P10 Scale Out

Download as pdf or txt
Download as pdf or txt
You are on page 1of 200

Front cover

IBM Power S1014, S1022s,


S1022, and S1024
Technical Overview and Introduction

Giuliano Anselmi
Young Hoon Cho
Andrew Laidlaw
Armin Rll
Tsvetomir Spasov

Redpaper
IBM Redbooks

IBM Power S1014, S1022s, S1022, and S1024 Technical


Overview and Introduction

August 2022

REDP-5675-00
Note: Before using this information and the product it supports, read the information in “Notices” on
page vii.

First Edition (August 2022)

This edition applies to the IBM Power S1014 (9105-41B), IBM Power S1022s (9105-22B), IBM Power S1022
(9105-22A), and IBM Power S1024 (9105-42A) servers.

This document was created or updated on August 24, 2022.

© Copyright International Business Machines Corporation 2022. All rights reserved.


Note to U.S. Government Users Restricted Rights -- Use, duplication or disclosure restricted by GSA ADP Schedule
Contract with IBM Corp.
Contents

Notices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Trademarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .x
Now you can become a published author, too! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
Comments welcome. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
Stay connected to IBM Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview . . . . . . . . . . . . . . . . 1


1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Systems overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.1 Power S1014 server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.2 Power S1022s server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.3 Power S1022 server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2.4 Power S1024 server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Operating environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4 Physical package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.4.1 Tower model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.4.2 Rack-mount models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.5 System features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.5.1 Power S1014 server features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.5.2 Power S1022s server features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.5.3 Power S1022 server features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.5.4 Power S1024 server features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.6 Minimum configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.7 PCIe adapter slots. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.8 Operating system support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.8.1 AIX operating system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.8.2 IBM i operating system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.8.3 Linux operating system distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.8.4 Red Hat OpenShift Container Platform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.8.5 Virtual I/O Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.8.6 Entitled System Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.8.7 Update Access Keys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.9 Hardware Management Console overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
1.9.1 HMC 7063-CR2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.9.2 Virtual HMC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
1.9.3 Baseboard management controller network connectivity rules for 7063-CR2 HMC .
30
1.9.4 High availability HMC configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
1.9.5 HMC code level requirements for the Power S1014, S1022s, S1022, and S1024 32
1.9.6 HMC currency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
1.10 IBM Power solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
1.10.1 IBM PowerSC 2.1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
1.10.2 IBM PowerSC Multi-Factor Authentication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
1.10.3 IBM Cloud PowerVC for Private Cloud . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
1.10.4 IBM VM Recovery Manager DR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
1.10.5 IBM Cloud Management Console . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

© Copyright IBM Corp. 2022. iii


1.11 IBM Power platform modernization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
1.11.1 IBM Power Virtual Servers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
1.11.2 Red Hat OpenShift Container Platform for Power. . . . . . . . . . . . . . . . . . . . . . . . 38
1.11.3 Hybrid Cloud Management Edition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

Chapter 2. Architecture and technical overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39


2.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.2 IBM Power10 processor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.2.1 Power10 processor overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.2.2 Processor modules for S1014, S1022s, S1022, and S1024 servers . . . . . . . . . . 47
2.2.3 Power10 processor core . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.2.4 Simultaneous multithreading. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.2.5 Matrix math accelerator AI workload acceleration . . . . . . . . . . . . . . . . . . . . . . . . 56
2.2.6 Power10 compatibility modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.2.7 Processor module options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.2.8 On-chip L3 cache and intelligent caching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
2.2.9 Open memory interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
2.2.10 Pervasive memory encryption. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2.2.11 Nest accelerator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
2.2.12 SMP interconnect and accelerator interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
2.2.13 Power and performance management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
2.2.14 Comparing Power10, Power9, and Power8 processors . . . . . . . . . . . . . . . . . . . 73
2.3 Memory subsystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
2.3.1 Memory feature and placement rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
2.3.2 Memory bandwidth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
2.4 Internal I/O subsystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
2.4.1 Slot configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
2.4.2 System port . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
2.5 Enterprise Baseboard Management Controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
2.5.1 Managing the system by using the ASMI GUI . . . . . . . . . . . . . . . . . . . . . . . . . . 100
2.5.2 Managing the system by using DMTF Redfish . . . . . . . . . . . . . . . . . . . . . . . . . . 107
2.5.3 Managing the system by using the Intelligent Platform Management Interface . 110

Chapter 3. Available features and options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111


3.1 Processor module features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
3.1.1 Power S1014 processor Feature Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
3.2 Memory features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
3.3 Power supply features. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
3.4 Peripheral Component Interconnect adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
3.4.1 Local area network adapters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
3.4.2 Fibre Channel adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
3.4.3 SAS adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
3.4.4 USB adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
3.4.5 Cryptographic coprocessor adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
3.5 Internal storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
3.5.1 S1022s and S1022 Backplane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
3.5.2 S1014 and S1024 Backplane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
3.5.3 NVMe JBOF to backplane cabling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
3.5.4 NVMe support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
3.5.5 RAID support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
3.6 External SAS ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
3.7 Media drawers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
3.7.1 External DVD drives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

iv IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
3.7.2 RDX removable disk drives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
3.8 Disk and media features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
3.9 External IO subsystems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
3.9.1 PCIe Gen3 I/O expansion drawer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
3.9.2 EXP24SX SAS Storage Enclosure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
3.9.3 IBM Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
3.10 System racks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
3.10.1 IBM Enterprise 42U Slim Rack 7965-S42 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
3.10.2 AC power distribution units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
3.10.3 Rack-mounting rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
3.10.4 Useful rack additions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
3.10.5 Original equipment manufacturer racks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

Chapter 4. Enterprise solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157


4.1 PowerVM virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
4.1.1 IBM Power Hypervisor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
4.1.2 Multiple shared processor pools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
4.1.3 Virtual I/O Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
4.1.4 Live Partition Mobility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
4.1.5 Active Memory Mirroring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
4.1.6 Remote Restart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
4.1.7 POWER processor modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
4.1.8 Single Root I/O virtualization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
4.2 IBM PowerVC overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
4.2.1 IBM PowerVC functions and advantages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
4.3 Digital transformation and IT modernization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
4.3.1 Application and services modernization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
4.3.2 System automation with Ansible . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
4.4 Protect trust from core to cloud . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
4.4.1 Power10 processor-based, technology-integrated security ecosystem . . . . . . . 175
4.4.2 Cryptographic engines and transparent memory encryption . . . . . . . . . . . . . . . 176
4.4.3 Quantum-safe cryptography support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
4.4.4 IBM PCIe3 Crypto Coprocessor BSC-Gen3 4769 . . . . . . . . . . . . . . . . . . . . . . . 177
4.4.5 IBM PowerSC support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
4.4.6 Secure and Trusted boot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
4.4.7 Enhanced CPU: BMC isolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
4.5 Running artificial intelligence where operational data is created, processed, and stored .
180
4.5.1 Train anywhere, deploy on Power10 processor-based server . . . . . . . . . . . . . . 181

Related publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183


IBM Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Online resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
Help from IBM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184

Contents v
vi IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Notices

This information was developed for products and services offered in the US. This material might be available
from IBM in other languages. However, you may be required to own a copy of the product or product version in
that language in order to access it.

IBM may not offer the products, services, or features discussed in this document in other countries. Consult
your local IBM representative for information on the products and services currently available in your area. Any
reference to an IBM product, program, or service is not intended to state or imply that only that IBM product,
program, or service may be used. Any functionally equivalent product, program, or service that does not
infringe any IBM intellectual property right may be used instead. However, it is the user’s responsibility to
evaluate and verify the operation of any non-IBM product, program, or service.

IBM may have patents or pending patent applications covering subject matter described in this document. The
furnishing of this document does not grant you any license to these patents. You can send license inquiries, in
writing, to:
IBM Director of Licensing, IBM Corporation, North Castle Drive, MD-NC119, Armonk, NY 10504-1785, US

INTERNATIONAL BUSINESS MACHINES CORPORATION PROVIDES THIS PUBLICATION “AS IS”


WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED
TO, THE IMPLIED WARRANTIES OF NON-INFRINGEMENT, MERCHANTABILITY OR FITNESS FOR A
PARTICULAR PURPOSE. Some jurisdictions do not allow disclaimer of express or implied warranties in
certain transactions, therefore, this statement may not apply to you.

This information could include technical inaccuracies or typographical errors. Changes are periodically made
to the information herein; these changes will be incorporated in new editions of the publication. IBM may make
improvements and/or changes in the product(s) and/or the program(s) described in this publication at any time
without notice.

Any references in this information to non-IBM websites are provided for convenience only and do not in any
manner serve as an endorsement of those websites. The materials at those websites are not part of the
materials for this IBM product and use of those websites is at your own risk.

IBM may use or distribute any of the information you provide in any way it believes appropriate without
incurring any obligation to you.

The performance data and client examples cited are presented for illustrative purposes only. Actual
performance results may vary depending on specific configurations and operating conditions.

Information concerning non-IBM products was obtained from the suppliers of those products, their published
announcements or other publicly available sources. IBM has not tested those products and cannot confirm the
accuracy of performance, compatibility or any other claims related to non-IBM products. Questions on the
capabilities of non-IBM products should be addressed to the suppliers of those products.

Statements regarding IBM’s future direction or intent are subject to change or withdrawal without notice, and
represent goals and objectives only.

This information contains examples of data and reports used in daily business operations. To illustrate them
as completely as possible, the examples include the names of individuals, companies, brands, and products.
All of these names are fictitious and any similarity to actual people or business enterprises is entirely
coincidental.

COPYRIGHT LICENSE:

This information contains sample application programs in source language, which illustrate programming
techniques on various operating platforms. You may copy, modify, and distribute these sample programs in
any form without payment to IBM, for the purposes of developing, using, marketing or distributing application
programs conforming to the application programming interface for the operating platform for which the sample
programs are written. These examples have not been thoroughly tested under all conditions. IBM, therefore,
cannot guarantee or imply reliability, serviceability, or function of these programs. The sample programs are
provided “AS IS”, without warranty of any kind. IBM shall not be liable for any damages arising out of your use
of the sample programs.

© Copyright IBM Corp. 2022. vii


Trademarks
IBM, the IBM logo, and ibm.com are trademarks or registered trademarks of International Business Machines
Corporation, registered in many jurisdictions worldwide. Other product and service names might be
trademarks of IBM or other companies. A current list of IBM trademarks is available on the web at “Copyright
and trademark information” at http://www.ibm.com/legal/copytrade.shtml

The following terms are trademarks or registered trademarks of International Business Machines Corporation,
and might also be trademarks or registered trademarks in other countries.
AIX® IBM Security® POWER9™
C3® IBM Spectrum® PowerHA®
Db2® IBM Watson® PowerPC®
DS8000® Instana® PowerVM®
Enterprise Storage Server® Interconnect® QRadar®
IBM® Micro-Partitioning® Redbooks®
IBM Cloud® PIN® Redbooks (logo) ®
IBM Cloud Pak® POWER® Storwize®
IBM Elastic Storage® Power Architecture® Turbonomic®
IBM FlashSystem® POWER8®

The following terms are trademarks of other companies:

Intel, Intel logo, Intel Inside logo, and Intel Centrino logo are trademarks or registered trademarks of Intel
Corporation or its subsidiaries in the United States and other countries.

The registered trademark Linux® is used pursuant to a sublicense from the Linux Foundation, the exclusive
licensee of Linus Torvalds, owner of the mark on a worldwide basis.

LTO, Ultrium, the LTO Logo and the Ultrium logo are trademarks of HP, IBM Corp. and Quantum in the U.S.
and other countries.

Ansible, OpenShift, Red Hat, are trademarks or registered trademarks of Red Hat, Inc. or its subsidiaries in
the United States and other countries.

VMware, and the VMware logo are registered trademarks or trademarks of VMware, Inc. or its subsidiaries in
the United States and/or other jurisdictions.

Other company, product, or service names may be trademarks or service marks of others.

viii IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Preface

This IBM® Redpaper™ publication is a comprehensive guide that covers the


IBM Power® S1014 (9105-41B), IBM Power S1022s (9105-22B), IBM Power S1022
(9105-22A), and IBM Power S1024 (9105-42A) servers that use the latest IBM Power10
processor-based technology and support the IBM AIX®, IBM i, and Linux operating systems.

The goal of this paper is to provide a hardware architecture analysis and highlight the
changes, new technologies, and major features that are introduced in these systems, such as
the following examples:
򐂰 The latest IBM Power10 processor design, including the Dual Chip Module (DCM) and
Entry Single Chip Module (eSCM) packaging, which is available in various configurations
of 4 - 24 cores per socket.
򐂰 Native Peripheral Component Interconnect® Express (PCIe) 5th generation (Gen5)
connectivity from the processor socket to deliver higher performance and bandwidth for
connected adapters.
򐂰 Open Memory Interface (OMI) connected differential DIMM (DDIMM) memory cards that
deliver increased performance, resilience, and security over industry-standard memory
technologies, including the implementation of transparent memory encryption.
򐂰 Enhanced internal storage performance with the use of native PCIe connected
Non-Volatile Memory Express (NVMe) devices in up to 16 internal storage slots to deliver
up to 102.4 TB of high-performance, low-latency storage in a single, two-socket system.
򐂰 Consumption-based pricing in the Power Private Cloud with Shared Utility Capacity
commercial model that allows customers to use resources more flexibly and efficiently,
including IBM AIX, IBM i, Red Hat Enterprise Linux, SUSE Linux Enterprise Server, and
Red Hat OpenShift Container Platform workloads.

This publication is intended for the following professionals who want to acquire a better
understanding of IBM Power server products:
򐂰 IBM Power customers
򐂰 Sales and marketing professionals
򐂰 Technical support professionals
򐂰 IBM Business Partners
򐂰 Independent software vendors (ISVs)

This paper expands the set of IBM Power documentation by providing a desktop reference
that offers a detailed technical description of the Power10 processor-based Scale Out server
models.

© Copyright IBM Corp. 2022. ix


Authors
This paper was produced by a team of specialists from around the world working at IBM
Redbooks, Poughkeepsie Center.

Giuliano Anselmi is an IBM® Power Digital Sales Technical Advisor in IBM Digital Sales
Dublin. He joined IBM and focused on Power processor-based technology. For almost 20
years, he covered several technical roles. He is an important resource for the mission of his
group and serves a reference for Business Partners and customers.

Young Hoon Cho is IBM Power Top Gun with the post-sales Technical Support Team for IBM
in Korea. He has over 10 years of experience working on IBM RS/6000, IBM System p, and
Power products. He provides second-line technical support to Field Engineers who are
working on IBM Power and system management.

Andrew Laidlaw is a Senior Power Technical Seller in the United Kingdom. He has 9 years of
experience in the IBM IT Infrastructure team, during which time he worked with the latest
technologies and developments. His areas of expertise include open source technologies,
including Linux and Kubernetes, open source databases, and artificial intelligence
frameworks, and tooling. His current focus area is on the Hybrid Cloud tools and capabilities
that support IBM customers in delivering modernization across their Power estate. He has
presented extensively on all of these topics across the world, including at the IBM Systems
Technical University conferences. He has been an author of many other IBM Redbooks®
publications.

Armin Röll works as a IBM Power IT specialist in Germany. He holds a degree in


Experimental Physics from the University of Hamburg, Germany. Armin has 25 years of
experience in IBM Power and AIX® pre-sales technical support. He has written several IBM
Redbooks publications about AIX and IBM Power systems hardware and is a Platinum
Redbooks author.

Tsvetomir Spasov is a IBM Power SME at IBM Bulgaria. His main area of expertise is FSP,
eBMC, HMC, POWERLC, and GTMS. He has been with IBM since 2016, providing reactive
break-fix, proactive, preventative, and cognitive support. He has conducted several technical
trainings and workshops.

The project that produced this material was managed by:

Scott Vetter
PMP, IBM Poughkeepsie, US

Thanks to the following people for their contributions to this project:

Ryan Achilles, Brian Allison, Ron Arroyo, Joanna Bartz, Bart Blaner, Gareth Coates,
Arnold Flores, Austin Fowler, George Gaylord, Douglas Gibbs, Nigel Griffiths,
Daniel Henderson, Markesha L Hill, Stephanie Jensen, Kyle Keaty,
Rajaram B Krishnamurthy, Charles Marino, Michael Mueller, Vincent Mulligan,
Hariganesh Muralidharan, Kaveh Naderi, Mark Nellen, Brandon Pederson,
Michael Quaranta, Hassan Rahimi, Ian Robinson, Todd Rosedahl, Bruno Spruth,
Nicole Schwartz, Bill Starke, Brian W. Thompto, Madhavi Valluri, Jacobo Vargas,
Madeline Vega, Russ Young
IBM

A special thanks to John Banchy for his relentless support of IBM Redbooks and his
contributions and corrections to them over the years.

x IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Now you can become a published author, too!
Here’s an opportunity to spotlight your skills, grow your career, and become a published
author—all at the same time! Join an IBM Redbooks residency project and help write a book
in your area of expertise, while honing your experience using leading-edge technologies. Your
efforts will help to increase product acceptance and customer satisfaction, as you expand
your network of technical contacts and relationships. Residencies run from two to six weeks
in length, and you can participate either in person or as a remote resident working from your
home base.

Find out more about the residency program, browse the residency index, and apply online at:
ibm.com/redbooks/residencies.html

Comments welcome
Your comments are important to us!

We want our papers to be as helpful as possible. Send us your comments about this paper or
other IBM Redbooks publications in one of the following ways:
򐂰 Use the online Contact us review Redbooks form found at:
ibm.com/redbooks
򐂰 Send your comments in an email to:
redbooks@us.ibm.com
򐂰 Mail your comments to:
IBM Corporation, IBM Redbooks
Dept. HYTD Mail Station P099
2455 South Road
Poughkeepsie, NY 12601-5400

Stay connected to IBM Redbooks


򐂰 Find us on LinkedIn:
http://www.linkedin.com/groups?home=&gid=2130806
򐂰 Explore new Redbooks publications, residencies, and workshops with the IBM Redbooks
weekly newsletter:
https://www.redbooks.ibm.com/Redbooks.nsf/subscribe?OpenForm
򐂰 Stay current on recent Redbooks publications with RSS Feeds:
http://www.redbooks.ibm.com/rss.html

Preface xi
xii IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
1

Chapter 1. IBM Power S1014, S1022s,


S1022, and S1024 overview
In this chapter, we introduce the IBM Power S1014 (9105-41B), IBM Power S1022s
(9105-22B), IBM Power S1022 (9105-22A), and IBM Power S1024 (9105-42A) server
models. These servers are built upon the Power10 processor architecture to deliver a
uniquely designed platform that is designed to help securely and efficiently scale core
operational and AI applications in a Hybrid Cloud environment.

This chapter includes the following topics:


򐂰 1.1, “Introduction” on page 2
򐂰 1.2, “Systems overview” on page 3
򐂰 1.3, “Operating environment” on page 8
򐂰 1.4, “Physical package” on page 10
򐂰 1.5, “System features” on page 11
򐂰 1.6, “Minimum configurations” on page 15
򐂰 1.7, “PCIe adapter slots” on page 16
򐂰 1.8, “Operating system support” on page 19
򐂰 1.9, “Hardware Management Console overview” on page 28
򐂰 1.10, “IBM Power solutions” on page 33
򐂰 1.11, “IBM Power platform modernization” on page 37

© Copyright IBM Corp. 2022. 1


1.1 Introduction
The Power10 processor-based Scale Out servers use the capabilities of the latest Power10
processor technology to deliver unprecedented security, reliability, and manageability for your
cloud and cognitive strategy and delivers industry-leading price and performance for your
mission-critical workloads.

The inclusion of PCIe Gen5 interconnects allows for high data transfer rates to provide higher
I/O performance or consolidation of the I/O demands of the system to fewer adapters running
at higher rates. This situation can result in greater system performance at a lower cost,
particularly when I/O demands are high.

The Power S1022s and S1022 servers deliver the performance of the Power10 processor
technology in a dense 2U (EIA units), rack-optimized form factor that is ideal for consolidating
multiple workloads with security and reliability. These systems are ready for hybrid cloud
deployment, with Enterprise grade virtualization capabilities built in to the system firmware
with the PowerVM hypervisor.

Figure 1-1 shows the Power S1022 server. The S1022s chassis is physically the same as the
S1022 server.

Figure 1-1 The Power S1022 is a 2U rack mount server

The Power S1014 server provides a powerful single-socket server that can be delivered in a
4U (EIA units) rack-mount form factor or as a desk-side tower model. It is ideally suited to the
modernization of IBM i, AIX, and Linux workloads to allow them to benefit from the
performance, security, and efficiency of the Power10 processor technology. This server easily
integrates into an organization’s cloud and cognitive strategy and delivers industry-leading
price and performance for your mission-critical workloads.

The Power S1024 server is a powerful one- or two-socket server that includes up to 48
Power10 processor cores in a 4U (EIA units) rack-optimized form factor that is ideal for
consolidating multiple workloads with security and reliability. With the inclusion of PCIe Gen5
connectivity and PCIe attached NVMe storage, this server maximizes the throughput of data
across multiple workloads to meet the requirements of modern hybrid cloud environments.

2 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Figure 1-2 show the Power S1024 server.

Figure 1-2 The Power S1024 is a 4U rack mount server

1.2 Systems overview


The following sections provide more information about the Power S1014 (9105-41B), Power
S1022s (9105-22B), Power S1022 (9105-22A), and Power S1024 (9105-42A) servers.

1.2.1 Power S1014 server


The Power S1014 (9105-41B) server is a powerful one-socket server that is available with a
4-core Power10 processor that is running at a typical 3.0 - 3.90 GHz (maximum) or an 8-core
Power10 processor that is running at a typical 3.0 - 3.90 GHz (maximum). All processor cores
can run up to eight simultaneous workload threads to deliver greater throughput.

This system is available in a rack-mount (4U EIA units) form factor, or as a desk-side tower
configuration, which offers flexibility in deployment models.

The Power S1014 server includes eight Differential DIMM (DDIMM) memory slots, each of
which can be populated with a DDIMM that is connected by using the new Open Memory
Interface (OMI). These DDIMMs incorporate DDR4 memory chips while delivering increased
memory bandwidth of up to 204 GBps peak transfer rates.

They also support transparent memory encryption to provide increased data security with no
management setup and no performance impact. The system supports up to 1 TB memory
capacity with the 8-core processor installed, with a minimum requirement of 32 GB memory
installed. The maximum memory capacity with the 4-core processor installed is 64 GB.

Note: The 128 GB DDIMMs will be made available on 18 November 2022; until that date,
the maximum memory capacity of an 8-core S1014 server is 512 GB.

Active Memory Mirroring is not supported on the Power S1014 model.

The Power S1014 server includes five usable PCIe adapter slots, four of which support PCIe
Gen5 adapters, while the fifth is a PCIe Gen4 adapter slot. These slots can be populated with
a range of adapters that cover LAN, Fibre Channel, SAS, USB, and cryptographic
accelerators. At least one network adapter must be included in each system. The 8-core
model can deliver more PCIe adapter slots through the addition of a PCIe Expansion drawer
(#EMX0) for a maximum of 10 PCIe adapter slots.

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 3


Internal storage for the Power S1014 is exclusively NVMe based, which connects directly into
the system PCIe lanes to deliver high performance and efficiency. A maximum of 16 U.2
form-factor NVMe devices can be installed, which offers a maximum storage capacity of
102.4 TB in a single server. More hard disk drive (HDD) or solid-state device (SSD) storage
can be connected to the 8-core system by way of SAS expansion drawers (the EXP24SX) or
Fibre Channel connectivity to an external storage array.

Note: The 4-core Power S1014 model does not support the connection of PCIe expansion
or storage expansion drawers.

The Power S1014 server includes PowerVM Enterprise Edition to deliver virtualized
environments and to support a frictionless hybrid cloud experience. Workloads can run the
AIX, IBM i, and Linux operating systems, including Red Hat OpenShift Container Platform.

1.2.2 Power S1022s server


The Power S1022s (9105-22B) server is a powerful one- or two-socket server that is available
with one or two processors per system, with an option of a 4-core eSCM Power10 processor
running at a typical 3.0 - 3.90 GHz (maximum) or an 8-core eSCM Power10 processor
running at a typical 3.0 - 3.90 GHz (maximum). All processor cores can run up to eight
simultaneous threads to deliver greater throughput. When two sockets are populated, both
must be the same processor model.

This system is a rack-mount (2U EIA units) form factor with an increased depth over previous
2U Power servers. A rack extension is recommended when installing the Power S1022s
server into an IBM Enterprise S42 rack.

The Power S1022s server includes 16 DDIMM memory slots, of which eight are usable when
only one processor socket is populated. Each of the memory slots can be populated with a
DDIMM that is connected by using the new OMI.

These DDIMMs incorporate DDR4 memory chips while delivering increased memory
bandwidth of up to 409 GBps peak transfer rates per socket. They also support transparent
memory encryption to provide increased data security with no management setup and no
performance impact.

The system supports up to 2 TB memory capacity with both sockets populated, with a
minimum requirement of 32 GB installed per socket.

Note: The 128 GB DDIMMs will be made available on 18 November 2022; until that date,
the maximum memory capacity of an S1022s server is 1 TB.

Active Memory Mirroring is available as an option to enhance resilience by mirroring critical


memory that is used by the PowerVM hypervisor.

The Power S1022s server includes 10 usable PCIe adapter slots, of which five are usable
when only one processor socket is populated. Eight of the PCIe adapter slots support PCIe
Gen5 adapters, while the remaining two (one per socket) are PCIe Gen4 adapter slots. These
slots can be populated with a range of adapters covering LAN, Fibre Channel, SAS, USB, and
cryptographic accelerators. At least one network adapter must be included in each system.

4 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
A system with one socket that is populated can deliver more PCIe adapter slots through the
addition of a PCIe expansion drawer (#EMX0) for a maximum of 10 PCIe adapter slots. A
system with two sockets that are populated can support up to 30 PCIe adapters with the
addition of PCIe expansion drawers.

Internal storage for the Power S1022s is exclusively NVMe-based, which connects directly
into the system PCIe lanes to deliver high performance and efficiency. A maximum of eight
U.2 form-factor NVMe devices can be installed, which offers a maximum storage capacity of
51.2 TB in a single server. More HDD or SSD storage can be connected to the 8-core system
by way of SAS expansion drawers (the EXP24SX) or Fibre Channel connectivity to an
external storage array.

The Power S1022s server includes PowerVM Enterprise Edition to deliver virtualized
environments and to support a frictionless hybrid cloud experience. Workloads can run the
AIX, IBM i, and Linux operating systems, including Red Hat OpenShift Container Platform.

Multiple IBM i partitions are supported to run on the Power S1022s server with the 8-core
processor feature, but each partition is limited to a maximum of four cores. These partitions
must use virtual I/O connections, and at least one VIOS partition is required. These partitions
can be run on systems that also run workloads that are based on the AIX and Linux operating
systems.

Note: The IBM i operating system is not supported on the Power S1022s model with
four-core processor option.

1.2.3 Power S1022 server


The Power S1022 (9105-22A) server is a powerful one- or two-socket server that is available
with one or two processors per system, with the following options:
򐂰 One or two 12-core Power10 processors running at a typical 2.90 - 4.0 GHz (maximum)
򐂰 Two 16-core Power10 processors running at a typical 2.75 - 4.0 GHz (maximum)
򐂰 Two 20 core Power10 processor running at a typical 2.45 - 3.90 GHz (maximum)

All processor cores can run up to eight simultaneous threads to deliver greater throughput.
When two sockets are populated, both must be the same processor model.

The Power S1022 supports Capacity Upgrade on Demand, where processor activations can
be purchased when they are required by workloads. A minimum of 50% of the installed
processor cores must be activated and available for use, with activations for the other
installed processor cores available to purchase as part of the initial order or as a future
upgrade. Static activations are linked only to the system for which they are purchased.

The Power S1022 server also can be purchased as part of a Power Private Cloud with
Shared Utility Capacity pool. In this case, the system can be purchased with one or more
base processor activations, which are shared within the pool of systems. More base
processor activations can be added to the pool in the future. A system with static activations
can be converted to become part of a Power Private Cloud with Shared Utility Capacity pool.

This system is a rack-mount (2U EIA units) form factor with an increased depth over previous
2U Power servers. A rack extension is recommended when installing the Power S1022 server
into an IBM Enterprise S42 rack.

The Power S1022 server includes 32 DDIMM memory slots, of which 16 are usable when
only one processor socket is populated.

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 5


Each of the memory slots can be populated with a DDIMM that is connected by using the new
OMI. These DDIMMs incorporate DDR4 memory chips while delivering increased memory
bandwidth of up to 409 GBps peak transfer rates per socket. They also support transparent
memory encryption to provide increased data security with no management setup and no
performance impact.

The system supports up to 4 TB memory capacity with both sockets populated, with a
minimum requirement of 32 GB installed per socket.

Note: The 128 GB DDIMMs will be made available on 18 November 2022; until that date,
the maximum memory capacity of an S1022 server is 2 TB.

Active Memory Mirroring is available as an option to enhance resilience by mirroring critical


memory that is used by the PowerVM hypervisor

The Power S1022 server includes 10 usable PCIe adapter slots, of which five are usable
when only one processor socket is populated. Eight of the PCIe adapter slots support PCIe
Gen5 adapters, while the remaining two (one per socket) are PCIe Gen4 adapter slots. These
slots can be populated with a range of adapters that covers LAN, Fibre Channel, SAS, USB,
and cryptographic accelerators. At least one network adapter must be included in each
system.

A system with one socket that is populated can deliver more PCIe adapter slots through the
addition of a PCIe expansion drawer (#EMX0) for a maximum of 10 PCIe adapter slots. A
system with two sockets that are populated can deliver up to 30 PCIe adapters with the
addition of PCIe expansion drawers.

Internal storage for the Power S1022 is exclusively NVMe based, which connects directly into
the system PCIe lanes to deliver high performance and efficiency. A maximum of eight U.2
form-factor NVMe devices can be installed, which offers a maximum storage capacity of
51.2 TB in a single server. More HDD or SSD storage can be connected to the system by
using SAS expansion drawers (the EXP24SX) or Fibre Channel connectivity to an external
storage array.

The Power S1022 server includes PowerVM Enterprise Edition to deliver virtualized
environments and to support a frictionless hybrid cloud experience. Workloads can run the
AIX, IBM i, and Linux operating systems, including Red Hat OpenShift Container Platform.

Multiple IBM i partitions are supported to run on the Power S1022 server, but each partition is
limited to a maximum of four cores. These partitions must use virtual I/O connections, and at
least one VIOS partition is required. These partitions can be run on systems that also run
workloads that are based on the AIX and Linux operating systems.

1.2.4 Power S1024 server


The Power S1024 (9105-42A) server is a powerful one- or two-socket server that is available
with one or two processors per system, with an option of a 12 core Power10 processor
running at a typical 3.40 - 4.0 GHz (maximum), a 16 core Power10 processor running at a
typical 3.10 - 4.0 GHz (maximum) or a 24 core Power10 processor running at a typical 2.75 -
3.90 GHz (maximum).

All processor cores can run up to eight simultaneous threads to deliver greater throughput.
When two sockets are populated, both must be the same processor model. A maximum of 48
Power10 cores are supported in a single system, which delivers up to 384 simultaneous
workload threads.

6 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
The Power S1024 supports Capacity Upgrade on Demand, where processor activations can
be purchased when they are required by workloads. A minimum of 50% of the installed
processor cores must be activated and available for use, with activations for the other
installed processor cores available to purchase as part of the initial order or as a future
upgrade. These static activations are linked only to the system for which they are purchased.

The Power S1024 server also can be purchased as part of a Power Private Cloud with
Shared Utility Capacity pool. In this case, the system can be purchased with one or more
base processor activations that are shared within the pool of systems. More base processor
activations can be added to the pool in the future. It is possible to convert a system with static
activations to become part of a Power Private Cloud with Shared Utility Capacity pool.

This system is a rack-mount (4U EIA units) form factor.

The Power S1024 server includes 32 DDIMM memory slots, of which 16 are usable when
only one processor socket is populated. Each of the memory slots can be populated with a
DDIMM that is connected by using the new OMI. These DDIMMs incorporate DDR4 memory
chips while delivering increased memory bandwidth of up to 409 GBps peak transfer rates
per socket.

They also support transparent memory encryption to provide increased data security with no
management setup and no performance impact. The system supports up to 8 TB memory
capacity with both sockets populated, with a minimum requirement of 32 GB installed per
socket.

Note: The 128 GB and 256 GB DDIMMs will be made available on November 2022; until
that date, the maximum memory capacity of an S1024 server is 2 TB.

Active Memory Mirroring is available as an option to enhance resilience by mirroring critical


memory that is used by the PowerVM hypervisor

The Power S1024 server includes 10 usable PCIe adapter slots, of which five are usable
when only one processor socket is populated. Eight of the PCIe adapter slots support PCIe
Gen5 adapters, while the remaining two (one per socket) are PCIe Gen4 adapter slots. These
slots can be populated with a range of adapters that covers LAN, Fibre Channel, SAS, USB,
and cryptographic accelerators. At least one network adapter must be included in each
system.

A system with one socket that is populated can deliver more PCIe adapter slots through the
addition of a PCIe expansion drawer (#EMX0) for a maximum of 10 PCIe adapter slots. A
system with two sockets that are populated can support up to 30 PCIe adapters with the
addition of PCIe expansion drawers.

Internal storage for the Power S1024 is exclusively NVMe-based, which connects directly into
the system PCIe lanes to deliver high performance and efficiency. A maximum of eight U.2
form-factor NVMe devices can be installed, which offers a maximum storage capacity of
102.4 TB in a single server. HDD or SSD storage can be connected to the system by using
SAS expansion drawers (the EXP24SX) or Fibre Channel connectivity to an external storage
array.

The Power S1024 server includes PowerVM Enterprise Edition to deliver virtualized
environments and support a frictionless hybrid cloud experience. Workloads can run the AIX,
IBM i, and Linux operating systems, including Red Hat OpenShift Container Platform.

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 7


1.3 Operating environment
The Power10 processor-based Scale Out servers are installed in industry-standard, 19-inch
racks that are housed in data centers. The exception is the Power S1014, which also is
available in a desk-side tower form factor.

Table 1-1 lists the electrical characteristics of the Power S1014, S1022s, S1022, and S1024
servers.

