM-106 manual UM-106 with Verilog UM-105 index checking UM-71 indexed arrays, escaping square brackets CR-16 $init_signal_driver UM-371 init_signal_driver UM-359 $init_signal_spy UM-374 init_signal_spy UM-88, UM-362 init_usertfs function UM-332, UM-476 initialization of SystemC state-based code UM-154 initialization sequence UM-529 inlining VHDL subprograms UM-71 instance code coverage UM-292 instantiation in mixed-language design Verilog from VHDL UM-188 VHDL from Verilog UM-192 instantiation in SystemC-Verilog design SystemC from Verilog UM-199 Verilog from SystemC UM-194 instantiation in SystemC-VHDL design VHDL from SystemC UM-202 instantiation in VHDL-SystemC design SystemC from VHDL UM-208 interconnect delays CR-309, UM-393 annotating per Verilog 2001 CR-319 internal signals, adding to a VCD file CR-227 interrupting design loading CR-305 IOPATH matching to specify path delays UM-387 iteration_limit, infinite zero-delay loops UM-78 IterationLimit .ini file variable UM-446
K
keyboard shortcuts List window UM-522 Main window UM-519 Source window UM-519 Wave window UM-523 keywords disabling 2001 keywords CR-299 enabling SystemVerilog keywords CR-298 SystemVerilog UM-103
Index
L
-L work UM-108 language templates GR-184 language versions, VHDL UM-72 layout command CR-147 libraries 64-bit and 32-bit in same library UM-62 archives CR-292 creating UM-56 dependencies, checking CR-264 design libraries, creating CR-292, UM-56 design library types UM-54 design units UM-54 group use, setting up UM-59 IEEE UM-61 importing FPGA libraries UM-65 including precompiled modules GR-80 listing contents CR-264 mapping from the command line UM-58 from the GUI UM-58 hierarchically UM-451 search rules UM-59 modelsim_lib UM-87 moving UM-59 multiple libraries with common modules UM-108 naming UM-58 predefined UM-60 refreshing library images CR-251, CR-298, UM-62 resource libraries UM-54 std library UM-60 Synopsys UM-61 vendor supplied, compatibility of CR-264 Verilog CR-316, UM-107, UM-177 VHDL library clause UM-60 working libraries UM-54 working vs resource UM-22 working with contents of UM-57 library map file, Verilog configurations UM-112 library mapping, overview UM-23 library maps, Verilog 2001 UM-112 library simulator state variable UM-455 library, definition UM-22 libsm UM-532 libswift UM-532 entry not found error UM-467 License .ini file variable UM-447 licensing
License variable in .ini file UM-447 linking SystemC source UM-151 lint-style checks CR-296 List pane see also pane, List pane List Signal Properties dialog GR-142 List Signal Search dialog GR-140 List window UM-231, GR-135 adding items to CR-44 context menu GR-138 GUI changes UM-265 setting triggers UM-254 waveform comparison UM-269 see also windows, List window LM_LICENSE_FILE environment variable UM-434 Load Coverage Data dialog GR-89 loading designs, interrupting CR-305 loading the design, overview UM-24 Locals window GR-148 see also windows, Locals window location maps, referencing source files UM-63 locations maps specifying source files with UM-63 lock message UM-463 locking cursors UM-233 log command CR-148 log file log command CR-148 nolog command CR-165 overview UM-211 QuickSim II format CR-333 redirecting with -l CR-308 virtual log command CR-279 virtual nolog command CR-282 see also WLF files Logic Modeling SmartModel command channel UM-536 SmartModel Windows lmcwin commands UM-537 memory arrays UM-538 long simulations saving at intervals UM-220 lshift command CR-150 lsublist command CR-151
Index
M
MacroNestingLevel simulator state variable UM-455 macros (DO files) UM-429 breakpoints, executing at CR-62 creating from a saved transcript GR-19 depth of nesting, simulator state variable UM-455 error handling UM-431 executing CR-125 forcing signals, nets, or registers CR-141 parameters as a simulator state variable (n) UM-455 passing CR-125, UM-429 total number passed UM-455 relative directories CR-125 shifting parameter values CR-208 startup macros UM-452 Main window GR-16 code coverage UM-297 GUI changes UM-256 see also windows, Main window manuals UM-31 mapping data types UM-176 libraries from the command line UM-58 hierarchically UM-451 symbols Dataflow window UM-287 SystemC in mixed designs UM-187 SystemC to Verilog UM-183 SystemC to VHDL UM-187 Verilog states in mixed designs UM-177 Verilog states in SystemC designs UM-182 Verilog to SytemC, port and data types UM-182 Verilog to VHDL data types UM-176 VHDL to SystemC UM-179 VHDL to Verilog data types UM-178 mapping libraries, library mapping UM-58 master slave library (SystemC), including CR-201 math_complex package UM-61 math_real package UM-61 +maxdelays CR-296 mc_scan_plusargs, PLI routine CR-318 MDI frame GR-20, UM-257 MDI pane tab groups GR-20 mem compare command CR-152 mem display command CR-153
mem list command CR-155 mem load command CR-156 mem save command CR-159 mem search command CR-161 memories displaying the contents of GR-151 initializing GR-157 loading memory patterns GR-157 MTI memory data file GR-161 navigating to memory locations GR-165 navigation GR-154 saving formats GR-153 saving memory data to a file GR-160 selecting memory instances GR-153 viewing contents GR-153 viewing multiple instances GR-153 memory modeling in VHDL UM-91 memory allocation profiler UM-338 Memory Declaration, View menu UM-273 memory leak, cancelling scheduled events UM-98 Memory pane GR-151 pane Memory pane see also Memory pane memory tab memories you can view GR-152 Memory window GR-151 GUI changes UM-270 modifying display GR-166 see also windows, Memory window memory window add memory command CR-47 adding items to CR-47 memory, comparing contents CR-152 memory, displaying contents CR-153 memory, listing CR-155 memory, loading contents CR-156 memory, saving contents CR-159 memory, searching for patterns CR-161 menu options grayed-out UM-257 menus Dataflow window GR-122 List window GR-137 Main window GR-23 Profiler windows GR-178 Source window GR-189 Wave window GR-199 merging coverage data UM-314 merging coverage reports CR-255
Index
