CSC 323 – Principles of Operating Systems

Instructor: Dr. M. Hasan Jamal

Lecture# 04: Threads

1
Processes
• Recall that a process includes many things
• An address space (code, data),
• OS resources (e.g., open files) and accounting information
• Execution state (program counter, registers, and stack pointer etc.).

• Creating a new process is costly

• because of all of the data structures that must be allocated and initialized

• Communicating between processes is also costly

• because most communication goes through the OS
• overhead of system calls and copying data

2
Example: Web Server
• How does a web server handle a request?

• A web server needs to handle many concurrent requests.

• Solution 1:
• forks off copies of itself to handle multiple simultaneous requests
• Processes communicate with each other via inter-process communication

3
Concurrent Programs
• To execute these programs, we need to
• Create several processes that execute in parallel
• Cause each to map to the same address space to share data
• They are all part of the same computation
• Have the OS schedule these processes in parallel (logically or physically)

• This situation is very inefficient

• Space: PCB, page tables, etc.
• Time: create data structures, fork and copy address space, etc.

4
Rethinking Processes
• What is similar in these cooperating processes?
• They all share the same code and data (address space)
• They all share the same privileges
• They all share the same resources (files, communication channels, etc.)

• What don’t they share?

• Each has its own execution state: PC, SP, and registers

• Idea: Why not separate the process concept from its execution state?
• Process: address space, privileges, resources, etc.
• Execution state: PC, SP, registers

• Execution state also called thread of control, or thread 5

Threads
• Modern OSes separate the concepts of processes and threads
• The thread defines a single sequential execution stream within a process
• The process defines an address space and general process attributes

• A thread is bound to a single process

• Processes, however, can have multiple threads

• Threads become the unit of scheduling

• Processes are now the containers in which threads execute
• Processes become static, threads are the dynamics entities

• Data structure: Thread Control Block (TCB)

6
Process Control Block (PCB)
• Process management information
• Process state (running, ready, waiting)
• PC & CPU registers, parents, etc.
• CPU scheduling information (priorities, etc.)

• Memory-management information
• memory allocated to the process (segments, page tables etc.)

• I/O and file management

• Communication ports, directories, file descriptors etc.

7
Thread Control Block (TCB)
• Shared information
• Process info: parent process
• Memory: code/data segment, page table, and stats
• I/O and file: communication ports, open file descriptors

• Private information
• State (ready, running and waiting)
• Registers, Program counter
• Execution stack

8
Threads in a Process

What about heap?

Threads in a Process

10
Threads vs. Processes
Process Thread
Processes are containers in which threads execute. Threads are the unit of scheduling.

Thread can share same set of open files and child

Each process executes the same code but has its own
processes. There can be several threads in a single
address space and file resources.
address space.

Each process operates independently of the others. A thread can access and alter another thread's data.

Process switching needs interaction with the operating

Thread switching doesn’t need interaction with the
system to save the states of address space, resources,
operating system as only the states of the PC, SP, and
accounting info. etc., along with PC, SP, and registers,
registers are saved, therefore, it is faster.
therefore, it is slower.

If a process is blocked by some I/O request, then no If a thread is blocked by some I/O request, another
further execution is possible until it gets unblocked. thread in the same task can continue its execution.

Thread, also known as “lightweight” process takes lesser 11

Process is heavyweight or resource intensive.
resources than a process.
Process/Thread Separation
• Easier to support multithreaded applications
• Concurrency does not require creating new processes

• Concurrency (multithreading) can be very useful

• Improving program structure
• Allowing one process to use multiple CPUs/cores
• Handling concurrent events (e.g., Web requests)
• Allowing program to overlap I/O and computation

• So multithreading is even useful on a uniprocessor

• Although today even cell phones are multicore

• But brings a whole new meaning to Spaghetti Code 12

• Forcing OS students to learn about synchronization
Multicore Programming
• Multicore or multiprocessor systems putting pressure on programmers,
challenges include:
• Dividing activities
• Balance
• Data splitting
• Data dependency
• Testing and debugging

• Parallelism implies a system can perform more than one task simultaneously

• Concurrency supports more than one task making progress

• Single processor / core, scheduler providing concurrency
13
Concurrency vs. Parallelism
• Concurrent execution on single-core system:

• Parallelism on a multi-core system:

14
Multicore Programming
• Types of parallelism
• Data parallelism – distributes subsets of the same data across multiple cores,
same operation on each
• Task parallelism – distributing threads across cores, each thread performing unique
operation