Table 1-1 Electrical characteristics for Power S1014, S1022s, S1022, and S1024 servers
Electrical Properties
characteristics
Power S1014 server Power S1022s server Power S1022 server Power S1024 server

Operating 1200 W power supply 1000 W power supply 2000 W power supply 1600 W power supply
voltage 100 - 127 V AC or 200 100 - 127 V AC or 200 - 200 - 240 V AC 200 - 240 V AC
- 240 V AC 240 V AC
or
1600 W power supply
200 - 240 V AC

Operating 50/60 ±3 Hz 50/60 ±3 Hz 50/60 ±3 Hz 50/60 ±3 Hz


frequency

Thermal output 3668 Btu/hour 7643 Btu/hour 7643 Btu/hour 9383 Btu/hour
(maximum) (maximum) (maximum) (maximum)

Power 1075 watts 2240 watts (maximum) 2240 watts 2750 watts
consumption (maximum) (maximum) (maximum)

Power-source 1.105 kVa (maximum 2.31 kVa (maximum 2.31 kVa (maximum 2.835 kVa (maximum
loading configuration) configuration) configuration) configuration)

Phase Single Single Single Single

Note: The maximum measured value is the worst-case power consumption that is
expected from a fully populated server under an intensive workload. The maximum
measured value also accounts for component tolerance and nonideal operating conditions.
Power consumption and heat load vary greatly by server configuration and utilization. The
IBM Systems Energy Estimator can be used to obtain a heat output estimate that is based
on a specific configuration.

8 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Table 1-2 lists the environment requirements for the Power10 processor-based Scale Out
servers.

Table 1-2 Environment requirements for Power S1014, S1022s, S1022 and S1024
Environment Recommended Allowable operating Nonoperating
operating

Temperature 18 - 27°C 5 - 40°C (41 - 104°F) 5 - 45°C (41 - 113°F)


(64.4 - 80.6°F)

Humidity range 9.0°C (48.2°F) dew 8% - 85% RH 8% - 85% RH


point (DP) to 60%
relative humidity (RH)
or 15°C (59°F) DP

Maximum DP N/A 24°C (75.2°F) 27°C (80.6°F)

Maximum operating N/A 3050 m (10000 ft) N/A


altitude

Table 1-3 lists the noise emissions for the systems.

Table 1-3 Noise emissions for Power S1014, S1022s, S1022 and S1024
Product Declared A-weighted sound Declared A-weighted sound
power level, LWAda (B)b pressure level, LpAm (dB)c

Operating Idle Operating Idle

Power S1014 (tower) 6.4 6.4 48 48

Power S1014 (rack) 7.5 7.5 62 62

Power S1022s 7.4 6.9 62 57

Power S1022 7.4 6.9 62 57

Power S1024 7.7 7.7 64 64


a. Declared level LWAd is the upper-limit A-weighted sound power level. Declared level LpAm is
the mean A-weighted emission sound pressure level that is measured at the 1-meter bystander
positions.
b. All measurements are made in conformance with ISO 7779 and declared in conformance with
ISO 9296.
c. 10 dB (decibel) equals 1 B (bel).

Important: NVMe PCIe adapters (EC5G, EC5B, EC5C, EC5D, EC5E, EC5F, EC6U, EC6V,
EC6W, EC6X, EC6Y, and EC6Z) require higher fan speeds to compensate for their higher
thermal output. This issue might affect the acoustic output of the server, depending on the
configuration. The e-config tool can be used to ensure a suitable configuration (log-in
required).

NVMe U.2 drives (ES1G, EC5V, ES1H, and EC5W) also require more cooling to
compensate for their higher thermal output, which might increase the noise emissions of
the servers.

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 9


1.4 Physical package
The Power S1014 server is available in rack-mount and tower form factors. The
Power S1022s, Power S1022, and Power S1024 servers are available in rack-mount form
factor only.

1.4.1 Tower model


Table 1-4 lists the physical dimensions of the Power S1014 tower chassis.

Table 1-4 Physical dimensions of the Power S1014 tower chassis


Dimension Power S1014 server (9105-41B)

Width 329 mm (13 in.)

Depth 815 mm (32 in.)

Height 522 mm (20.6 in.)

Weight 47.6 kg (105 lb)

1.4.2 Rack-mount models


Table 1-5 lists the physical dimensions of the Power S1022s and S1022 rack-mounted
chassis. These server models are available only in a rack-mounted form factor and take 2U of
rack space each.

Table 1-5 Physical dimensions of the Power S1022 rack-mounted chassis


Dimension Power S1022s server (9105-22B) Power S1022 server (9105-22A)

Width 482 mm (18.97 in.) 482 mm (18.97 in.)

Depth 813 mm (32 in.) 813 mm (32 in.)

Height 86.5 mm (3.4 in.) 86.5 mm (3.4 in.)

Weight 31.75 kg (70 lb) 31.75 kg (70 lb)

Figure 1-3 shows the front view of the Power S1022 server.

Figure 1-3 Front view of the Power S1022 server

10 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Table 1-6 lists the physical dimensions of the rack-mounted Power S1014 and Power S1024
chassis. The server is available only in a rack-mounted form factor and takes 4U of rack
space.

Table 1-6 Physical dimensions of the rack-mounted Power S1014 and Power S1024 chassis
Dimension Power S1014 server (9105-41B) Power S1024 server (9105-42A)

Width 482 mm (18.97 in.) 482 mm (18.97 in.)

Depth 712 mm (28 in.) 712 mm (28 in.)

Height 173 mm (6.48 in.) 173 mm (6.48 in.)

Weight 42.6 kg (94 lb) 42.6 kg (94 lb)

Figure 1-4 shows the front view of the Power S1024 server.

Figure 1-4 Front view of the Power S1024 server

1.5 System features


This section covers the standard system features that are included in the Power10
processor-based Scale Out server models. Each model has a different base feature set.

1.5.1 Power S1014 server features


The following standard features are available on the Power S1014 (9105-41B) server model:
򐂰 One entry single-chip module (eSCM) processor per system server:
– 3.0 - 3.90 GHz, 4-core Power10 processor
– 3.0 - 3.90 GHz, 8-core Power10 processor
򐂰 Processor core activation features available on a per-core basis: Static processor core
activations for all installed cores
򐂰 Up to 1 TB of system memory distributed across eight DDIMM slots per system server,
made up of one to eight DDR4 memory features. Each memory feature includes two
memory DDIMM parts:
– 32 GB (2 x 16 GB DDIMMs)
– 64 GB (2 x 32 GB DDIMMs)
– 128 GB (2 x 64 GB DDIMMs)
– 256 GB (2 x 128 GB DDIMMs)1

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 11


򐂰 Active Memory Mirroring for Hypervisor is available as an option to enhance resilience by
mirroring critical memory that is used by the PowerVM hypervisor
򐂰 PCIe slots:
– One x16 Gen4 or x8 Gen5 full-height, half-length slot (CAPI)
– Two x8 Gen5 full-height, half-length slots (with x16 connectors) (CAPI)
– One x8 Gen5 full-height, half-length slot (with x16 connector)
– One x8 Gen4 full-height, half-length slot (with x16 connector) (CAPI)
򐂰 Up to two storage backplanes each with eight NVMe U.2 drive slots:
– Up to 16 NMVe U.2 cards (800 GB, 1.6 TB, 3.2 TB)
– Optional internal RDX drive
򐂰 Integrated:
– Baseboard management/service processor
– EnergyScale technology
– Hot-swap and redundant cooling
– Redundant hot-swap AC power supplies
– One front and two rear USB 3.0 ports
– Two 1 GbE RJ45 ports for HMC connection
– One system port with RJ45 connector
– 19-inch rack-mounting hardware (4U)
򐂰 Optional PCIe I/O expansion drawer with PCIe slots on eight-core model only:
– One PCIe Gen3 I/O Expansion Drawer
– I/O drawer holds one 6-slot PCIe fan-out module
– Fanout module attaches to the system node through a PCIe optical cable adapter

1.5.2 Power S1022s server features


The following standard features are available on the Power S1022s (9105-22B) server model:
򐂰 One or two entry single-chip module (eSCM) processors per system server:
– One 3.0 - 3.90 GHz, 4-core Power10 processor
– One or two 3.0 - 3.90 GHz, 8-core Power10 processors
򐂰 Processor core activation features available on a per-core basis: Static processor core
activations for all installed cores
򐂰 Up to 2 TB of system memory that is distributed across 16 DDIMM slots per system
server, made up of one to eight DDR4 memory features per populated socket. Each
memory feature includes two memory DDIMM parts:
– 32 GB (2 x 16 GB DDIMMs)
– 64 GB (2 x 32 GB DDIMMs)
– 128 GB (2 x 64 GB DDIMMs)
– 256 GB (2 x 128 GB DDIMMs)2
򐂰 Active Memory Mirroring for Hypervisor is available as an option to enhance resilience by
mirroring critical memory used by the PowerVM hypervisor
򐂰 PCIe slots with a single processor socket populated:
– One x16 Gen4 or x8 Gen5 half-height, half-length slot (CAPI)
– Two x8 Gen5 half-height, half-length slots (with x16 connector) (CAPI)
– One x8 Gen5 half-height, half-length slot (with x16 connector)
– One x8 Gen4 half-height, half-length slot (with x16 connector) (CAPI)
1
The 128 GB DDIMMs will be made available on 18 November 2022
2 The 128 GB DDIMMs will be made available on 18 November 2022

12 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
򐂰 PCIe slots with two processor sockets populated:
– Two x16 Gen4 or x8 Gen5 half-height, half-length slots (CAPI)
– Two x16 Gen4 or x8 Gen5 half-height, half-length slots
– Two x8 Gen5 half-height, half-length slots (with x16 connectors) (CAPI)
– Two x8 Gen5 half-height, half-length slots (with x16 connectors)
– Two x8 Gen4 half-height, half-length slots (with x16 connectors) (CAPI)
򐂰 Up to two storage backplanes each with four NVMe U.2 drive slots: Up to eight NMVe U.2
cards (800 GB, 1.6 TB, 3.2 TB, and 6.4 TB)
򐂰 Integrated:
– Baseboard management/service processor
– EnergyScale technology
– Hot-swap and redundant cooling
– Redundant hot-swap AC power supplies
– One front and two rear USB 3.0 ports
– Two 1 GbE RJ45 ports for HMC connection
– One system port with RJ45 connector
– 19-inch rack-mounting hardware (2U)
򐂰 Optional PCIe I/O expansion drawer with PCIe slots on eight-core model only:
– Up to two PCIe Gen3 I/O Expansion Drawers
– Each I/O drawer holds up to two 6-slot PCIe fan-out modules
– Each fanout module attaches to the system node through a PCIe optical cable adapter

1.5.3 Power S1022 server features


The following standard features are available on the Power S1022 (9105-22A) server model:
򐂰 One or two dual-chip module (DCM) processors per system server:
– One or two 2.90 - 4.0 GHz, 12-core Power10 processors
– Two 2.75 - 4.0 GHz, 16-core Power10 processor
– Two 2.45 - 3.90 GHz, 20-core Power10 processor
򐂰 Processor core activation features available on a per-core basis. These exclusive options
cannot be mixed in a single server:
– Static processor core activations for all installed cores
– Capacity Upgrade on-Demand core activations for a minimum of half the installed
processor cores
– Base processor activations for Pools 2.0 for between one and all installed cores
򐂰 Up to 4 TB of system memory distributed across 32 DDIMM slots per system server,
which consist of 1 - 16 DDR4 memory features per populated socket. Each memory
feature includes two memory DDIMM parts:
– 32 GB (2 x 16 GB DDIMMs)
– 64 GB (2 x 32 GB DDIMMs)
– 128 GB (2 x 64 GB DDIMMs)
– 256 GB (2 x 128 GB DDIMMs)3
򐂰 Active Memory Mirroring for Hypervisor is available as an option to enhance resilience by
mirroring critical memory used by the PowerVM hypervisor

3 The 128 GB DDIMMs will be made available on 18 November 2022.

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 13


򐂰 PCIe slots with a single processor socket populated:
– One x16 Gen4 or x8 Gen5 half-height, half-length slot (CAPI)
– Two x8 Gen5 half-height, half-length slots (with x16 connector) (CAPI)
– One x8 Gen5 half-height, half-length slot (with x16 connector)
– One x8 Gen4 half-height, half-length slot (with x16 connector) (CAPI)
򐂰 PCIe slots with two processor sockets populated:
– Two x16 Gen4 or x8 Gen5 half-height, half-length slots (CAPI)
– Two x16 Gen4 or x8 Gen5 half-height, half-length slots
– Two x8 Gen5 half-height, half-length slots (with x16 connectors) (CAPI)
– Two x8 Gen5 half-height, half-length slots (with x16 connectors)
– Two x8 Gen4 half-height, half-length slots (with x16 connectors) (CAPI)
򐂰 Up to two storage backplanes each with four NVMe U.2 drive slots: Up to eight NMVe U.2
cards (800 GB, 1.6 TB, 3.2 TB, and 6.4 TB)
򐂰 Integrated:
– Baseboard management/service processor
– EnergyScale technology
– Hot-swap and redundant cooling
– Redundant hot-swap AC power supplies
– One front and two rear USB 3.0 ports
– Two 1 GbE RJ45 ports for HMC connection
– One system port with RJ45 connector
– 19-inch rack-mounting hardware (2U)
򐂰 Optional PCIe I/O expansion drawer with PCIe slots:
– Up to two PCIe Gen3 I/O Expansion Drawers
– Each I/O drawer holds up to two 6-slot PCIe fan-out modules
– Each fanout module attaches to the system node through a PCIe optical cable adapter

1.5.4 Power S1024 server features


The following standard features are available on the Power S1024 (9105-42A) server model:
򐂰 One or two dual-chip module (DCM) processors per system server:
– One or two 3.40 - 4.0 GHz, 12-core Power10 processors
– Two 3.10 - 4.0 GHz, 16-core Power10 processor
– Two 2.75 - 3.90 GHz, 24-core Power10 processor
򐂰 Processor core activation features available on a per-core basis. These exclusive options
cannot be mixed in a single server:
– Static processor core activations for all installed cores
– Capacity Upgrade on-Demand core activations for a minimum of half the installed
processor cores
– Base processor activations for Pools 2.0 for between one and all installed cores
򐂰 Up to 8 TB of system memory that is distributed across 32 DDIMM slots per system
server, made up of one to 16 DDR4 memory features per populated socket. Each memory
feature includes two memory DDIMM parts:
– 32 GB (2 x 16 GB DDIMMs)
– 64 GB (2 x 32 GB DDIMMs)
– 128 GB (2 x 64 GB DDIMMs)
– 256 GB (2 x 128 GB DDIMMs)4
– 512 GB (2 x 256 GB DDIMMs)4

14 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
򐂰 Active Memory Mirroring for Hypervisor is available as an option to enhance resilience by
mirroring critical memory used by the PowerVM hypervisor
򐂰 PCIe slots with a single processor socket populated:
– One x16 Gen4 or x8 Gen5 full-height, half-length slot (CAPI)
– Two x8 Gen5 full-height, half-length slots (with x16 connector) (CAPI)
– One x8 Gen5 full-height, half-length slot (with x16 connector)
– One x8 Gen4 full-height, half-length slot (with x16 connector) (CAPI)
򐂰 PCIe slots with two processor sockets populated:
– Two x16 Gen4 or x8 Gen5 full-height, half-length slots (CAPI)
– Two x16 Gen4 or x8 Gen5 full-height, half-length slots
– Two x8 Gen5 full-height, half-length slots (with x16 connectors) (CAPI)
– Two x8 Gen5 full-height, half-length slots (with x16 connectors)
– Two x8 Gen4 full-height, half-length slots (with x16 connectors) (CAPI)
򐂰 Up to two storage backplanes each with eight NVMe U.2 drive slots:
– Up to 16 NMVe U.2 cards (800 GB, 1.6 TB, 3.2 TB, and 6.4 TB)
– Optional internal RDX drive
򐂰 Integrated:
– Baseboard management/service processor
– EnergyScale technology
– Hot-swap and redundant cooling
– Redundant hot-swap AC power supplies
– One front and two rear USB 3.0 ports
– Two 1 GbE RJ45 ports for HMC connection
– One system port with RJ45 connector
– 19-inch rack-mounting hardware (2U)
򐂰 Optional PCIe I/O expansion drawer with PCIe slots:
– Up to two PCIe Gen3 I/O Expansion Drawers
– Each I/O drawer holds up to two 6-slot PCIe fan-out modules
– Each fanout module attaches to the system node through a PCIe optical cable adapter

1.6 Minimum configurations


The minimum Power_S1014, S1022s, S1022, and S1024 servers initial order must include
the following components:
򐂰 Processor module
򐂰 One LAN adapter
򐂰 At least two power supplies and power cords
򐂰 An operating system indicator
򐂰 An Operator-panel (required only for IBM i; optional for AIX/Linux)
򐂰 A cover set indicator
򐂰 Language Group Specify

The minimum initial order also must include one of the following memory options and one of
the following power supply options:
򐂰 Memory options:
– One processor module: Minimum of two DDIMMs (one memory feature)
– Two processor modules: Minimum of four DDIMMs (two memory features)
4 The 128 GB and 256 GB DDIMMs will be made available on 18 November 2022.

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 15


򐂰 Storage options:
– For boot from NVMe for AIX/Linux: One NVMe drive slot and one NVMe drive or one
PCIe NVMe add-in adapter.
– For boot from NVMe for IBM i: Two NVMe drive slots and two NVMe drives or two PCIe
NVMe add-in adapters.
– For boot from SAN: Internal NVMe drive and RAID card are not required if Feature
Code 0837 (boot from SAN) is selected. An FC adapter must be ordered if Feature
Code 0837 is selected.
򐂰 Power supply options (for more information, see 3.3, “Power supply features” on
page 119):
– S1022 and S1022s need two power supplies.
– S1014 needs two power supplies for the rack-mounted version; four power supplies for
the tower version.
– S1024 needs four power supplies.

1.7 PCIe adapter slots


The Power10 processor-based Scale Out server models each include 10 physical PCIe
adapter slots in the chassis. The number of slots that are available for use depends on the
number of processor sockets in the system that are populated; for example:
򐂰 With one processor socket populated, five of the PCIe adapter slots are available for use
򐂰 With two processor sockets populated, all 10 of the PCIe adapter slots can be used

These internal PCIe adapter slots support a range of different adapters. More information
about the available adapters, see 3.4, “Peripheral Component Interconnect adapters” on
page 119.

The adapter slots are a mix of PCIe Gen5 and PCIe Gen4 slots, with some running at x8
speed and others at x16. Some of the PCIe adapter slots also support OpenCAPI functions
when OpenCAPI is used enabled adapter cards. All PCIe adapter slots support hot-plug
capability when used with Hardware Management Console (HMC) or eBMC based
maintenance procedures.

Two other slots are available in the rear of each server. One of these slots is dedicated to the
eBMC management controller for the system, and the other is a dedicated slot for OpenCAPI
connected devices. These slots cannot be used for any other PCIe adapter type.

Each system requires at least one LAN adapter to support connection to local networks. This
requirement allows for initial system testing and configuration, and the preinstallation of any
operating systems, if required. By default, this server is the #5899 in the S1014 server, the
#EC2T in the S1022s or S1022 servers, or the #EC2U in the S1024 server. Alternative LAN
adapters can be installed instead. This required network adapter is installed by default in slot
C10.

16 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Table 1-7 lists the adapter slots that are available in the Power10 processor-based Scale Out
servers in various configurations.

Table 1-7 PCIe slot details for Power S1014, S1022s, S1022, and S1024 servers
Adapter slot Type Sockets populated OpenCAPI enabled

P0-C0 PCIe4 x16 or PCIe5 x8 2 only No


slots

P0-C1 PCIe4 x8 with x16 2 only Yes


connector

P0-C2 PCIe5 x8 with x16 2 only No


connector

P0-C3 PCIe4 x16 or PCIe5 x8 2 only Yes


slots

P0-C4 PCIe4 x16 or PCIe5 x8 2 only No


slots

P0-C5a eBMC All systems N/A

P0-C6b OpenCAPI x16 1 or 2 Yes


connector

P0-C7 PCIe5 x8 with x16 1 or 2 Yes


connector

P0-C8 PCIe4 x8 with x16 1 or 2 Yes


connector

P0-C9 PCIe5 x8 with x16 1 or 2 Yes


connector

P0-C10 PCIe4 x16 or PCIe5 x8 1 or 2 Yes


slots

P0-C11 PCIe5 x8 with x16 1 or 2 No


connector
a. Used for eBMC only.
b. Used for devices with OpenCAPI connections only.

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 17


The Power S1014 and S1024 servers are 4U (EIA units), and support the installation of
full-height PCIe adapters. Figure 1-5 shows the PCIe adapter slot locations for the
Power S1014 and S1024 server models.

Figure 1-5 PCIe adapter slot locations on the Power S1014 and S1024 server models

The Power S1022s and S1022 servers are 2U (EIA units), and support the installation of
low-profile PCIe adapters. Figure 1-6 shows the PCIe adapter slot locations for the
Power S1022s and S1022 server models.

Figure 1-6 PCIe adapter slot locations on the Power S1022sand S1022 server models

The total number of PCIe adapter slots available can be increased by adding PCIe Gen3 I/O
expansion drawers. With one processor socket populated (except the S1014 four core
option), one I/O expansion drawer that is installed with one fan-out module is supported.

When two processor sockets are populated, up to two I/O expansion drawers with up to four
fanout modules are supported. The connection of each fan out module in a PCIe Gen3
expansion drawer requires the installation of a PCIe optical cable adapter in one of the
internal PCIe x16 adapter slots (C0, C3®, C4, or C10).

For more information about the connectivity of the internal I/O bus and the PCIe adapter slots,
see 2.4, “Internal I/O subsystem” on page 92.

18 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
1.8 Operating system support
The Power10 processor-based Scale Out servers support the following families of operating
systems:
򐂰 AIX
򐂰 IBM i
򐂰 Linux

In addition, the Virtual I/O Server (VIOS) can be installed in special partitions that provide
virtualization of I/O capabilities, such as network and storage connectivity. Multiple VIOS
partitions can be installed to provide support and services to other partitions running AIX,
IBM i, or Linux, such as virtualized devices and Live Partition Mobility capabilities.

For more information about the Operating System and other software that is available on
Power, see this IBM Infrastructure web page.

The minimum supported levels of IBM AIX, IBM i, and Linux at the time of announcement are
described in the following sections. For more information about hardware features and
Operating System level support, see this IBM Support web page.

This tool helps to plan a successful system upgrade by providing the prerequisite information
for features that are in use or planned to be added to a system. A machine type and model
can be selected and the prerequisites, supported operating system levels and other
information can be determined.

The machine types and model for the Power10 processor-based Scale Out systems are listed
given in Table 1-8.

Table 1-8 Machine types and models of S1014, S1022s, S1022, and S1024 server models
Server name Machine type and model

S1014 9105-41B

S1022s 9105-22B

S1022 9105-22A

S1024 9105-42A

1.8.1 AIX operating system


At announcement, the Power S1014, S1022s, S1022, and S1024 servers support the
following minimum levels of the AIX operating system when installed with virtual I/O:
򐂰 AIX Version 7.3 with Technology Level 7300-00 and Service Pack 7300-00-01-2148
򐂰 AIX Version 7.2 with Technology Level 7200-05 and Service Pack 7200-05-01-2038
򐂰 AIX Version 7.2 with Technology Level 7200-04 and Service Pack 7200-04-02-2016
򐂰 AIX Version 7.1 with Technology Level 7100-05 and Service Pack 7100-05-06-2016

At announcement, the Power S1014, S1022s, S1022, and S1024 servers support the
following minimum levels of the AIX operating system when installed by using direct I/O
connectivity:
򐂰 AIX Version 7.3 with Technology Level 7300-00 and Service Pack 7300-00-02-2220
򐂰 AIX Version 7.2 with Technology Level 7200-05 and Service Pack 7200-05-04-2220
򐂰 AIX Version 7.2 with Technology Level 7200-04 and Service Pack 7200-04-06-2220

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 19


Note: AIX Version 7.2 with Technology Level 7200-04 and Service Pack 7200-04-06-2220
is planned to be available from 16 September 2022.

Consider the following points:


򐂰 AIX 7.1 instances must run in an LPAR in IBM POWER8® compatibility mode with
VIOS-based virtual storage and networking.
򐂰 AIX 7.2 instances can use physical and virtual I/O adapters, but must run in an LPAR in
IBM POWER9™ compatibility mode.
򐂰 AIX 7.3 instances can use physical and virtual I/O adapters, and can run in an LPAR in
native Power10 mode.

IBM periodically releases maintenance packages (service packs or technology levels) for the
AIX operating system. For more information about these packages, downloading, and
obtaining the installation packages, see this IBM Support Fix Central web page.

For more information about hardware features compatibility and the corresponding AIX
Technology Levels, see this IBM Support web page.

The Service Update Management Assistant (SUMA), which can help you automate the task
of checking and downloading operating system downloads, is part of the base operating
system. For more information about the suma command, see this IBM Documentation web
page.

The AIX Operating System can be licensed by using different methods, including the following
examples:
򐂰 Stand-alone as AIX Standard Edition
򐂰 With other software tools as part of AIX Enterprise Edition
򐂰 As part of the IBM Power Private Cloud Edition software bundle

Subscription licensing model


AIX Standard Edition also is available under a subscription licensing model that provides
access to an IBM program and IBM software maintenance for a specified subscription term
(one or three years). The subscription term begins on the start date and ends on the
expiration date, which is reflected in Entitled Systems Support (IBM Enterprise Storage
Server®).

Customers are licensed to run the product through the expiration date of the 1- or 3-year
subscription term. Then, they can renew at the end of the subscription to continue the use
other product. This model provides flexible and predictable pricing over a specific term, with
lower up-front costs of acquisition.

Another benefit of this model is that the licenses are customer-number entitled, which means
they are not tied to a specific hardware serial number as with a standard license grant.
Therefore, the licenses can be moved between on-premises and cloud if needed, something
that is becoming more of a requirement with hybrid workloads.

The subscription licenses are orderable through IBM configurator. The standard AIX license
grant and monthly term licenses for standard edition are still available.

20 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
1.8.2 IBM i operating system
At announcement, the Power S1014, S1022s, S1022, and S1024 servers support the
following minimum levels of IBM i:
򐂰 IBM i 7.5
򐂰 IBM i 7.4 Technology Release 6 or later
򐂰 IBM i 7.3 Technology Release 12 or later

Some limitations exist when running the IBM i operating system on the Power S1022s or
Power S1022 servers. Virtual I/O by way of VIOS is required, and partitions must be set to
“restricted I/O” mode.

The maximum size of the partition also is limited. Up to four cores (real or virtual) per IBM i
partition are supported. Multiple IBM i partitions can be created and run concurrently, and
individual partitions can have up to four cores.

Note: The IBM i operating system is not supported on the Power S1022s model with the
four-core processor option.

IBM periodically releases maintenance packages (service packs or technology releases) for
the IBM i. For more information about these packages, downloading, and obtaining the
installation packages, see this IBM Support Fix Central web page.

For more information about hardware feature compatibility and the corresponding IBM i
Technology Releases, see this IBM Support web page.

IBM i operating system transfer


IBM i customers can move IBM i entitlements to new Power10 processor-based Scale Out
servers, just as with previous new system, nonserial preserving upgrades or replacements if
the required criteria are met.

IBM i license terms and conditions require that IBM i operating system license entitlements
remain with the machine for which they were originally purchased. Under qualifying
conditions, IBM allows the transfer of IBM i processor and user entitlements from one
machine to another. This capability helps facilitate machine replacement, server
consolidation, and load rebalancing while protecting a customer’s investment in IBM i
software.

When requirements are met, IBM i license transfer can be configured by using IBM
configurator tools.

The following prerequisites must be met for transfers:


򐂰 The IBM i entitlements are owned by the user’s enterprise.
򐂰 The donor machine and receiving machine are owned by the user’s enterprise.
򐂰 The donor machine was owned in the same user’s enterprise as the receiving machine for
a minimum of one year.
򐂰 Current Software Maintenance (SWMA) is on the donor machine. Each software
entitlement that is to be transferred includes SWMA coverage.
򐂰 An electronic Proof of Entitlement (ePoE) exists for the entitlements to be transferred.
򐂰 The donor machine entitlements are for IBM i 5.4 or later.
򐂰 The receiving machine includes activated processors that are available to accommodate
the transferred entitlements.

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 21


A charge is incurred for the transfer of IBM i entitlements between servers. Each IBM i
processor entitlement that is transferred to a target machine includes one year of new SWMA
at no extra charge. Extra years of coverage or 24x7 support are available options for an extra
charge.

IBM i virtual serial numbers


IBM now offers a customer the ability to acquire a Virtual Serial Number (VSN) and assign it
to a logical partition. IBM i software can then be ordered against or transferred to the VSN
instead of being tied to a physical IBM Power machine’s serial number.

Having the IBM i entitlement, keys, and support entitlement on a VSN provides the flexibility
to move the partition to a different Power machine without transferring the entitlement.

Note: VSNs can be ordered in specific countries. For more information, see the local
announcement letters.

With VSNs, each partition can have its own serial number that is not tied to the hardware
serial number. If VSNs are not used, an IBM i partition still defaults to the use of the physical
host serial number.

In the first phase of VSN deployment, only one partition can use a single VSN at any time;
therefore, multiple IBM i LPARs cannot use the same VSN. In the first phase, VSNs are not
supported within Power Private Cloud (Power Enterprise Pools 2.0) environments.

VSNs are supported for partitions that are running any version of IBM i that is supported on
the Power10 processor-based Scale Out servers, although some other PTFs might be
required.

22 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
IBM i software tiers
The IBM i operating system is licensed per processor core that is used by the operating
system, and by the number of users that are interacting with the system. Different licensing
requirements depend on the capability of the server model and processor performance.
These systems are designated with a software tier that determines the licensing that is
required for workloads running on each server, as listed in Table 1-9.

Table 1-9 IBM i software tiers for the Power S1014, S1022s, S1022, and S1024
Server model Processor IBM i software tier

S1014 4-core (#EPG0) P05

S1014 8-core (#EPG2) P10

S1022s 4-core (#EPGR) N/Aa

S1022s 8-core (#EPGQ) P10b

S1022 12-core (#EPG9) P10b

S1022 16-core (#EPG8) P10b

S1022 20-core (#EPPGA) P10b

S1024 12-core (#EPGM) P20

S1024 16-core (#EPGC) P30

S1024 24-core (#EPGE) P30


a. IBM i does not support the 4-core processor (#EPGR) in the S1022s server.
b. IBM i partitions are limited to four cores on the S1022s and S1022 models.

1.8.3 Linux operating system distributions


Linux is an open source, cross-platform operating system that runs on numerous platforms
from embedded systems to mainframe computers. It provides an UNIX-like implementation
across many computer architectures.

The Linux distributions that are described next are supported on the Power S1014, S1022s,
S1022, and S1024 server models. Other distributions, including open source releases, can
run on these servers, but do not include any formal Enterprise Grade support.

Red Hat Enterprise Linux


The latest version of the Red Hat Enterprise Linux (RHEL) distribution from Red Hat is
supported in native Power10 mode, which allows it to access all of the features of the
Power10 processor and platform.

At announcement, the Power S1014, S1022s, S1022, and S1024 servers support the
following minimum levels of the Red Hat Enterprise Linux operating system:
򐂰 Red Hat Enterprise Linux 8.4 for Power LE, or later
򐂰 Red Hat Enterprise Linux for SAP with Red Hat Enterprise Linux 8.4 for Power LE, or later
򐂰 Red Hat Enterprise Linux 9.0 for Power LE, or later

Red Hat Enterprise Linux is sold on a subscription basis, with initial subscriptions and support
available for one, three, or five years. Support is available directly from Red Hat or IBM
Technical Support Services.

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 23


Red Hat Enterprise Linux 8 for Power LE subscriptions cover up to four cores and up to four
LPARs, and can be stacked to cover a larger number of cores or LPARs.

When you order RHEL from IBM, a subscription activation code is automatically published in
Enterprise Storage Server. After retrieving this code from Enterprise Storage Server, you use
it to establish proof of entitlement and download the software from Red Hat.

SUSE Linux Enterprise Server


The latest version of the SUSE Linux Enterprise Server distribution of Linux from SUSE is
supported in native Power10 mode, which allows it to access all of the features of the
Power10 processor and platform.

At announcement, the Power S1014, S1022s, S1022, and S1024 servers support the
following minimum levels of the SUSE Linux Enterprise Server operating system:
򐂰 SUSE Linux Enterprise Server 15 Service Pack 3, or later
򐂰 SUSE Linux Enterprise Server for SAP with SUSE Linux Enterprise Server 15 Service
Pack 3, or later

SUSE Linux Enterprise Server is sold on a subscription basis, with initial subscriptions and
support available for one, three, or five years. Support is available directly from SUSE or from
IBM Technical Support Services.

SUSE Linux Enterprise Server 15 subscriptions cover up to one socket or one LPAR, and can
be stacked to cover a larger number of sockets or LPARs.

When you order SLES from IBM, a subscription activation code is automatically published in
Enterprise Storage Server, you use it to establish proof of entitlement and download the
software from SUSE.

Linux and Power10 technology


The Power10 specific toolchain is available in the IBM Advance Toolchain for Linux version
15.0, which allows customers and developers to use all new Power10 processor-based
technology instructions when programming. Cross-module function call overhead was
reduced because of a new PC-relative addressing mode.

One specific benefit of Power10 technology is a 10x to 20x advantage over Power9
processor-based technology for AI inferencing workloads because of increased memory
bandwidth and new instructions. One example is the new special purpose-built matrix math
accelerator (MMA) that was tailored for the demands of machine learning and deep learning
inference. It also supports many AI data types.

Network virtualization is an area with significant evolution and improvements, which benefit
virtual and containerized environments. The following recent improvements were made for
Linux networking features on Power10 processor-based servers:
򐂰 SR-IOV allows virtualization of network adapters at the controller level without the need to
create virtual Shared Ethernet Adapters in the VIOS partition. It is enhanced with virtual
Network Interface Controller (vNIC), which allows data to be transferred directly from the
partitions to or from the SR-IOV physical adapter without transiting through a VIOS
partition.
򐂰 Hybrid Network Virtualization (HNV) allows Linux partitions to use the efficiency and
performance benefits of SR-IOV logical ports and participate in mobility operations, such
as active and inactive Live Partition Mobility (LPM) and Simplified Remote Restart (SRR).
HNV is enabled by selecting a new migratable option when an SR-IOV logical port is
configured.

24 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Security
Security is a top priority for IBM and our distribution partners. Linux security on IBM Power is
a vast topic; however, improvements in the areas of hardening, integrity protection,
performance, platform security, and certifications are introduced with this section.

Hardening and integrity protection deal with protecting the Linux kernel from unauthorized
tampering while allowing upgrading and servicing to the kernel. These topics become even
more important when in a containerized environment is run with an immutable operating
system, such as RHEL CoreOS, as the underlying operating system for the Red Hat
OpenShift Container Platform.