message system UM-458 messages UM-457 bad magic number UM-213 echoing CR-128 empty port name warning UM-463 exit codes UM-461 getting more information CR-265, UM-458 loading, disabling with -quiet CR-298 loading, disbling with -quiet CR-251 lock message UM-463 long description UM-458 message system variables UM-449 metavalue detected UM-464 redirecting UM-448 sensitivity list warning UM-464 suppressing warnings from arithmetic packages UM-452 Tcl_init error UM-464 too few port connections UM-466 turning off assertion messages UM-452 VSIM license lost UM-467 warning, suppressing UM-460 metavalue detected warning UM-464 -mfcu CR-296 MGC_LOCATION_MAP env variable UM-63 MGC_LOCATION_MAP variable UM-434 +mindelays CR-296 MinGW gcc UM-485, UM-492 missed coverage branches GR-113 Missed Coverage pane GR-113 mixed-language simulation UM-172 access limitations UM-173 mnemonics, assigning to signal values CR-290 mode, setting default coverage UM-310 MODEL_TECH environment variable UM-434 MODEL_TECH_TCL environment variable UM-434 modeling memory in VHDL UM-91 MODELSIM environment variable UM-435 modelsim.ini found by the tool UM-529 default to VHDL93 UM-453 delay file opening with UM-453 environment variables in UM-451 force command default, setting UM-453 hierarchical library mapping UM-451 opening VHDL files UM-453 restart command defaults, setting UM-453 startup file, specifying with UM-452 transcript file created from UM-451
turning off arithmetic package warnings UM-452 turning off assertion messages UM-452 modelsim.tcl GR-253 modelsim_lib UM-87 path to UM-439 MODELSIM_PREFERENCES variable UM-435 MODELSIM_TCL environment variable UM-435 modes of operation UM-25 Modified field, Project tab UM-43 Modify Breakpoints dialog GR-96 Modify Display Properties dialog GR-145 modules handling multiple, common names UM-108 with unnamed ports UM-191 Monitor window grouping/ungrouping objects GR-192 monitor window GR-191 monitors, dual, font scaling GR-33 mouse shortcuts Main window UM-519 Source window UM-519 Wave window UM-523 .mpf file UM-36 loading from the command line UM-52 order of access during startup UM-526 MTI memory data file GR-161 mti_cosim_trace environment variable UM-435 mti_inhibit_inline attribute UM-71 MTI_SYSTEMC macro UM-147 MTI_TF_LIMIT environment variable UM-435 MultiFileCompilationUnit .ini file variable UM-440 multiple document interface GR-20, UM-257 multiple drivers on unresolved signal GR-59, GR-68 Multiple simulations UM-211 multi-source interconnect delays CR-309
N
n simulator state variable UM-455 name case sensitivity, VHDL vs. Verilog CR-16 Name field Project tab UM-43 name visibility in Verilog generates UM-113 names, modules with the same UM-108 negative pulses driving an error state CR-319 negative timing $setuphold/$recovery UM-130 algorithm for calculating delays UM-121
Index
check limits UM-121 extending check limits CR-316 nets Dataflow window, displaying in UM-274, GR-121 drivers of, displaying CR-126 readers of, displaying CR-191 stimulus CR-141 values of displaying in Objects window GR-167 examining CR-132 saving as binary log file UM-212 waveforms, viewing GR-194 new features UM-255 next and previous edges, finding UM-524 Nlview widget Symlib format UM-287 no space in time literal GR-59, GR-68 -no_risefall_delaynets CR-317 NoCaseStaticError .ini file variable UM-441 NoDebug .ini file variable (VCOM) UM-441 NoDebug .ini file variable (VLOG) UM-440 -nodebug argument (vcom) CR-249 -nodebug argument (vlog) CR-296 noforce command CR-164 NoIndexCheck .ini file variable UM-441 +nolibcell CR-297 nolog command CR-165 NOMMAP environment variable UM-435 non-blocking assignments UM-119 NoOthersStaticError .ini file variable UM-441 NoRangeCheck .ini file variable UM-441 Note .ini file variable UM-450 notepad command CR-167 Notepad windows, text editing UM-519 -notrigger argument UM-256 noview command CR-168 NoVital .ini file variable UM-441 NoVitalCheck .ini file variable UM-441 Now simulator state variable UM-455 now simulator state variable UM-455 +nowarn<CODE> CR-297 nowhen command CR-169 numeric_bit package UM-61 numeric_std package UM-61 disabling warning messages UM-452 NumericStdNoWarnings .ini file variable UM-447
O
object defined UM-30 object_list_file, WLF files CR-336 Objects window GR-167 see also windows, Objects window observe function, SystemC UM-175 observe_foreign_signal() function UM-173 onbreak command CR-170 onElabError command CR-171 onerror command CR-172 Open File dialog GR-46 opening files GR-46 operating systems supported, See Installation Guide optimizations disabling for Verilog designs CR-297 disabling for VHDL designs CR-250 disabling process merging CR-246 VHDL subprogram inlining UM-71 optimize for std_logic_1164 GR-59, GR-68 Optimize_1164 .ini file variable UM-441 OptionFile entry in project files GR-62, GR-71 order of events changing in Verilog CR-293 ordering files for compile UM-44 organizing projects with folders UM-48 organizing windows, MDI pane GR-20 OSCI simulator, differences with vsim UM-162 others .ini file variable UM-439 overview, simulation tasks UM-21
P
packages standard UM-60 textio UM-60 util UM-87 VITAL 1995 UM-84 VITAL 2000 UM-84 page setup Dataflow window UM-286 Wave window UM-250, GR-212 pan, Dataflow window UM-279 panes docking and undocking GR-245 Memory panes GR-151 parameter support SystemC instantiating Verilog UM-196
Index
Verilog instantiating SystemC UM-199 parameters making optional UM-430 using with macros CR-125, UM-429 path delay mode UM-124 path delays,matching to IOPATH statements UM-387 pathnames comparisons UM-267 hiding in Wave window UM-243 in VSIM commands CR-12 spaces in CR-11 PathSeparator .ini file variable UM-447 pause command CR-173 PedanticErrors .ini file variable UM-441 performance cancelling scheduled events UM-98 platforms supported, See Installation Guide PLI loading shared objects with global symbol visibility CR-308, UM-498 specifying which apps to load UM-477 Veriuser entry UM-477 PLI/VPI UM-135, UM-475 debugging UM-319 tracing UM-514 PLIOBJS environment variable UM-435, UM-477 pop command CR-174 popup toggling waveform popup on/off UM-268, GR-238 port collapsing, toggle coverage UM-302 Port driver data, capturing UM-408 ports, unnamed, in mixed designs UM-191 ports, VHDL and Verilog UM-176 Postscript saving a waveform in UM-250 saving the Dataflow display in UM-284 pragmas UM-305 precedence of variables UM-454 precision, simulator resolution UM-114, UM-174 PrefCoverage(DefaultCoverageMode) UM-310 PrefCoverage(pref_InitFilterFrom) UM-307 Preference dialog GR-103 preference variables .ini files, located in UM-438 editing GR-251 saving GR-251 preferences drag and drop GR-102 saving GR-251 PrefMain(EnableCommandHelp) GR-19