• As number of threads grows, so does architectural support for threading

• CPUs have cores as well as hardware threads
• Consider Oracle SPARC T4 with 8 cores, and 8 hardware threads per core

15
Amdahl’s Law
• Identifies performance gains from adding additional cores to an application that has
both serial and parallel components
• S is serial portion with N processing cores

• That is, if application is 75% parallel / 25% serial, moving from 1 to 2 cores results
in speedup of 1.6 times
• As N approaches infinity, speedup approaches 1/S
• Serial portion of an application has disproportionate effect on performance
gained by adding additional cores 16

• But does the law take into account contemporary multicore systems?
Amdahl’s Law

17
Motivation
• Most modern applications are multithreaded
• Threads run within application
• Multiple tasks with the application can be implemented by separate threads
• Update display
• Fetch data
• Spell checking
• Answer a network request
• Process creation is heavy-weight while thread creation is light-weight
• Can simplify code, increase efficiency
• Kernels are generally multithreaded

18
Benefits of Threads
• Responsiveness – may allow continued execution if part of process is
blocked, especially important for user interfaces.
• E.g., user interaction and image loading in multi-threaded web browser:

• Resource Sharing – threads share resources of process, easier than shared

memory or message passing

• Economy – cheaper than process creation, thread switching lower overhead

than context switching
• Solaris: creating process is thirty times slower
• Context-switch is about five times slower

19
• Scalability – process can take advantage of multiprocessor architectures
Kernel-Level Threads
• All thread operations are implemented in the kernel

• The OS schedules all the threads in the system

• A blocking system call will not block the whole process

• Examples:
• Windows
• Solaris
• Linux
• Tru64 UNIX
• Mac OS X 20
Kernel Thread Limitations
• Every thread operation must go through kernel
• create, join, exit, synchronize, or switch for any reason
• Generally: threads 10x to 30x slower when implemented in kernel

• One-size fits all thread implementation

• Kernel threads must please all people
• Maybe pay for fancy features (priority etc.) you don’t need

• General heavy-weight memory requirements

• e.g., requires a fixed-size stack within kernel
• other data structures designed for heavier-weight processes
21
Alternative: User-Level Threads
• Implement as user-level library
• One kernel thread per process
• thread_create, thread_exit, etc., just library functions
• Library does thread context switch

• Usually faster to create and manage

• Examples:
• POSIX Pthreads,
• Java threads

22
User-Level Thread Limitations
• Cannot take advantage of multiple CPUs or cores

• User-level threads are invisible to the OS

• They are not well integrated with the OS

• As a result, OS can make poor decisions

• Scheduling a process with idle threads
• A blocking system call (e.g., disk read) blocks all threads
• Even if the process has other threads that can execute
• Unscheduling a process with a thread holding a lock

• How to solve this?

• Communication between the kernel and the user-level thread manager (Windows 8)
• Scheduler Activation 23
Kernel vs. User Threads
• Kernel-level threads
• Integrated with OS (informed scheduling)
• Slower to create, manipulate, synchronize

• User-level threads
• Faster to create, manipulate, synchronize
• Not integrated with OS (uninformed scheduling)

• Understanding their differences is important

• Correctness, performance

• Or use both kernel and user-level threads

• Can associate a user-level thread with a kernel-level thread 24
• Or multiplex user-level threads on top of kernel-level threads
Multithreading Models
• Three common ways to establishing a relation between user and kernel
threads are as follow:
• Many-to-One
• One-to-One
• Many-to-Many

• In user thread implementation, all user threads of the same process are
effectively mapped to one kernel thread

25
Many-to-One
• Many user-level threads mapped to single kernel thread
• One thread blocking causes all to block
• Thread management is done by a thread library in user space
• Multiple threads may not run in parallel on multi-core system because only one
may be in kernel at a time
• Few systems currently use this model

• Examples:
• Solaris Green Threads
• GNU Portable Threads

26
One-to-One
• Each user-level thread maps to a kernel thread
• Creating a user-level thread creates a kernel thread
• More concurrency than many-to-one model
• System keeps running while one thread executes a blocking system call
• It also allow multiple threads to run in parallel on multiprocessor system
• Number of threads per process sometimes restricted due to overhead
• May lead to too many kernel threads

• Examples:
• Windows
• Linux
27
Many-to-Many
• Allows many user level threads to be mapped to many kernel threads