Performance also is an important security topic because specific hardening mitigation


strategies (for example, against side-channel attacks) can have a significant performance
effect. In addition, cryptography can use significant compute cycles.

1.8.4 Red Hat OpenShift Container Platform


The Red Hat OpenShift Container Platform (OCP) is supported to run on IBM Power servers,
including the Power10 processor-based Scale Out server models. OCP is a container
orchestration and management platform that provides a resilient and flexible environment in
which to develop and deploy applications. It extends the open source Kubernetes project to
provide an Enterprise grade platform to run and manage workloads that uses Linux container
technology.

A Red Hat OpenShift Container Platform cluster consists of several nodes, which can run on
physical or virtual machines. A minimum of three control plane nodes are required to support
the cluster management functions. At least two compute nodes are required to provide the
capacity to run workloads. During installation, another bootstrap node is required to host the
files that are required for installation and initial setup.

The bootstrap and control plane nodes are all based on the Red Hat Enterprise Linux
CoreOS operating system, which is a minimal immutable container host version of the Red
Hat Enterprise Linux distribution. It and inherits the associated hardware support statements.
The compute nodes can run on Red Hat Enterprise Linux or RHEL CoreOS.

The Red Hat OpenShift Container Platform is available on a subscription basis, with initial
subscriptions and support available for one, three, or five years. Support is available directly
from Red Hat or from IBM Technical Support Services. Red Hat OpenShift Container Platform
subscriptions cover two processor cores each, and can be stacked to cover many cores. Only
the compute nodes require subscription coverage.

At announcement, the Power S1014, S1022s, S1022, and S1024 servers support the
following minimum levels of the operating systems that are supported for Red Hat OpenShift
Container Platform:
򐂰 Red Hat Enterprise Linux CoreOS 4.9 for Power LE, or later
򐂰 Red Hat Enterprise Linux 8.4 for Power LE, or later

Red Hat OpenShift Container Platform 4.9 for IBM Power is the minimum level of the Red Hat
OpenShift Container Platform on the Power10 processor-based Scale Out servers.

When you order Red Hat OpenShift Container Platform from IBM, a subscription activation
code automatically is published in Enterprise Storage Server, you use it to establish proof of
entitlement and download the software from Red Hat.

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 25


For more information about running Red Hat OpenShift Container Platform on IBM Power,
see this Red Hat OpenShift Documentation web page.

1.8.5 Virtual I/O Server


The minimum required level of VIOS for the Power S1014, S1022s, S1022 and S1024 server
models is VIOS 3.1.3.21 or later.

IBM regularly updates the VIOS code. For more information, see this IBM Fix Central web
page.

1.8.6 Entitled System Support


IBM’s Enterprise Storage Server web page is available to view and manage Power and
Storage software and hardware. In general, most products that are offered by IBM Systems
that are purchased through our IBM Digital Sales representatives or Business Partners are
accessed on this site when the IBM Configurator is used.

The site features the following three main sections:


򐂰 My entitled software: Activities are listed that are related to Power and Storage software,
including the ability to download licensed, free, and trial software media, place software
update orders, and manage software keys.
򐂰 My entitled hardware: Activities are listed related to Power and Storage hardware
including the ability to renew Update Access Keys, buy and use Elastic Capacity on
Demand, assign or buy credits for new and existing pools in a Power Private Cloud
environment (Enterprise Pools 2.0), download Storage Capacity on-Demand codes, and
manage Hybrid Capacity credits.
򐂰 My inventory: Activities related to Power and Storage inventory including the ability to
browse software license, software maintenance, and hardware inventory, manage
inventory retrievals by way of Base Composer or generate several types of reports.

1.8.7 Update Access Keys


Since the introduction of the Power8 processor-based servers, IBM uses the concept of an
Update Access Key (UAK) for each server.

When system firmware updates are applied to the system, the UAK and its expiration date are
checked. System firmware updates include a release date. If the release date for the firmware
updates is past the expiration date for the update access key when attempting to apply
system firmware updates, the updates are not processed.

As update access keys expire, they must be replaced by using the Hardware Management
Console (HMC) or the ASMI on the eBMC.

By default, newly delivered systems include an UAK that expires after three years. Thereafter,
the UAK can be extended every six months, but only if a current hardware maintenance
contract exists for that server. The contract can be verified on the Enterprise Storage Server
web page.

Checking the validity and expiration date of the current UAK can be done through the HMC or
eBMC graphical interfaces or command-line interfaces. However, the expiration date also can
be displayed by using the suitable AIX or IBM i command.

26 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
UAK expiration date by using AIX 7.1
In the case of AIX 7.1, use the following command:
lscfg -vpl sysplanar0 | grep -p "System Firmware"

The output is similar to the output that is shown in Example 1-1 (the Microcode Entitlement
Date represents the UAK expiration date).

Example 1-1 Output of the command to check UAK expiration date by way of AIX 7.1
$ lscfg -vpl sysplanar0 | grep -p "System Firmware"
System Firmware:
...
Microcode Image.............NL1020_035 NL1020_033 NL1020_035
Microcode Level.............FW1020.00 FW1020.00 FW1020.00
Microcode Build Date........20220527 20220527 20220527
Microcode Entitlement Date..20220515
Hardware Location Code......U9105.42A.XXXXXXX-Y1
Physical Location: U9105.42A.XXXXXXX-Y1

UAK expiration date by using AIX 7.2 and 7.3


In the case of AIX 7.2 and 7.3, the output is slightly different from AIX 7.1. Use the following
command:
lscfg -vpl sysplanar0 |grep -p "System Firmware"

The output is similar to the output that is shown in Example 1-2 (the Update Access Key Exp
Date represents the UAK expiration date).

Example 1-2 Output of the command to check UAK expiration date by way of AIX 7.2 and 7.3
$ lscfg -vpl sysplanar0 |grep -p "System Firmware"
System Firmware:
...
Microcode Image.............NL1020_035 NL1020_033 NL1020_035
Microcode Level.............FW1020.00 FW1020.00 FW1020.00
Microcode Build Date........20220527 20220527 20220527
Update Access Key Exp Date..20220515
Hardware Location Code......U9105.42A.XXXXXXX-Y1
Physical Location: U9105.42A.XXXXXXX-Y1

UAK expiration date by using IBM i


When IBM i is used as an operating system, the status of the UAK can be checked by using
the Display Firmware Status window.

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 27


If the update access key expired, proceed to the Enterprise Storage Server web page to
replace your update access key. Figure 1-7 shows the output in the IBM i 7.1 and 7.2
releases. In the 7.3 and 7.4 releases, the text changes to Update Access Key Expiration
Date. The line that is highlighted in Figure 1-7 is displayed whether the system is operating
system managed or HMC managed.

Figure 1-7 Display Firmware Status window

1.9 Hardware Management Console overview


The HMC can be a hardware or virtual appliance that can be used to configure and manage
your systems. The HMC connects to one or more managed systems and provides capabilities
for the following primary functions:
򐂰 Provide systems management functions, including the following examples:
– Power off
– Power on
– System settings
– Capacity on Demand
– Enterprise Pools
– Shared Processor Pools
– Performance and Capacity Monitoring
– Starting Advanced System Management Interface (ASMI) for managed systems
򐂰 Deliver virtualization management through support for creating, managing, and deleting
Logical Partitions, Live Partition Mobility, Remote Restart, configuring SRIOV, managing
Virtual IO Servers, dynamic resource allocation, and operating system terminals.
򐂰 Acts as the service focal point for systems and supports service functions, including call
home, dump management, guided repair and verify, concurrent firmware updates for
managed systems, and around-the-clock error reporting with Electronic Service Agent for
faster support.

28 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
򐂰 Provides appliance management capabilities for configuring network, users on the HMC,
and updating and upgrading the HMC.

We discuss the available HMC offerings next.

1.9.1 HMC 7063-CR2


The 7063-CR2 IBM Power HMC (see Figure 1-8) is a second-generation Power
processor-based HMC. It includes the following features:
򐂰 6-core Power9 130W processor chip
򐂰 64 GB (4x16 GB) or 128 GB (4x32 GB) of memory
򐂰 1.8 TB of internal disk capacity with RAID1 protection
򐂰 4-ports 1 Gbps Ethernet (RJ-45), 2-ports 10 Gbps Ethernet (RJ-45), two USB 3.0 ports
(front side) and two USB 3.0 ports (rear side), and 1 Gbps IPMI Ethernet (RJ-45)
򐂰 Two 900W power supply units
򐂰 Remote Management Service: IPMI port (OpenBMC) and Redfish application
programming interface (API)
򐂰 Base Warranty is 1-year 9x5 with available optional upgrades

A USB Smart Drive is not included.

Note: The recovery media for V10R1 is the same for 7063-CR2 and 7063-CR1.

Figure 1-8 HMC 7063-CR2

The 7063-CR2 is compatible with flat panel console kits 7316-TF3, TF4, and TF5.

Note: The 7316-TF3 and TF4 were withdrawn from marketing.

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 29


1.9.2 Virtual HMC
Initially, the HMC was sold only as a hardware appliance, including the HMC firmware
installed. However, IBM extended this offering to allow the purchase of the hardware
appliance or a virtual appliance that can be deployed on ppc64le architectures or x86
platforms.

Any customer with a valid contract can download the HMC from the Enterprise Storage
Server web page, or it can be included within an initial Power S1014, S1022s, S1022, or
S1024 order.

The virtual HMC supports the following hypervisors:


򐂰 On x86 processor-based servers:
– KVM
– Xen
– VMware
򐂰 On Power processor-based servers: IBM PowerVM®

The following minimum requirements must be met to install the virtual HMC:
򐂰 16 GB of Memory
򐂰 4 virtual processors
򐂰 2 network interfaces (maximum 4 allowed)
򐂰 1 disk drive (500 GB available disk drive)

For an initial Power S1014, S1022s, S1022 or S1024 order with the IBM configurator
(e-config), HMC virtual appliance can be found by selecting add software → Other System
Offerings (as product selections) and then:
򐂰 5765-VHP for IBM HMC Virtual Appliance for Power V10
򐂰 5765-VHX for IBM HMC Virtual Appliance x86 V10

For more information and an overview of the Virtual HMC, see this web page.

For more information about how to install the virtual HMC appliance and all requirements, see
this IBM Documentation web page.

1.9.3 Baseboard management controller network connectivity rules for


7063-CR2 HMC
The 7063-CR2 HMC features a baseboard management controller (BMC), which is a
specialized service processor that monitors the physical state of the system by using sensors.
OpenBMC that is used on 7063-CR2 provides a graphical user interface (GUI) that can be
accessed from a workstation that includes network connectivity to the BMC. This connection
requires an Ethernet port to be configured for use by the BMC.

The 7063-CR2 provides two network interfaces (eth0 and eth1) for configuring network
connectivity for BMC on the appliance.

Each interface maps to a different physical port on the system. Different management tools
name the interfaces differently. The HMC task Console Management → Console
Settings → Change BMC/IPMI Network Settings modifies only the Dedicated interface.

30 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
The BMC ports are listed in Table 1-10.

Table 1-10 BMC ports


Management tool Logical port Shared/Dedicated CR2 physical port

OpenBMC UI eth0 Shared eth0

OpenBMC UI eth1 Dedicated Management port only

ipmitool lan1 Shared eth0

ipmitool lan2 Dedicated Management port only

HMC task (change lan2 Dedicated Management port only


BMC/IPMI Network
settings

Figure 1-9 shows the BMC interfaces of the HMC.

Figure 1-9 BMC interfaces

The main difference is the shared and dedicated interface to the BMC can coexist. Each has
its own LAN number and physical port. Ideally, the customer configures one port, but both can
be configured. The rules for connecting Power Systems to the HMC remain the same as for
previous versions.

1.9.4 High availability HMC configuration


For the best manageability and redundancy, a dual HMC configuration is suggested. This
configuration can be two hardware appliances, or one hardware appliance and one virtual
appliance or two virtual appliances.

The following requirements must be met:


򐂰 Two HMCs are at the same version.
򐂰 The HMCs use different subnets to connect to the BMCs.
򐂰 The HMCs can communicate with the servers’ partitions over a public network to allow for
full synchronization and function.

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 31


1.9.5 HMC code level requirements for the Power S1014, S1022s, S1022, and
S1024
The minimum required HMC version for the Power S1014, S1022s, S1022, and S1024 are
V10R1.1020. V10R1 is supported on 7063-CR1, 7063-CR2, and Virtual HMC appliances
only. It is not supported on 7042 machine types. HMC with V10R1 cannot manage POWER7
processor-based systems.

An HMC that is running V10R1 M1020 includes the following features:


򐂰 HMC OS Secure Boot support for the 7063-CR2 appliance
򐂰 Ability to configure login retries and suspended time and support for inactivity expiration in
password policy
򐂰 Ability to specify HMC location and data replication for groups
򐂰 VIOS Management Enhancements:
– Prepare for VIOS Maintenance:
• Validation for redundancy for the storage and network that is provided by VIOS to
customer partitions
• Switch path of redundant storage and network to start failover
• Roll back to original configuration on failure of prepare
• Audit various validation and prepare steps performed
• Report any failure that is seen during prepare
– Command Line and Scheduled operations support for VIOS backup or restore VIOS
Configuration and SSP Configuration
– Option to back up or restore Shared Storage Pool configuration in HMC
– Options to import or export the backup files to external storage
– Option to fail over all Virtual NICs from one VIOS to another
򐂰 Support 128 and 256 MB LMB sizes
򐂰 Automatically choose the fastest network for LPM memory transfer
򐂰 HMC user experience enhancements:
– Usability and performance
– Help connect global search
– Quick view of serviceable events
– More progress information for UI wizards
򐂰 Allow LPM/Remote Restart when virtual optical device is assigned to a partition
򐂰 UAK support

32 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
򐂰 Scheduled operation function: In the Electronic Service Agent, a new feature that allows
customers to receive message alerts only if scheduled operations fail (see Figure 1-10).

Figure 1-10 HMC alert feature

Log retention of the HMC audit trail also is increased.

1.9.6 HMC currency


In recent years, cybersecurity emerged as a national security issue and an increasingly
critical concern for CIOs and enterprise IT managers.

The IBM Power processor-based architecture always ranked highly in terms of end-to-end
security, which is why it remains a platform of choice for mission-critical enterprise workloads.

A key aspect of maintaining a secure Power environment is ensuring that the HMC (or virtual
HMC) is current and fully supported (including hardware, software, and Power firmware
updates).

Outdated or unsupported HMCs represent a technology risk that can quickly and easily be
mitigated by upgrading to a current release.

1.10 IBM Power solutions


The Power10 processor-based, scale-out servers are delivered cloud-enabled with integrated
PowerVM Enterprise capabilities.

The IBM Power Private Cloud Edition V1.8 is a complete package that adds flexible licensing
models in the cloud era. It helps you to rapidly deploy multi-cloud infrastructures with a
compelling set of cloud-enabled capabilities. The Power Private Cloud Edition primarily
provides value for clients that use AIX and Linux on Power, with simplified licensing models
and advanced features.

The IBM Private Cloud Edition (5765-ECB) includes:


򐂰 IBM PowerSC 2.1
򐂰 IBM Cloud® PowerVC for Private Cloud
򐂰 IBM VM Recovery Manager DR
򐂰 IBM Cloud Management Console for IBM Power

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 33


If you use IBM AIX as a primary OS, a specific offering is available: IBM Private Cloud Edition
with AIX 7 1.8.0 (5765-CBA). The offering includes:
򐂰 IBM AIX 7.3 or IBM AIX 7.2
򐂰 IBM PowerSC 2.1
򐂰 IBM Cloud PowerVC for Private Cloud
򐂰 IBM VM Recovery Manager DR
򐂰 IBM Cloud Management Console for IBM Power

1.10.1 IBM PowerSC 2.1


IBM PowerSC 2.1 (5765-SC2) provides a security and compliance solution that is optimized
for virtualized environments on IBM Power running IBM PowerVM and IBM AIX, IBM i, or
Linux on Power. Security control and compliance are some of the key components that are
needed to defend virtualized data centers and a cloud infrastructure against evolving data
security threats.

The PowerSC 2.1 product contains the following enhancements:


򐂰 Blocklisting anti-virus feature to allow selective, on-demand hash searches across
endpoints that are managed through PowerSC.
򐂰 Linux on Intel support for PowerSC endpoints, including Multi-Factor Authentication on
Power (MFA).
򐂰 Single sign-on (SSO) support
Users can log in to PowerSC through SSO with their OpenID Connect (OIDC) Enterprise
identity provider and MFA, which enables integration with any application user interface
(UI).
򐂰 MFA support for RSA web API
User MFA includes RSA through the web API and no longer uses the access control entry
(ACE) protocol.
򐂰 User-defined alt-disk for technology level (TL) and service pack (SP) updates
Users can specify alt-disk through Trusted Network Connect (TNC) for TL and SP updates
on AIX endpoints.

For more information, see this Announcement letter.

1.10.2 IBM PowerSC Multi-Factor Authentication


IBM PowerSC Multi-Factor Authentication (MFA) provides alternative authentication
mechanisms for systems that are used with RSA SecurID-based authentication systems, and
certificate authentication options, such as Common Access Card (CAC) and Personal
Identification Verification (PIV) cards. IBM PowerSC MFA allows the use of alternative
authentication mechanisms instead of the standard password.

You can use IBM PowerSC MFA with various applications, such as Remote Shell (RSH),
Telnet, and Secure Shell (SSH).

IBM PowerSC Multi-Factor Authentication raises the level of assurance of your mission-critical
systems with a flexible and tightly integrated MFA solution for IBM AIX and Linux on Power
virtual workloads that are running on Power servers.

IBM PowerSC MFA is part of the PowerSC 2.1 software offering; therefore, it also is included
in the IBM Power Private Cloud Edition software bundle.

34 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
For more information, see this Announcement Letter.

1.10.3 IBM Cloud PowerVC for Private Cloud


IBM PowerVC for Private Cloud (PowerVC for Private Cloud) works with IBM Power
Virtualization Center to provide an end-to-end cloud solution. PowerVC for Private Cloud
allows you to provision workloads and manage virtual images.

With PowerVC for Private Cloud, you can perform several operations, depending on your role
within a project.

Administrators can perform the following tasks:


򐂰 Create projects and assign images to projects to give team-specific access to images
򐂰 Set policies on projects to specify default virtual machine expiration, and so on
򐂰 Authorize users to projects
򐂰 Create expiration policies to reduce abandoned virtual machines
򐂰 Specify which actions require approvals and approving requests
򐂰 Create, edit, delete, and deploy templates
򐂰 Deploy an image from a template
򐂰 Approve or deny user requests
򐂰 Perform life-cycle operations on virtual machines, such as capture, start, stop, delete,
resume, and resize
򐂰 Monitor usage (metering) data across the project or per user

Users can perform the following tasks on resources to which they are authorized. Some
actions might require administrator approval. When a user tries to perform a task for which
approval is required, the task moves to the request queue before it is performed (or rejected):
򐂰 Perform life-cycle operations on virtual machines, such as capture, start, stop, delete,
resume, and resize
򐂰 Deploy an image from a deploy template
򐂰 View and withdraw outstanding requests
򐂰 Request virtual machine expiration extension
򐂰 View their own usage data

PowerVC version 2.0.0 UI


IBM Power Virtualization Center version 2.0.0 introduces an all new user interface that is
based on the Carbon framework. Carbon is an open source design system from IBM for
products and digital experiences. With the IBM Design Language as its foundation, the
system consists of working code, design tools and resources, human interface guidelines,
and a vibrant community of contributors.

IBM Power Virtualization Center version 2.0.0 features a new UI, and many new features and
enhancements. IBM listens to our client requirements and implements them along with our
benchmark innovation strategies that take PowerVC to the next level every release.

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 35


Because IBM Power Virtualization Center is built on the OpenStack technology, you might
see some terminology in messages or other text that is not the same as you might see
elsewhere in PowerVC. Some terminology also might be used that can be different from what
you are used to seeing in other IBM Power software products.

Feature support for PowerVC editions


PowerVC offers different functions, depending on the edition that is installed.

IBM Cloud PowerVC for Private Cloud includes all the functions of the PowerVC Standard
Edition plus the following features:
򐂰 A self-service portal that allows the provisioning of new VMs without direct system
administrator intervention. An option is for policy approvals for the requests that are
received from the self-service portal.
򐂰 Templates can be deployed that simplify cloud deployments.
򐂰 Cloud management policies are available that simplify managing cloud deployments.
򐂰 Metering data is available that can be used for chargeback.

For more information, see this Announcement Letter.

1.10.4 IBM VM Recovery Manager DR


IBM VM Recovery Manager DR is an automated DR solution that enables Power users to
achieve low recovery times for both planned and unplanned outages. It introduces support for
more storage replication solutions and for an extra guest operating system, which broadens
the offering’s applicability to a wider range of client requirements.

IBM VM Recovery Manager DR offers support for:


򐂰 IBM i as a guest operating system, which adds to the current support for IBM AIX and
Linux
򐂰 IBM DS8000® Global Mirror
򐂰 IBM SAN Volume Controller and IBM Storwize® Metro and Global Mirror as used in IBM
FlashSystem® storage arrays
򐂰 EMC Symmetrix Remote Data Facility (SRDF) synchronous replication
򐂰 Hitachi Virtual Storage Platform (VSP) G1000 and Hitachi VSP G400
򐂰 Boot device selection for IBM Power8 and later processor-based systems

For more information, see this Announcement Letter.

36 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
1.10.5 IBM Cloud Management Console
IBM Cloud Management Console for Power (CMC) runs as a hosted service in the IBM
Cloud. It provides a view of the entire IBM Power estate that is managed by a customer,
covering traditional and private cloud deployments of workloads.

The CMC interface collates and presents information about the IBM Power hardware
environment and the virtual machines that are deployed across that infrastructure. The CMC
provides access to tools for:
򐂰 Monitor the status of your IBM Power inventory
򐂰 Access insights from consolidated logging across all workloads
򐂰 Monitor the performance and see use trends across the estate
򐂰 Perform patch planning for hardware, operating systems, and other software
򐂰 Manage the use and credits for a Power Private Cloud environment

Data is collected from on-premises HMC devices by using a secure cloud connector
component. This configuration ensures that the CMC provides accurate and current
information about your IBM Power environment.

For more information, see IBM Power Systems Private Cloud with Shared Utility Capacity:
Featuring Power Enterprise Pools 2.0, SG24-8478.

1.11 IBM Power platform modernization


Cloud capabilities are a prerequisite for the use of enterprise-level IT resources. In addition to
the Power Private Cloud Edition offering that is described in 1.10, “IBM Power solutions” on
page 33, a rich infrastructure around IBM Power is available to help modernize services with
the strategic initiatives of your business.

The most important products are:


򐂰 IBM Power Virtual Servers
򐂰 Red Hat OpenShift Container Platform for Power

1.11.1 IBM Power Virtual Servers


IBM Power Virtual Server (PowerVS) on IBM Cloud is an infrastructure as a service (IaaS)
offering that you can use to deploy a virtual server, also known as a logical partition (LPAR),
in a matter of minutes.

Power clients who often relied on an on-premises only infrastructure can now quickly and
economically extend their Power IT resources into the cloud. The use of IBM Power Virtual
Server on IBM Cloud is an alternative to the large capital expense or added risk when
replatforming and moving your essential workloads to another public cloud.

PowerVS on IBM Cloud integrates your IBM AIX and IBM i capabilities into the IBM Cloud
experience, which means you get fast, self-service provisioning, flexible management
on-premises and off, and access to a stack of enterprise IBM Cloud services all with
pay-as-you-use billing that lets you easily scale up and out.

You can quickly deploy an IBM Power Virtual Server on IBM Cloud instance to meet your
specific business needs. With IBM Power Virtual Server on IBM Cloud, you can create a
hybrid cloud environment that allows you to easily control workload demands.

Chapter 1. IBM Power S1014, S1022s, S1022, and S1024 overview 37


For more information, see this IBM Cloud Docs web page.

1.11.2 Red Hat OpenShift Container Platform for Power


Red Hat OpenShift Container Platform for Power (5639-OCP) provides a secure,
enterprise-grade platform for on-premises, private platform-as-a-service (PaaS) clouds on
IBM Power servers. It brings together industry-leading container orchestration from
Kubernetes, advanced application build and delivery automation, and Red Hat Enterprise
Linux certified containers for Power.

Red Hat OpenShift Container Platform brings developers and IT operations together with a
common platform. It provides applications, platforms, and services for creating and delivering
cloud-native applications and management so IT can ensure that the environment is secure
and available.

Red Hat OpenShift Container Platform for Power provides enterprises the same functions as
the Red Hat OpenShift Container Platform offering on other platforms. Key features include:
򐂰 A self-service environment for application and development teams.
򐂰 Pluggable architecture that supports a choice of container run times, networking, storage,
Continuous Integration/Continuous Deployment (CI-CD), and more.
򐂰 Ability to automate routine tasks for application teams.

Red Hat OpenShift Container Platform subscriptions are offered in two core increments that
are designed to run in a virtual guest.

For more information, see 5639-RLE Red Hat Enterprise Linux for Power, little endian V7.0.

1.11.3 Hybrid Cloud Management Edition


Many enterprises continue to develop hybrid cloud environments as they host many of their
core business workloads on-premises while creating or migrating newer workloads to public
cloud or private cloud environments.

Managing these environments can be a daunting task. Organizations need the right tools to
tackle the challenges that are posed by these heterogeneous environments to accomplish
their objectives.

Collectively, the capabilities that are listed in this section work well together to create a
consistent management platform between client data centers, public cloud providers, and
multiple hardware platforms (fully inclusive of IBM Power), all of which provide all of the
necessary elements for a comprehensive hybrid cloud platform.

The following capabilities are available:


򐂰 IBM Cloud Pak® for Watson AIOps
򐂰 IBM Observability by Instana®
򐂰 IBM Turbonomic®
򐂰 Red Hat Advanced Cluster Management for Kubernetes (RHACM)
򐂰 IBM Power servers
򐂰 IBM Power Virtual Server

38 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
2

Chapter 2. Architecture and technical


overview
This chapter describes the overall system architecture for the Power10 processor-based
scale-out servers IBM Power S1014 (9105-41B), IBM Power S1022s (9105-22B),
IBM Power S1022 (9105-22A), and IBM Power S1024 (9105-42A).

This chapter includes the following topics:


򐂰 2.1, “Overview” on page 40
򐂰 2.2, “IBM Power10 processor” on page 44
򐂰 2.3, “Memory subsystem” on page 75
򐂰 2.4, “Internal I/O subsystem” on page 92
򐂰 2.5, “Enterprise Baseboard Management Controller” on page 100

© Copyright IBM Corp. 2022. 39


2.1 Overview
All scale-out servers share a common system planar design that provides the physical
resources for the following features and options:
򐂰 Two sockets to support 1- or 2-socket configurations
򐂰 32 physical slots to accommodate memory modules of various capacity specifications
򐂰 10 physical slots to support PCI Express (PCIe) adapter cards
򐂰 One physical slot that is dedicated to the enterprise baseboard management controller
(eBMC)
򐂰 One physical internal connector to hold the system’s Trusted Platform Module (TPM) card
򐂰 Six open coherent accelerator processor interface (OpenCAPI) ports
򐂰 Two internal receptacles to hold the voltage regulator modules (VRMs) for configured
sockets

The Power10 processor-based scale-out servers introduce two new Power10 processor
module packages. System planar sockets of a scale-out server are populated with dual-chip
models (DCMs) or entry single chip modules (eSCMs).

The DCM module type combines two Power10 processor chips in a tightly integrated unit and
each chip contributes core, memory interface, and PCIe interface resources.

The eSCM also consists of two Power10 chips, but differs from the DCM by the fact that core
and memory resources are provided by only one (chip-0) of the two chips. The other
processor chip (chip-1) on the eSCM only facilitates access to more PCIe interfaces,
essentially a switch.

Figure 2-1 on page 41 shows a logical diagram of Power S1022 and Power S1024 in a
2-socket DCM configuration.

40 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
P0-C39 DDR4 Differential DIMM OMI2B
P0-C38 DDR4 Differential DIMM OMI2A
P0-C40 DDR4 Differential DIMM OMI1B
P0-C21 DDR4 Differential DIMM OMI1A
P0-C20 DDR4 Differential DIMM OMI0B
P0-C19 DDR4 Differential DIMM OMI0A
P0-C29
P0-C28
P0-C23
P0-C27
P0-C13 DDR4 Differential DIMM OMI0B
P0-C12 DDR4 Differential DIMM OMI0A

P0-C42 DDR4 Differential DIMM OMI3B


P0-C41 DDR4 Differential DIMM OMI3A
P0-C30 DDR4 Differential DIMM OMI3B
P0-C31 DDR4 Differential DIMM OMI3A
DDR4
DDR4
DDR4
DDR4
Differential
Differential
Differential
Differential
DIMM
DIMM
DIMM
DIMM
OMI2B
OMI2A
OMI1B
OMI1A
8 OMI links 8 OMI links
@ 32 Gbps @ 32 Gbps
OpenCAPI x8 ports @ 25 Gbps OpenCAPI x8 port @ 25 Gbps

OpenCAPI SlimSAS connector OpenCAPI SlimSAS connector

OMI0
OMI0
OMI1

OMI3
OMI2
OMI1
OMI3
OMI2
OP3A OP3A
OpenCAPI SlimSAS connector OP3B
SMP 2 x9
C10 - PCIe Gen4 x16 or PCIe Gen5 x8 FHHL slot E1 OP4 OP7 E0 C4 - PCIe Gen4 x16 or PCIe Gen5 x8 FHHL slot
@ 32 Gbps
DCM-0 DCM-1
C11 - PCIe Gen5 x8 FHHL slot w/ x16 connectors P0 / Chip-0 P2 / Chip-0
E0 OP7 OP4 E1 C3 - PCIe Gen4 x16 or PCIe Gen5 x8 FHHL slot
C5
RJ45 (rear) 1 GbE Controller
OP2 OP6 OP2 OP6
BMC Display Port (rear)
eBMC
@ 32 Gbps

@ 32 Gbps

@ 32 Gbps
@ 32 Gbps
SMP 2 x9

SMP 2 x9

SMP 2 x9
SMP 2 x9
Chip
BMC USB / UPS (rear)

RJ45 (rear) 1 GbE Controller


USB 3.0 (rear) OP1 OP4 OP1 OP4

USB 3.0 (front)


Host USB 3.0 C2 - PCIe Gen5 x8 FHHL slot w/ x16 connectors
RDX (internal)
E1 OP7 OP6 E0
C7 - PCIe Gen5 x8 FHHL slot w/ x16 connectors C1 - PCIe Gen4 x8 FHHL slot w/ x16 connectors
DCM-0 DCM-1
P1 / Chip-1 P3 / Chip-1
C9 - PCIe Gen5 x8 FHHL slot w/ x16 connectors SMP 2 x9
E0 OP6 OP7 E1 C0 - PCIe Gen4 x16 or PCIe Gen5 x8 FHHL slot
C8 - PCIe Gen4 x8 FHHL slot w/ x16 connectors @ 32 Gbps

OP0A
OMI6
OMI5
OMI4
OMI7

OMI6
OMI5
OMI4
OMI7

C6 - OpenCAPI only FHHL slot w/ x16 connectors


OP0B OP0B
OpenCAPI SlimSAS connector OpenCAPI SlimSAS connector
OpenCAPI SlimSAS connector OpenCAPI x8 port @ 25 Gbps
8 OMI links 8 OMI links
OpenCAPI x8 ports @ 25 Gbps @ 32 Gbps @ 32 Gbps
P0-C18 DDR4 Differential DIMM OMI4A
P0-C17 DDR4 Differential DIMM OMI4B
P0-C34 DDR4 Differential DIMM OMI7A
P0-C33 DDR4 Differential DIMM OMI7B
OMI6A
OMI6B
OMI5A
OMI5B

P0-C44 DDR4 Differential DIMM OMI7A


P0-C45 DDR4 Differential DIMM OMI7B
P0-C47 DDR4 Differential DIMM OMI6A
P0-C46 DDR4 Differential DIMM OMI6B
P0-C48 DDR4 Differential DIMM OMI5A
P0-C43 DDR4 Differential DIMM OMI5B
P0-C25 DDR4 Differential DIMM OMI4A
P0-C26 DDR4 Differential DIMM OMI4B
DIMM
DIMM
DIMM
DIMM
Differential
Differential
Differential
Differential
DDR4
DDR4
DDR4
DDR4
P0-C37
P0-C36
P0-C16
P0-C35

Figure 2-1 Logical diagram of Power S1022 or Power S1024 servers in 2-socket configurations

Figure 2-1 also shows the following components:


򐂰 All interface ports (OP) that are used for intra-DCM and inter-DCM chip-to-chip
connections
򐂰 Open memory interfaces (OMIs), which includes the location codes for the associated 32
DIMM slots
򐂰 OpenCAPI connectors
򐂰 PCIe slots (C0 - C4 and C6 - C11) and their respective number of PCIe lanes
򐂰 eBMC card in slot C5 and all available 1 Gbps Ethernet and USB ports

The relevant busses and links are labeled with their respective speeds.

The logical diagram of Power S1022 and Power S1024 1-socket configurations can be
deduced by conceptually omitting the second socket (DCM-1). The number of memory slots
and PCIe slots is reduced by half if only one socket (DCM-0) is populated.

Chapter 2. Architecture and technical overview 41


The logical architecture of a 2-socket Power S1022s configuration is shown in Figure 2-2.