PrefMain(ShowFilePane) preference variable GR-18 PrefMemory(ExpandPackedMem) variable GR-152 primitives, symbols in Dataflow window UM-287 Print dialog GR-128 Print Postscript dialog GR-130 printenv command CR-175, CR-176 printing waveforms in the Wave window UM-250 processes optimizations, disabling merging CR-246 without wait statements GR-59, GR-68 profile clear command CR-177 profile interval command CR-178 profile off command CR-179 profile on command CR-180 profile option command CR-181 profile reload command CR-182 profile report command CR-183, UM-352 Profile Report dialog GR-94, GR-180 Profiler UM-337 %parent fields UM-345 clear profile data UM-341 enabling memory profiling UM-339 enabling statistical sampling UM-341 getting started UM-339 handling large files UM-340 Hierarchical View UM-345 interpreting data UM-343 memory allocation UM-338 memory allocation profiling UM-341 profile report command UM-352 Profile Report dialog UM-353, GR-94 Ranked View UM-344 report option UM-352 reporting GR-94 results, viewing UM-344 statistical sampling UM-338 Structural View UM-347 unsupported on Opteron UM-337 view_profile command UM-344 viewing profile details UM-348 Programming Language Interface UM-135, UM-475 Project Compiler Settings dialog GR-56 Project Settings dialog GR-63 project tab information in UM-43 sorting UM-43 projects UM-35 accessing from the command line UM-52 adding files to UM-39
Index
benefits UM-36 close UM-42 code coverage settings UM-295 compile order UM-44 changing UM-44 compiler properties in UM-50 compiling files UM-41 creating UM-38 creating simulation configurations UM-46 delete UM-42 folders in UM-48 grouping files in UM-45 loading a design UM-42 MODELSIM environment variable UM-435 open and existing UM-42 override mapping for work directory with vcom CR201, CR-252 override mapping for work directory with vlog CR299 overview UM-36 propagation, preventing X propagation CR-310 Properties (memory) dialog GR-166 Protect .ini file variable (VLOG) UM-440 protect compiler directive UM-133 protected types UM-91 pulse error state CR-319 push command CR-186 pwd command CR-187
Q
quick reference table of simulation tasks UM-21 QuickSim II logfile format CR-333 Quiet .ini file variable VCOM UM-441 Quiet .ini file variable (VLOG) UM-440 quietly command CR-188 quit command CR-189
R
race condition, problems with event order UM-117 radix changing in Objects, Locals, Dataflow, List, and Wave windows CR-190 character strings, displaying CR-290 default, DefaultRadix variable UM-446 List window UM-247
of signals being examined CR-133 of signals in Wave window CR-51 specifying in Memory window GR-166 Wave window UM-243 radix command CR-190 range checking UM-71 disabling CR-250 enabling CR-251 readers and drivers UM-277 readers command CR-191 real type, converting to time UM-90 rebuilding supplied libraries UM-61 reconstruct RTL-level design busses UM-223 record field selection, syntax CR-13 records, values of, changing GR-149 $recovery UM-130 redirecting messages, TranscriptFile UM-448 reference region UM-263 refreshing library images CR-251, CR-298, UM-62 registered function calls UM-327 registers values of displaying in Objects window GR-167 saving as binary log file UM-212 waveforms, viewing GR-194 report simulator control UM-434 simulator state UM-434 report command CR-192 reporting code coverage UM-309 variable settings CR-18 RequireConfigForAllDefaultBinding variable UM-441 resolution in SystemC simulation UM-153 mixed designs UM-174 returning as a real UM-87 specifying with -t argument CR-312 verilog simulation UM-114 VHDL simulation UM-75 Resolution .ini file variable UM-447 resolution simulator state variable UM-455 resource libraries UM-60 restart command CR-194 defaults UM-453 in GUI GR-31 toolbar button GR-41, GR-120, GR-205 Restart dialog GR-88 results, saving simulations UM-211 resume command CR-196
Index
RTL-level design busses reconstructing UM-223 run command CR-197 RunLength .ini file variable UM-447 Runtime Options dialog GR-85
S
saving simulation options in a project UM-46 waveforms UM-211 saving GUI preferences GR-252 saving memory, dialog GR-160 sc_argc() function UM-163 sc_argv() function UM-163 sc_clock() functions, moving UM-143 sc_cycle() function UM-162 sc_fifo UM-161 sc_fix and sc_ufix UM-164 sc_fixed and sc_ufixed UM-164 sc_foreign_module UM-203 and parameters UM-196 sc_initialize(), removing calls UM-162 sc_main() function UM-162 sc_main() function, converting UM-143 SC_MODULE_EXPORT macro UM-144 sc_signed and sc_unsigned UM-164 sc_start() function UM-162 sc_start() function, replacing in SystemC UM-162 sc_start(), replacing UM-143 scaling fonts GR-15 sccom using sccom vs. raw C++ compiler UM-149 sccom command CR-199 sccom -link command UM-151, UM-209 sccomLogfile .ini file variable (sccom) UM-442 sccomVerbose .ini file variable (sccom) UM-442 scgenmod command CR-203 scgenmod, using UM-194, UM-202 -sclib command CR-319 scope, setting region environment CR-131 ScTimeUnit .ini file variable UM-447 SCV library, including CR-201 SDF controlling missing instance messages CR-311 disabling timing checks UM-393 errors and warnings UM-383 errors on loading, disabling CR-311 instance specification UM-382
interconnect delays UM-393 mixed VHDL and Verilog designs UM-392 specification with the GUI UM-383 troubleshooting UM-394 Verilog $sdf_annotate system task UM-386 optional conditions UM-391 optional edge specifications UM-390 rounded timing values UM-391 SDF to Verilog construct matching UM-387 VHDL resolving errors UM-385 SDF to VHDL generic matching UM-384 warning messages, disabling CR-311 $sdf_done UM-129 search libraries CR-316, GR-80 searching binary signal values in the GUI CR-30 Expression Builder UM-241 in the source window GR-187 List window signal values, transitions, and names CR-23 Verilog libraries UM-107, UM-192 Wave window signal values, edges and names GR-215 searchlog command CR-205 sensitivity list warning UM-464 setenv command CR-207 $setuphold UM-130 severity, changing level for errors UM-458 shared library building in SystemC UM-151, GR-30 shared objects loading FLI applications see FLI Reference manual loading PLI/VPI C applications UM-484 loading PLI/VPI C++ applications UM-491 loading with global symbol visibility CR-308, UM498 shift command CR-208 Shortcuts text editing UM-519 shortcuts command history CR-20, UM-517 command line caveat CR-19, UM-517 List window UM-522 Main window UM-519 Source window UM-519 Wave window UM-523 show command CR-209