• Allows the operating system to create a sufficient number of kernel threads

• Examples:
• Windows ThreadFiber package
• Otherwise not very common

28
Two-level Model
• Similar to many-to-many model, except that it allows a user thread to be bound
to kernel thread

• Examples:
• IRIX
• HP-UX
• Tru64 UNIX
• Solaris 8 and earlier

29
Thread Context Switch
• The context switch routine does all of the magic
• Saves context of the currently running thread
• Push all machine state onto its stack
• Restore context of the next thread
• Pop all machine state from the next thread’s stack
• The next thread becomes the current thread
• Returns to caller as new thread

• This is all done in assembly language

• It works at the level of the procedure calling convention, so it cannot be implemented
using procedure calls

30
Context Switching of User-Level Threads
• If both threads belong to the same process
• Handled by the dispatcher in the library
• Only need to store/load the TCB information
• OS does not do anything

• If both threads belong to different processes

• Like an ordinary context switch of two processes
• Handled by OS (drop in/out of the kernel)
• OS needs to load/store PCB and TCB information

31
Programming with Threads: Thread
Libraries
• Thread library provides programmer with API for creating & managing threads

• Two primary ways of implementing

• Library entirely in user space
• Kernel-level library supported by the OS

32
Programming with Threads: Pthreads
• May be provided either as user-level or kernel-level

• A POSIX standard (IEEE 1003.1c) API for thread creation and synchronization

• Specification, not implementation

• API specifies behavior of the thread library, implementation is up to

development of the library

• Common in UNIX operating systems (Solaris, Linux, Mac OS X)

33
Programming with Threads: Pthreads

34
Programming with Threads: Windows

35
Programming with Threads: Java Threads
• Java threads are managed by the JVM

• Typically implemented using the threads model provided by underlying OS

• Java threads may be created by:

• Extending Thread class
• Implementing the Runnable interface

• Standard practice is to implement Runnable interface

36
Programming with Threads: Java Threads
• Implementing Runnable interface:

• Creating a thread:

• Waiting on a thread:

37
Programming with Threads: Java Threads
• Rather than explicitly creating threads, Java also allows thread creation around
the Executor interface:

• The Executor is used as follows:

38
Programming with Threads: Java Threads

39
Programming with Threads: Implicit
Threading
• Growing in popularity as numbers of threads increase, program correctness
more difficult with explicit threads

• Creation and management of threads done by compilers and run-time libraries

rather than programmers

• Five methods explored

• Thread Pools
• Fork-Join
• OpenMP
• Grand Central Dispatch
• Intel Threading Building Blocks (TBB)
40
Programming with Threads: Thread Pools
• Create a number of threads in a pool where they await work

• Advantages:
• Usually slightly faster to service a request with an existing thread than to create a
new thread
• Allows the number of threads in the application(s) to be bound to the size of the pool
• Separating task to be performed from mechanics of creating task allows different
strategies for running task
• i.e., tasks could be scheduled to run periodically

• Windows API supports thread pools:

41
Programming with Threads: Java Thread
Pools
• Three factory methods for creating thread pools in Executors class:

42
Programming with Threads: Fork-Join
Parallelism
• Multiple threads (tasks) are forked, and then joined.

• General algorithm for fork-join strategy:

43
Programming with Threads: Fork-Join
Parallelism

44
Programming with Threads: Fork-Join in Java

45
Programming with Threads: Fork-Join in Java
• The ForkJoinTask is an abstract base class
• RecursiveTask and RecursiveAction classes extend ForkJoinTask
• RecursiveTask returns a result (via the return value from compute() method)
• RecursiveAction does not return a result

46
Programming with Threads: OpenMP
• Set of compiler directives and an API for C, C++, FORTRAN
• Provides support for parallel programming in shared-memory environments
• Identifies parallel regions – blocks of code that can run in parallel

#pragma omp parallel

Create as many threads as there are cores

Run the for loop in parallel

47
Programming with Threads: Grand Central
Dispatch
• Apple technology for Mac OS X and iOS operating systems
• Extensions to C, C++ and Objective-C languages, API, and run-time library
• Allows identification of parallel sections and manages most details of threading
• Block is in “^{ }” - ˆ{ printf("I am a block"); }
• Blocks placed in dispatch queue and assigned to available thread in thread pool
when removed from queue
• Two types of dispatch queues:
• Serial: blocks removed in FIFO order, queue is per process, called main queue
• Programmers can create additional serial queues within program
• concurrent – removed in FIFO order but several may be removed at a time
• Four system wide queues divided by quality of service
QOS_CLASS_USER_INTERACTIVE
48
QOS_CLASS_USER_INITIATED
QOS_CLASS_USER_UTILITY
QOS_CLASS_USER_BACKGROUND
Programming with Threads: Intel TBB
• Template library for designing parallel C++ programs