P0-C42 DDR4 Differential DIMM


P0-C41 DDR4 Differential DIMM
P0-C39 DDR4 Differential DIMM
P0-C38 DDR4 Differential DIMM
P0-C40 DDR4 Differential DIMM
P0-C21 DDR4 Differential DIMM
P0-C20 DDR4 Differential DIMM OMI0B
P0-C19 DDR4 Differential DIMM OMI0A
P0-C30 DDR4 Differential DIMM OMI3B
P0-C31 DDR4 Differential DIMM OMI3A
P0-C29 DDR4 Differential DIMM OMI2B
P0-C28 DDR4 Differential DIMM OMI2A
P0-C23 DDR4 Differential DIMM OMI1B
P0-C27 DDR4 Differential DIMM OMI1A
P0-C13 DDR4 Differential DIMM OMI0B
P0-C12 DDR4 Differential DIMM OMI0A

OMI3B
OMI3A
OMI2B
OMI2A
OMI1B
OMI1A
8 OMI links 8 OMI links
@ 32 Gbps @ 32 Gbps

OMI0
OMI0
OMI1

OMI3
OMI2
OMI1
OMI3
OMI2
SMP 2 x9
C10 - PCIe Gen4 x16 or PCIe Gen5 x8 FHHL slot E1 OP4 OP7 E0 C4 - PCIe Gen4 x16 or PCIe Gen5 x8 FHHL slot
@ 32 Gbps
eSCM-0 eSCM-1
C11 - PCIe Gen5 x8 FHHL slot w/ x16 connectors P0 / Chip-0 P2 / Chip-0
E0 OP7 OP4 E1 C3 - PCIe Gen4 x16 or PCIe Gen5 x8 FHHL slot
C5
RJ45 (rear) 1 GbE Controller
OP2 OP6 OP2 OP6
BMC Display Port (rear)
eBMC @ 32 Gbps

@ 32 Gbps
@ 32 Gbps

@ 32 Gbps
SMP 2 x9

SMP 2 x9
SMP 2 x9

SMP 2 x9
Chip
BMC USB / UPS (rear)

RJ45 (rear) 1 GbE Controller


USB 3.0 (rear) OP1 OP4 OP1 OP4

USB 3.0 (front)


Host USB 3.0 C2 - PCIe Gen5 x8 FHHL slot w/ x16 connectors
RDX (internal)
E1 OP7 OP6 E0
C7 - PCIe Gen5 x8 FHHL slot w/ x16 connectors eSCM-0 eSCM-1 C1 - PCIe Gen4 x8 FHHL slot w/ x16 connectors
Switch Switch
C9 - PCIe Gen5 x8 FHHL slot w/ x16 connectors P1 / Chip-1 SMP 2 x9 P3 / Chip-1
E0 OP6 OP7 E1 C0 - PCIe Gen4 x16 or PCIe Gen5 x8 FHHL slot
C8 - PCIe Gen4 x8 FHHL slot w/ x16 connectors @ 32 Gbps

Figure 2-2 Logical diagram of the Power S1022s server in a 2-socket configuration

Unlike the Power S1022 and Power S1024 servers, the sockets do not host DCM modules;
instead, they are occupied by eSCM models. This configuration implies that the number of
active memory interfaces decreases from 16 to 8, the number of available memory slots
decreases from 32 to 16, and all memory DDIMMs are connected to the first Power10 chip
(chip-0) of each eSCM.

Also, the eSCM-based systems do not support OpenCAPI ports. However, the PCIe
infrastructure of the Power S1022s is identical to PCIe layout of the DCM-based Power S1022
and Power S1024 servers and the number and the specification of the PCIe slots is the same.

The logical diagram of Power S1022s 1-socket configurations can be deducted by


conceptually omitting the second socket (eSCM-1) from the logical diagram. The number of
memory slots and the number of PCIe slots is reduced by half if only one socket (eSCM-0) is
populated.

42 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
By design, the Power S1014 is a 1-socket server is based on eSCM modules. As shown in
Figure 2-3, only four memory interfaces with the associated eight DDIMM slots are present to
provide main memory access and memory capacity. Also, the number of available PCIe slots
is reduced to five, which is half of the PCIe slots that are offered by Power10 scale-out servers
in 2-socket configurations.

P0-C30 DDR4 Differential DIMM OMI3B


P0-C31 DDR4 Differential DIMM OMI3A
P0-C29 DDR4 Differential DIMM OMI2B
P0-C28 DDR4 Differential DIMM OMI2A
P0-C23 DDR4 Differential DIMM OMI1B
P0-C27 DDR4 Differential DIMM OMI1A
P0-C13 DDR4 Differential DIMM OMI0B
P0-C12 DDR4 Differential DIMM OMI0A
8 OMI links
@ 32 Gbps

OMI0
OMI3
OMI2
OMI1
C10 - PCIe Gen4 x16 or PCIe Gen5 x8 FHHL slot E1
eSCM-0
C11 - PCIe Gen5 x8 FHHL slot w/ x16 connectors P0 / Chip-0
E0
C5
RJ45 (rear) 1 GbE Controller
OP2 OP6
BMC Display Port (rear)
eBMC
@ 32 Gbps

@ 32 Gbps
SMP 2 x9

SMP 2 x9
Chip
BMC USB / UPS (rear)

RJ45 (rear) 1 GbE Controller


USB 3.0 (rear) OP1 OP4

USB 3.0 (front)


Host USB 3.0
RDX (internal)
E1
C7 - PCIe Gen5 x8 FHHL slot w/ x16 connectors eSCM-0
Switch
C9 - PCIe Gen5 x8 FHHL slot w/ x16 connectors P1 / Chip-1
E0
C8 - PCIe Gen4 x8 FHHL slot w/ x16 connectors

Figure 2-3 Logical diagram of the Power S1014 server

Note: The bandwidths that are provided throughout the chapter are theoretical maximums
that are used for reference. The speeds that are shown are at an individual component
level. Multiple components and application implementation are key to achieving the best
performance. Always conduct the performance sizing at the application workload
environment level and evaluate performance by using real-world performance
measurements and production workloads.

Chapter 2. Architecture and technical overview 43


2.2 IBM Power10 processor
The IBM Power10 processor was introduced to the general public on 17 August 2020 at the
32nd HOT CHIPS1 semiconductor conference. At that meeting, the new capabilities and
features of the latest POWER processor microarchitecture and the Power Instruction Set
Architecture (ISA) v3.1B were revealed and categorized according to the following Power10
processor design priority focus areas:
򐂰 Data plane bandwidth
Terabyte per second signaling bandwidth on processor functional interfaces, petabyte
system memory capacities, 16-socket symmetric multiprocessing (SMP) scalability, and
memory clustering and memory inception capability.
򐂰 Powerful enterprise core
New core micro-architecture, flexibility, larger caches, and reduced latencies.
򐂰 End-to-end security
Hardware-enabled security features that are co-optimized with PowerVM hypervisor
support.
򐂰 Energy efficiency
Up to threefold energy-efficiency improvement in comparison to Power9 processor
technology.
򐂰 Artificial intelligence (AI)-infused core
A 10-20x matrix math performance improvement per socket compared to the Power9
processor technology capability.

The remainder of this section provides more specific information about the Power10
processor technology as it is used in the Power S1014, S1022s, S1022, and S1024 server.

The IBM Power10 Processor session material as presented at the 32nd HOT CHIPS
conference is available at this web page.

2.2.1 Power10 processor overview


The Power10 processor is a superscalar symmetric multiprocessor that is manufactured in
complimentary metal-oxide-semiconductor (CMOS) 7 nm lithography with 18 layers of metal.
The processor contains up to 15 cores that support eight simultaneous multithreading (SMT8)
independent execution contexts.

Each core has private access to 2 MB L2 cache and local access to 8 MB of L3 cache
capacity. The local L3 cache region of a specific core also is accessible from all other cores
on the processor chip. The cores of one Power10 processor share up to 120 MB of latency
optimized nonuniform cache access (NUCA) L3 cache.

1 https://hotchips.org/

44 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
The processor supports the following three distinct functional interfaces that all can run with a
signaling rate of up to 32 Gigatransfers per second (GTps):
򐂰 Open memory interface
The Power10 processor has eight memory controller unit (MCU) channels that support
one open memory interface (OMI) port with two OMI links each2. One OMI link aggregates
eight lanes that are running at 32 GTps and connects to one memory buffer-based
differential DIMM (DDIMM) slot to access main memory.
Physically, the OMI interface is implemented in two separate die areas of eight OMI links
each. The maximum theoretical full-duplex bandwidth aggregated over all 128 OMI lanes
is 1 TBps.
򐂰 SMP fabric interconnect (PowerAXON)
A total of 144 lanes are available in the Power10 processor to facilitate the connectivity to
other processors in a symmetric multiprocessing (SMP) architecture configuration. Each
SMP connection requires 18 lanes, eight data lanes plus one spare lane per direction
(2 x(8+1)). In this way, the processor can support a maximum of eight SMP connections
with a total of 128 data lanes per processor. This configuration yields a maximum
theoretical full-duplex bandwidth aggregated over all SMP connections of 1 TBps.
The generic nature of the interface implementation also allows the use of 128 data lanes
to potentially connect accelerator or memory devices through the OpenCAPI protocols.
Also, it can support memory cluster and memory interception architectures.
Because of the versatile characteristic of the technology, it is also referred to as
PowerAXON interface (Power A-bus/X-bus/OpenCAPI/Networking3). The OpenCAPI and
the memory clustering and memory interception use cases can be pursued in the future
and as of this writing are not used by available technology products.
򐂰 PCIe Version 5.0 interface
To support external I/O connectivity and access to internal storage devices, the Power10
processor provides differential Peripheral Component Interconnect Express version 5.0
interface busses (PCIe Gen 5) with a total of 32 lanes.
The lanes are grouped in two sets of 16 lanes that can be used in one of the following
configurations:
– 1 x16 PCIe Gen 4
– 2 x8 PCIe Gen 4
– 1 x8, 2 x4 PCIe Gen 4
– 1 x8 PCIe Gen 5, 1 x8 PCIe Gen 4
– 1 x8 PCIe Gen 5, 2 x4 PCIe Gen 4

2
The OMI links are also referred to as OMI subchannels.
3 A-busses (between CEC drawers) and X-busses (within CEC drawers) provide SMP fabric ports.

Chapter 2. Architecture and technical overview 45


Figure 2-4 shows the Power10 processor chip with several functional units labeled. A total of
16 SMT8 processor cores are shown, but only 4-, 6-, 8-,10-, or 15-core processor options are
available for Power10 processor-based scale-out server configurations.

Figure 2-4 The Power10 processor chip

46 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Important Power10 processor characteristics are listed in Table 2-1.

Table 2-1 Summary of the Power10 processor chip and processor core technology
Technology Power10 processor

Processor die size 602 mm2

Fabrication technology 򐂰 CMOSa 7-nm lithography


򐂰 18 layers of metal

Maximum processor cores per chip 15

Maximum execution threads per core / chip 8/120

Maximum L2 cache core 2 MB

Maximum On-chip L3 cache per core / chip 8 MB/120 MB

Number of transistors 18 billion

Processor compatibility modes Support for Power ISAb of Power8 and Power9
a. Complimentary metal-oxide-semiconductor (CMOS)
b. Power instruction set architecture (Power ISA)

2.2.2 Processor modules for S1014, S1022s, S1022, and S1024 servers

For the Power10 processor-based scale-out servers, the Power10 processor is packaged as
a DCM or as an eSCM:
򐂰 The DCM contains two directly coupled Power10 processor chips (chip-0 and chip-1) plus
more logic that is needed to facilitate power supply and external connectivity to the
module.
򐂰 The eSCM is a special derivative of the DCM where all active compute cores run on the
first chip (chip-0) and the second chip (chip-1) contributes only extra PCIe connectivity,
essentially a switch:
– Power S1022 and the Power S1024 servers use DCM modules
– Power S1014 and the Power S1022s servers are based on eSCM technology

Chapter 2. Architecture and technical overview 47


Power10 dual-chip module
Figure 2-5 shows the logical diagram of the Power10 DCM.

216 A-bus/X-bus/OpenCAPI lanes to bottom of module

OP0
OP2
OP3
OP5

OP7
OP1
OP3
OP4

OP7
OP0

OP6
OP5
74.5 mm x 85.75 mm

OMI0 OMI4
OP2
64 OMI lanes 64 OMI lanes
OMI1 Power10 OP1 Power10 OMI5
to bottom to bottom
of module
OMI2 Chip-0 OP6 Chip-1 OMI6
of module
OMI3 OP4 OMI7

E0 E1 E0 E1

64 PCIe5 lanes
to bottom of module

Figure 2-5 Power10 dual-chip module logical diagram

A total of 36 X-bus lanes are used for two chip-to-chip, module internal connections. Each
connection runs at 32 GTps (32 Gbps) speed and bundles 18 lanes, eight data lanes plus one
spare lane per direction (2 x(8+1)).

In this way, the DCM’s internal total aggregated full-duplex bandwidth between chip-0 and
chip-1 culminates at 256 GBps.

The DCM internal connections are implemented by using the interface ports OP2 and OP6 on
chip-0 and OP1 and OP4 on chip-1:
򐂰 2 × 9 OP2 lanes of chip-0 connect to 2 x 9 OP1 lanes of chip-1
򐂰 2 × 9 OP6 lanes of chip-0 connect to 2 × 9 OP4 lanes of chip-1

In addition to the interface ports OP2 and OP6 on chip-0 and OP1 and OP4 on chip-1, the
DCM offers 216 A-bus/X-bus/OpenCAPI lanes that are grouped in 12 other interface ports:
򐂰 OP0, OP1, OP3, OP4, OP5, OP7 on chip-0
򐂰 OP0, OP2, OP3, OP5, OP6, OP7 on chip-1

In 2-socket configurations of the Power S1022 or Power S1024 server, the interface ports
OP4 and OP7 on chip-0 and OP6 and OP7 on chip-1 are used to implement direct
chip-to-chip SMP connections across the two DCM modules.

The interface port OP3 on chip-0 and OP0 on chip-1 implement OpenCAPI interfaces that are
accessible through connectors that are on the mainboard of Power S1022 and Power S1024
servers.

Note: Although the OpenCAPI interfaces likely can be used in the future, they are not used
by available technology products as of this writing.

48 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
The interface ports OP0, OP1, and OP5 on chip-0 and OP2, OP3, and OP5 on chip-1 are
physically present, but not used by DCMs in Power S1022 and Power S1024 servers. This
status is indicated by the dashed lines that are shown in Figure 2-1 on page 41.

In addition to the chip-to-chip DCM internal connections, the cross DCM SMP links, and the
OpenCAPI interfaces, the DCM facilitates eight open memory interface ports (OMI0 - OMI7)
with two OMI links each to provide access to the buffered main memory differential DIMMs
(DDIMMs):
򐂰 OMI0 - OMI3 of chip-0
򐂰 OMI4 - OMI7 of chip-1

Note: The OMI interfaces are driven by eight on-chip memory controller units (MCUs) and
are implemented in two separate physical building blocks that lie in opposite areas at the
outer edge of the Power10 processor chip. One MCU directly controls one OMI port.
Therefore, a total of 16 OMI ports (OMI0 - OMI7 on chip-0 and OMI0 - OMI7 on chip-1) are
physically present on a Power10 DCM. However, because the chips on the DCM are tightly
integrated and the aggregated memory bandwidth of eight OMI ports culminates at
1 TBps, only half of the OMI ports are active. OMI4 to OMI7 on chip-0 and OMI0 to OMI3 of
chip-1 are disabled.

Finally, the DCM also offers differential Peripheral Component Interconnect Express version
5.0 interface busses (PCIe Gen 5) with a total of 64 lanes. Every chip of the DCM contributes
32 PCIe Gen5 lanes, which are grouped in two PCIe host bridges (E0, E1) with 16 PCIe Gen5
lanes each:
򐂰 E0, E1 on chip-0
򐂰 E0, E1 on chip-1

Chapter 2. Architecture and technical overview 49


Figure 2-6 shows the physical diagram of the Power10 DCM. Interface ports that are not used
by Power S1022 and Power S1024 servers (OP0, OP1, and OP5 on chip-0 and OP2, OP3,
and OP5 on chip-1) are shown, but no specification labels are shown.

64 open memory interface


lanes (8 x8) 32 Gbps
2 x9 SMP 32 Gbps

1 x16 Gen4 or
2 x8 Gen4 or
OP4 OP1 1 x8, 2 x4 Gen4 or
AXON OMI [0:3] AX 1 x8 Gen5, 1 x8 Gen4 or

AXON

AXON
OP0
OP5
1 x8 Gen5, 2 x4 Gen4

PCIe
1 x16 Gen4 or

E0
Power10

I/O MISC
2 x8 Gen4 or
Chip-0 1 x8, 2 x4 Gen4 or

PCIe
E1
1 x8 Gen5, 1 x8 Gen4 or
2 x9 SMP 32 Gbps
1 x8 Gen5, 2 x4 Gen4

AXON
AXON
OP7

OP3
AXON OMI AX 2 x8 OpenCAPI 32 Gbps
OP6 OP2

OP4 OP1
AXON OMI AX 2 x8 OpenCAPI 32 Gbps
AXON

AXON
OP0
OP5

1 x16 Gen4 or
PCIe
E0
Power10
I/O MISC

2 x8 Gen4 or
1 x8, 2 x4 Gen4 or
Chip-1
PCIe
E1
1 x8 Gen5, 1 x8 Gen4 or
1 x8 Gen5, 2 x4 Gen4
AXON
AXON
OP7

OP3

AXON OMI [4:7] AX 1 x16 Gen4 or


2 x9 SMP 32 Gbps 2 x8 Gen4 or
OP6 OP2
1 x8, 2 x4 Gen4 or
1 x8 Gen5, 1 x8 Gen4 or
1 x8 Gen5, 2 x4 Gen4
2 x9 SMP 32 Gbps

64 open memory interface


lanes (8 x8) 32 Gbps

Figure 2-6 Power10 dual-chip module physical diagram

To conserve energy, the unused OMI on the lower side of chip-0 and on the upper side of
chip-1 are grounded on the DCM package. For the same reason, the interface ports OP0 and
OP1 in the upper right corner of chip-0 and OP2 and OP3 in the lower right corner of chip-1
are grounded on the system planar.

Entry single chip module


The Power S1014 and Power S1022s server are based on the eSCM processor package.
Figure 2-7 on page 51 shows the logical diagram of the Power10 eSCM.

The main differences between the eSCM and the DCM structure include the following
examples:
򐂰 All active cores are on chip-0 and no active cores are on chip-1.
򐂰 Chip-1 works with chip-0 as a switch to facilitate more I/O connections.
򐂰 All active OMI interfaces are on chip-0 and no active OMI interfaces on chip-1.
򐂰 No OpenCAPI connectors are supported through any of the interface ports.

The eSCM internal chip-to-chip connectivity, the SMP links across the eSCM in 2-socket
configurations, and the PCIe Gen5 bus structure are identical to the Power10 DCM
implementation.

50 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
As with the Power10 DCM 36 X-bus, lanes are used for two chip-to-chip connections. These
eSCM internal connections are implemented by the interface ports OP2 and OP6 on chip-0
and OP1 and OP4 on chip-1:
򐂰 2 × 9 OP2 lanes of chip-0 connect to 2 x 9 OP1 lanes of chip-1
򐂰 2 × 9 OP6 lanes of chip-0 connect to 2 × 9 OP4 lanes of chip-1

The eSCM module internal chip-to-chip links exhibit the theoretical maximum full-duplex
bandwidth of 256 GBps.

72 A-bus/X-bus/OpenCAPI lanes to bottom of module

OP6
OP7
OP4
OP7
74.5 mm x 85.75 mm

OMI0
OP2 Power10
64 OMI lanes OMI1 Power10 OP1
to bottom Switch
of module
OMI2 Chip-0 OP6
OP4 Chip-1
OMI3

E0 E1 E0 E1

64 PCIe5 lanes
to bottom of module

Figure 2-7 Power10 entry single chip module logical diagram

The Power S1014 servers are available only in 1-socket configurations and no other interface
ports than OP2 and OP6 on chip-0 and OP1 and OP4 on chip-1 are operational. The same
interface port constellation applies to 1-socket configurations of the Power S1022s server.
Figure 2-3 on page 43 shows the logical system diagram of the Power S1014 1-socket server
based on a single eSCM.

However, in 2-socket eSCM configurations of the Power S1022s server, the interface ports
OP4 and OP7 on chip-0 and OP6 and OP7 on chip-1 of the processor module are active and
used to implement direct chip-to-chip SMP connections between the two eSCM modules.

Figure 2-2 on page 42 shows logical system diagram of the Power S1022s 2-socket server
that is based on two eSCM modules. (The 1-socket constellation can easily be deduced from
Figure 2-2 on page 42 if eSCM-1 is conceptually omitted.)

As with the DCM, the eSCM offers differential PCIe Gen 5 with a total of 64 lanes. Every chip
of the eSCM contributes 32 PCIe Gen5 lanes, which are grouped in two PCIe host bridges
(E0, E1) with 16 PCIe Gen5 lanes each:
򐂰 E0, E1 on chip-0
򐂰 E0, E1 on chip-1

Chapter 2. Architecture and technical overview 51


Figure 2-8 shows the physical diagram of the Power10 entry single chip module. The most
important difference in comparison to the physical diagram of the Power10 DCM that is
shown in Figure 2-6 on page 50 is that chip-1 has no active cores or memory interfaces. Also,
because the eSCM does not support any OpenCAPI connectivity, the interface port OP3 on
chip-0 and OP0 on chip-1 are disabled.

64 open memory interface


lanes (8 x8) 32 Gbps

2 x9 SMP 32 Gbps

1 x16 Gen4 or
2 x8 Gen4 or
OP4 OP1 1 x8, 2 x4 Gen4 or
AXON OMI [0:3] AX 1 x8 Gen5, 1 x8 Gen4 or

OP0
AXON

AXON
OP5
1 x8 Gen5, 2 x4 Gen4

PCIe
1 x16 Gen4 or

E0
Power10

I/O MISC
2 x8 Gen4 or
Chip-0 1 x8, 2 x4 Gen4 or

PCIe
E1
1 x8 Gen5, 1 x8 Gen4 or
2 x9 SMP 32 Gbps
1 x8 Gen5, 2 x4 Gen4

AXON
AXON
OP7

OP3
AXON OMI AX

OP6 OP2

OP4 OP1
AXON OMI AX

OP0
AXON

AXON
OP5

Power10 1 x16 Gen4 or


PCIe
E0
I/O MISC

2 x8 Gen4 or
Switch 1 x8, 2 x4 Gen4 or
PCIe
E1

1 x8 Gen5, 1 x8 Gen4 or
Chip-1 1 x8 Gen5, 2 x4 Gen4
AXON
AXON

OP3
OP7

AXON OMI [4:7] AX 1 x16 Gen4 or


2 x9 SMP 32 Gbps 2 x8 Gen4 or
OP6 OP2
1 x8, 2 x4 Gen4 or
1 x8 Gen5, 1 x8 Gen4 or
1 x8 Gen5, 2 x4 Gen4

2 x9 SMP 32 Gbps

Figure 2-8 Power10 entry single chip module physical diagram

2.2.3 Power10 processor core


The Power10 processor core inherits the modular architecture of the Power9 processor core,
but the redesigned and enhanced micro-architecture that significantly increases the
processor core performance and processing efficiency.

The peak computational throughput is markedly improved by new execution capabilities and
optimized cache bandwidth characteristics. Extra matrix math acceleration engines can
deliver significant performance gains for machine learning, particularly for AI inferencing
workloads.

The Power10 processor-based scale-out servers uses the Power10 enterprise-class


processor variant in which each core can run with up to eight independent hardware threads.
If all threads are active, the mode of operation is referred to as 8-way simultaneous
multithreading (SMT8) mode. A Power10 core with SMT8 capability is named Power10 SMT8
core (SMT8 core). The Power10 core also supports modes with four active threads (SMT4),
two active threads (SMT2), and one single active thread (ST).

52 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
The SMT8 core includes two execution resource domains. Each domain provides the
functional units to service up to four hardware threads.

Figure 2-9 shows the functional units of an SMT8 core where all eight threads are active. The
two execution resource domains are highlighted with colored backgrounds in two different
shades of blue.

Figure 2-9 Power10 SMT8 core

Each of the two execution resource domains supports 1 - 4 threads and includes four vector
scalar units (VSU) of 128-bit width, two matrix math accelerator (MMA) units, and one
quad-precision floating-point (QP) and decimal floating-point (DF) unit.

One VSU and the directly associated logic are called an execution slice. Two neighboring
slices also can be used as a combined execution resource, which is then named super-slice.
When operating in SMT8 mode, eight SMT threads are subdivided in pairs that collectively
run on two adjacent slices, as indicated by colored backgrounds in different shades of green
in Figure 2-9.

In SMT4 or lower thread modes, one to two threads each share a four-slice resource domain.
Figure 2-9 also shows other essential resources that are shared among the SMT threads,
such as instruction cache, instruction buffer, and L1 data cache.

The SMT8 core supports automatic workload balancing to change the operational SMT
thread level. Depending on the workload characteristics, the number of threads that is running
on one chiplet can be reduced from four to two and even further to only one active thread. An
individual thread can benefit in terms of performance if fewer threads run against the core’s
executions resources.

Micro-architecture performance and efficiency optimization lead to a significant improvement


of the performance per watt signature compared with the previous Power9 core
implementation. The overall energy efficiency per socket is better by a factor of approximately
2.6, which demonstrates the advancement in processor design that is manifested by the
Power10 processor.

Chapter 2. Architecture and technical overview 53


The Power10 processor core includes the following key features and improvements that affect
performance:
򐂰 Enhanced load and store bandwidth
򐂰 Deeper and wider instruction windows
򐂰 Enhanced data prefetch
򐂰 Branch execution and prediction enhancements
򐂰 Instruction fusion

Enhancements in the area of computation resources, working set size, and data access
latency are described next. The change in relation to the Power9 processor core
implementation is provided in square parentheses.

Enhanced computation resources


The following computational resource enhancements are available:
򐂰 Eight vector scalar unit (VSU) execution slices, each supporting 64-bit scalar or 128-bit
single instructions multiple data (SIMD) [+100% for permute, fixed-point, floating-point,
and crypto (Advanced Encryption Standard (AES)/SHA) +400% operations].
򐂰 Four units for matrix math accelerator (MMA) acceleration each capable of producing a
512-bit result per cycle (new) [+400% Single and Double precision FLOPS plus support for
reduced precision AI acceleration].
򐂰 Two units for quad-precision floating-point and decimal floating-point operations instruction
types

Larger working sets


The following major changes were implemented in working set sizes:
򐂰 L1 instruction cache: 2 x 48 KB 6-way (96 KB total); +50%
򐂰 L2 cache: 2 MB 8-way; +400%
򐂰 L2 translation lookaside buffer (TLB): 2 x 4K entries (8K total); +400%

Data access with reduced latencies


The following major changes reduce latency for load data:
򐂰 L1 data cache access at four cycles nominal with zero penalty for store-forwarding;
(- 2 cycles) for store forwarding
򐂰 L2 data access at 13.5 cycles nominal (-2 cycles)
򐂰 L3 data access at 27.5 cycles nominal (-8 cycles)
򐂰 Translation lookaside buffer (TLB) access at 8.5 cycles nominal for effective-to-real
address translation (ERAT) miss, including for nested translation (-7 cycles)

Micro-architectural innovations that complement physical and logic design techniques and
specifically address energy efficiency include the following examples:
򐂰 Improved clock-gating
򐂰 Reduced flush rates with improved branch prediction accuracy
򐂰 Fusion and gather operating merging
򐂰 Reduced number of ports and reduced access to selected structures
򐂰 Effective address (EA)-tagged L1 data and instruction cache yield ERAT access on a
cache miss only

54 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
In addition to significant improvements in performance and energy efficiency, security
represents a major architectural focus area. The Power10 processor core supports the
following security features:
򐂰 Enhanced hardware support that provides improved performance while mitigating for
speculation-based attacks
򐂰 Dynamic Execution Control Register (DEXCR) support
򐂰 Return oriented programming (ROP) protection

2.2.4 Simultaneous multithreading


Each core of the Power10 processor supports multiple hardware threads that represent
independent execution contexts. If only one hardware thread is used, the processor core runs
in single-threaded (ST) mode.

If more than one hardware thread is active, the processor runs in SMT mode. In addition to
the ST mode, the Power10 processor core supports the following SMT modes:
򐂰 SMT2: Two hardware threads active
򐂰 SMT4: Four hardware threads active
򐂰 SMT8: Eight hardware threads active

SMT enables a single physical processor core to simultaneously dispatch instructions from
more than one hardware thread context. Computational workloads can use the processor
core’s execution units with a higher degree of parallelism. This ability significantly enhances
the throughput and scalability of multi-threaded applications and optimizes the compute
density for single-threaded workloads.

SMT is primarily beneficial in commercial environments where the speed of an individual


transaction is not as critical as the total number of transactions that are performed. SMT
typically increases the throughput of most workloads, especially those workloads with large or
frequently changing working sets, such as database servers and web servers.

Chapter 2. Architecture and technical overview 55


Table 2-2 lists a historic account of the SMT capabilities that are supported by each
implementation of the IBM Power Architecture® since Power4.

Table 2-2 SMT levels that are supported by IBM POWER® processors
Technology Maximum cores Supported hardware Maximum hardware
per system threading modes threads per partition

IBM Power4 32 ST 32

IBM Power5 64 ST, SMT2 128

IBM POWER6 64 ST, SMT2 128

IBM POWER7 256 ST, SMT2, SMT4 1024

IBM Power8 192 ST, SMT2, SMT4, SMT8 1536

IBM Power9 192 ST, SMT2, SMT4, SMT8 1536

IBM Power10 240 ST, SMT2, SMT4, SMT8 1920a


a. The PHYP hypervisor supports a maximum 240 x SMT8 = 1920. AIX supports up to 1920
(240xSMT8) total threads in a single partition, starting with AIX v7.3 on Power10 based
systems.

All Power10 processor-based scale-out servers support the ST, SMT2, SMT4, and SMT8
hardware threading modes. Table 2-3 lists the maximum hardware threads per partition for
each scale-out server model.

Table 2-3 Maximum hardware threads supported by Power10 processor-based scale-out servers
Server Maximum cores per system Maximum hardware threads
per partition

Power S1014 8 64

Power S1022s 16 128

Power S1022 40 320

Power S1024 48 384

2.2.5 Matrix math accelerator AI workload acceleration


The matrix math accelerator (MMA) facility was introduced by the Power Instruction Set
Architecture (ISA) v3.1. The related instructions implement numerical linear algebra
operations on small matrixes and are meant to accelerate computation-intensive kernels,
such as matrix multiplication, convolution, and discrete Fourier transform.

To efficiently accelerate MMA operations, the Power10 processor core implements a dense
math engine (DME) microarchitecture that effectively provides an accelerator for cognitive
computing, machine learning, and AI inferencing workloads.

The DME encapsulates compute efficient pipelines, a physical register file, and associated
data-flow that keeps resulting accumulator data local to the compute units. Each MMA
pipeline performs outer-product matrix operations, reading from and writing back a 512-bit
accumulator register.

Power10 implements the MMA accumulator architecture without adding a designed state.
Each designed 512-bit accumulator register is backed by four 128-bit Vector Scalar eXtension
(VSX) registers.

56 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Code that uses the MMA instructions is included in OpenBLAS and Eigen libraries. This
library can be built by using the most recent versions of GNU Compiler Collection (GCC)
compiler. The latest version of OpenBLAS is available at this web page.

OpenBLAS is used by Python-NumPy library, PyTorch, and other frameworks, which makes it
easy to use the performance benefit of the Power10 MMA accelerator for AI workloads.

The Power10 MMA accelerator technology also is used by the IBM Engineering and Scientific
Subroutine Library for AIX on POWER 7.1 (program number 5765-EAP).

Program code that is written in C/C++ or Fortran can benefit from the potential performance
gains by using the MMA facility if compiled by the following IBM compiler products:
򐂰 IBM Open XL C/C++ for AIX 17.1 (program numbers 5765-J18, 5765-J16, and 5725-C72)
򐂰 IBM Open XL Fortran for AIX 17.1 (program numbers 5765-J19, 5765-J17, and 5725-C74)

For more information about the implementation of the Power10 processor’s high throughput
math engine, see the white paper A matrix math facility for Power ISA processors.

For more information about fundamental MMA architecture principles with detailed instruction
set usage, register file management concepts, and various supporting facilities, see
Matrix-Multiply Assist Best Practices Guide, REDP-5612.

2.2.6 Power10 compatibility modes


The Power10 core implements the Processor Compatibility Register (PCR) as described in
the Power instruction set architecture (ISA) version 3.1, primarily to facilitate live partition
mobility (LPM) to and from previous generations of IBM Power hardware.

Depending on the specific settings of the PCR, the Power10 core runs in a compatibility mode
that pertains to Power9 (Power ISA version 3.0) or Power8 (Power ISA version 2.07)
processors. The support for processor compatibility modes also enables older operating
systems versions of AIX, IBM i, Linux, or Virtual I/O server environments to run on Power10
processor-based systems.

The Power10 processor-based scale-out servers support the Power8, Power9 Base, Power9,
and Power10 compatibility modes.

2.2.7 Processor module options


Power10 processor-based scale-out servers use DCMs or eSCMs:
򐂰 Power S1014 and Power S1022s server sockets are populated with eSCM modules.
򐂰 Power S1022 and Power S1024 server sockets are populated with DCM modules.

DCM modules cannot be configured for Power S1014 or Power S1022s servers. eSCM
modules are not supported in Power S1022 of Power S1024 systems.

Depending on the scale-out server model and the number of populated sockets, the following
core densities are available for the supported processor module types:
򐂰 Power S1014 and Power S1022s server are offered with four or eight functional cores per
eSCM. The Power S1014 is available only as 1-socket server. The Power S1022s
supports the 4-core eSCM only in a 1-socket configuration, and the 8-core eSCM in 1- and
2-socket configurations.

Chapter 2. Architecture and technical overview 57


򐂰 Power S1022 servers can be configured with 12, 16, or 20 functional cores per DCM. The
12-core DCM module is available for 1-socket and 2-socket configurations. The 16-core
and 20-core DCM modules are supported only in configurations in which both sockets are
populated.
򐂰 Power S1024 servers support 12, 16, or 24 functional cores per DCM. Regarding the
12-core DCM, the Power S1024 allows configurations with one or two populated sockets.
However, both sockets of the Power S1024 server must be configured if 16-core or 24-core
DCMs are chosen.

Note: In 2-socket Power10 processor-based scale-out servers, the processor module


types must be identical in terms of the number of functional cores and the related
processor typical frequency range of operation.

The supported processor activation types and use models vary with the Power10
processor-based scale-out server model type:
򐂰 Static processor activations
Only the eSCM models with 4-core or 8-core processor density in Power S1014 and
Power S1022s servers support the classical static processor activation model. All
functional cores of the configured eSCM modules are delivered with processor activation
features at initial order. This use model provides static and permanent processor
activations and is the default for the named eSCM-based servers.
򐂰 Capacity Upgrade on Demand (CUoD) processor activations
The Power S1022 and Power S1024 servers, which are based on DCM modules, support
the Capacity Upgrade on Demand (CUoD) technology option. For these servers, a
minimum of 50% of the configured total processor capacity must be activated through the
related CUoD processor activation features at the time of initial order.
Later, more CUoD processor activations can be purchase through a miscellaneous
equipment specification (MES) upgrade order. The CUoD is the default use model of
Power S1022 and Power S1024 servers. It offers static and permanent processor
activations with the added flexibility to adjust the processor capacity between half of the
physically present cores and the maximum of the configured processor module capacity
as required by the workload demand.
򐂰 Power Private Cloud with Shared Utility Capacity use model
The Power S1022 and Power S1024 servers also support the IBM Power Private Cloud
with Shared Utility Capacity solution (Power Enterprise Pools 2.0), which is an
infrastructure offering model that enables cloud agility and cost optimization with
pay-for-use pricing.
This use model requires the configuration of the Power Enterprise Pools 2.0 Enablement
feature (#EP20) for the specific server and a minimum of one Base Processor Activation
for Pools 2.0 feature is needed. The base processor activations are permanent and shared
within a pool of servers. More processor resources that are needed beyond the capacity
that is provided by the base processor activations are metered by the minute and paid
through capacity credits.