Index
show drivers Dataflow window UM-277 Wave window UM-257 show source lines with errors GR-58, GR-67 Show_ WarnMatchCadence .ini file variable UM-440 Show_BadOptionWarning .ini file variable UM-440 Show_Lint .ini file variable VCOM UM-441 Show_Lint .ini file variable (VLOG) UM-440 Show_source .ini file variable VCOM UM-441 Show_source .ini file variable (VLOG) UM-440 Show_VitalChecksOpt .ini file variable UM-441 Show_VitalChecksWarning .ini file variable UM-442 Show_WarnCantDoCoverage .ini file variable UM-440 Show_WarnCantDoCoverage variable UM-442 Show_Warning1 .ini file variable UM-442 Show_Warning10 .ini file variable UM-442 Show_Warning2 .ini file variable UM-442 Show_Warning3 .ini file variable UM-442 Show_Warning4 .ini file variable UM-442 Show_Warning5 .ini file variable UM-442 Show_Warning9 .ini file variable UM-442 Show_WarnLocallyStaticError variable UM-442 ShowUnassociatedScNameWarning variable UM-447 ShowUndebuggableScTypeWarning variable UM-448 Signal Breakpoints dialog GR-98 signal interaction Verilog and SystemC UM-179 Signal Spy UM-88, UM-362 disable UM-357, UM-369 enable UM-358, UM-370 overview UM-356 $signal_force UM-377 signal_force UM-88, UM-365 $signal_release UM-379 signal_release UM-88, UM-367 signals alternative names in the Wave window (-label) CR50 applying stimulus to GR-169 attributes of, using in expressions CR-25 breakpoints CR-327 combining into a user-defined bus CR-50, UM-252 Dataflow window, displaying in UM-274, GR-121 drivers of, displaying CR-126 driving in the hierarchy UM-359 environment of, displaying CR-131 filtering in the Objects window GR-168 finding CR-137
force time, specifying CR-142 hierarchy driving in UM-359, UM-371 referencing in UM-88, UM-362, UM-374 releasing anywhere in UM-367 releasing in UM-88, UM-379 log file, creating CR-148 names of, viewing without hierarchy GR-237 pathnames in VSIM commands CR-12 radix specifying for examine CR-133 specifying in List window CR-45 specifying in Wave window CR-51 readers of, displaying CR-191 sampling at a clock change UM-256 states of, displaying as mnemonics CR-290 stimulus CR-141 transitions, searching for UM-237 types, selecting which to view GR-168 unresolved, multiple drivers on GR-59, GR-68 values of displaying in Objects window GR-167 examining CR-132 forcing anywhere in the hierarchy UM-88, UM-365, UM-377 replacing with text CR-290 saving as binary log file UM-212 waveforms, viewing GR-194 Signals (Objects) window UM-274 simulating batch mode UM-25 command-line mode UM-25 Comparing simulations UM-211 default run length GR-86 delays, specifying time units for CR-19 design unit, specifying CR-305 graphic interface to GR-76 iteration limit GR-86 mixed language designs compilers UM-173 libraries UM-173 resolution limit in UM-174 mixed Verilog and SystemC designs channel and port type mapping UM-179 SystemC sc_signal data type mapping UM-180 Verilog port direction UM-182 Verilog state mapping UM-182 mixed Verilog and VHDL designs Verilog parameters UM-176 Verilog state mapping UM-177
Index
VHDL and Verilog ports UM-176 VHDL generics UM-178 mixed VHDL and SystemC designs SystemC state mapping UM-187 VHDL port direction UM-186 VHDL port type mapping UM-184 VHDL sc_signal data type mapping UM-184 saving dataflow display as a Postscript file UM-284 saving options in a project UM-46 saving simulations CR-148, CR-313, UM-211 saving waveform as a Postscript file UM-250 speeding-up with the Profiler UM-337 stepping through a simulation CR-213 stimulus, applying to signals and nets GR-169 stopping simulation in batch mode CR-330 SystemC UM-137, UM-152 usage flow for SystemC only UM-142 time resolution GR-77 Verilog UM-114 delay modes UM-123 hazard detection UM-120 resolution limit UM-114 XL compatible simulator options UM-121 VHDL UM-75 viewing results in List pane GR-135 viewing results in List window UM-231 VITAL packages UM-86 simulating the design, overview UM-24 simulation basic steps for UM-22 Simulation Configuration creating UM-46 dialog GR-52 simulation task overview UM-21 simulations event order in UM-117 saving results CR-119, CR-120, UM-211 saving results at intervals UM-220 simulator resolution mixed designs UM-174 returning as a real UM-87 SystemC UM-153 Verilog UM-114 VHDL UM-75 vsim -t argument CR-312 simulator state variables UM-455 simulator version CR-312, CR-322 simulator, difference from OSCI UM-162 simultaneous events in Verilog changing order CR-293
sizetf callback function UM-504 sm_entity UM-533 SmartModels creating foreign architectures with sm_entity UM533 invoking SmartModel specific commands UM-536 linking to UM-532 lmcwin commands UM-537 memory arrays UM-538 Verilog interface UM-539 VHDL interface UM-532 so, shared object file loading PLI/VPI C applications UM-484 loading PLI/VPI C++ applications UM-491 software version GR-39 source balloon C Debug GR-100 source code pragmas UM-305 source code, security UM-66, UM-133 source directory, setting GR-24, GR-189 source files, referencing with location maps UM-63 source files, specifying with location maps UM-63 source highlighting, customizing GR-188 source libraries arguments supporting UM-109 source lines with errors showing GR-58, GR-67 Source window GR-182 code coverage data UM-298 colorization GR-188 tab stops in GR-188 see also windows, Source window source-level debug SystemC, enabling UM-157 spaces in pathnames CR-11 sparse memories listing with write report CR-345 specify path delays CR-319 matching to IOPATH statements UM-387 speeding-up the simulation UM-337 square brackets, escaping CR-16 Standard Developers Kit User Manual UM-31 standards supported UM-27 Start Simulation dialog GR-76 start_of_simulation() function UM-163 startup alternate to startup.do (vsim -do) CR-306 environment variables access during UM-527 files accessed during UM-526 macro in the modelsim.ini file UM-448
Index