• A serial version of a simple for loop

• The same for loop written using TBB with parallel_for statement:

49
Programming with Threads: Summary
• Flexible, but error-prone, since there is no protection between threads

• In C/C++
• Automatic variables are private to each thread
• Global variables and dynamically allocated memory are shared

• Need Synchronization!!!

50
Threading Issues: Semantics of fork() and
exec()
• Does fork( ) duplicate only the calling thread or all threads?
• Some UNIXes have two versions of fork

• exec( ) usually works as normal – replace the running process including

all threads

51
Threading Issues: Signal Handling
• UNIX systems use signals to notify occurrence of a particular event to a process

• A signal handler is used to process signals

• Signal is generated by particular event
• Signal is delivered to a process
• Signal is handled by one of two signal handlers:
• Every signal has default handler that kernel runs when handling signal
• User-defined signal handler can override default hander

• For single-threaded, signal delivered to process

• Where should a signal be delivered for multi-threaded?

• Deliver the signal to the thread to which the signal applies
• Deliver the signal to every thread in the process 52
• Deliver the signal to certain threads in the process
• Assign a specific thread to receive all signals for the process
Threading Issues: Thread Cancellation
• Two general approaches of terminating a target thread before it has finished:
• Asynchronous cancellation terminates the target thread immediately
• Deferred cancellation allows the target thread to periodically check if it should be
cancelled

• Pthread code to create and cancel a thread:

53
Threading Issues: Thread Cancellation
• Invoking thread cancellation requests cancellation, but actual cancellation
depends on thread state

• If thread has cancellation disabled, cancellation remains pending until thread

enables it
• Default type is deferred
• Cancellation only occurs when thread reaches cancellation point
• i.e., pthread_testcancel()
• Then cleanup handler is invoked
54
• On Linux systems, thread cancellation is handled through signals
Threading Issues: Thread Cancellation in
Java
• Deferred cancellation uses the interrupt() method, which sets the interrupted
status of a thread.

• A thread can then check to see if it has been interrupted:

55
Threading Issues: Thread Specific Data
• Thread-local storage (TLS) allows each thread to have its own copy of data

• Useful when you do not have control over the thread creation process (i.e., when
using a thread pool)

• Different from local variables

• Local variables visible only during single function invocation
• TLS visible across function invocations

• Similar to static data

• TLS is unique to each thread

56
Threading Issues: Scheduler Activations
• Many-to-Many models require communication to maintain the appropriate
number of kernel threads allocated to the application

• Typically use an intermediate data structure between user and kernel threads –
lightweight process (LWP)
• Appears to be a virtual processor on which process can schedule user thread to run
• Each LWP attached to kernel thread
• How many LWPs to create?

• Scheduler activations provide upcalls - a

communication mechanism from the kernel to
the upcall handler in the thread library

• This communication allows an application to 57

maintain the correct number kernel threads
Operating System Examples
• Windows Threads
• Linux Threads

58
Windows Threads
• Windows API – primary API for Windows applications
• Implements the one-to-one mapping, kernel-level
• Each thread contains
• A thread ID
• Register set representing state of processor
• Separate user and kernel stacks for when thread runs in user mode or kernel mode
• Private data storage area used by run-time libraries and dynamic link libraries
• The register set, stacks, and private storage area are known as the context of
the thread

59
Windows Threads
• The primary data structures of a thread include:
• ETHREAD (executive thread block) – includes pointer to process to which thread
belongs and to KTHREAD, in kernel space
• KTHREAD (kernel thread block) – scheduling and synchronization info, kernel-mode
stack, pointer to TEB, in kernel space
• TEB (thread environment block) – thread id, user-mode stack, thread-local storage,
in user space

60
Linux Threads
• Linux refers to them as tasks rather than threads

• Thread creation is done through clone() system call

• clone() allows a child task to share address space of the parent task (process)
• Flags control behavior

61
• struct task_struct points to process data structures (shared or unique)