To assist with the optimization of software licensing, the factory deconfiguration feature
(#2319) is available at initial order for all scale-out server models to permanently reduce the
number of active cores that is less than the minimum processor core activation requirement.
Factory deconfigurations are permanent and they are available only in the context of the static
processor activation use model and the CUoD processor activation use model.

58 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Note: The static activation usage model, the CUoD technology usage model, and the
Power Private Cloud Shared Utility Capacity (Power Enterprise Pools 2.0) offering models
are all mutually exclusive in respect to each other.

Table 2-4 lists the processor module options that are available for Power10 processor-based
scale-out servers. The list is sorted by increasing order of the processor module capacity.

Table 2-4 Processor module options for Power10 processor-based scale-out servers
Module Module CUoD Pools 2.0 Typical Minimum Power Power Power Power
capacity type support option frequency quantity S1014 S1022s S1022 S1024
range per
[GHz] server

4-core eSCM No No 3.0 - 3.9 1 X X — —

8-core eSCM No No 3.0 - 3.9 1 X X — —

12-core DCM Yes Yes 2.9 - 4.0 1 — — X —

3.4 - 4.0 1 — — — X

16-core DCM Yes Yes 2.75 - 4.0 2 — — X —

3.1 - 4.0 2 — — — X

20-core DCM Yes Yes 2.45 - 3.9 2 — — X —

24-core DCM Yes Yes 2.75 - 3.9 2 — — — X

For each processor module option the module type (eSCM / DCM), the support for CUoD, the
availability of the Pools 2.0 option, and the minimum number of sockets that must be
populated are indicated.

Power10 processors automatically optimize their core frequencies based on workload


requirements, thermal conditions, and power consumption characteristics. Therefore, each
processor module option that is listed is associated with a processor core frequency range
within which the DCM or eSCM cores typically operate.

Depending on the different physical characteristics of the Power S1022 and Power S1024
servers. two distinct, model-specific frequency ranges are available for processor modules
with 12- and 16-core density.

The last four columns of Table 2-4 list the availability matrix between a specific processor
module capacity and frequency specification on one side and the Power10 processor-base
scale-out server models on the other side. (Available combinations are labeled with “X” and
unavailable combinations are indicated by a “—” hyphen.)

Chapter 2. Architecture and technical overview 59


2.2.8 On-chip L3 cache and intelligent caching
The Power10 processor includes a large on-chip L3 cache of up to 120 MB with a NUCA
architecture that provides mechanisms to distribute and share cache footprints across a set of
L3 cache regions. Each processor core can access an associated local 8 MB of L3 cache. It
also can access the data in the other L3 cache regions on the chip and throughout the
system.

Each L3 region serves as a victim cache for its associated L2 cache and can provide
aggregate storage for the on-chip cache footprint.

Intelligent L3 cache management enables the Power10 processor to optimize the access to
L3 cache lines and minimize cache latencies. The L3 includes a replacement algorithm with
data type and reuse awareness.

It also supports an array of prefetch requests from the core, including instruction and data,
and works cooperatively with the core, memory controller, and SMP interconnection fabric to
manage prefetch traffic, which optimizes system throughput and data latency.

The L3 cache supports the following key features:


򐂰 Enhanced bandwidth that supports up to 64 bytes per core processor cycle to each SMT8
core.
򐂰 Enhanced data prefetch that is enabled by 96 L3 prefetch request machines that service
prefetch requests to memory for each SMT8 core.
򐂰 Plus-one prefetching at the memory controller for enhanced effective prefetch depth and
rate.
򐂰 Power10 software prefetch modes that support fetching blocks of data into the L3 cache.
򐂰 Data access with reduced latencies.

2.2.9 Open memory interface


The Power10 processor introduces a new and innovative OMI. The OMI is driven by eight
on-chip MCUs and is implemented in two separate physical building blocks that lie in opposite
areas at the outer edge of the Power10 die. Each area supports 64 OMI lanes that are
grouped in four OMI ports. One port in turn consists of two OMI links with eight lanes each,
which operate in a latency-optimized manner with unprecedented bandwidth and scale at
32 Gbps speed.

One Power10 processor chip supports the following functional elements to access main
memory:
򐂰 Eight MCUs
򐂰 Eight OMI ports that are controlled one-to-one through a dedicated MCU
򐂰 Two OMI links per OMI port for a total of 16 OMI links
򐂰 Eight lanes per OMI link for a total of 128 lanes, all running at 32 Gbps speed

The Power10 processor provides natively an aggregated maximum theoretical full-duplex


memory interface bandwidth of 1 TBps per chip.

Memory interface architecture for dual-chip modules


The DCM, which is used in Power S1022 and Power S1024 servers, combines two Power10
processor chips in one processor package. Therefore, a total of 2 x 8 = 16 OMI ports and
2 x 16 = 32 OMI links are physically present on a Power10 DCM.

60 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
However, because the chips on the DCM are tightly integrated and the aggregated memory
bandwidth of eight OMI ports culminates at a maximum theoretical full-duplex bandwidth of
1 TBps, only half of the OMI ports are active.

Each chip of the DCM contributes four OMI ports and eight OMI links to facilitate main
memory access. For more information about the OMI port designation and the physical
location of the active OMI units of a DCM, see Figure 2-5 on page 48 and Figure 2-6 on
page 50.

In summary, one DCM supports the following functional elements to access main memory:
򐂰 Four active MCUs per chip for a total of eight MCUs per module
򐂰 Each MCU maps one-to-one to an OMI port
򐂰 Four OMI ports per chip for at total of eight OMI ports per module
򐂰 Two OMI links per OMI port for a total of eight OMI links per chip and 16 OMI links per
module
򐂰 Eight lanes per OMI link for a total of 128 lanes per module, all running at 32 Gbps

The Power10 DCM provides an aggregated maximum theoretical full-duplex memory


interface bandwidth of 512 GBps per chip and 1 TBps per module.

Memory interface architecture of single-chip modules


Because the eSCM as used in Power S1014 and Power S1022s servers is a derivative of the
DCM module, it combines two Power10 processor chips in one processor package. However,
unlike the DCM package. only one of the chips (chip-0) hosts active processor cores and
active memory interfaces. This configuration implies that 16 OMI ports and 32 OMI links are
physically present on the eSCM, but only the first Power10 chip (chip-0) contributes four OMI
ports and eight OMI links.

The second Power10 chip (chip-1) is dedicated to drive PCIe Gen5 and Gen4 interfaces
exclusively. For more information about for the OMI port designation and physical location of
the active OMI units of an eSCM, see Figure 2-7 on page 51 and Figure 2-8 on page 52.

In summary, one eSCM supports the following elements to access main memory:
򐂰 Four active MCUs per module
򐂰 Each MCU maps one-to-one to an OMI port
򐂰 Four OMI ports per module
򐂰 Two OMI links per OMI port for a total of eight OMI links per module
򐂰 Eight lanes per OMI link for a total of 64 lanes, all running at 32 Gbps speed

The Power10 eSCM provides an aggregated maximum theoretical full-duplex memory


interface bandwidth of 512 GBps per module.

The OMI physical interface enables low-latency, high-bandwidth, technology-agnostic host


memory semantics to the processor and allows attaching established and emerging memory
elements.

With the Power10 processor-based scale-out servers, OMI initially supports one main tier,
low-latency, enterprise-grade Double Data Rate 4 (DDR4) DDIMMs per OMI link. This
architecture yields a total memory module capacity of:
򐂰 8 DDIMMs per socket for eSCM-based Power S1014 and Power S1022s server
򐂰 16 DDIMMs per socket for DCM-based Power S1022 and Power S1024 servers

Chapter 2. Architecture and technical overview 61


The memory bandwidth and the total memory capacity depend on the DDIMM density and
the associated DDIMM frequency that is configured for a specific Power10 processor-base
scale-out server.

Table 2-5 list the maximum memory bandwidth for Power S1014, Power S1022s,
Power S1022, and Power S1024 servers under the assumption that the maximum number of
supported sockets are configured and all available slots are populated with DDIMMs of the
named density and speed.

Table 2-5 Maximum theoretical memory bandwidth for Power10 processor-based scale-out servers
Server model DDIMM DDIMM Maximum memory Maximum theoretical
density frequency capacity memory bandwidth
(GB) (MHz) (GB) (GBps)

Power S1014 128a 2666 1024 170

64, 32, 16 3200 512b 204

Power S1022s 128a 2666 2048 341

64, 32, 16 3200 1024b 409

Power S1022 128a 2666 4096 682


b
64, 32, 16 3200 2048 818

Power S1024 256a , 128a 2933 8192c 750

64, 32, 16 3200 4096b 818


a. The 128 GB and 256 GB DDIMMs are planned to be available from 18 November 2022.
b. Based on DDIMMs of 64 GB density.
c. Based on DDIMMs of 256 GB density.

2.2.10 Pervasive memory encryption


The Power10 MCU provides the system memory interface between the on-chip symmetric
multiprocessing (SMP) interconnect fabric and the OMI links. This design qualifies the MCU
as ideal functional unit to implement memory encryption logic. The Power10 on-chip MCU
encrypts and decrypts all traffic to and from system memory that is based on the AES
technology.

The Power10 processor supports the following modes of operation:


򐂰 AES XTS mode
The xor–encrypt–xor (XTS)-based tweaked-codebook mode features ciphertext stealing.
AES XTS provides a block cipher with strong encryption, which is useful to encrypt
persistent memory.
Persistent DIMM technology retains the data that is stored inside the memory DIMMs,
even if the power is turned off. A malicious attacker who gains physical access to the
DIMMs can steal memory cards. The data that is stored in the DIMMs can leave the data
center in the clear, if not encrypted.
Also, memory cards that leave the data center for repair or replacement can be a potential
security breach. Because the attacker might have arbitrary access to the persistent DIMM
data, the stronger encryption of the AES XTS mode is required for persistent memory. The
AES XTS mode of the Power10 processor is supported for future use if persistent memory
solutions become available for IBM Power servers.

62 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
򐂰 AES CTR mode
The Counter (CTR) mode of operation designates a low-latency AES bock cipher.
Although the level of encryption is not as strong as with the XTS mode, the low-latency
characteristics make it the preferred mode for memory encryption for volatile memory.
AES CTR makes it more difficult to physically gain access to data through the memory
card interfaces. The goal is to protect against physical attacks, which becomes
increasingly important in the context of cloud deployments.
The Power10 processor-based scale-out servers support the AES CTR mode for
pervasive memory encryption. Each Power10 processor holds a 128-bit encryption key
that is used by the processor’s MCU to encrypt the data of the differential DIMMs that are
attached to the OMI links.
The MCU crypto engine is transparently integrated into the data path, which ensures that
the data fetch and store bandwidth are not compromised by the AES CTR encryption
mode. Because the encryption has no noticeable performance effect and because of the
obvious security benefit, the pervasive memory encryption is enabled by default and
cannot be switched off through any administrative interface.

Note: Consider the following points:


򐂰 The pervasive memory encryption of the Power10 processor does not affect the
encryption status of a system memory dump content. All data from the DDIMMs is
decrypted by the memory controller unit before it is passed onto the memory dump
devices under the control of the memory dump program code. This statement applies to
the traditional system dump under the operating system control and the firmware assist
dump utility.
򐂰 The PowerVM Live Partition Mobility (LPM) data encryption does not interfere with the
pervasive memory encryption. Data transfer that occurs during an LPM operation uses
the following general flow:
a. On the source server, the Mover Server Partition (MSP) provides the hypervisor with
a buffer.
b. The hypervisor of the source system copies the partition memory into the buffer.
c. The MSP transmits the data over the network.
d. The data is received by the MSP on the target server and copied in to the related
buffer.
e. The hypervisor of the target system copies the data from the buffer into the memory
space of the target partition.

To facilitate LPM data compression and encryption, the hypervisor on the source system
presents the LPM buffer to the on-chip nest accelerator (NX) unit as part of process in
Step b. The reverse decryption and decompress operation is applied on the target server
as part of process in Step d.

The pervasive memory encryption logic of the MCU decrypts the memory data before it is
compressed and encrypted by the NX unit on the source server. It also encrypts the data
before it is written to memory, but after it is decrypted and decompressed by the NX unit of
the target server.

Chapter 2. Architecture and technical overview 63


2.2.11 Nest accelerator
The Power10 processor features an on-chip accelerator that is called the nest accelerator
unit (NX unit). The coprocessor features that are available on the Power10 processor are
similar to the features of the Power9 processor. These coprocessors provide specialized
functions, such as the following examples:
򐂰 IBM proprietary data compression and decompression
򐂰 Industry-standard Gzip compression and decompression
򐂰 AES and Secure Hash Algorithm (SHA) cryptography
򐂰 Random number generation

Figure 2-10 shows a block diagram of the NX unit.

Figure 2-10 Block diagram of the NX unit

Each one of the AES/SHA engines, data compression, and Gzip units consist of a
coprocessor type and the NX unit features three coprocessor types. The NX unit also
includes more support hardware to support coprocessor invocation by user code, use of
effective addresses, high-bandwidth storage access, and interrupt notification of job
completion.

The direct memory access (DMA) controller of the NX unit helps to start the coprocessors
and move data on behalf of coprocessors. SMP interconnect unit (SIU) provides the interface
between the Power10 SMP interconnect and the DMA controller.

The NX coprocessors can be started transparently through library or operating system kernel
calls to speed up operations that are related to:
򐂰 Data compression
򐂰 Live partition mobility migration
򐂰 IPsec
򐂰 JFS2 encrypted file systems

64 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
򐂰 PKCS11 encryption
򐂰 Random number generation
򐂰 The most recently announced logical volume encryption

In effect, this on-chip NX unit on Power10 systems implements a high throughput engine that
can perform the equivalent work of multiple cores. The system performance can benefit by
off-loading these expensive operations to on-chip accelerators, which in turn can greatly
reduce the CPU usage and improve the performance of applications.

The accelerators are shared among the logical partitions (LPARs) under the control of the
PowerVM hypervisor and accessed by way of a hypervisor call. The operating system, along
with the PowerVM hypervisor, provides a send address space that is unique per process that
is requesting the coprocessor access. This configuration allows the user process to directly
post entries to the first in-first out (FIFO) queues that are associated with the NX accelerators.
Each NX coprocessor type features a unique receive address space that corresponds to a
unique FIFO for each of the accelerators.

For more information about the use of the xgzip tool that uses the Gzip accelerator engine,
see the following resources:
򐂰 IBM support article: Using the POWER9 NX (gzip) accelerator in AIX
򐂰 IBM Power community article: Power9 GZIP Data Acceleration with IBM AIX
򐂰 AIX community article: Performance improvement in openssh with on-chip data
compression accelerator in power9
򐂰 IBM Documentation: nxstat Command

2.2.12 SMP interconnect and accelerator interface


The Power10 processor provides a highly optimized, 32 Gbps differential signaling
technology interface that is structured in 16 entities. Each entity consists of eight data lanes
and one spare lane. This interface can facilitate the following functional purposes:
򐂰 First- or second-tier, symmetric multiprocessing link interface, which enables up to 16
Power10 processors to be combined into a large, robustly scalable, single-system image.
򐂰 Open Coherent Accelerator Processor Interface (OpenCAPI) to attach cache coherent
and I/O-coherent computational accelerators, load or store addressable host memory
devices, low latency network controllers, and intelligent storage controllers.
򐂰 Host-to-host integrated memory clustering interconnect, which enabling multiple Power10
systems to directly use memory throughout the cluster.

Note: The OpenCAPI interface and the memory clustering interconnect are Power10
technology options for future use.

Because of the versatile nature of signaling technology, the 32 Gbps interface also is referred
to as Power/A-bus/X-bus/OpenCAPI/Networking (PowerAXON) interface. The IBM
proprietary X-bus links connect two processors on a board with a common reference clock.
The IBM proprietary A-bus links connect two processors in different drawers on different
reference clocks by using a cable.

Chapter 2. Architecture and technical overview 65


OpenCAPI is an open interface architecture that allows any microprocessor to attach to the
following components:
򐂰 Coherent user-level accelerators and I/O devices
򐂰 Advanced memories that are accessible through read/write or user-level DMA semantics

The OpenCAPI technology is developed, enabled, and standardized by the OpenCAPI


Consortium. For more information about the consortium’s mission and the OpenCAPI
protocol specification, see OpenCAPI Consortium.

The PowerAXON interface is implemented on dedicated areas that are at each corner of the
Power10 processor die.

The Power10 processor-based scale-out servers use this interface to implement:


򐂰 DCM internal chip-to-chip connections
򐂰 eSCM internal chip-to-chip connections
򐂰 Chip-to-chip SMP interconnects between DCMs and eSCMs in 2-socket configurations
򐂰 OpenCAPI accelerator interface connections

The chip-to-chip DCM internal (see Figure 2-5 on page 48) and eSCM internal (see
Figure 2-7 on page 51) chip-to-chip connections are implemented by using the interface ports
OP2 and OP6 on chip-0 and OP1 and OP4 on chip-1:
򐂰 2 × 9 OP2 lanes of chip-0 connect to 2 x 9 OP1 lanes of chip-1
򐂰 2 × 9 OP6 lanes of chip-0 connect to 2 × 9 OP4 lanes of chip-1

The processor module internal chip-to-chip connections feature the following common
properties:
򐂰 Two (2 x 9)-bit buses implement two independent connections between the module chips
򐂰 Eight data lanes, plus one spare lane in each direction per chip-to-chip connection
򐂰 32 Gbps signaling rate that provides 128 GBps per chip-to-chip connection bandwidth,
which yields a maximum theoretical full-duplex bandwidth between the two processor
module chips of 256 GBps

In addition to the interface ports OP2 and OP6 on chip-0 and OP1 and OP4 on chip-1, the
DCM offers 216 A-bus/X-bus/OpenCAPI lanes that are grouped in 12 other interface ports:
򐂰 OP0, OP1, OP3, OP4, OP5, OP7 on chip-0
򐂰 OP0, OP2, OP3, OP5, OP6, OP7 on chip-1

Each OP1 and OP2 interface port runs as a 2 × 9 SMP bus at 32 Gbps whereas the OP0,
OP3, OP4, OP5, OP6, and OP7 interface ports can run in one of the following two modes:
򐂰 2 × 9 SMP at 32 Gbps
򐂰 2 × 8 OpenCAPI at 32 Gbps

66 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
SMP topology and accelerator interfaces for DCM-based servers
Figure 2-11 shows the flat, one-hop SMP topology and the associated interface ports for
Power S1022 and Power S1024 servers in 2-socket configurations (all interfaces that do not
contribute to the SMP fabric were omitted for clarity).

SMP 2 x9
OP4 OP7
@ 32 Gbps
DCM-0 DCM-1
P0 / Chip-0 P2 / Chip-0
OP7 OP4

OP2 OP6 OP2 OP6

@ 32 Gbps
@ 32 Gbps

@ 32 Gbps

@ 32 Gbps
SMP 2 x9

SMP 2 x9

SMP 2 x9
SMP 2 x9
OP1 OP4 OP1 OP4

OP7 OP6
DCM-0 DCM-1
P1 / Chip-1 P3 / Chip-1
SMP 2 x9
OP6 OP7
@ 32 Gbps

Figure 2-11 SMP connectivity for Power S1022 or Power S1024 servers in 2-socket configurations

The interface ports OP4, OP6, and OP7 are used to implement direct SMP connections
between the first DCM chip (DCM-0) and the second DCM chip (DCM-1), as shown in the
following example:
򐂰 2 x 9 OP4 lanes of chip-0 on DCM-0 connect to 2 x 9 OP7 lanes of chip-0 on DCM-1
򐂰 2 x 9 OP7 lanes of chip-0 on DCM-0 connect to 2 x 9 OP6 lanes of chip-1 on DCM-1
򐂰 2 x 9 OP7 lanes of chip-1 on DCM-0 connect to 2 x 9 OP4 lanes of chip-0 on DCM-1
򐂰 2 x 9 OP6 lanes of chip-1 on DCM-0 connect to 2 x 9 OP7 lanes of chip-1 on DCM-1

Each inter-DCM chip-to-chip SMP link provides a maximum theoretical full-duplex bandwidth
of 128 GBps.

The interface port OP3 on chip-0 and OP0 on chip-1 of the DCM are used to implement
OpenCAPI interfaces through connectors that are on the mainboard of Power S1022 and
Power S1024 servers. The relevant interface ports are subdivided in two bundles of eight
lanes, which are designated by the capital letters A and B respectively. Therefore, the named
ports OP3A, OP3B, OP0A, and OP0B represent one bundle of eight lanes that can support
one OpenCAPI interface in turn.

In a 1-socket Power S1022 or Power S1024 server, a total of 4 OpenCAPI interfaces are
implemented through DCM-0, as shown in the following example:
򐂰 OP3A and OP3B on chip-0 of DCM-0
򐂰 OP0A and OP0B on chip-1 of DCM-0

In a 2-socket Power S1022 or Power S1024 server, two other OpenCAPI interfaces are
provided through DCM-1, as shown in the following example:
򐂰 OP3A on chip-0 of DCM-1
򐂰 OP0B on chip-1 of DCM-1

Chapter 2. Architecture and technical overview 67


The 2-socket logical diagram of the Power S1022 and the Power S1024 server that is shown
in Figure 2-1 on page 41 shows the OpenCAPI interfaces that are represented by their
SlimSAS connectors. The 1-socket constellation can be deduced from Figure 2-1 on page 41
if DCM-1 is conceptually omitted.

Note: The implemented OpenCAPI interfaces can be used in the future and are not used
by available technology products as of this writing.

SMP topology for eSCM based servers


The flat, one-hop SMP topology and the associated interface ports for the Power S1022s
2-socket configuration is shown in Figure 2-12. All active interface ports contribute to the SMP
fabric. Unlike the DCM, no other interface ports are enabled to support OpenCAPI interfaces.

SMP 2 x9
OP4 OP7
@ 32 Gbps
eSCM-0 eSCM-1
P0 / Chip-0 P2 / Chip-0
OP7 OP4

OP2 OP6 OP2 OP6


@ 32 Gbps

@ 32 Gbps

@ 32 Gbps
@ 32 Gbps

SMP 2 x9
SMP 2 x9

SMP 2 x9

SMP 2 x9

OP1 OP4 OP1 OP4

OP7 OP6
eSCM-0 eSCM-1
Switch Switch
P1 / Chip-1 SMP 2 x9 P3 / Chip-1
OP6 OP7
@ 32 Gbps

Figure 2-12 SMP connectivity for a Power S1022s server in 2-socket configuration

In 2-socket eSCM configurations of the Power S1022s server, the interface ports OP4 and
OP7 on chip-0 and OP6 and OP7 on chip-1 of the processor module are active. They are
used to implement direct SMP connections between the first eSCM (eSCM-0) and the second
eSCM (eSCM-1) in the same way for the 2-socket DCM configurations of the Power S1022
and Power S1024 servers.

However, the eSCM constellation differs by the fact that no active cores (0-cores) are on
chip-1 of eSCM-0 and chip-1 of eSCM-1. These chips operate as switches. For more
information about the Power S1022s 2-socket server that is based on two eSCM modules,
see Figure 2-2 on page 42.

In summary, the SMP interconnect between the eSCMs of a Power S1022s server in 2-socket
configuration and between the DCMs of a Power S1022 or Power S1024 server in 2-socket
configuration features the following properties:
򐂰 One (2 x 9)-bit buses per chip-to-chip connection across the processor modules
򐂰 Eight data lanes plus one spare lane in each direction per chip-to-chip connection

68 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
򐂰 Flat, 1-hop SMP topology through direct connection between all chips
򐂰 32 Gbps signaling rate, which provides 128 GBps bandwidth per chip-to-chip connection
an increase of 33% compared to the Power9 processor-based scale-out servers
implementation

2.2.13 Power and performance management


Power10 processor-based servers implement an enhanced version of the power
management EnergyScale technology. As in the previous Power9 EnergyScale
implementation, the Power10 EnergyScale technology supports dynamic processor
frequency changes that depend on several factors, such as workload characteristics, the
number of active cores, and environmental conditions.

Based on the extensive experience that was gained over the past few years, the Power10
EnergyScale technology evolved to use the following effective and simplified set of
operational modes:
򐂰 Power saving mode
򐂰 Static mode (nominal frequency)
򐂰 Maximum performance mode (MPM)

The Power9 dynamic performance mode (DPM) has many features in common with the
Power9 maximum performance mode (MPM). Because of this redundant nature of
characteristics, the DPM for Power10 processor-based systems was removed in favor of an
enhanced MPM. For example, the maximum frequency is now achievable in the Power10
enhanced maximum performance mode (regardless of the number of active cores), which
was not always the case with Power9 processor-based servers.

The Power10 processor-based scale-out servers feature MPM enabled by default. This mode
dynamically adjusts processor frequency to maximize performance and enable a much higher
processor frequency range. Each of the power saver modes deliver consistent system
performance without any variation if the nominal operating environment limits are met.

For Power10 processor-based systems that are under control of the PowerVM hypervisor, the
MPM is a system-wide configuration setting, but each processor module frequency is
optimized separately.

The following factors determine the maximum frequency at which a processor module can
run:
򐂰 Processor utilization: Lighter workloads run at higher frequencies.
򐂰 Number of active cores: Fewer active cores run at higher frequencies.
򐂰 Environmental conditions: At lower ambient temperatures, cores are enabled to run at
higher frequencies.

Chapter 2. Architecture and technical overview 69


The following Power10 EnergyScale modes are available:
򐂰 Power saving mode
The frequency is set to the minimum frequency to reduce energy consumption. Enabling
this feature reduces power consumption by lowering the processor clock frequency and
voltage to fixed values. This configuration reduces power consumption of the system while
delivering predictable performance.
򐂰 Static mode
The frequency is set to a fixed point that can be maintained with all normal workloads and
in all normal environmental conditions. This frequency is also referred to as nominal
frequency.
򐂰 Maximum performance mode
Workloads run at the highest frequency possible, depending on workload, active core
count, and environmental conditions. The frequency does not fall below the static
frequency for all normal workloads and in all normal environmental conditions.
In MPM, the workload is run at the highest frequency possible. The higher power draw
enables the processor modules to run in an MPM typical frequency range (MTFR), where
the lower limit is greater than the nominal frequency and the upper limit is given by the
system’s maximum frequency.
The MTFR is published as part of the system specifications of a specific Power10 system
if it is running by default in MPM. The higher power draw potentially increases the fan
speed of the respective system node to meet the higher cooling requirements, which in
turn causes a higher noise emission level of up to 15 decibels.
The processor frequency typically stays within the limits that are set by the MTFR, but can
be lowered to frequencies between the MTFR lower limit and the nominal frequency at
high ambient temperatures greater than 27 °C (80.6 °F).
If the data center ambient environment is less than 27 °C, the frequency in MPM is
consistently in the upper range of the MTFR (roughly 10% - 20% better than nominal). At
lower ambient temperatures (less than 27 °C, or 80.6 °F), MPM mode also provides
deterministic performance. As the ambient temperature increases above 27 °C,
determinism can no longer be ensured. This mode is the default mode for all Power10
processor-based scale-out servers.
򐂰 Idle power saver mode (IPS)
IPS mode lowers the frequency to the minimum if the entire system (all cores of all
sockets) is idle. It can be enabled or disabled separately from all other modes.

70 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Figure 2-13 shows the comparative frequency ranges for the Power10 power saving mode,
static or nominal mode, and the maximum performance mode. The frequency adjustments for
different workload characteristics, ambient conditions, and idle states are also indicated.

Figure 2-13 Power10 power management modes and related frequency ranges

Table 2-6, Table 2-7, Table 2-8 on page 72, and Table 2-9 on page 72 show the power saving
mode, the static mode frequencies, and the frequency ranges of the MPM for all processor
module types that are available for the Power S1014, Power S1022s, Power S1022, and
Power S1024 servers.

Note: For all Power10 processor-based scale-out systems, the MPM is enabled by default.

Table 2-6 Characteristic frequencies and frequency ranges for Power S1014 servers
Feature Cores per Power saving Static mode Maximum performance mode
code single-chip mode frequency frequency frequency range
module [GHz] [GHz] [GHz]

#EPG0 4 2.0 3.0 3.0 to 3.90 (max)

#EPG2 8 2.0 3.0 3.0 to 3.90 (max)

Table 2-7 Characteristic frequencies and frequency ranges for Power S1022s servers
Feature Cores per Power saving Static mode Maximum performance mode
code single-chip mode frequency frequency frequency range
module [GHz] [GHz] [GHz]

#EPGR 4 2.0 3.0 3.0 to 3.90 (max)

#EPGQ 8 2.0 3.0 3.0 to 3.90 (max)

Chapter 2. Architecture and technical overview 71


Table 2-8 Characteristic frequencies and frequency ranges for Power S1022 servers
Feature Cores per Power saving Static mode Maximum performance mode
code single-chip mode frequency frequency frequency range
module [GHz] [GHz] [GHz]

#EPG9 12 2.0 2.90 2.90 to 4.0 (max)

#EPG8 16 2.0 2.75 2.75 to 4.0 (max)

#EPGA 20 2.0 2.45 2.45 to 3.9 (max)

Table 2-9 Characteristic frequencies and frequency ranges for Power S1024 servers
Feature Cores per Power saving Static mode Maximum performance mode
code single-chip mode frequency frequency frequency range
module [GHz] [GHz] [GHz]

#EPGM 12 2.0 3.40 3.40 - 4.0

#EPGC 16 2.0 3.10 3.10 - 4.0

#EPGD 24 2.0 2.75 2.75 - 3.9

The controls for all power saver modes are available on the Advanced System Management
Interface (ASMI) and can be dynamically modified. A system administrator can also use the
Hardware Management Console (HMC) to set power saver mode or to enable static mode or
MPM.

Figure 2-14 shows the ASM interface menu for Power and Performance Mode Setup on a
Power10 processor-based scale-out server.

Figure 2-14 ASMI menu for Power and Performance Mode setup

72 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Figure 2-15 shows the HMC menu for power and performance mode setup.

Figure 2-15 HMC menu for Power and Performance Mode setup

2.2.14 Comparing Power10, Power9, and Power8 processors


The Power10 processor-based systems are available in three different processor module
packages:
򐂰 Single-chip modules (SCMs), which are based on one Power10 chip. These modules
allow for the highest core performance.
򐂰 Dual-chip modules (DCMs), which combine two Power10 chips where both Power10 chips
contribute active processor cores. This configuration allows for processors with the
memory and I/O bandwidth of two Power10 chips in a single socket.
򐂰 eSCMs, which combine two Power10 chips but all active processor core resources are
bundled on one of the two chips. This configuration allows for processors with a lower core
count, but the full I/O bandwidth of both Power10 chips.

The Power E1080 enterprise class systems exclusively use SCM modules with up to 15
active SMT8 capable cores. These SCM processor modules are structural and performance
optimized for usage in scale-up multi-socket systems.

The Power E1050 enterprise class system exclusively uses DCM modules with up to 24
active SMT8 capable cores. This configuration maximizes the core density and I/O
capabilities of these servers.

DCM modules with up to 24 active SMT8-capable cores are used in 1- or 2-socket


Power S1022 and Power S1024 servers. eSCMs with up to eight active SMT8-capable cores
are used in 1-socket Power S1014 and 1- or 2-socket Power S1022s servers.

DCM and eSCM modules are designed to support scale-out 1- to 4-socket Power10
processor-based servers.

Chapter 2. Architecture and technical overview 73


Table 2-10 compares key features and characteristics of the Power10, Power9, and Power8
processor implementations as used in the range of Power10 processor-based servers.

Table 2-10 Comparison of the Power10 processor technology to prior processor generations
Characteristics Power10 Power9 Power8

DCM eSCM SCM

Technology 7 nm 14 nm 22 nm

Die size 2 x 602 mm2 2 x 602 mm2 602 mm2 693 mm2 649 mm2

Processor module size 74.5 mm x 74.5 mm x 68.5 mm x 68.5 mm x 71.5 mm x


85.75 mm 85.75 mm 77.5 mm 68.5 mm 71.5 mm

Number of transistors 2 x 18 billion 2 x 18 billion 18 billion 8 billion 4.2 billion

Maximum cores 24 8 15 12 12

Maximum hardware threads 8 8 8


per core

Maximum static frequency / 3.4 - 4.0 GHz 3.0 - 3.9 GHz 3.6 - 4.15 GHz 3.9 - 4.0 GHz 4.15 GHz
high-performance frequency
rangea

L2 Cache per core 2048 KB 512 KB 512 KB

L3 Cache 8 MB of L3 cache per core with each core having 10 MB of L3 8 MB of L3


access to the full 120 MB of L3 cache, on-chip cache per core cache per core
high-efficiency SRAMb with each core with each core
having access having access
to the full to the full
120 MB of L3 96 MB of L3
cache, on-chip cache, on-chip
eDRAMc eDRAM

Supported DDR4d: Packaged on differential DIMMs with more DDR4 and DDR3 and
memory technology performance and resilience capabilities DDR3e DDR4

I/O bus PCIe Gen 5 PCIe Gen 4 PCIe Gen 3


a. Best of class typical frequency range where the lower limit of the range coincides with the maximum
static/nominal frequency.
b. Static random-access memory.
c. Embedded dynamic random-access memory.
d. The Power10 processor memory logic and the memory subsystem of Power10 processor-based servers can
support DDR5 or other memory technologies in the future.
e. Only DDR3 memory CDIMMs, which are transferred in the context of a model upgrade from Power E870, Power
E870C, Power E880, or Power E880C systems to a Power E980 server, are supported.

74 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
2.3 Memory subsystem
The Power10 processor contains eight independent MCUs that provide the system memory
interface between the on-chip SMP interconnect fabric and the OMI links. Each MCU maps in
a one-to-one relation to an OMI port, which is also referred to as memory channel. Each OMI
port in turn supports two OMI links, for a total of 16 OMI links per chip. The OMI links of a
specific OMI port are also referred to as memory subchannel A and B.

As used in Power S1022 and Power S1024 servers, the Power10 DCM only half of the MCUs
and OMI links on each Power10 chip are used, which results in 16 total OMI links per DCM.
One IBM DDIMM connects to each OMI link, for a total of 32 DDIMMs when two DCM
modules are configured.