macros UM-452 startup macro in command-line mode UM-25 using a startup file UM-452 Startup .ini file variable UM-448 state variables UM-455 statistical sampling profiler UM-338 status bar Main window GR-22 status command CR-212 Status field Project tab UM-43 std .ini file variable UM-439 std_arith package disabling warning messages UM-452 std_developerskit .ini file variable UM-439 Std_logic mapping to binary radix CR-30 std_logic_arith package UM-61 std_logic_signed package UM-61 std_logic_textio UM-61 std_logic_unsigned package UM-61 StdArithNoWarnings .ini file variable UM-448 STDOUT environment variable UM-435 step command CR-213 steps for simulation, overview UM-22 stimulus applying to signals and nets GR-169 stop command CR-214 struct of sc_signal<T> UM-159 subprogram inlining UM-71 subprogram write is ambiguous error, fixing UM-81 Support UM-32 Suppress .ini file variable UM-450 symbol mapping Dataflow window UM-287 symbolic constants, displaying CR-290 symbolic link to design libraries (UNIX) UM-59 symbolic names, assigning to signal values CR-290 synopsys .ini file variable UM-439 Synopsys libraries UM-61 syntax highlighting GR-188 synthesis rule compliance checking CR-247, UM-441, GR58, GR-67 system calls VCD UM-402 Verilog UM-125 system commands UM-421 system tasks proprietary UM-128
VCD UM-402 Verilog UM-125 Verilog-XL compatible UM-129 SystemC aggregates of signals/ports UM-159 class and structure member naming syntax CR-13 compiling for source level debug UM-146 compiling optimized code UM-146 component declaration for instantiation UM-208 construction parameters UM-164 control function UM-175 converting sc_main() UM-143 exporting sc_main, example UM-144 exporting top level module UM-144 fixed-point types UM-164 foreign module declaration UM-194 generic support, instantiating VHDL UM-204 hierarchical references in mixed designs UM-175 instantiation criteria in Verilog design UM-199 instantiation criteria in VHDL design UM-208 Link dialog GR-74 linking the compiled source UM-151 maintaining design portability UM-147 mapping states in mixed designs UM-187 VHDL UM-187 master slave library, including CR-201 mixed designs with Verilog UM-172 mixed designs with VHDL UM-172 observe function UM-175 parameter support, Verilog instances UM-196 prim channel aggregates UM-159 replacing sc_start() UM-143 sc_clock(), moving to SC_CTOR UM-143 sc_fifo UM-161 simulating UM-152 source code, modifying for vsim UM-143 specifying shared library path, command CR-319 stack space for threads UM-166 state-based code, initializing and cleanup UM-154 troubleshooting UM-166 unsupported functions UM-162 verification library, including CR-201 viewable/debuggable objects UM-155 viewing FIFOs UM-161 virtual functions UM-154 SystemC modules exporting for use in Verilog UM-199 exporting for use in VHDL UM-209 SystemVerilog enabling with -sv argument CR-298
Index
keyword considerations UM-103 multiple files on vlog command line CR-296 suppported implementation details UM-27 SystemVerilog DPI registering DPIapplications UM-480 specifying the DPI file to load UM-497
T
tab groups GR-20 tab stops Source window GR-188 tb command CR-215 Tcl UM-414UM-424 command separator UM-420 command substitution UM-418 command syntax UM-416 evaluation order UM-420 Man Pages in Help menu GR-39 preference variables GR-251 relational expression evaluation UM-420 time commands UM-423 variable in when commands CR-328 substitution UM-421 VSIM Tcl commands UM-422 Tcl_init error message UM-464 Technical support and updates UM-32 temp files, VSOUT UM-437 test signal delaying GR-226 testbench, accessing internal objectsfrom UM-355 text and command syntax UM-30 Text editing UM-519 TEXTIO buffer, flushing UM-83 TextIO package alternative I/O files UM-83 containing hexadecimal numbers UM-82 dangling pointers UM-82 ENDFILE function UM-82 ENDLINE function UM-82 file declaration UM-79 implementation issues UM-81 providing stimulus UM-83 standard input UM-80 standard output UM-80 WRITE procedure UM-81 WRITE_STRING procedure UM-81
TF routines UM-510, UM-512 TFMPC disabling warning CR-317 explanation UM-466 time absolute, using @ CR-19 measuring in Wave window UM-233 resolution in SystemC UM-153 simulation time units CR-19 time resolution as a simulator state variable UM-455 time collapsing CR-313, UM-221 time literal, missing space GR-59, GR-68 time resolution in mixed designs UM-174 in Verilog UM-114 in VHDL UM-75 setting with the GUI GR-77 with vsim command CR-312 time type converting to real UM-89 time unit in SystemC UM-153 time, time units, simulation time CR-19 timescale directive warning disabling CR-317 investigating UM-115 timing $setuphold/$recovery UM-130 annotation UM-381 differences shown by comparison UM-267 disabling checks CR-297, UM-393 disabling checks for entire design CR-310 negative check limits described UM-121 extending CR-316 title, Main window, changing CR-312 TMPDIR environment variable UM-435 to_real VHDL function UM-89 to_time VHDL function UM-90 toggle add command CR-216 toggle coverage UM-300 enabling UM-300 excluding enum signals UM-303 excluding signals CR-218 max VHDL integer values UM-448 port collapsing UM-302 reporting UM-301 viewing in Signals window UM-300 toggle disable command CR-218
Index
toggle enable command CR-219 toggle report command CR-220 toggle reset command CR-222 toggle statistics enabling CR-216 reporting CR-220 resetting CR-222 toggling waveform popup on/off UM-268, GR-238 tolerance leading edge UM-264 trailing edge UM-264 too few port connections, explanation UM-466 tool structure UM-20 toolbar Dataflow window GR-125 Main window GR-40 Wave window GR-203 tooltip, toggling waveform popup GR-238 tracing events UM-280 source of unknown UM-281 transcript disable file creation UM-451, GR-19 file name, specifed in modelsim.ini UM-451 redirecting with -l CR-308 reducing file size CR-224 saving GR-19 using as a DO file GR-19 transcript command CR-223 transcript file command CR-224 TranscriptFile .ini file variable UM-448 TreeUpdate command CR-342 triggers, in the List window UM-254 triggers, in the List window, setting UM-253, GR-146 troubleshooting SystemC UM-166 unexplained behaviors, SystemC UM-166 TSCALE, disabling warning CR-317 TSSI CR-348 in VCD files UM-408 tssi2mti command CR-225 type converting real to time UM-90 converting time to real UM-89 Type field, Project tab UM-43 types, fixed-point in SystemC UM-162
U