As used in Power S1014 and Power S1022s servers, the Power10 eSCM supports only eight
configured OMI links per module, which resulting in a total of 16 DDIMMs when two eSCM
modules are configured.

The DDIMM cards are available in two rack unit (2U) and four rack unit (4U) form factors and
are based on DDR4 DRAM technology. Depending on the form factor and the module
capacity of 16 GB, 32 GB, 64 GB, 128 GB, or 256 GB data rates of 2666 MHz, 2933 MHz, or
3200 MHz are supported.

DDIMM cards contain an OMI attached memory buffer, power management interface
controllers (PMICs), an Electrically Erasable Programmable Read-only Memory (EEPROM)
chip for vital product data, and the DRAM elements.

The PMICs supply all voltage levels that are required by the DDIMM card so that no separate
voltage regulator modules are needed. For each 2U DDIMM card, one PMIC plus one spare
are used.

For each 4U DDIMM card, two PMICs plus two spares are used. Because the PMICs operate
as redundant pairs, no DDIMM is called for replacement if one PMIC in each of the redundant
pairs is still functional.

Chapter 2. Architecture and technical overview 75


Figure 2-16 shows the memory logical diagram for DCM-based Power10 scale-out servers.

D1P1-OMI4A DDR4 Differential DIMM: P0-C25


D1P1-OMI4B DDR4 Differential DIMM: P0-C26

OMI4A
OMI4B
OMI5A P1D1-OMI5A DDR4 Differential DIMM: P0-C48
DCM-1 OMI5B P1D1-OMI5B DDR4 Differential DIMM: P0-C43
P1 / Chip-1 OMI6A D1P1-OMI6A DDR4 Differential DIMM: P0-C47
OMI6B D1P1-OMI6B DDR4 Differential DIMM: P0-C46
OMI7A D1P1-OMI7A DDR4 Differential DIMM: P0-C44
OP4
OMI7B
OP1
D1P1-OMI7B DDR4 Differential DIMM: P0-C45

OP6 OP2
OMI3B D1P0-OMI3B DDR4 Differential DIMM: P0-C42
OMI3A D1P0-OMI3A DDR4 Differential DIMM: P0-C41
OMI2B D1P0-OMI2B DDR4 Differential DIMM: P0-C39
DCM-1 OMI2A D1P0-OMI2A DDR4 Differential DIMM: P0-C38
P0 / Chip-0 OMI1B D1P0-OMI1B DDR4 Differential DIMM: P0-C40
OMI1A
OMI0B
OMI0A D1P0-OMI1A
P0-C21DDR4 Differential DIMM: P0-C21
D1P0-OMI0B
P0-C20 DDR4 Differential DIMM: P0-C20
D1P0-OMI0A DDR4 Differential DIMM: P0-C19
D0P1-OMI4A DDR4 Differential DIMM: P0-C18
P0-C17 DDR4 Differential DIMM: P0-C17
D0P1-OMI4B
OMI4A P0-C16DDR4 Differential DIMM: P0-C16
D0P1-OMI5A
OMI4B
OMI5A
DCM-0 OMI5B D0P1-OMI5B DDR4 Differential DIMM: P0-C35
P1 / Chip-1 OMI6A D0P1-OMI6A DDR4 Differential DIMM: P0-C37
OMI6B D0P1-OMI6B DDR4 Differential DIMM: P0-C36
OMI7A D0P1-OMI7A DDR4 Differential DIMM: P0-C34
OMI7B D0P1-OMI7B DDR4 Differential DIMM: P0-C33
OP4 OP1

OP6 OP2
OMI3B D0P0-OMI3B DDR4 Differential DIMM: P0-C30
OMI3A D0P0-OMI3A DDR4 Differential DIMM: P0-C31
OMI2B D0P0-OMI2B DDR4 Differential DIMM: P0-C29
DCM-0 OMI2A D0P0-OMI2A DDR4 Differential DIMM: P0-C28
P0 / Chip-0 OMI1B D0P0-OMI1B DDR4 Differential DIMM: P0-C32
OMI1A D0P0-OMI1A DDR4 Differential DIMM: P0-C27
OMI0B
OMI0A

D0P0-OMI0B DDR4 Differential DIMM: P0-C13


D0P0-OMI0A DDR4 Differential DIMM: P0-C12

Figure 2-16 Memory logical diagram of DCM-based Power S1022 and Power S1024 servers

All active OMI subchannels are indicated by the labels OMI1A/OMI1B to OMI7A/OMI7B for
the respective DCM chips.

The DDIMM label begins with the DCM-chip-link designation. For example, D1P1-OMI4A
refers to a memory module that is connected to the OMI link OMI4A on chip-1 (processor-1)
of DCM-1.

The DDIMM label concludes with the physical location code of the memory slot. In our
example of the D1P1-OMI4A connected DDIMM, the location code P0-C25 reveals that the
DDIMM is plugged into slot connector 25 (C25) on the main board (P0). Although Figure 2-16
resembles the physical placement and the physical grouping of the memory slots, some slot
positions were moved for the sake of improved clarity.

The memory logical diagram for 1-socket DCM-based Power10 scale-out servers easily can
be seen in Figure 2-16 if you conceptually omit the DCM-1 processor module, including all of
the attached DDIMM memory modules.

76 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Figure 2-17 shows the memory logical diagram for eSCM-based Power10 scale-out servers.
Only half of the OMI links are available for eSCMs in comparison to DCMs and that all active
OMI links are on chip-0 of each eSCM.

eSCM-1
P1 / Chip-1

OP4 OP1

OP6 OP2
OMI3B S1P0-OMI3B DDR4 Differential DIMM: P0-C42
OMI3A S1P0-OMI3A DDR4 Differential DIMM: P0-C41
OMI2B S1P0-OMI2B DDR4 Differential DIMM: P0-C39
eSCM-1 OMI2A S1P0-OMI2A DDR4 Differential DIMM: P0-C38
P0 / Chip-0 OMI1B S1P0-OMI1B DDR4 Differential DIMM: P0-C40
OMI1A
OMI0B
OMI0A S1P0-OMI1A
P0-C21DDR4 Differential DIMM: P0-C21
S1P0-OMI0B
P0-C20 DDR4 Differential DIMM: P0-C20
S1P0-OMI0A DDR4 Differential DIMM: P0-C19

eSCM-0
P1 / Chip-1

OP4 OP1

OP6 OP2
OMI3B S0P0-OMI3B DDR4 Differential DIMM: P0-C30
OMI3A S0P0-OMI3A DDR4 Differential DIMM: P0-C31
OMI2B S0P0-OMI2B DDR4 Differential DIMM: P0-C29
eSCM-0 OMI2A S0P0-OMI2A DDR4 Differential DIMM: P0-C28
P0 / Chip-0 OMI1B S0P0-OMI1B DDR4 Differential DIMM: P0-C32
OMI1A S0P0-OMI1A DDR4 Differential DIMM: P0-C27
OMI0B
OMI0A

S0P0-OMI0B DDR4 Differential DIMM: P0-C13


S0P0-OMI0A DDR4 Differential DIMM: P0-C12

Figure 2-17 Memory logical diagram of eSCM-based Power S1022s server

Again, the memory logical diagram for 1-socket eSCM-based Power10 scale-out servers can
easily be deduced from Figure 2-17 if you conceptually omit the eSCM-1 processor module
including all of the attached DDIMM memory modules.

Physically, the memory slots are organized into the following groups, as shown in Figure 2-18
on page 78:
򐂰 C12 and C13 are placed at the outward-facing side of eSCM-0/DCM-0 and are connected
to chip-0 of the named processor modules.
򐂰 C25 and C26 are placed at the outward-facing side of eSCM-1/DCM-1 and are connected
to chip-1 of the named processor modules.
򐂰 C27 to C37 (11 slots) are placed toward the front of the server and are assigned to the first
processor module (eSCM-0/DCM-0).
򐂰 C38 to C48 (11 slots) are placed toward the front of the server and are assigned to the
second processor module (eSCM-1/DCM-1).
򐂰 C16 to C21 (six slots) are placed between the processor modules where the first half (C16
to C18) is wired to eSCM-0/DCM-0 and the second half (C19 to C21) to eSCM-1/DCM-1.

Chapter 2. Architecture and technical overview 77


P0-C0 P0-C25 P0-C26 P0-C48
P0-C47
P0-C1 P0-C46
P0-C45
P0-C2
DCM-1 / P0-C44
P0-C3 eSCM-1 P0-C43
P0-C42
P0-C4 P0-C41
P0-C40
P0-C39
P0-C21 P0-C38
Power Supply Units P0-C20
P0-C19 Optional
P0-C18 NVMe
P0-C17 P0-C37 Devices
P0-C5 eBMC P0-C16 P0-C36
P0-C35
P0-C6 OpenCAPI P0-C34
DCM-0 / P0-C33
P0-C7
eSCM-0 P0-C32
P0-C8 P0-C31 Optional
P0-C30 NVMe
P0-C9 P0-C29 Devices

P0-C10 P0-C28
P0-C12 P0-C13 P0-C27
P0-C11

Figure 2-18 Memory module physical slot locations and DDIMM location codes

Figure 2-18 also shows the physical location of the ten PCIe adapter slots C0 to C4 and C7 to
C11. Slot C5 is always occupied by the eBMC and slot C6 reserves the option to establish an
external OpenCAPI based connection in the future.

2.3.1 Memory feature and placement rules


This section discusses the memory feature codes, memory feature order rules, and memory
placement rules for the Power10 processor-based scale-out servers:
򐂰 Power S1014 memory feature and placement rules
򐂰 Power S1022s memory feature and placement rules
򐂰 Power S1022 and Power S1024 memory feature and placement rules

In general, the preferred approach is to install memory evenly across all processor modules in
the system. Balancing memory across the installed processor modules enables memory
access in a consistent manner and typically results in the best possible performance for your
configuration. Account for any plans for future memory upgrades when you decide which
memory feature size to use at the time of the initial system order.

78 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Power S1014 memory feature and placement rules
Table 2-11 lists the available memory feature codes for Power S1014 servers. No specific
memory enablement features are required and the entire physical DDIMM capacity of a
memory feature is enabled by default.

Table 2-11 Memory feature codes for Power S1014 servers


Feature code Description

#EM6N 32 GB (2x16GB) DDIMMs, 3200 MHz, 8 Gbit DDR4 memory

#EM6W 64 GB (2x32 GB) DDIMMs, 3200 MHz, 8 Gbit DDR4 memory

#EM6X 128 GB (2x64 GB) DDIMMs, 3200 MHz, 16 Gbit DDR4 memory

#EM6Ya 256 GB (2x128 GB) DDIMMs, 2666 MHz, 16 Gbit DDR4 memory
a. The 128 GB DDIMM parts in feature code #EM6Y are planned to be available on 18 November
2022.

The memory DDIMMs must be ordered in pairs by using the following feature codes:
򐂰 16 GB: #EM6N
򐂰 32 GB: #EM6W
򐂰 64 GB: EM6X
򐂰 128 GB: EM6Y

The minimum ordering granularity is one memory feature and all DDIMMs must be of the
same feature code type for a Power S1014 server. A maximum of four memory feature codes
can be configured to cover all of the available eight memory slots.

The minimum memory capacity requirement of the Power S1014 server is 32 GB, which can
be fulfilled by one #EM6N feature.

The maximum memory capacity is 64 GB if the 4-core eSCM module (#EPG0) was chosen
and IBM i is the primary operating system for the server. This configuration can be realized by
using one #EM6W memory feature or two #EM6N memory features.

If the Power S1014 server is based on the 8-core eSCM module, a maximum memory
capacity of 1 TB is supported. This specific maximum configuration requires four #EM6Y
memory features. Until the availability of the 128 GB memory DDIMMs (planned for
18 November 2022), the maximum memory capacity is 512 GB.

Figure 2-19 shows the DDIMM plug sequence for Power S1014 servers.

Power10 eSCM-0 Chip-0


OMI0 OMI1 OMI2 OMI3
A B A B A B A B
C12 C13 C27 C32 C28 C29 C31 C30
1 1
2 2
3 3
4 4
Figure 2-19 DDIMM plug sequence for Power S1014 servers

Chapter 2. Architecture and technical overview 79


All memory modules are attached to the first chip (chip-0) of the single eSCM (eSCM-0) and
are of the same type as highlighted by the cells that are shaded in green in Figure 2-19 on
page 79.

The memory controllers and the related open memory interface (OMI) channels are
highlighted in bright yellow in Figure 2-19 on page 79 and labeled OMI0, OMI1, OMI2, and
OMI3.

The related OMI links (subchannels A and B) are highlighted in light yellow in Figure 2-19 on
page 79 and the physical memory slot location codes are highlighted in light blue:
򐂰 First DDIMM pair is installed on links OMI1A and OMI1B in slots C27 and C32
򐂰 Second DDIMM pair is installed on links OMI1A and OMI1B in slots C12 and C13
򐂰 Third DDIMM pair is installed on links OMI2A and OMI2B in slots C28 and C29
򐂰 Fourth DDIMM pair is installed on links OMI3A and OMI3B in slots C31 and C30

Power S1022s memory feature and placement rules


Table 2-12 lists the available memory feature codes for Power S1022s servers. No specific
memory enablement features are required and the entire physical DDIMM capacity of a
memory feature is enabled by default.

Table 2-12 Memory feature codes for Power S1022s servers


Feature code Description

#EM6N 32 GB (2x16GB) DDIMMs, 3200 MHz, 8 Gbit DDR4 memory

#EM6W 64 GB (2x32GB) DDIMMs, 3200 MHz, 8 Gbit DDR4 memory

#EM6X 128 GB (2x64GB) DDIMMs, 3200 MHz, 16 Gbit DDR4 memory

#EM6Ya 256 GB (2x128GB) DDIMMs, 2666 MHz, 16 Gbit DDR4 memory


a. The 128 GB DDIMM parts in feature code #EM6Y are planned to be available on 18 November
2022

The memory DDIMMs are bundled in pairs by using the following feature codes:
򐂰 16 GB: #EM6N
򐂰 32 GB: #EM6W
򐂰 64 GB: #EM6X
򐂰 128 GB: #EM6Y

The Power S1022s server supports the Active Memory Mirroring (AMM) feature #EM8G.
AMM requires a minimum four configured memory feature codes with a total of eight DDIMM
modules.

80 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Memory rules for 1-socket Power S1022s servers
The minimum ordering granularity is one memory feature and all DDIMMs must be of the
same feature code type for a Power S1022s server in 1-socket configuration. A maximum of
four memory feature codes can be configured to cover all of the available eight memory slots.

The minimum memory capacity limit of the Power S1022s 1-socket server is 32 GB, which
can be fulfilled by one #EM6N feature.

The maximum memory capacity of the 1-socket Power S1022s is 1 TB. This specific
maximum configuration requires four #EM6Y memory features. Until the availability of the
128 GB memory DDIMMs (planned for 18 November 2022), the maximum memory capacity
is 512 GB.

Figure 2-20 shows the DDIMM plug sequence for Power S1022s servers in 1-socket
configurations (the rules are identical to the previously described for Power S1014 servers).

Power10 eSCM-0 Chip-0


OMI0 OMI1 OMI2 OMI3
A B A B A B A B
C12 C13 C27 C32 C28 C29 C31 C30
1 1
2 2
3 3
4 4
Figure 2-20 DDIMM plug sequence for Power S1022s 1-socket servers

All memory modules are attached to the first chip (chip-0) of the single eSCM (eSCM-0) and
are of the same type as highlighted in green in Figure 2-20.

The memory controllers and the related open memory interface (OMI) channels are
highlighted in bright yellow in Figure 2-20 and labeled OMI0, OMI1, OMI2, and OMI3.

The related OMI links (subchannels A and B) are highlighted in light yellow in Figure 2-20 and
the physical memory slot location codes are highlighted in light blue:
򐂰 First DDIMM pair is installed on links OMI1A and OMI1B in slots C27 and C32
򐂰 Second DDIMM pair is installed on links OMI1A and OMI1B in slots C12 and C13
򐂰 Third DDIMM pair is installed on links OMI2A and OMI2B in slots C28 and C29
򐂰 Fourth DDIMM pair is installed on links OMI3A and OMI3B in slots C31 and C30

Memory rules for 2-socket Power S1022s servers


If only one memory feature type is used throughout the Power S1022s server, the minimum
ordering granularity is one memory feature. In this scenario, all of the DDIMMs adhere to the
same technical specification.

Chapter 2. Architecture and technical overview 81


Figure 2-21 shows the DDIMM plug sequence for Power S1022s servers in 2-socket
configuration when only a single memory feature code type is used.

Power10 eSCM-0 Chip-0 Power10 eSCM-1 Chip-0


OMI0 OMI1 OMI2 OMI3 OMI0 OMI1 OMI2 OMI3
A B A B A B A B A B A B A B A B
C12 C13 C27 C32 C28 C29 C31 C30 C19 C20 C21 C40 C38 C39 C41 C42
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
Figure 2-21 DDIMM plug sequence for Power S1022s 2-socket servers with single memory feature

The memory modules are attached to the first chip (chip-0) of the first eSCM (eSCM-0) or to
the first chip (chip-0) of the second eSCM (eSCM-1) and are of the same type, as highlighted
in green in Figure 2-21.

The memory controllers and the related open memory interface (OMI) channels are
highlighted in bright yellow in Figure 2-21 and labeled OMI0, OMI1, OMI2, and OMI3.

The related OMI links (subchannels A and B) are highlighted in light yellow in Figure 2-21 and
the physical memory slot location codes are highlighted in light blue:
򐂰 First DDIMM pair is installed on links OMI1A and OMI1B in slots C27 and C32 of eSCM-0
򐂰 Second DDIMM pair is installed on links OMI1A and OMI1B in slots C21 and C40 of
eSCM-1
򐂰 Third DDIMM pair is installed on links OMI0A and OMI0B in slots C12 and C13 of eSCM-0
򐂰 Fourth DDIMM pair is installed on links OMI0A and OMI0B in slots C19 and C20 of
eSCM-1
򐂰 Fifth DDIMM pair is installed on links OMI2A and OMI2B in slots C28 and C29 of eSCM-0
򐂰 Sixth DDIMM pair is installed on links OMI2A and OMI2B in slots C38 and C39 of eSCM-1
򐂰 Seventh DDIMM pair is installed on links OMI3A and OMI3B in slots C31 and C30 of
eSCM-0
򐂰 Eighth DDIMM pair is installed on links OMI3A and OMI3B in slots C41 and C42 of
eSCM-1

If the 2-socket configuration is based on two different memory feature types, the minimum
ordering granularity is two identical memory feature codes (4 DDIMMs). All DDIMMs that are
attached to a eSCM must be of the same technical specification, which implies that they are
of the same memory feature code type.

It is not required to configure equal quantities of the two memory feature types. A maximum of
four configured entities of each memory feature type (eight DDIMMs of equal specification)
can be used.

Configurations with more than two memory feature types are not supported.

82 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Figure 2-22 shows the DDIMM plug sequence for Power S1022s servers in 2-socket
configuration when that two different memory feature code types are used.

Power10 eSCM-0 Chip-0 Power10 eSCM-1 Chip-0


OMI0 OMI1 OMI2 OMI3 OMI0 OMI1 OMI2 OMI3
A B A B A B A B A B A B A B A B
C12 C13 C27 C32 C28 C29 C31 C30 C19 C20 C21 C40 C38 C39 C41 C42
1 1
1 1
2 2
2 2
3 3
3 3
4 4
4 4
Figure 2-22 DDIMM plug sequence for Power S1022s 2-socket servers with two memory feature types

The memory modules of the first feature type are attached to the first chip (chip-0) of the first
eSCM (eSCM-0) and are highlighted in green in Figure 2-22. The memory modules of the
second feature type are attached to the first chip (chip-0) of the second eSCM (eSCM-1) and
are highlighted in purple.

The memory controllers and the related open memory interface (OMI) channels are
highlighted in bright yellow in Figure 2-22 and labeled OMI0, OMI1, OMI2, and OMI3 for both
eSCMs.

The related OMI links (subchannels A and B) are highlighted in light yellow in Figure 2-22 and
the physical memory slot location codes are highlighted in light blue. Each eSCM can be
viewed as an independent memory feature type domain with its own inherent plug sequence.

The following plug sequence is used for the memory feature type for eSCM-0:
򐂰 First DDIMM pair is installed on links OMI1A and OMI1B in slots C27 and C32 of eSCM-0
򐂰 Second DDIMM pair is installed on links OMI1A and OMI1B in slots C12 and C13 of
eSCM-0
򐂰 Third DDIMM pair is installed on links OMI2A and OMI2B in slots C28 and C29 of eSCM-0
򐂰 Fourth DDIMM pair is installed on links OMI3A and OMI3B in slots C31 and C30 of
eSCM-0

The following plug sequence is used for the memory feature type for eSCM-1:
򐂰 First DDIMM pair is installed on links OMI1A and OMI1B in slots C21 and C40 of eSCM-1
򐂰 Second DDIMM pair is installed on links OMI0A and OMI0B in slots C19 and C20 of
eSCM-1
򐂰 Third DDIMM pair is installed on links OMI2A and OMI2B in slots C38 and C39 of eSCM-1
򐂰 Fourth DDIMM pair is installed on links OMI3A and OMI3B in slots C41 and C42 of
eSCM-1

The maximum memory capacity of the 2-socket Power S1022s is 2 TB. This specific
maximum configuration requires eight #EM6Y memory features with a total of 16 128-GB
DDIMM modules. Until the availability of the 128 GB memory DDIMMs (planned for
18 November 2022), the maximum memory capacity is 1 TB.

Chapter 2. Architecture and technical overview 83


Power S1022 and Power S1024 memory feature and placement rules
Table 2-13 lists the available memory feature codes for Power S1022 servers. No specific
memory enablement features are required and the entire physical DDIMM capacity of a
memory feature is enabled by default.

Table 2-13 Memory feature codes for Power S1022 servers


Feature code Description

#EM6N 32 GB (2x16GB) DDIMMs, 3200 MHz, 8 Gbit DDR4 memory

#EM6W 64 GB (2x32GB) DDIMMs, 3200 MHz, 8 Gbit DDR4 memory

#EM6X 128 GB (2x64GB) DDIMMs, 3200 MHz, 16 Gbit DDR4 memory

#EM6Ya 256 GB (2x128GB) DDIMMs, 2666 MHz, 16 Gbit DDR4 memory


a. The 128 GB DDIMM parts in feature code #EM6Y are planned to be available on November
18th 2022

The 16 GB, 32 GB, 64 GB, and 128 GB memory DDIMMs for Power S1022 servers are
bundled in pairs through the related memory feature codes #EM6N, #EM6W, EM6X, and
EM6Y.

The DDIMMs of all memory features are in a form factor suitable for two rack units (2U) high
Power S1022 servers.

Table 2-14 lists the available memory feature codes for Power S1024 servers. No specific
memory enablement features are required and the entire physical DDIMM capacity of a
memory feature is enabled by default.

Table 2-14 Memory feature codes for Power S1024 servers


Feature code Description

#EM6N 32 GB (2x16GB) DDIMMs, 3200 MHz, 8-bit DDR4 memory

#EM6W 64 GB (2x32GB) DDIMMs, 3200 MHz, 8-Gbit DDR4 memory

#EM6X 128 GB (2x64GB) DDIMMs, 3200 MHz, 16-Gbit DDR4 memory

#EM6Ua 256 GB (2x128GB) DDIMMs, 2933 MHz, 16-Gbit DDR4 memory

#EM78a 512 GB (2x256GB) DDIMM, 2933 MHz, 16-Gbit DDR4 memory


a. The 128 GB DDIMM parts in feature code #EM6U and 256 GB DDIMM parts in feature code
#EM78 are planned to be available on November 18th 2022

The memory DDIMMs for Power S1024 servers are bundled by using the following memory
feature codes:
򐂰 16 GB: #EM6N
򐂰 32 GB: #EM6W
򐂰 64 GB: #EM6X
򐂰 128 GB: #EM6U
򐂰 256 GB: #EM78

The DDIMMs of the memory features #EM6N, #EM6W, and #EM6X are in a form factor of two
rack units (2U). The DDIMMs of this types are extended by spacers to fit in four rack units
(4U) high Power S1024 servers.

84 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
The 128 GB and 256 GB DDIMMs of memory features #EM6U and #EM78 are of higher
capacity that is compared with their 16 GB, 32 GB, and 64 GB counterparts; therefore, they
must fully use the 4U height of Power S1024 servers.

The Power S1024 server does not support a memory configuration that includes DDIMMs of
different form factors. All memory modules must be 2U DDIMM memory feature codes
(#EM6N, EM6W, and EM6X) or all memory modules must be 4U DDIMM memory feature
codes (EM6U and EM78).

Note: Power S1024 servers in 2-socket configuration do not support the 4U DDIMM
memory feature codes #EM6U and #EM78 if the RDX USB Internal Docking Station for
Removable Disk Cartridge feature is installed.

The Power S1022 and Power S1024 servers support the Active Memory Mirroring (AMM)
Feature Code #EM8G. AMM requires a minimum four configured memory feature codes with
a total of eight DDIMM modules.

The Power S1022 and Power S1024 server share most of the memory feature and placement
rules, which are described next.

Memory rules for 1-socket Power S1022 and Power S1024 servers
The minimum ordering granularity is one memory feature (two DDIMMs) and all DDIMMs
must be of the same Feature Code type for a Power S1022 or Power S1024 server in
1-socket configuration. A maximum of eight memory feature codes can be configured to cover
all of the available (16) memory slots.

The minimum memory capacity limit of the Power S1022 or the Power S1024 1-socket server
is 32 GB, which can be fulfilled by one #EM6N feature.

The maximum memory capacity of the Power S1022 in 1-socket configuration is 2 TB. This
specific maximum configuration requires eight #EM6Y memory features. Until the availability
of the 128 GB memory DDIMMs (planned for 18 November 2022), the maximum memory
capacity is 1 TB.

The maximum memory capacity of the Power S1024 in 1-socket configuration is 4 TB. This
specific maximum configuration requires eight #EM78 memory features. Until the availability
of the 128 GB memory DDIMMs and 256 GB memory DDIMMs (planned for 18 November
2022), the maximum memory capacity is 1 TB.

Chapter 2. Architecture and technical overview 85


Figure 2-23 shows the DDIMM plug sequence for Power S1022 or Power S1024 servers in
1-socket configuration.

Power10 DCM-0
Chip-0 Chip-1
OMI0 OMI1 OMI2 OMI3 OMI4 OMI5 OMI6 OMI7
A B A B A B A B A B A B A B A B
C12 C13 C27 C32 C28 C29 C31 C30 C18 C17 C16 C35 C37 C36 C34 C33
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
Figure 2-23 DDIMM plug sequence for Power S1022 and Power S1024 1-socket servers

The memory modules are attached to the first chip (chip-0) or the second chip (chip-1) of the
configured DCM (DCM-0). All memory modules are of the same type as highlighted in green
in Figure 2-23.

The memory controllers and the related OMI channels are highlighted in bright yellow in
Figure 2-23 and labeled OMI0, OMI1, OMI2, OMI3, OMI4, OMI5, OMI6, OMI7, and OMI8.

The related OMI links (subchannels A and B) are highlighted in light yellow in Figure 2-23 and
the physical memory slot location codes are highlighted in light blue:
򐂰 First DDIMM pair is installed on links OMI1A and OMI1B in slots C27 and C32 of chip-0
򐂰 Second DDIMM pair is installed on links OMI5A and OMI5B in slots C16 and C35 of chip-1
򐂰 Third DDIMM pair is installed on links OMI0A and OMI0B in slots C12 and C13 of chip-0
򐂰 Fourth DDIMM pair is installed on links OMI4A and OMI4B in slots C18 and C17 of chip-1
򐂰 Fifth DDIMM pair is installed on links OMI2A and OMI2B in slots C28 and C29 of chip-0
򐂰 Sixth DDIMM pair is installed on links OMI6A and OMI6B in slots C37 and C36 of chip-1
򐂰 Seventh DDIMM pair is installed on links OMI3A and OMI3B in slots C31 and C30 of
chip-0
򐂰 Eighth DDIMM pair is installed on links OMI7A and OMI7B in slots C34 and C33 of chip-1

Memory rules for 2-socket Power S1022 and Power S1024 servers
The minimum ordering granularity is two identical memory feature codes (four DDIMMs) for
Power S1022 or Power S1024 server in 2-socket configuration.

The minimum memory capacity limit of the Power S1022 or the Power S1024 2-socket server
is 64 GB, which can be fulfilled by two #EM6N features.

The maximum memory capacity of the Power S1022 in 2-socket configuration is 4 TB. This
specific maximum configuration requires 16 #EM6Y memory features. Until the availability of
the 128 GB memory DDIMMs (planned for 18 November 2022), the maximum memory
capacity is 2 TB.

86 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
The maximum memory capacity of the Power S1024 in 2-socket configuration is 8 TB. This
specific maximum configuration requires 16 #EM78 memory features. Until the availability of
the 128 GB memory DDIMMs and 256 GB memory DDIMMs (planned for 18 November
2022), the maximum memory capacity is 2 TB.

Regarding the memory plugging rules, the following configuration scenarios are supported
and must be considered separately:
򐂰 Only one memory feature type is used across both sockets and all of the DDIMMs adhere
to the same technical specification.
򐂰 Two different memory feature codes with the corresponding different DDIMM
characteristics are configured. Each memory feature code type is assigned in a
one-to-one relation to one of the two DCM sockets.
It is not required to configure equal quantities of the two memory feature types. A
maximum of eight configured entities of each memory feature type (16 DDIMMs of equal
specification) are allowed.

Note: The Power S1022 nor the Power S1024 servers support memory configurations that
are based on more than two memory feature types.

Figure 2-24 shows the DDIMM plug sequence for Power S1022 and Power S1024 servers in
2-socket configuration when only a single memory feature code type is used. Each chip
(chip-0 and chip-1) of each DCM (DCM-0 and DCM-1) provide four memory channels for
memory module access. All memory DDIMMs are of the same type, as highlighted in green in
Figure 2-24.

Power10 DCM-0 Power10 DCM-1


Chip-0 Chip-1 Chip-0 Chip-1
OMI0 OMI1 OMI2 OMI3 OMI4 OMI5 OMI6 OMI7 OMI0 OMI1 OMI2 OMI3 OMI4 OMI5 OMI6 OMI7
A B A B A B A B A B A B A B A B A B A B A B A B A B A B A B A B
C12 C13 C27 C32 C28 C29 C31 C30 C18 C17 C16 C35 C37 C36 C34 C33 C19 C20 C21 C30 C38 C39 C41 C42 C25 C26 C48 C43 C47 C46 C44 C45

1 1 1 1
2 2 2 2
3 3 3 3
4 4 4 4
5 5 5 5
6 6 6 6
7 7 7 7
8 8 8 8
Figure 2-24 DDIMM plug sequence for Power S1022 and Power S1024 2-socket servers

The memory controllers and the related OMI channels are highlighted in bright yellow in
Figure 2-24 and labeled OMI0, OMI1, OMI2, OMI3, OMI4, OMI5, OMI6, OMI7, and OMI8 for
each configured DCM. The related OMI links (subchannels A and B) are highlighted in light
yellow in Figure 2-24 and the physical memory slot location codes are highlighted in light
blue:
򐂰 First double DDIMM pair is installed on links OMI1A and OMI1B in slots C27 and C32 of
chip-0 on DCM-0 and OMI1A and OMI1B in slots C21 and C30 of chip-0 on DCM-1
򐂰 Second double DDIMM pair is installed on links OMI5A and OMI5B in slots C16 and C35
of chip-1 on DCM-0 and OMI5A and OMI5B in slots C48 and C43 of chip-1 on DCM-1
򐂰 Third double DDIMM pair is installed on links OMI0A and OMI0B in slots C12 and C13 of
chip-0 on DCM-0 and OMI0A and OMI0B in slots C19 and C20 of chip-0 on DCM-1

Chapter 2. Architecture and technical overview 87


򐂰 Fourth double DDIMM pair is installed on links OMI4A and OMI4B in slots C18 and C17 of
chip-1 on DCM-0 and OMI4A and OMI4B in slots C25 and C26 of chip-1 on DCM-1
򐂰 Fifth double DDIMM pair is installed on links OMI2A and OMI2B in slots C28 and C29 of
chip-0 on DCM-0 and OMI2A and OMI2B in slots C38 and C39 of chip-0 on DCM-1
򐂰 Sixth double DDIMM pair is installed on links OMI6A and OMI6B in slots C37 and C36 of
chip-1 on DCM-0 and OMI6A and OMI6B in slots C47 and C46 of chip-1 on DCM-1
򐂰 Seventh double DDIMM pair is installed on links OMI3A and OMI3B in slots C31 and C30
of chip-0 on DCM-0 and OMI3A and OMI3B in slots C41 and C42 of chip-0 on DCM-1
򐂰 Eighth double DDIMM pair is installed on links OMI7A and OMI7B in slots C34 and C33 of
chip-1 on DCM-0 and OMI7A and OMI7B in slots C44 and C45 of chip-1 on DCM-1

Figure 2-25 shows the DDIMM plug sequence for Power S1022 and Power S1024 servers in
2-socket configuration when two different memory feature code types are used.

P o we r 10 D C M -0 P o we r 10 D C M -1
C hip-0 C hip-1 C hip-0 C hip-1
OM I0 OM I1 OM I2 OM I3 OM I4 OM I5 OM I6 OM I7 OM I0 OM I1 OM I2 OM I3 OM I4 OM I5 OM I6 OM I7
A B A B A B A B A B A B A B A B A B A B A B A B A B A B A B A B
C12 C13 C27 C32 C28 C29 C31 C30 C18 C17 C16 C35 C37 C36 C34 C33 C19 C20 C21 C30 C38 C39 C41 C42 C25 C26 C48 C43 C47 C46 C44 C45

1 1 1 1
2 2 2 2
3 3 3 3
4 4 4 4
5 5 5 5
6 6 6 6
7 7 7 7
8 8 8 8

P o we r 10 D C M -0 P o we r 10 D C M -1
C hip-0 C hip-1 C hip-0 C hip-1
OM I0 OM I1 OM I2 OM I3 OM I4 OM I5 OM I6 OM I7 OM I0 OM I1 OM I2 OM I3 OM I4 OM I5 OM I6 OM I7
A B A B A B A B A B A B A B A B A B A B A B A B A B A B A B A B
C12 C13 C27 C32 C28 C29 C31 C30 C18 C17 C16 C35 C37 C36 C34 C33 C19 C20 C21 C30 C38 C39 C41 C42 C25 C26 C48 C43 C47 C46 C44 C45

1 1 1 1
1 1 1 1
2 2 2 2
2 2 2 2
3 3 3 3
3 3 3 3
4 4 4 4
4 4 4 4

Figure 2-25 DDIMM plug sequence for Power S1022 and Power S1024 2-socket servers

The memory modules of the first memory feature type are attached to the first chip (chip-0)
and second chip (chip-1) of the first DCM (DCM-0) as highlighted in green in FIGURE. The
memory modules of the second memory feature type are attached to the first chip (chip-0)
and second chip (chip-1) of the second DCM (DCM-1) as highlighted purple in Figure 2-25.