-u CR-299 unbound component GR-59, GR-68 UnbufferedOutput .ini file variable UM-448 undeclared nets, reporting an error CR-296 undefined symbol, error UM-166 unexplained behavior during simulation UM-166 unexplained simulation behavior UM-166 ungrouping objects, Monitor window GR-192 unit delay mode UM-124 unknowns, tracing UM-281 unnamed ports, in mixed designs UM-191 unresolved signals, multiple drivers on GR-59, GR-68 unsetenv command CR-226 unsupported functions in SystemC UM-162 use 1076-1993 language standard GR-57, GR-67 use clause, specifying a library UM-60 use explicit declarations only GR-58, GR-67 use flow Code Coverage UM-292 SystemC-only designs UM-142 UseCsupV2 .ini file variable UM-448 user-defined bus CR-50, UM-222, UM-252 UserTimeUnit .ini file variable UM-448 UseScv .ini file variable (sccom) UM-443 util package UM-87
V
-v CR-299 v2k_int_delays CR-319 values describe HDL items CR-123 examine HDL item values CR-132 of HDL items GR-186 replacing signal values with strings CR-290 variable settings report CR-18 variables describing CR-123 environment variables UM-434 LM_LICENSE_FILE UM-434 precedence between .ini and .tcl UM-454 referencing in commands CR-18 setting environment variables UM-434 simulator state variables current settings report UM-434 iteration number UM-455 name of entity or module as a variable UM-455
Index
resolution UM-455 simulation time UM-455 value of changing from command line CR-67 changing with the GUI GR-149 examining CR-132 values of displaying in Objects window GR-167 saving as binary log file UM-212 Variables (Locals) window UM-278 vcd add command CR-227 vcd checkpoint command CR-228 vcd comment command CR-229 vcd dumpports command CR-230 vcd dumpportsall command CR-232 vcd dumpportsflush command CR-233 vcd dumpportslimit command CR-234 vcd dumpportsoff command CR-235 vcd dumpportson command CR-236 vcd file command CR-237 VCD files UM-397 adding items to the file CR-227 capturing port driver data CR-230, UM-408 case sensitivity UM-398 converting to WLF files CR-245 creating CR-227, UM-398 dumping variable values CR-228 dumpports tasks UM-402 flushing the buffer contents CR-241 from VHDL source to VCD output UM-404 generating from WLF files CR-335 inserting comments CR-229 internal signals, adding CR-227 specifying maximum file size CR-242 specifying name of CR-239 specifying the file name CR-237 state mapping CR-237, CR-239 stimulus, using as UM-399 supported TSSI states UM-408 turn off VCD dumping CR-243 turn on VCD dumping CR-244 VCD system tasks UM-402 viewing files from another tool CR-245 vcd files command CR-239 vcd flush command CR-241 vcd limit command CR-242 vcd off command CR-243 vcd on command CR-244 vcd2wlf command CR-245 vcom
enabling code coverage UM-295 vcom command CR-246 vcover command UM-315 vcover convert command CR-254 vcover merge command CR-255 vcover rank command CR-257 vcover report command CR-258 vcover utility UM-315 vdel command CR-263 vdir command CR-264 vector elements, initializing CR-67 vendor libraries, compatibility of CR-264 Verilog ACC routines UM-508 capturing port driver data with -dumpports CR-237, UM-408 cell libraries UM-123 compiler directives UM-132 compiling and linking PLI C applications UM-484 compiling and linking PLI C++ applications UM491 compiling design units UM-103 compiling with XL uselib compiler directive UM110 component declaration UM-189 configurations UM-112 event order in simulation UM-117 generate statements UM-113 instantiation criteria in mixed-language design UM188 instantiation criteria in SystemC design UM-194 instantiation of VHDL design units UM-192 language templates GR-184 library usage UM-107 mapping states in mixed designs UM-177 mapping states in SystemC designs UM-182 mixed designs with SystemC UM-172 mixed designs with VHDL UM-172 parameter support, instantiating SystemC UM-199 parameters UM-176 port direction UM-182 sc_signal data type mapping UM-180 SDF annotation UM-386 sdf_annotate system task UM-386 simulating UM-114 delay modes UM-123 XL compatible options UM-121 simulation hazard detection UM-120 simulation resolution limit UM-114 SmartModel interface UM-539
Index
source code viewing GR-182 standards UM-27 system tasks UM-125 TF routines UM-510, UM-512 to SystemC, channel and port type mapping UM179 XL compatible compiler options UM-109 XL compatible routines UM-512 XL compatible system tasks UM-129 verilog .ini file variable UM-439 Verilog 2001 disabling support CR-299, UM-440 Verilog PLI/VPI 64-bit support in the PLI UM-513 compiling and linking PLI/VPI C applications UM484 compiling and linking PLI/VPI C++ applications UM-491 debugging PLI/VPI code UM-514 PLI callback reason argument UM-502 PLI support for VHDL objects UM-507 registering PLI applications UM-476 registering VPI applications UM-478 specifying the PLI/VPI file to load UM-497 Verilog-XL compatibility with UM-101, UM-473 Veriuser .ini file variable UM-448, UM-477 Veriuser, specifying PLI applications UM-477 veriuser.c file UM-506 verror command CR-265 version obtaining via Help menu GR-39 obtaining with vsim command CR-312 obtaining with vsim<info> commands CR-322 vgencomp command CR-266 VHDL compiling design units UM-70 creating a design library UM-70 delay file opening UM-453 dependency checking UM-70 field naming syntax CR-13 file opening delay UM-453 instantiation criteria in SystemC design UM-202 instantiation from Verilog UM-192 instantiation of Verilog UM-176 language templates GR-184 language versions UM-72 library clause UM-60 mixed designs with SystemC UM-172 mixed designs with Verilog UM-172
object support in PLI UM-507 optimizations inlining UM-71 port direction UM-186 port type mapping UM-184 sc_signal data type mapping UM-184 simulating UM-75 SmartModel interface UM-532 source code viewing GR-182 standards UM-27 timing check disabling UM-75 VITAL package UM-61 VHDL utilities UM-87, UM-88, UM-362, UM-374 get_resolution() UM-87 to_real() UM-89 to_time() UM-90 VHDL-1987, compilation problems UM-72 VHDL-1993, enabling support for CR-246, UM-442 VHDL-2002, enabling support for CR-246, UM-442 VHDL93 .ini file variable UM-442 view command CR-268 view_profile command UM-344 viewing library contents UM-57 waveforms CR-313, UM-211 viewing FIFOs UM-161 virtual compare signal, restrictions UM-252 virtual count commands CR-270 virtual define command CR-271 virtual delete command CR-272 virtual describe command CR-273 virtual expand commands CR-274 virtual function command CR-275 virtual functions in SystemC UM-154 virtual hide command CR-278, UM-223 virtual log command CR-279 virtual nohide command CR-281 virtual nolog command CR-282 virtual objects UM-222 virtual functions UM-223 virtual regions UM-224 virtual signals UM-222 virtual types UM-224 virtual region command CR-284, UM-224 virtual regions reconstruct the RTL hierarchy in gate-level design UM-224 virtual save command CR-285, UM-223 virtual show command CR-286 virtual signal command CR-287, UM-222