Each DCM can be viewed as an independent memory feature type domain with its own
inherent plug sequence.

The following plug sequence is used for the memory feature type for DCM-0:
򐂰 First double DDIMM pair is installed on links OMI1A and OMI1B in slots C27 and C32 of
chip-0 and OMI5A and OMI5B in slots C16 and C35 of chip-1 on DCM-0
򐂰 Second double DDIMM pair is installed on links OMI0A and OMI0B in slots C12 and C13
of chip-0 and OMI4A and OMI4B in slots C16 and C35 of chip-1 on DCM-0
򐂰 Third double DDIMM pair is installed on links OMI2A and OMI2B in slots C28 and C29 of
chip-0 and OMI6A and OMI6B in slots C37 and C36 of chip-1 on DCM-0
򐂰 Fourth double DDIMM pair is installed on links OMI3A and OMI3B in slots C31 and C30 of
chip-0 and OMI7A and OMI7B in slots C34 and C33 of chip-1 on DCM-0

88 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
The following plug sequence is used for the memory feature type for DCM-1:
򐂰 First double DDIMM pair is installed on links OMI1A and OMI1B in slots C21 and C40 of
chip-0 and OMI5A and OMI5B in slots C48 and C43 of chip-1 on DCM-1
򐂰 Second double DDIMM pair is installed on links OMI0A and OMI0B in slots C19 and C20
of chip-0 and OMI4A and OMI4B in slots C25 and C26 of chip-1 on DCM-1
򐂰 Third double DDIMM pair is installed on links OMI2A and OMI2B in slots C38 and C39 of
chip-0 and OMI6A and OMI6B in slots C47 and C46 of chip-1 on DCM-1
򐂰 Fourth double DDIMM pair is installed on links OMI3A and OMI3B in slots C41 and C42 of
chip-0 and OMI7A and OMI7B in slots C44 and C45 of chip-1 on DCM-1

2.3.2 Memory bandwidth


The Power10 DCM supports 16 open memory interface (OMI) links, and the Power10 eSCM
supports eight OMI links. One DDIMM is driven by each OMI link in turn. One OMI link
represents a bundle of eight lanes that can transfer 1 byte with one transaction per direction.

The Power10 processor-based scale-out servers offers four different DDIMM sizes for all
server models: 16 GB, 32 GB, 64 GB, and 128 GB. The 16 GB, 32 GB, and 64 GB DDIMMs
run at a data rate of 3200 Mbps.

The DDIMMs of 128 GB capacity and 2U form factor are configurable for Power S1014,
Power S1022s, and Power S1024 servers and run at a data rate of 2666 Mbps.

The 128 GB DDIMMs of 4U form factor are exclusively available for Power S1024 servers,
which run at a slightly higher data rate of 2933 Mbps. Only Power S1024 servers can use an
other 4U form factor DDIMM type that holds 256 GB of data and is running at 2933 Mbps.

Table 2-15 lists the available DDIMM capacities and their related maximum theoretical
bandwidth figures per OMI link, Power10 eSCM, and Power10 DCM.

Table 2-15 Memory bandwidth of supported DDIMM sizes


DDIMM capacity DRAM data Maximum Maximum Maximum
rate theoretical theoretical theoretical
bandwidth per bandwidth per bandwidth per
OMI link eSCMa DCMa

16 GB, 32 GB, 64 GB 3200 Mbps 25.6 GBps 204.8 GBps 409.6 GBps

128 GB 2666 Mbps 21.3 GBps 170.4 GBps 340.8 GBps

128 GB, 256 GB 2933 Mbps 23.5 GBps 187.7 GBps 375.4 GBps
a. DDIMM modules that are attached to one DCM or eSCM must be all of the same size.

Memory bandwidth considerations


Power10 processor-based scale-out servers are memory performance-optimized with eight
DDIMM slots that are available per eSCM processor module and 16 DDIMM slots per DCM
processor module.

Each DDIMM slot is serviced by one OMI link (memory subchannel). The maximum
bandwidth of the system depends on the number of OMI links that are used and the data rate
of the DDIMMs that populate the configured links.

Chapter 2. Architecture and technical overview 89


To calculate the maximum memory bandwidth of a system, use the following formula:
򐂰 1-socket and 2-socket configurations are based on one memory feature code type:
Maximum Bandwidth = Number of DDIMMs x maximum bandwidth per OMI link as
determined by the DRAM data rate
򐂰 2-socket configurations based on two memory feature code type:
Maximum Bandwidth = Number of DDIMMs of the first memory feature code type x
maximum bandwidth per OMI link as determined by the related DRAM data rate + Number
of DDIMMs of the second memory feature code type x maximum bandwidth per OMI link
as determined by the related DRAM data rate.

Important: For the best possible performance, it is generally recommended that memory
is installed evenly in all memory slots and across all configured processor modules.

Balancing memory across the installed system planar cards enables memory access in a
consistent manner and typically results in better performance for your configuration.

Table 2-16 lists the maximum memory bandwidth for the eSCM-based Power S1014 and
Power S1022s servers, depending on the number of DDIMMs that are used and the DRAM
data rate of the selected DDIMM type. The listing accounts for the minimum memory feature
code order granularity. Unsupported configurations are indicated by a “—” hyphen.

Table 2-16 Maximum memory bandwidth for the Power S1014 and Power S1022s servers
DDIMM Maximum bandwidth based on Maximum bandwidth based on
quantity 3200 Mbps data rate DDIMMs (GBps)a 2666 Mbps data rate DDIMMs (GBps)a

Power S1014 Power S1022s Power S1014 Power S1022s

2 51 51 43 43

4 102 102 85 85

6 154 154 128 128

8 205 205 170 170

10 — 256 — 213

12 — 307 — 256

14 — 358 — 298

16 — 410 — 341
a. Numbers are rounded to the nearest integer.

90 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Table 2-17 lists the maximum memory bandwidth for the DCM-based Power S1022 and
Power S1024 servers, depending on the number of DDIMMs that are used, the DRAM data
rate of the selected DDIMM type, and the number of configured sockets. The listing accounts
for the minimum memory feature code order granularity and pertains to configurations that
are based on only one single memory feature code type. Unsupported configurations are
indicated by a “—” hyphen.

Table 2-17 Maximum memory bandwidth for the Power S1022 and Power S1024 servers
DDIMM Power S1022 and Power S1024 Power S1024 maximum bandwidth based
quantity maximum bandwidth based on on 2933 Mbps data rate DDIMMs (GBs)a
3200 Mbps data rate DDIMMs (GBs)a

1-socket 2-socket 1-socket 2-socket


configuration configuration configuration configuration

2 51 — 47 —

4 102 102 94 94

6 154 — 141 —

8 205 205 188 188

10 256 — 235 —

12 307 307 282 282

14 358 — 329 —

16 410 410 376 376

18 — — — —

20 — 512 — 470

22 — — — —

24 — 614 — 564

26 — — — —

28 — 717 — 658

30 — — — —

32 — 819 — 752
a. Numbers are rounded to the nearest integer.

Chapter 2. Architecture and technical overview 91


2.4 Internal I/O subsystem
The internal I/O subsystem of the Power S1014, S1022s, S1022, and S1024 servers is
connected to the PCIe Express controllers on the Power10 chips in the system. Each chip
features two PCI Express controllers (PECs), which support up to three PCI host bridges
(PHBs) that directly connect to a PCIe slot or device.

PEC0 and PEC1 can be configured as shown in the following example:


򐂰 1 x16 Gen4 PHB/1 x8 Gen5 PHB
򐂰 OR 1 x8 Gen5 and 1 x8 Gen4 PHB
򐂰 OR 1 x8 Gen5 PHB and 2 x4 Gen4 PHBs

The Power10 chip is installed in pairs in a DCM or eSCM that plugs into a socket in the
system board of the systems.

The following versions of Power10 processor modules are used on the Power10
processor-based scale-out servers:
򐂰 A DCM in which both chips are fully functional with cores, memory, and I/O.
򐂰 An eSCM in which the first chip (P0) is fully function with cores, memory, and I/O and the
second chip (P1) supports I/O only.

The PCIe slot internal connections of 2 DCM server are shown in Figure 2-26.

C 4 = G 4 x 1 6 o r G 5 x 8 s lo t
C 1 1 = G 5 x 8 w / x 1 6 c s lo t PEC 0 PEC 0
C 3 = G 4 x 1 6 o r G 5 x 8 s lo t
D C M -0 D C M -1
P 0 / C h i p -0 P 2 / C h i p -0

C 1 0 = G 4 x 1 6 o r G 5 x 8 s lo t PEC 1 PEC 1

C 9 = G 5 x 8 w / x 1 6 c s lo t C 2 = G 5 x 8 w / x 1 6 c s lo t
PEC 0 PEC 0
C 8 = G 4 x 8 w / x 1 6 c s lo t C 1 = G 4 x 8 w / x 1 6 c s lo t
D C M -0 D C M -1
P 1 / C h i p -1 P 3 / C h i p -1
( f u lly f u n c t io n o r I/O o n ly) ( f u lly f u n c t io n o r I/O o n ly)

C 7 = G 5 x 8 w / x 1 6 c s lo t PEC 1 PEC 1
C 0 = G 4 x 1 6 o r G 5 x 8 s lo t

Figure 2-26 PCIe slot internal connections of a 2-socket DCM server

All PCIe slots support hot-plug adapter installation and maintenance when service
procedures are used that are activated by way of the eBMC or HMC interfaces, and enhanced
error handling (EEH).

PCIe EEH-enabled adapters respond to a special data packet that is generated from the
affected PCIe slot hardware by calling system firmware, which examines the affected bus,
allows the device driver to reset it, and continues without a system restart.

For Linux, EEH support extends to most of the frequently used devices, although some
third-party PCI devices might not provide native EEH support.

92 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
All PCIe adapter slots support hardware-backed network virtualization through single root IO
virtualization (SR-IOV) technology. Configuring an SR-IOV adapter into SR-IOV shared mode
might require more hypervisor memory. If sufficient hypervisor memory is not available, the
request to move to SR-IOV shared mode fails. The user is then instructed to free up extra
memory and attempt the operation again.

The server PCIe slots are allocated DMA space that use the following algorithm:
򐂰 All slots are allocated a 2 GB default DMA window.
򐂰 All I/O adapter slots (except the embedded USB) are allocated Dynamic DMA Window
(DDW) capability that is based on installed platform memory. DDW capability is calculated
assuming 4 K I/O mappings. Consider the following points:
– For systems with less than 64 GB of memory, slots are allocated 16 GB of DDW
capability.
– For systems with at least 64 GB of memory, but less than 128 GB of memory, slots are
allocated 32 GB of DDW capability.
– For systems with 128 GB or more of memory, slots are allocated 64 GB of DDW
capability.
– Slots can be enabled with Huge Dynamic DMA Window capability (HDDW) by using
the I/O Adapter Enlarged Capacity setting in the ASMI.
– HDDW enabled slots are allocated enough DDW capability to map all of installed
platform memory by using 64 K I/O mappings.
– Minimum DMA window size for HDDW enabled slots is 32 GB.
– Slots that are HDDW enabled are allocated the larger of the calculated DDW and
HDDW capability.

The x16 slots can provide up to twice the bandwidth of x8 slots because they offer twice as
many PCIe lanes. PCIe Gen5 slots can support up to twice the bandwidth per lane of a PCIe
Gen4 slot, and PCIe Gen4 slots can support up to twice the bandwidth per lane of a PCIe
Gen3 slot.

The servers are smart about energy efficiency when cooling the PCIe adapter environment.
They sense which IBM PCIe adapters are installed in their PCIe slots and, if an adapter
requires higher levels of cooling, they automatically speed up fans to increase airflow across
the PCIe adapters. Faster fans increase the sound level of the server. Higher wattage PCIe
adapters include the PCIe3 SAS adapters and SSD/flash PCIe adapters (#EJ10, #EJ14, and
#EJ0J).

Chapter 2. Architecture and technical overview 93


2.4.1 Slot configuration
The various slot configurations are described in this section. For each server model, the PCIe
slot locations are listed by the slot type and the OpenCAPI capability. The I/O adapter
enlarged capacity enablement order is provided for each individual PCIe adapter slot.

Slot configuration for the Power S1014 server


The Power S1014 server features up to 21 PCIe hot-plug slots (16 U.2 NVMe plus up to five
PCIe add-in cards), which provides excellent configuration flexibility and expandability. The
PCIe slots are enabled to support the cable adapter (#EJ2A) that is used to attach the EMX0
PCIe Gen3 I/O expansion drawer.

With one Power10 processor, five PCIe slots are available:


򐂰 One PCIe x16 Gen4 or x8 Gen5, full-height, half length slot (OpenCAPI)
򐂰 Two PCIe x8 (x16 connector) Gen5, full-height, half-length slots (OpenCAPI)
򐂰 One PCIe x8 (x16 connector) Gen5, full-height, half-length slot
򐂰 One PCIe x8 (x16 connector) Gen4, full-height, half-length slot (OpenCAPI)

Table 2-18 lists the available PCIe slot types and the related slot location codes in Power
S1014 server.

Table 2-18 PCIe slot locations for a slot type in the Power S1014 server
Slot type Number of slots Location codes Adapter size

PCIe4 x16 or PCIe5 x8 1 P0-C10 Full-height, half-length


slots

PCIe5 x8 with x16 3 P0-C7, P0-C9, and Full-height, half-length


connector P0-C11

PCIe4 x8 with x16 1 P0-C8 Full-height, half-length


connector

eBMC 1 P0-C5

OpenCAPI Only 1 P0-C6 Full-height, half-length

Table 2-19 lists the PCIe adapter slot locations and related characteristics for the Power
S1014 server.

Table 2-19 PCIe slot locations and capabilities for the Power S1014 server
Location code Description Processor OpenCAPI I/O adapter
module capable enlarged
capacity
enablement
ordera

P0-C5b eBMC

P0-C6c OpenCAPI x16 Yes


connector

P0-C7 PCIe5 x8 with DCM0-P1-E1-PH Yes 3


x16 connector B3

P0-C8d PCIe4 x8 with DCM0-P1-E0-PH Yes 4


x16 connector B1

94 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Location code Description Processor OpenCAPI I/O adapter
module capable enlarged
capacity
enablement
ordera

P0-C9 PCIe5 x8 with DCM0-P1-E0-PH Yes 5


x16 connector B0

P0-C10d PCIe4 x16 or DCM0-P0-E1-PH Yes 1


PCIe5 x8 slots B3

P0-C11e PCIe5 x8 with DCM0-P0-E0-PH No 2


x16 connector B0
a. Enabling the I/O adapter enlarged capacity option affects only Linux partitions.
b. Only used for eBMC.
c. OpenCAPI only.
d. The NVMe JBOF adapter that is cabled to drive backplane P1 is supported in slot P0-C8 or
P0-C10.
e. The NVMe JBOF adapter that is connected to drive backplane P2 is supported in slot P0-C11.

Note: Consider the following points:


򐂰 Slot P0-C7 is a slot with PCIe x8 buses direct from the system processor modules and
can be used to install high-performance adapters. The adapter priority for this slot is for
the high-performance adapters, followed by any other adapters.
򐂰 Slot P0-C8 is a slot with PCIe x16 buses direct from the system processor modules and
can be used to install high-performance adapters. The adapter priority for these slots is
for CAPI accelerator adapters, PCI accelerator adapters, high-performance adapters,
followed by any other adapters.

Figure 2-27 shows the rear view of the Power S1014 server with the location codes for the
PCIe adapter slots.

Figure 2-27 Rear view of a Power S1014 server with PCIe slots location codes

Chapter 2. Architecture and technical overview 95


Slot configuration for the Power S1022s and S1022 servers
The Power S1022s and S1022 servers feature up to 18 PCIe hotplug slots (eight U.2 NVMe
plus up to 10 PCIe add-in cards), which provides excellent configuration flexibility and
expandability. The PCIe slots are enabled to support the cable adapter (#EJ24) that is used to
attach the EMX0 PCIe Gen3 I/O expansion drawer.

With two Power10 processors, 10 PCIe slots are available:


򐂰 Two x16 Gen4 or x8 Gen5 half-height, half-length slots (OpenCAPI)
򐂰 Two x16 Gen4 or x8 Gen5 half-height, half-length slots
򐂰 Two x8 Gen5 half-height, half-length slots (with x16 connectors) (OpenCAPI)
򐂰 Two x8 Gen5 half-height, half-length slots (with x16 connectors)
򐂰 Two x8 Gen4 half-height, half-length slots (with x16 connectors) (OpenCAPI)

With one Power10 processor, five PCIe slots are available:


򐂰 One PCIe x16 Gen4 or x8 Gen5, half-height, half-length slots (OpenCAPI)
򐂰 Two PCIe x8 (x16 connector) Gen5, half-height, half-length slots (OpenCAPI)
򐂰 One PCIe x8 (x16 connector) Gen5, half-height, half-length slots
򐂰 One PCIe x8 (x16 connector) Gen4, half-height, half-length slots (OpenCAPI)

Table 2-20 lists the available PCIe slot types and the related slot location codes in Power
S1022s and S1022 servers.

Table 2-20 PCIe slot locations for a slot type in the Power S1022s and S1022 servers
Slot type Number of slots Location codes Adapter size

PCIe4 x16 or PCIe5 x8 4 P0-C0, P0-C3, P0-C4, Half-height, half-length


slots and P0-C10

PCIe5 x8 with x16 4 P0-C2, P0-C7, P0-C9, Half-height, half-length


connector and P0-C11

PCIe4 x8 with x16 2 P0-C1 and P0-C8 Half-height, half-length


connector

eBMC 1 P0-C5

OpenCAPI Only 1 P0-C6 Half-height, half-length

Table 2-21 lists the PCIe adapter slot locations and related characteristics for the Power
S1022s and S1022 servers.

Table 2-21 PCIe slot locations and capabilities for the Power S1022s and S1022 servers
Location code Description Processor OpenCAPI I/O adapter
module capable enlarged
capacity
enablement
ordera

P0-C0 PCIe4 x16 or DCM1-P1-E1-PH No 2


PCIe5 x8 slots B3

P0-C1 PCIe4 x8 with DCM1-P1-E0-PH Yes 8


x16 connector B1

P0-C2 PCIe5 x8 with DCM1-P1-E0-PH No 10


x16 connector B0

96 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Location code Description Processor OpenCAPI I/O adapter
module capable enlarged
capacity
enablement
ordera

P0-C3 PCIe4 x16 or DCM1-P0-E1-PH Yes 3


PCIe5 x8 slots B3

P0-C4 PCIe4 x16 or DCM1-P0-E0-PH No 4


PCIe5 x8 slots B0

P0-C5b eBMC

P0-C6c OpenCAPI x16 Yes


connector

P0-C7 PCIe5 x8 with DCM0-P1-E1-PH Yes 6


x16 connector B3

P0-C8d PCIe4 x8 with DCM0-P1-E0-PH Yes 7


x16 connector B1

P0-C9 PCIe5 x8 with DCM0-P1-E0-PH Yes 9


x16 connector B0

P0-C10d PCIe4 x16 or DCM0-P0-E1-PH Yes 1


PCIe5 x8 slots B3

P0-C11e PCIe5 x8 with DCM0-P0-E0-PH No 5


x16 connector B0
a. Enabling the I/O adapter enlarged capacity option affects only Linux partitions.
b. Only used for eBMC.
c. OpenCAPI only.
d. The NVMe JBOF adapter that is cabled to drive backplane P1 is supported in slot P0-C8 or
P0-C10.
e. The NVMe JBOF adapter that is connected to drive backplane P2 is supported in slot P0-C11.

Note: Consider the following points:


򐂰 Slots P0-C1 and P0-C7 are slots with PCIe x8 buses direct from the system processor
modules and can be used to install high-performance adapters. The adapter priority for
this slot is for the high-performance adapters, followed by any other adapters.
򐂰 Slots P0-C2, P0-C3, and P0-C8 are slots with PCIe x16 buses direct from the system
processor modules and can be used to install high-performance adapters. The adapter
priority for these slots is for CAPI accelerator adapters, PCI accelerator adapters,
high-performance adapters, followed by any other adapters.

Figure 2-28 shows the rear view of the Power S1022s and S1022 servers with the location
codes for the PCIe adapter slots.

Figure 2-28 Rear view of Power S1022s and S1022 servers with PCIe slots location codes

Chapter 2. Architecture and technical overview 97


Slot configuration for the Power S1024 server
The Power S1024 server features up to 26 PCIe hotplug slots (16 U.2 NVMe plus up to 10
PCIe add-in cards), which provides excellent configuration flexibility and expandability. The
PCIe slots are enabled to support the cable adapter (#EJ2A) that is used to attach the EMX0
PCIe Gen3 I/O expansion drawer.

With two Power10 processor DCMs, 10 PCIe slots are available:


򐂰 Two x16 Gen4 or x8 Gen5 full-height, half-length slots (CAPI)
򐂰 Two x16 Gen4 or x8 Gen5 full-height, half-length slots
򐂰 Two x8 Gen5 full-height, half-length slots (with x16 connectors) (CAPI)
򐂰 Two x8 Gen5 full-height, half-length slots (with x16 connectors)
򐂰 Two x8 Gen4 full-height, half-length slots (with x16 connectors) (CAPI)

With one Power10 processor DCM, five PCIe slots are available:
򐂰 One PCIe x16 Gen4 or x8 Gen5, full-height, half-length slots (CAPI)
򐂰 Two PCIe x8 Gen5, full-height, half-length slots (with x16 connector) (CAPI)
򐂰 One PCIe x8 Gen5, full-height, half-length slots (with x16 connector)
򐂰 One PCIe x8 Gen4, full-height, half-length slots (with x16 connector) (CAPI)

Table 2-22 lists the available PCIe slot types and related slot location codes in the Power
S1024 server.

Table 2-22 PCIe slot locations for each slot type in the Power S1024 server
Slot type Number of slots Location codes Adapter size

PCIe4 x16 or PCIe5 x8 4 P0-C0, P0-C3, P0-C4, Full-height, half-length


slots and P0-C10

PCIe5 x8 with x16 4 P0-C2, P0-C7, P0-C9, Full-height, half-length


connector and P0-C11

PCIe4 x8 with x16 2 P0-C1 and P0-C8 Full-height, half-length


connector

eBMC 1 P0-C5

OpenCAPI Only 1 P0-C6 Full-height, half-length

Table 2-23 lists the PCIe adapter slot locations and related characteristics for the Power
S1024 server.

Table 2-23 PCIe slot locations and capabilities for the Power S1024 servers
Location code Description Processor OpenCAPI I/O adapter
module capable enlarged
capacity
enablement
ordera

P0-C0 PCIe4 x16 or DCM1-P1-E1-PH No 2


PCIe5 x8 slots B3

P0-C1 PCIe4 x8 with DCM1-P1-E0-PH Yes 8


x16 connector B1

P0-C2 PCIe5 x8 with DCM1-P1-E0-PH No 10


x16 connector B0

98 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Location code Description Processor OpenCAPI I/O adapter
module capable enlarged
capacity
enablement
ordera

P0-C3 PCIe4 x16 or DCM1-P0-E1-PH Yes 3


PCIe5 x8 slots B3

P0-C4 PCIe4 x16 or DCM1-P0-E0-PH No 4


PCIe5 x8 slots B0

P0-C5b eBMC

P0-C6c OpenCAPI x16 Yes


connector

P0-C7 PCIe5 x8 with DCM0-P1-E1-PH Yes 6


x16 connector B3

P0-C8d PCIe4 x8 with DCM0-P1-E0-PH Yes 7


x16 connector B1

P0-C9 PCIe5 x8 with DCM0-P1-E0-PH Yes 9


x16 connector B0

P0-C10d PCIe4 x16 or DCM0-P0-E1-PH Yes 1


PCIe5 x8 slots B3

P0-C11e PCIe5 x8 with DCM0-P0-E0-PH No 5


x16 connector B0
a. Enabling the I/O adapter enlarged capacity option affects only Linux partitions.
b. Only used for eBMC.
c. OpenCAPI only.
d. The NVMe JBOF adapter that is cabled to drive backplane P1 is supported in either slot P0-C8
or P0-C10.
e. The NVMe JBOF adapter that is connected to drive backplane P2 is supported in slot P0-C11.

Note: Consider the following points:


򐂰 Slots P0-C1 and P0-C7 are slots with PCIe x8 buses direct from the system processor
modules and can be used to install high-performance adapters. The adapter priority for
this slot is for the high-performance adapters, followed by any other adapters.
򐂰 Slots P0-C2, P0-C3, and P0-C8 are slots with PCIe x16 buses direct from the system
processor modules and can be used to install high-performance adapters. The adapter
priority for these slots is for CAPI accelerator adapters, PCI accelerator adapters,
high-performance adapters, followed by any other adapters.

Chapter 2. Architecture and technical overview 99


Figure 2-29 shows the rear view of the Power S1024 server with the location codes for the
PCIe adapter slots.

Figure 2-29 Rear view of a Power S1024 server with PCIe slots location codes

2.4.2 System port


The Power S1014, S1022s, S1022, and S1024 servers do not include a system port.

2.5 Enterprise Baseboard Management Controller


The Power10 scale-out systems use an eBMC for system service management, monitoring,
maintenance, and control. The eBMC also provides access to the system event log files
(SEL).

The eBMC is a specialized service processor that monitors the physical state of the system
by using sensors. A system administrator or service representative can communicate with the
eBMC through an independent connection.

2.5.1 Managing the system by using the ASMI GUI


The ASMI is the GUI to the eBMC. It is similar in terms of its function to the ASMI of Flexible
Service Processor (FSP)-managed servers (for example, the Power E1080 server), but it is a
complete redesign of the UI that was driven by customer feedback that was received during a
Design Thinking workshop.

To enter the ASMI GUI, you can use the HMC by selecting the server and then selecting
Operations → Launch Advanced System Management. A window opens that displays the
name of the system; model, type, and serial; and the IP of the service processor (eBMC).
Click OK and the ASMI window opens.

If the eBMC is connected to a network that also is accessible from your workstation, you can
connect directly by entering https://<eBMC IP> in your web browser.

100 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Figure 2-30 show the ASMI login window.

Figure 2-30 ASMI login window

When you log in for the first time, the default username and password is admin, but
invalidated. That is, after the first login, you must immediately change the admin password.

This change also must be made after a factory reset of the system. This policy change helps
to enforce that the eBMC is not left in a state with a well-known password, which improves the
security posture of the system.

The password must meet specific criteria (for example, a password of abcd1234 is invalid).

For more information about password rules, see this IBM Documentation web page.

Chapter 2. Architecture and technical overview 101


After you log in, the Overview window is shown, which includes server, firmware, network,
power, and status information, as shown in Figure 2-31.

Figure 2-31 ASMI overview window

The new ASMI for eBMC managed servers feature some important differences from the ASMI
version that is used by FSP-based systems. It also delivers some valuable new features:
򐂰 Update system firmware
A firmware update can be installed for the server by using the ASMI GUI, even if the
system is managed by an HMC. In this case, the firmware update always is disruptive.
To install a concurrent firmware update, the HMC must be used, which is not possible by
using the ASMI GUI.
򐂰 Download memory dumps
Memory dumps can be downloaded by using the HMC. Also, they also download them
from the ASMI menu if necessary.
It also is possible to start a memory dump from the ASMI. Click Logs → Dumps and then,
select the memory dump type and click Initiate memory dump. The following memory
dump types are available:
– BMC memory dump (nondisruptive)
– Resource memory dump
– System memory dump (disruptive)
򐂰 Network Time Protocol (NTP) server support

102 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
򐂰 Lightweight directory access protocol (LDAP) for user management
򐂰 Host console
By using the host console, you can monitor the server’s start process. The host console
also can be used to access the operating system when only a single LPAR uses all of the
resources.

Note: The host console also can be accessed by using an SSH client over port 2200
and logging in as the admin user.

򐂰 User management
You can create your own users in the eBMC. This feature also can be used to create an
individual user that can be used for the HMC to access the server.
A user features the following types privileges:
– Administrator
– ReadOnly (you cannot modify anything (except the password of that user); therefore, a
user with this privilege level cannot be used for HMC access to the server.
򐂰 IBM security by way of Access Control Files
To get “root access” to the service processor by using the user celogin in FSP-managed
servers, the IBM support team generated a password by using the serial number and the
date.
In eBMC managed systems, the support team generates an Access Control File (ACF).
This file must be uploaded to the server to get access. This procedure is needed (for
example) if the admin password must be reset. This process requires physical access to
the system.
򐂰 Jumper reset
Everything on the server on be reset by using a physical jumper. This factory reset
process resets everything on the server, such as LPAR definitions, eBMC settings, and the
NVRAM.

Next, we discuss some functions of the ASMI.

Realtime progress indicator


The ASMI of an eBMC server also provides a real-time progress indicator to see the operator
panel codes. To open the window that shows the codes, click Logs → Progress logs and
then, click View code in real time.

Inventory and LEDs


Under Hardware status → Inventory and LEDs, you can find most of the hardware
components with their current state, and controls for an identification LED for each
component, the system identification LED, and the system attention LED. You can switch all
identification LEDs on and off individually. The system attention LED can be turned off only.

A component also can be displayed. This feature is helpful to see details; for example, the
size of a DDIMM or the part number of a component if something must be exchanged.

Sensors
The ASMI features data from various sensors that are available within the server and many
components by clicking Hardware status → Sensors. The loading of the sensor data takes
some time, during which you see a progress bar on the top of the window.

Chapter 2. Architecture and technical overview 103


Note: Although the progress bar might be finished, it can take some extra time until the
sensor data is shown.

Figure 2-32 shows an example of the sensors window.

Figure 2-32 ASMI sensors

Network settings
The default network settings for the two eBMC ports are to use DHCP. Therefore, when you
connect a port to a private HMC network with the HMC as a DHCP server, the new system
receives its IP address from the HMC during the start of the firmware. Then, the new system
automatically appears in the HMC and can be configured.

DHCP is the recommended way to attach the eBMC of a server to the HMC.

If you do not use DHCP and want to use a static IP, you can set the IP in the ASMI GUI.
However, before you can make this change, you must connect to the ASMI. Because no
default IPs are the same for every server, you first must determine the configured IP.

To determine the configured IP, use the operator window. This optional component includes
the recommendation that one operator window is purchased per rack of Power10
processor-based scale-out servers.

In the control window, complete the following steps:


1. Use the Increment or Decrement options to scroll to function 02.
2. Click Enter until the value changes from N (normal) to M (manual).
This process activates access to function 30.
3. Scroll to function 30 and click Enter until 30** appears.
4. Scroll to 3000 and click Enter, which displays the IP of the ETH0 port. If you scroll to 3001
and click Enter, the IP of ETH1 is displayed.
5. After you determine the IP, scroll again to function 02 and set the value back from M to N.

104 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
For more information about function 30 in the operator window, see this IBM Documentation
web page.

Now that you determined the IP, you can configure any computer with a web browser to an IP
in the same subnet (class C) and connect the computer to the correct Ethernet port of the
server.

Hint: Most connections work by using a standard Ethernet cable. If you do not see a link
with the standard Ethernet cable, use a crossover cable where the send and receive wires
are crossed.

After connecting the cable, you can now use a web browser to access the ASMI with
https://<IP address> and then, configure the network port address settings.

To configure the network ports, click Settings → Network and select the correct adapter to
configure.

Figure 2-33 shows and example of changing eth1. Before you can configure a static IP
address, switch off DHCP. Several static IPs can be configured on one physical Ethernet port.

Figure 2-33 ASMI network settings

In the ASMI network settings window, you cannot configure the VMI address. The VMI
address is another IP that is configured on the physical eBMC Ethernet port of the server to
mange the Virtualization of the server. The VMI address can be configured in the HMC only.

Chapter 2. Architecture and technical overview 105


Using an Access Control File
If you lost the access password for the ASMI service user, you can access the ASMI by using
an ACF. The ACF is a digital certificate that is provided by IBM support when you open a
support case. To use the ACF, the system must be enabled at the server by using the operator
panel.

Complete the following steps:


1. On the operator panel, use the Increment or Decrement buttons to scroll to function 74.
2. Click Enter and then, select 00 to accept the function (FF rejects it). Now, the ACF
function is active for 30 minutes. To use it, complete the following steps:
a. Enter the ASMI login window.
b. Click Upload service login certificate to upload the ACF into the system and allow
the user to enter the ASMI with the associated password that aso is supplied by IBM
support.

For more information, see this IBM Documentation web page.

Policies
In Security and access → Policies, you can switch security related functions on and off; for
example, whether management over Intelligent Platform Management Interface (IPMI) is
enabled.

Some customers require that the USB ports of the server must be disabled. This change can
be made in the Policies window. Switch off Host USB enablement, as shown in Figure 2-34.

Figure 2-34 Switching off the USB ports of the server

106 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
2.5.2 Managing the system by using DMTF Redfish
eBMC-based systems also can be managed by using the DMTF Redfish APIs. Redfish is a
REST API that is used for platform management and is standardized by the Distributed
Management Task Force, Inc. For more information, see this web page.

You can work with Redfish by using several methods, ll of which require an https connection
to the eBMC. One possibility is to use the curl operating system command. The following
examples show how to work with curl and Redfish.

Before you can acquire data from the server or run systems management tasks by using
Redfish, you must authenticate against the server. In return for supplying a valid username
and password, you receive a token that is used to authenticate requests (see Example 2-1).

Example 2-1 Receiving a token from Redfish


# export eBMC=<IP>
# export USER=admin
# export PASSWORD=<PW>

# export TOKEN=`curl -k -H "Content-Type: application/json" -X POST https://${eBMC}/login


-d "{\"username\" : \"${USER}\", \"password\" : \"${PASSWORD}\"}" | grep token | awk
'{print $2;}' | tr -d '"'`

With this token, you now can receive data from the server. You start by requesting data of the
Redfish root with /Redfish/v1. You receive data with other branches in the Redfish tree; for
example, Chassis.

For more data, you can use the newly discovered odata.id field information
/Redfish/v1/Chassis, as shown in Example 2-2.