Index
virtual signals reconstruct RTL-level design busses UM-223 reconstruct the original RTL hierarchy UM-223 virtual hide command UM-223 virtual type command CR-290 VITAL compiling and simulating with accelerated VITAL packages UM-86 disabling optimizations for debugging UM-86 specification and source code UM-84 VITAL packages UM-84 vital95 .ini file variable UM-439 vlib command CR-292 vlog enabling code coverage UM-295 multiple file compilation CR-296 vlog command CR-293 vlog.opt file GR-62, GR-71 vlog95compat .ini file variable UM-440 vmake command CR-302 vmap command CR-304 VPI, registering applications UM-478 VPI/PLI UM-135, UM-475 compiling and linking C applications UM-484 compiling and linking C++ applications UM-491 vsim build date and version CR-322 vsim command CR-305 VSIM license lost UM-467 vsim, differences with OSCI simulator UM-162 VSOUT temp file UM-437
W
WarnConstantChange .ini file variable UM-448 Warning .ini file variable UM-450 WARNING[8], -lint argument to vlog CR-296 warnings empty port name UM-463 exit codes UM-461 getting more information UM-458 messages, long description UM-458 metavalue detected UM-464 SDF, disabling CR-311 severity level, changing UM-458 suppressing VCOM warning messages CR-250, CR-297, UM-460 suppressing VLOG warning messages CR-297, UM-460 suppressing VSIM warning messages CR-317, UM-
460 Tcl initialization error 2 UM-464 too few port connections UM-466 turning off warnings from arithmetic packages UM452 waiting for lock UM-463 watch window add watch command CR-48 adding items to CR-48 watching a signal value GR-191 watching signal values CR-48 wave commands CR-324 Wave Log Format (WLF) file UM-211 wave log format (WLF) file CR-313 of binary signal values CR-148 see also WLF files wave viewer, Dataflow window UM-278 Wave window UM-228, GR-194 adding items to CR-49 compare waveforms UM-267 docking and undocking UM-229, GR-195 in the Dataflow window UM-278 saving layout UM-249 toggling waveform popup on/off UM-268, GR-238 values column UM-268 see also windows, Wave window WaveActivateNextPane command CR-342 Waveform Comparison CR-69 add region UM-263 adding signals UM-262 annotating differences UM-268 clocked comparison UM-264 compare by region UM-263 compare by signal UM-262 compare options UM-266 compare tab UM-261 comparison method UM-264 comparison method tab UM-264 delaying the test signal GR-226 difference markers UM-267 flattened designs UM-271 hierarchical designs UM-271 icons UM-269 introduction UM-258 leading edge tolerance UM-264 List window display UM-269 pathnames UM-267 reference dataset UM-260 reference region UM-263 test dataset UM-261
Index
timing differences UM-267 trailing edge tolerance UM-264 values column UM-268 Wave window display UM-267 waveform logfile log command CR-148 overview UM-211 see also WLF files waveform popup UM-268, GR-238 waveforms UM-211 optimize viewing of UM-449 optimizing viewing of CR-314 saving and viewing CR-148, UM-212 viewing GR-194 WaveRestoreCursors command CR-342 WaveRestoreZoom command CR-342 WaveSignalNameWidth .ini file variable UM-448 when command CR-327 when statement time-based breakpoints CR-331 where command CR-332 wildcard characters for pattern matching in simulator commands CR-18 windows Active Processes pane GR-107 code coverage statistics UM-297 Dataflow window UM-274, GR-121 toolbar GR-125 zooming UM-279 List window UM-231, GR-135 display properties of UM-247 formatting HDL items UM-247 output file CR-343 saving data to a file UM-251 saving the format of CR-341 setting triggers UM-253, UM-254, GR-146 Locals window GR-148 Main window GR-16 status bar GR-22 text editing UM-519 time and delta display GR-22 toolbar GR-40 Memory window GR-151 monitor GR-191 Objects window GR-167 opening from command line CR-268 with the GUI GR-28 Signals window VHDL and Verilog items viewed in GR-167
Source window GR-182 text editing UM-519 viewing HDL source code GR-182 Variables window VHDL and Verilog items viewed in GR-148 Wave window UM-228, GR-194 adding HDL items to UM-232 cursor measurements UM-233 display properties UM-243 display range (zoom), changing UM-237 format file, saving UM-249 path elements, changing CR-100, UM-448 time cursors UM-233 zooming UM-237 WLF files collapsing deltas CR-313 collapsing events UM-221 collapsing time steps CR-313 converting to VCD CR-335 creating from VCD CR-245 filtering, combining CR-336 limiting size CR-314 log command CR-148 optimizing waveform viewing CR-314, UM-449 overview UM-212 repairing CR-340 saving CR-119, CR-120, UM-213 saving at intervals UM-220 specifying name CR-313 wlf2log command CR-333 wlf2vcd command CR-335 WLFCacheSize .ini file variable UM-449 WLFCollapseMode .ini file variable UM-449 WLFFilename .ini file variable UM-449 wlfman command CR-336 wlfrecover command CR-340 work library UM-54 creating UM-56 workspace GR-17 code coverage GR-109 Files tab GR-109 write format command CR-341 write list command CR-343 write preferences command CR-344 WRITE procedure, problems with UM-81 write report command CR-345 write timing command CR-346 write transcript command CR-347 write tssi command CR-348 write wave command CR-350
Index
X
X tracing unknowns UM-281 .Xdefaults file, controlling fonts GR-15 X propagation disabling for entire design CR-310 xml format coverage reports UM-310 X-session controlling fonts GR-15
Y
-y CR-299
Z
zero delay elements UM-77 zero delay mode UM-124 zero-delay loop, infinite UM-78 zero-delay oscillation UM-78 zero-delay race condition UM-117 zoom Dataflow window UM-279 from Wave toolbar buttons UM-237 saving range with bookmarks UM-238 wave window returning current range CR-324 with the mouse UM-237 zooming window panes GR-247

Man

Uploaded by

Copyright:

Available Formats

Man

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Man

Uploaded by

Copyright:

Available Formats

ModelSim Users Manual

Version 6.1a Published: 12/Jul/05

Copyright Mentor Graphics Corporation 2005 All rights reserved.

ModelSim Users Manual

ModelSim Users Manual

ModelSim Users Manual

Installation directory pathnames

ModelSim Users Manual

3 - Design libraries (UM-53)

4 - VHDL simulation (UM-67)

ModelSim Users Manual

5 - Verilog and SystemVerilog simulation (UM-101)

ModelSim Users Manual

UM-132 UM-132 UM-133 UM-134

Verilog PLI/VPI and SystemVerilog DPI . . . . . . . . . . . . . . . . . . . . . UM-135

6 - SystemC simulation (UM-137)

ModelSim Users Manual

7 - Mixed-language simulation (UM-171)

ModelSim Users Manual

UM-208 UM-209 UM-209 UM-209

8 - WLF files (datasets) and virtuals (UM-211)

9 - Waveform analysis (UM-225)

ModelSim Users Manual

10 - Tracing signals with the Dataflow window (UM-273)

ModelSim Users Manual

11 - Measuring code coverage (UM-291)

ModelSim Users Manual

Debugging functions when quitting simulation . . . . . . . . . . . . . . . . . . . UM-334 . . . . . . . . . . . . . . . . . . . . . . . . . UM-335

13 - Profiling performance and memory use (UM-337)

ModelSim Users Manual

14 - Signal Spy (UM-355)

15 - Standard Delay Format (SDF) Timing Annotation (UM-381)

SDF for mixed VHDL and Verilog designs . . . . . . . . . . . . . . . . . . . . UM-392 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-393

ModelSim Users Manual

Disabling timing checks

16 - Value Change Dump (VCD) Files (UM-397)

17 - Tcl and macros (DO files) (UM-413)

ModelSim Users Manual

A - Simulator variables (UM-433)

B - Error and warning messages (UM-457)

Exit codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UM-461

ModelSim Users Manual

C - Verilog PLI / VPI / DPI (UM-473)

ModelSim Users Manual

D - ModelSim shortcuts (UM-517)

E - System initialization (UM-525)

F - Logic Modeling SmartModels (UM-531)

ModelSim Users Manual

Sections in this document . What is an "object". Text conventions . . . . .

ModelSim Users Manual

Tool structure and flow

VHDL Design Libraries

vlog/ vcom/ sccom

.ini or .mpf file

HDL/SystemC Analyze/ Compile

ModelSim Users Manual

Simulation task overview UM-21

Simulation task overview

Compile or Compile All icons:

Step 3: Load the design into the simulator