Example 2-2 Get chassis data from Redfish


#curl -s -k -H "X-Auth-Token: $TOKEN" -X GET https://${eBMC}/redfish/v1
{
"@odata.id": "/redfish/v1",
"@odata.type": "#ServiceRoot.v1_12_0.ServiceRoot",
"AccountService": {
"@odata.id": "/redfish/v1/AccountService"
},
"Cables": {
"@odata.id": "/redfish/v1/Cables"
},
"CertificateService": {
"@odata.id": "/redfish/v1/CertificateService"
},
"Chassis": {
"@odata.id": "/redfish/v1/Chassis"
},

# curl -k -H "X-Auth-Token: $TOKEN" -X GET https://${eBMC}/Redfish/v1/Chassis


{
"@odata.id": "/Redfish/v1/Chassis",
"@odata.type": "#ChassisCollection.ChassisCollection",
"Members": [
{
"@odata.id": "/Redfish/v1/Chassis/chassis"
}
],

Chapter 2. Architecture and technical overview 107


"Members@odata.count": 1,
"Name": "Chassis Collection"

# curl -k -H "X-Auth-Token: $TOKEN" -X GET https://${eBMC}/Redfish/v1/Chassis/chassis


{
"@odata.id": "/Redfish/v1/Chassis/chassis",
"@odata.type": "#Chassis.v1_16_0.Chassis",
"Actions": {
...
"PCIeSlots": {
"@odata.id": "/Redfish/v1/Chassis/chassis/PCIeSlots"
},
...
"Sensors": {
"@odata.id": "/Redfish/v1/Chassis/chassis/Sensors"
},
...

Under Chassis, another chassis is available (with lower case c). We can now use the tree with
both; that is, /Redfish/v1/Chassis/chassis. After running the tool, you can see in
Example 2-2 on page 107 PCIeSlots and Sensors are available as examples of other
resources on the server.

In Example 2-3, you see what is available through the Sensors endpoint. Here, you can find
the same sensors as in the ASMI GUI (see Figure 2-32 on page 104).

For example, in the output, you find the sensor total_power. When you ask for more
information about that sensor (see Example 2-3), you can see that the server needed 1.426
watts at the time of running the command. Having programmatic access to this type of data
allows you to build a view of the electrical power consumption of your Power environment in
real time, or to report usage over a period.

Example 2-3 Sensor data from Redfish


# curl -k -H "X-Auth-Token: $TOKEN" -X GET
https://${eBMC}/Redfish/v1/Chassis/chassis/Sensors
{
"@odata.id": "/Redfish/v1/Chassis/chassis/Sensors",
"@odata.type": "#SensorCollection.SensorCollection",
"Description": "Collection of Sensors for this Chassis",
"Members": [
{
"@odata.id": "/Redfish/v1/Chassis/chassis/Sensors/Altitude"
},
{
"@odata.id": "/Redfish/v1/Chassis/chassis/Sensors/1V1CS_0167_p1_rail_iout"
},
...
{
"@odata.id": "/Redfish/v1/Chassis/chassis/Sensors/total_power"
},

# curl -k -H "X-Auth-Token: $TOKEN" -X GET \


https://${eBMC}/Redfish/v1/Chassis/chassis/Sensors/total_power
{
"@odata.id": "/Redfish/v1/Chassis/chassis/Sensors/total_power",
"@odata.type": "#Sensor.v1_0_0.Sensor",
"Id": "total_power",
"Name": "total power",

108 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
"Reading": 1426.0,
"ReadingRangeMax": null,
"ReadingRangeMin": null,
"ReadingType": "Power",
"ReadingUnits": "W",
"Status": {
"Health": "OK",
"State": "Enabled"
}

The following other strings are useful:


򐂰 Type/Model of the server:
# curl -s -k -H "X-Auth-Token: $TOKEN" -X GET
https://${eBMC}/Redfish/v1/Systems/system | grep Model | grep -v SubModel |
grep \ -v \"\"
"Model": "9043-MRX",
򐂰 Serial number:
# curl -s -k -H "X-Auth-Token: $TOKEN" -X GET
https://${eBMC}/redfish/v1/Systems/system | grep SerialNumber \
"SerialNumber": "<SN>",
򐂰 Type/Model/SN:
#curl -s -k -H "X-Auth-Token: $TOKEN" -X GET
https://${eBMC}/redfish/v1/Systems/system | grep AssetTag
"AssetTag": "Server-9043-MRX-<SN>",
򐂰 System indicator LED:
#curl -s -k -H "X-Auth-Token: $TOKEN" -X GET
https://${eBMC}/redfish/v1/Systems/system | grep IndicatorLED
"IndicatorLED": "Off",
򐂰 Total Memory:
# curl -s -k -H "X-Auth-Token: $TOKEN" -X GET
https://${eBMC}/redfish/v1/Systems/system | grep TotalSystemMemoryGiB
"TotalSystemMemoryGiB": 8192
򐂰 Power State:
# curl -s -k -H "X-Auth-Token: $TOKEN" -X GET
https://${eBMC}/redfish/v1/Systems/system | grep PowerState
"PowerState": "On",

Operations also can be run on the server by using the POST method to the Redfish API
interface. The following curl commands can be used to start or stop the server (these
commands work only if you are authenticated as a user with administrator privileges):
򐂰 Power on server:
# curl -k -H "X-Auth-Token: $TOKEN" -X POST https://${eBMC}/redfish/v1/Systems/
system/Actions/Reset -d '{"ResetType":"On"}'
򐂰 Power off server:
# curl -k -H "X-Auth-Token: $TOKEN" -X POST https://${eBMC}/redfish/v1/Systems/
system/Actions/Reset -d '{"ResetType":"ForceOff"}'

Chapter 2. Architecture and technical overview 109


For more information about Redfish, see this IBM Documentation web page.

For more information about how to work with Redfish in Power systems, see this IBM
Documentation web page.

2.5.3 Managing the system by using the Intelligent Platform Management


Interface
The server can also be managed by using the IPMI, although for security reasons the IPMI is
disabled by default on the system.

Because inherent security vulnerabilities are associated with the IPMI, consider the use of
Redfish APIs or the GUI to manage your system.

If you want to use IPMI, this service must be enabled first. This process can be done by
clicking Security and access → Policies. There, you find the policy Network IPMI
(out-of-band IPMI) that must be enabled to support IPMI access.

For more information about common IPMI commands, see this IBM Documentation web
page.

110 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
3

Chapter 3. Available features and options


In this chapter, we discuss the major features and options that are available.

This chapter includes the following topics:


򐂰 3.1, “Processor module features” on page 112
򐂰 3.2, “Memory features” on page 118
򐂰 3.3, “Power supply features” on page 119
򐂰 3.4, “Peripheral Component Interconnect adapters” on page 119
򐂰 3.5, “Internal storage” on page 126
򐂰 3.6, “External SAS ports” on page 132
򐂰 3.7, “Media drawers” on page 133
򐂰 3.8, “Disk and media features” on page 134
򐂰 3.9, “External IO subsystems” on page 138
򐂰 3.10, “System racks” on page 149

© Copyright IBM Corp. 2022. 111


3.1 Processor module features
This section describes all processor-related Feature Codes for the Power10 processor-based
scale-out servers Power S1014, Power S1022s, Power S1022, and Power S1024.

3.1.1 Power S1014 processor Feature Codes


The Power S1014 provides one socket to accommodate one Power10 entry single-chip
module (eSCM). Two different eSCM types with a core density of four (Feature Code #EPG0)
or eight (Feature Code #EPG2) functional processors are offered.

Power S1014 servers do not support any Capacity on Demand (CoD) capability; therefore, all
available functional cores of an eSCM type are activated by default.

The 4-core eSCM #EPG0 requires four static processor activation features #EPFT and the
8-core eSCM #EPG2 requires eight static processor activation features #EPF6. To assist with
the optimization of software licensing, the factory deconfiguration feature #2319 is available at
initial order to permanently reduce the number of active cores, if wanted.

Table 3-1 lists the processor card Feature Codes that are available at initial order for
Power S1014 servers.

Table 3-1 Processor card Feature Code specification for the Power S1014 server
Processor Processor Number Typical Static processor core
card feature module of cores frequency activation Feature
code type range (GHz) Code

#EPG0 eSCM 4 3.0 - 3.9 #EPFT

#EPG2 eSCM 8 3.0 - 3.9 #EPF6

Table 3-2 lists all processor-related Feature Codes for Power S1014 servers.

Table 3-2 Processor-related features of the Power S1014 server


Feature code Description

#EPG0 4-core typical 3.0 to 3.90 GHz (maximum) Power10 processor card

#EPG2 8-core typical 3.0 to 3.90 GHz (maximum) Power10 processor card

#EPFT Entitlement for one processor core activation for #EPG0

#EPF6 Entitlement for one processor core activation for #EPG2

#2319 Factory deconfiguration of one core for EPG0 or #EPG2

Power S1022s processor Feature Codes


The Power S1022s provides two sockets to accommodate one or two Power10 eSCMs. Two
eSCM types with a core density of four (Feature Code #EPGR) or eight (Feature Code
#EPGQ) functional cores are offered.

Power S1022s servers do not support any CoD capability; therefore, all available functional
cores of an eSCM type are activated by default.

112 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
The 4-core eSCM processor module #EPGR requires four static processor activation features
#EPFR, and the 8-core eSCM processor module #EPGQ requires eight static processor
activation features #EPFQ. To assist with the optimization of software licensing, the factory
deconfiguration Feature Code #2319 is available at initial order to permanently reduce the
number of active cores. if wanted.

The Power S1022s server can be configured with one 4-core processor, one 8-core
processor, or two 8-core processors. An option for a system with two 4-core processors that
are installed is not available.

Table 3-3 lists the processor card Feature Codes that are available at initial order for
Power S1022s servers.

Table 3-3 Processor card Feature Code specification for the Power S1022s server
Processor Processor Number of Typical Static processor core
card Feature module type cores frequency activation Feature
Code range (GHz) Code

#EPGR eSCM 4 3.0 to 3.9 #EPFR

#EPGQ eSCM 8 3.0 to 3.9 #EPFQ

Table 3-4 lists all processor-related Feature Codes for Power S1022s servers.

Table 3-4 Processor-related features of the Power S1022s server


Feature code Description

#EPGR 4-core typical 3.0 to 3.90 GHz (maximum) Power10 processor card

#EPGQ 8-core typical 3.0 to 3.90 GHz (maximum) Power10 processor card

#EPFR Entitlement for one processor core activation for #EPGR

#EPFQ Entitlement for one processor core activation for #EPGQ

#2319 Factory deconfiguration of one core for #EPGR or #EPGQ

Power S1022 processor Feature Codes


The Power S1022 provides two sockets to accommodate one or two Power10 dual-chip
modules (DCMs). The following DCM core densities are available:
򐂰 12-core: Feature Code #EPG9
򐂰 16-core: Feature Code #EPG8
򐂰 20-core: Feature Code #EPGA

The 12-core #EPG9 DCM can be used in 1-socket or 2-socket Power S1022 configurations.
The higher core density modules with 16 or 20 functional cores are available only in 2-socket
configurations and both sockets must be populated by the same processor feature.

Power S1022 servers support the Capacity Upgrade on Demand (CUoD) capability by
default. At an initial order, a minimum of 50% of configured physical processor cores must be
covered by CUoD static processor core activations:
򐂰 The 12-core DCM processor module #EPG9 requires a minimum of six CUoD static
processor activation features #EPF9 in a 1-socket and a minimum of 12 #EPF9 features in
a 2-socket configuration.

Chapter 3. Available features and options 113


򐂰 The 16-core DCM processor module #EPG8 is supported only in 2-socket configurations.
It requires a minimum of eight CUoD static processor activation features #EPF8.
Therefore, a minimum of 16 CUoD static processor activations are needed per server.
򐂰 The 20-core DCM processor module #EPGA is supported only in 2-socket configurations.
It requires a minimum of 10 CUoD static processor activation features #EPFA. Therefore,
a minimum of 20 CUoD static processor activations are needed per server.

Extra CUoD static activations can be purchased later after the initial order until all physically
present processor cores are entitled.

To assist with the optimization of software licensing, the factory deconfiguration feature #2319
is available at initial order to permanently reduce the number of active cores that are below
the imposed minimum of 50% CUoD static processor activations, if wanted.

As an alternative to the CUoD processor activation use model and to enable cloud agility and
cost optimization with pay-for-use pricing, the Power S1022 server supports the IBM Power
Private Cloud with Shared Utility Capacity solution (also known as Power Enterprise Pools 2.0
or Pools 2.0). This solution is configured at initial system order by including Feature Code
#EP20.

When configured as a Power Private Cloud system, each Power S1022 server requires a
minimum of one base processor core activation. The maximum number of base processor
activations is limited by the physical capacity of the server.

Although configured against a specific server, the base activations can be aggregated across
a pool of servers and used on any of the systems in the pool. When a system is configured in
this way, all processor cores that are installed in the system become available for use. Any
usage above the base processor activations that are purchased across a pool is monitored by
the IBM Cloud Management Console for Power and is debited from the customers cloud
capacity credits, or is invoiced monthly for total usage across a pool of systems.

A system that is initially ordered with a configuration that is based on the CUoD processor
activations can be converted to the Power Private Cloud with Shared Utility Capacity model
later. This process requires the conversion of existing CUoD processor activations to base
activations, which include different feature codes. The physical processor feature codes do
not change.

A system cannot be converted from the Power Private Cloud with Shared Utility Capacity
model to CUoD activations.

Table 3-5 lists the processor card feature codes that are available at initial order for
Power S1022 servers.

Table 3-5 Processor feature code specification for the Power 1022 server

Processor Processor Number Typical CUoDa static Base processor Base core
card module of cores frequency processor core core activation activations
feature type range activation Feature Code for converted
code [GHz] Feature Code Pools 2.0 from CUoD
static
activations

#EPG9 DCM 12 2.90 to 4.0 #EPF9 #EUCB #EUCH

#EPG8 DCM 16 2.75 to 4.0 #EPF8 #EUCA #EUCG

#EPGA DCM 20 2.45 to 3.9 #EPFA #EUCC #EUCJ


a. Capacity Upgrade on Demand

114 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Table 3-6 lists all processor-related feature codes for Power S1022 servers.

Table 3-6 Processor-related features of the Power S1022 server


Feature Code Description

#EPG9 12-core typical 2.90 to 4.0 GHz (maximum) Power10 processor card, available in
quantity of one (1-socket configuration) or two (2-socket configuration)

#EPG8 16-core typical 2.75 to 4.0 GHz (maximum) Power10 processor card, available in
quantity of two (2-socket configuration) only

#EPGA 20-core typical 2.45 to 3.90 GHz (maximum) Power10 processor card, available in
quantity of two (2-socket configuration) only

#EPF9 One CUoD static processor core activation for #EPG9

#EPF8 One CUoD static processor core activation for #EPG8

#EPFA One CUoD static processor core activation for #EPGA

#2319 Factory deconfiguration of one core for #EPG9, #EPG8, or #EPGA

#EP20 Power Enterprise Pools 2.0 enablement

#EUCB One base processor core activation on processor card #EPG9 for Pools 2.0 to
support any operating system

#EUCA One base processor core activation on processor card #EPG8 for Pools 2.0 to
support any operating system

#EUCC One base processor core activation on processor card #EPGA for Pools 2.0 to
support any operating system

#EUCH One base processor core activation on processor card #EPG9 for Pools 2.0 to
support any operating system (converted from #EPF9)

#EUCG One base processor core activation on processor card #EPG8for Pools 2.0 to
support any operating system (converted from #EPF8)

#EUCJ One base processor core activation on processor card #EPGA for Pools 2.0 to
support any operating system (converted from #EPFA)

Power S1024 processor feature codes


The Power S1024 provides two sockets to accommodate one or two Power10 DCMs. The
following DCM core densities are available:
򐂰 12-core: Feature Code #EPGM
򐂰 16-core: Feature Code #EPGC
򐂰 24-core Feature Code #EPGD

The 12-core #EPGM DCM can be used in 1-socket or 2-socket Power S1024 configurations.
The higher core density modules with 16 or 24 functional cores are available only for 2-socket
configurations and both sockets must be populated by the same processor feature.

Power S1024 servers support the CUoD capability by default. At an initial order, a minimum of
50% of configured physical processor cores must be covered by CUoD static processor core
activations:
򐂰 The 12-core DCM processor module #EPGM requires a minimum of six CUoD static
processor activation features #EPFM in a 1-socket and 12 #EPFM features in a 2-socket
configuration.

Chapter 3. Available features and options 115


򐂰 The 16-core DCM processor module #EPGC is supported only in 2-socket configurations
and requires a minimum of eight CUoD static processor activation features #EPFC.
Therefore, a minimum of 16 CUoD static processor activations are needed per server.
򐂰 The 24-core DCM processor module #EPGD is supported only in 2-socket configurations
and requires a minimum of 12 CUoD static processor activation features #EPFD.
Therefore, a minimum of 24 CUoD static processor activations are needed per server.

To assist with the optimization of software licensing, the factory deconfiguration feature #2319
is available at initial order to permanently reduce the number of active cores that are below
the imposed minimum of 50% CUoD static processor activations, if wanted.

As an alternative to the CUoD processor activation use model and to enable cloud agility and
cost optimization with pay-for-use pricing, the Power S1024 server also supports the IBM
Power Private Cloud with Shared Utility Capacity solution (also known as Power Enterprise
Pools 2.0, or just Pools 2.0). This solution is configured at initial system order by including
Feature Code #EP20.

When configured as a Power Private Cloud system, each Power S1024 server requires a
minimum of one base processor core activation. The maximum number of base processor
activations is limited by the physical capacity of the server.

Although configured against a specific server, the base activations can be aggregated across
a pool of servers and used on any of the systems in the pool. When a system is configured in
this way, all processor cores that are installed in the system become available for use. Any
usage above the base processor activations that are purchased across a pool is monthly for
total usage across a pool of systems.

A system that is initially ordered with a configuration that is based on the CUoD processor
activations can be converted to the Power Private Cloud with Shared Utility Capacity model
later. This process requires the conversion of existing CUoD processor activations to base
activations, which include different feature codes. The physical processor feature codes do
not change.

A system cannot be converted from the Power Private Cloud with Shared Utility Capacity
model to CUoD activations.

Table 3-7 lists the processor card feature codes that are available at initial order for
Power S1024 servers.

Table 3-7 Processor feature code specification for the Power S1024 server
Processor Processor Number Typical CUoD static Base processor Base core
card module of cores frequency processor core core activation activations
feature type range activation Feature Code for converted
code [GHz] Feature Code Pools 2.0 from CUoD
static
activations

#EPGM DCM 12 3.40 - 4.0 #EPFM #EUBX #EUBZ

#EPGC DCM 16 3.10 - 4.0 #EPFC #EUCK #EUCR

#EPGD DCM 24 2.75 - 3.9 #EPFD #EUCS #EUCT

116 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Figure 3-8 lists all processor-related feature codes for Power S1024 servers.

Table 3-8 Processor-related features of the Power S1024 server


Feature code Description

#EPGM 12-core typical 3.40 to 4.0 GHz (maximum) Power10 processor card, available in
quantity of one (1-socket configuration) or two (2-socket configuration)

#EPGC 16-core typical 3.10 to 4.0 GHz (maximum) Power10 processor card, available in
quantity of two (2-socket configuration) only

#EPGD 24-core typical 2.75 to 3.9 GHz (maximum) Power10 processor card, available in
quantity of two (2-socket configuration) only

#EPFM One CUoD static processor core activation for #EPGM

#EPFC One CUoD static processor core activation for #EPGC

#EPFD One CUoD static processor core activation for #EPGD

#2319 Factory deconfiguration of one core for #EPGM, #EPGC, or #EPGD

#EP20 Power Enterprise Pools 2.0 enablement

#EUBX One base processor core activation on processor card #EPGM for Pools 2.0 to
support any operating system

#EUCK One base processor core activation on processor card #EPGC for Pools 2.0 to
support any operating system

#EUCL One base processor core activation on processor card #EPGD for Pools 2.0 to
support any operating system

#EUBZ One base processor core activation on processor card #EUGM for Pools 2.0 to
support any operating system converted from #EPFM

#EUCR One base processor core activation on processor card #EUGC for Pools 2.0 to
support any operating system converted from #EPFC

#EUCT One base processor core activation on processor card #EUGD for Pools 2.0 to
support any operating system converted from #EPFD

Chapter 3. Available features and options 117


3.2 Memory features
All available memory feature codes for the Power10 processor-based scale-out servers
Power S1014, Power S1022s, Power S1022, and Power S1024 are listed by Table 3-9. Each
memory feature code relates to two differential DIMM (DDIMM) cards of identical
specifications.

Table 3-9 Memory Feature Codes for Power10 processor-based scale-out servers
Feature Capacity Packaging DRAM DRAM Form Supported
code density data rate factor servers

#EM6N 32 GB 2 x 16 GB 8 Gbit 3200 MHz 2U All scale-out


DDIMMs models

#EM6W 64 GB 2 x 32 GB 8 Gbit 3200 MHz 2U All scale-out


DDIMM models

#EM6X 128 GB 2 x 64 GB 16 Gbit 3200 MHz 2U All scale-out


DDIMM models

#EM6Y 256 GB 2 x 128 GB 16 Gbit 2666 MHz 2U 򐂰 Power S1014


DDIMM 򐂰 Power S1022s
򐂰 Power S1022

#EM6U 256 GB 2 x 128 GB 16 Gbit 2933 MHz 4U Power S1024


DDIMM

#EM78 512 GB 2 x 256 GB 16 Gbit 2933 MHz 4U Power S1024


DDIMM

The memory module cards for the scale-out servers are manufactured in two different form
factors, which are used in servers with 2 rack units (2U) or 4 rack units (4U). The 2U memory
cards can be extended through spacers for use in 4U servers, but the 4U high cards do not fit
in 2U servers.

All Power10 processor-based scale-out servers can use the following configurations:
򐂰 2U 16 GB capacity DDIMMs of memory feature #EN6N
򐂰 2U high 32 GB capacity DDIMMs of memory feature #EM6W
򐂰 2U high 64 GB capacity DDIMMs of memory feature #EM6X.

The 2U 128 GB capacity DDIMMs of feature #EM6Y can be used in all of the Power10
scale-out servers except for Power S1024 systems. The 4U high 128 GB capacity DDIMMs of
feature #EM6U and the 4U high 256 GB capacity DDIMMs of feature #EM78 are exclusively
provided for Power S1024 servers.

All memory slots that are connected to a DCM or an eSCM must be fitted with DDIMMs of the
same memory feature code:
򐂰 For 1-socket Power10 scale-out server configurations, all memory modules must be of the
same capacity, DRAM density, DRAM data rate and form factor.
򐂰 For 2-socket Power10 scale-out server configurations two different memory feature codes
can be selected, but the memory slots that are connected to a socket must be filled with
DDIMMs of the same memory feature code, which implies that they are of identical
specifications.

The minimum memory capacity limit is 32 GB per eSCM or DCM processor module that can
be fulfilled by one #EM6N memory feature.

118 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
No specific memory enablement features are required for any of the supported Power10
scale-out server memory features. The entire physical DDIMM capacity of a memory
configuration is enabled by default.

All Power10 processor-based scale-out servers (except the Power S1014) support the Active
Memory Mirroring (AMM) feature #EM8G. AMM is available as an optional feature to enhance
resilience by mirroring critical memory that is used by the PowerVM hypervisor so that it can
continue operating if a memory failure occurs.

A portion of available memory can be operatively partitioned such that a duplicate set can be
used if noncorrectable memory errors occur. This partitioning can be implemented at the
granularity of DDIMMs or logical memory blocks.

3.3 Power supply features


The Power S1014 server supports the following power supply options:
򐂰 Four 1200 W 100 - 127 V AC or 200 - 240 V AC power supplies (#EB3W) that support a
tower chassis. Two power supplies are required during the boot phase and for normal
system operation; and the third and fourth are for redundancy.
򐂰 Two 1600 W 200 - 240 V AC power supplies (#EB3S) that support a rack chassis. One
power supply is required during the boot phase and for normal system operation; the
second is for redundancy.

The Power S1022s server supports two 2000 W 200 - 240 V AC power supplies (#EB3N).
Two power supplies are always installed. One power supply is required during the boot phase
and for normal system operation, and the second is for redundancy.

The Power S1022 server supports two 2000 W 200 - 240 V AC power supplies (#EB3N). Two
power supplies are always installed. One power supply is required during the boot phase and
for normal system operation, and the second is for redundancy.

The Power S1024 server supports four 1600 W 200 - 240 V AC (#EB3S) power supplies. Four
power supplies are always installed. Two power supplies are required during the boot phase
and for normal system operation, and the third and fourth are for redundancy.

3.4 Peripheral Component Interconnect adapters


This section discusses the various types and functions of the PCIe adapters that are
supported by the following servers:
򐂰 Power S1014 (9105-41B)
򐂰 Power S1022s (9105-22B)
򐂰 Power S1022 (9105-22A)
򐂰 Power S1024 (9105-42A) servers

This list is based on the PCIe adapters available on the General Availability (GA) date of
these systems, but is subject to change as more PCIe adapters are tested and certified, or
listed adapters are no longer available.

For more information about the supported adapters, see the IBM Offering Information web
page.

Chapter 3. Available features and options 119


The following sections describe the supported adapters and provide tables of orderable and
supported feature numbers. The tables indicate operating system support (AIX, IBM i, and
Linux) for each of the adapters.

The Order type table column in the following subsections is defined as:
Initial Denotes the orderability of a feature only with the purchase of a new
system.
MES Denotes the orderability of a feature only as part of an MES upgrade
purchase for an existing system.
Both Denotes the orderability of a feature as part of new and MES upgrade
purchases.
Supported Denotes that feature is not orderable with a system, but is supported;
that is, the feature can be migrated from existing systems, but cannot
be ordered new.

3.4.1 Local area network adapters


To connect the IBM Power S1014, S1022s, S1022, and S1024 server models to a local area
network (LAN), you can use the LAN adapters that are supported in the PCIe slots of the
system. Various connection speeds and physical connections are supported.

Table 3-10 lists the low profile (LP) LAN adapters that are supported within the Power S1022s
and Power S1022 server models.

Table 3-10 Low profile LAN adapters that are supported in the S1022s and S1022
Feature CCIN Description Operating Order type
code system support

5260 576F PCIe2 LP 4-port 1 GbE Adapter AIX, Linux, IBM ia Supported

EC2T 58FB PCIe3 LP 2-Port 25/10 Gb NIC&ROCE AIX, Linuxc, Both


SR/Cu Adapterb IBM ia

EC67 2CF3 PCIe4 LP 2-port 100 Gb ROCE EN LP AIX, Linuxc , Both


adapterd IBM ia

EN0T 2CC3 PCIe2 LP 4-Port (10 Gb+1 GbE) AIX, Linux, IBM ia Supported
SR+RJ45 Adapter

EN0V 2CC3 PCIe2 LP 4-port (10 Gb+1 GbE) Copper AIX, Linux, IBM ia Supported
SFP+RJ45 Adapter

EN0X 2CC4 PCIe2 LP 2-port 10/1 GbE BaseT RJ45 AIX, Linux, IBM ia Both
Adapter
a. The IBM i operating system is supported through VIOS only.
b. The #EC2T adapter requires one or two suitable transceivers to provide 10 Gbps SFP+
(#EB46), 25 Gbps SFP28 (#EB47), or 1 Gbps RJ45 (#EB48) connectivity as required.
c. Linux support requires Red Hat Enterprise Linux 8.4 or later, Red Hat Enterprise Linux for SAP
8.4 or later, SUSE Linux Enterprise Server 15 Service Pack 3 or later, SUSE Linux Enterprise
Server for SAP with SUSE Linux Enterprise Server 15 Service Pack 3 or later, or Red Hat
OpenShift Container Platform 4.9 or later. All require Mellanox OFED 5.5 drivers or later.
d. To deliver the full performance of both ports, each 100 Gbps Ethernet adapter must be
connected to a PCIe slot with 16 lanes (x16) of PCIe Gen4 connectivity. In the Power S1022s
and Power S1022 server models this limits placement to PCIe slots C0, C3, C4, and C10. In
systems with only a single socket populated, a maximum of one 100 Gbps Ethernet adapter is
supported. The 100 Gbps Ethernet adapters are not supported in PCIe expansion drawers.

120 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Table 3-11 lists the full-height LAN adapters that are supported within the Power S1014 and
Power S1024 server models, and within the PCIe expansion drawer (EMX0) connected to any
of the Power10 processor-based scale-out server models.

Table 3-11 Full-height LAN adapters supported in the S1014, S1024, and PCIe expansion drawers
Feature CCIN Description Operating Order type
code system support

5899 576F PCIe2 4-port 1 GbE Adapter AIX, Linux, IBM ia Supported

EC2U 58FB PCIe3 2-Port 25/10 Gb NIC&ROCE AIX, Linuxc, Both


SR/Cu Adapterb IBM ia

EN0S 2CC3 PCIe2 4-Port (10 Gb+1 GbE) SR+RJ45 AIX, Linux, IBM i Supported
Adapter (through VIOS)

EN0U 2CC3 PCIe2 L4-port (10 Gb+1 GbE) Copper AIX, Linux, IBM i Supported
SFP+RJ45 Adapter (through VIOS)

EN0W 2CC4 PCIe2 2-port 10/1 GbE BaseT RJ45 AIX, Linux, IBM i Both
Adapter (through VIOS)
a. When this adapter is installed in an expansion drawer that is connected to an S1022s or S1022
server, IBM i is supported through VIOS only.
b. The #EC2U adapter requires one or two suitable transceivers to provide 10 Gbps SFP+
(#EB46), 25 Gbps SFP28 (#EB47), or 1 Gbps RJ45 (#EB48) connectivity as required.
c. Linux support covers Requires Red Hat Enterprise Linux 8.4 or later, Requires Red Hat
Enterprise Linux for SAP 8.4 or later, SUSE Linux Enterprise Server 15 Service Pack 3 or later,
SUSE Linux Enterprise Server for SAP with SUSE Linux Enterprise Server 15 Service Pack 3
or later, or Red Hat OpenShift Container Platform 4.9 or later. All require Mellanox OFED 5.5
drivers or later.

Two full-height LAN adapters with 100 Gbps connectivity are available that are supported only
when they are installed within the Power S1014 or Power S1024 server models. To deliver the
full performance of both ports, each 100 Gbps Ethernet adapter must be connected to a PCIe
slot with 16 lanes (x16) of PCIe Gen4 connectivity.

In the Power S1014 or the Power S1024 with only a single socket that is populated, this
requirement limits placement to PCIe slot C10. In the Power S1024 with both sockets
populated, this requirement limits placement to PCIe slots C0, C3, C4, and C10. These 100
Gbps Ethernet adapters are not supported in PCIe expansion drawers.

Chapter 3. Available features and options 121


Table 3-12 lists the 100 Gbps LAN adapters that are supported within the Power S1014 and
Power S1024 servers only.

Table 3-12 Full-height 100 Gbps LAN adapters that are supported in the S1014 and S1024 only
Feature CCIN Description Operating Order type
code system support

EC66 2CF3 PCIe4 2-port 100 Gb ROCE EN adapter AIX, Linux, IBM i Both
(through VIOS)a

EC78 2CFA PCIe4 2-port 100 Gb Crypto ConnectX-6 IBM i onlyb


DX QFSP56
a. Linux support covers Requires Red Hat Enterprise Linux 8.4 or later, Requires Red Hat
Enterprise Linux for SAP 8.4 or later, SUSE Linux Enterprise Server 15 Service Pack 3 or later,
SUSE Linux Enterprise Server for SAP with SUSE Linux Enterprise Server 15 Service Pack 3
or later, or Red Hat OpenShift Container Platform 4.9 or later. All require Mellanox OFED 5.5
drivers or later.
b. The #EC78 adapter is supported only by IBM i 7.3 or later and runs in dedicated mode only (no
PowerVM virtualization). RoCE and IP Security (IPsec) are supported only by IBM i 7.4 or later
(RoCE and IPsec are supported only by IBM i Db2® Mirror).

3.4.2 Fibre Channel adapters


The Power10 processor-based scale-out servers support connection to devices that use
Fibre Channel connectivity directly or through a Storage Area Network (SAN). A range of
PCIe-connected Fibre Channel adapters are available in low profile and full-height form
factors.

All supported Fibre Channel adapters feature LC connections. If you are attaching a switch or
a device with an SC type fiber connector, an LC-SC 50-Micron Fibre Converter Cable or an
LC-SC 62.5-Micron Fiber Converter Cable is required.

Table 3-13 lists the low profile Fibre Channel adapters that are supported within the
Power S1022s and Power S1022 server models.

Table 3-13 Low profile FC adapters that are supported in the S1022s and S1022
Feature CCIN Description Operating Order type
code system support

EN1B 578F PCIe3 LP 32 Gb 2-port Fibre Channel AIX, Linux, IBM i Both
Adapter (through VIOS)

EN1D 578E PCIe3 LP 16 Gb 4-port Fibre Channel AIX, Linux, IBM i


Adapter (through VIOS)

EN1F 579A PCIe3 LP 16 Gb 4-port Fibre Channel AIX, Linux, IBM i


Adapter (through VIOS)

EN1H 579B PCIe3 LP 2-Port 16 Gb Fibre Channel AIX, Linux


Adapter

EN1K 579C PCIe4 LP 32 Gb 2-port Optical Fibre AIX, Linux, IBM i


Channel Adapter (through VIOS)

EN2B 579D PCIe3 LP 16 Gb 2-port Fibre Channel AIX, Linux, IBM i


Adapter (through VIOS)

122 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Table 3-14 lists the full-height Fibre Channel adapters that are supported within the
Power S1014 and Power S1024 server models, and within the PCIe expansion drawer
(EMX0) that is connected to any of the Power10 processor-based scale-out server models.

Table 3-14 Full-height FC adapters supported in the S1014, S1024, and PCIe expansion drawers
Feature CCIN Description Operating Order type
code system support

EN1A 578F PCIe3 LP 32 Gb 2-port Fibre Channel AIX, Linux, IBM i Both
Adapter

EN1C 578E PCIe3 LP 16 Gb 4-port Fibre Channel AIX, Linux, IBM i


Adapter

EN1E 579A PCIe3 LP 16 Gb 4-port Fibre Channel AIX, Linux, IBM i


Adapter (through VIOS)

EN1G 579B PCIe3 LP 2-Port 16 Gb Fibre Channel AIX, Linux


Adapter

EN1J 579C PCIe4 LP 32 Gb 2-port Optical Fibre AIX, Linux, IBM ia


Channel Adapter

EN2A 579D PCIe3 LP 16 Gb 2-port Fibre Channel AIX, Linux, IBM i


Adapter
a. IBM i support is limited to IBM i 7.5 or later, or IBM i 7.4 TR6 or later.

3.4.3 SAS adapters


The internal storage in the Power10 processor-based, scale-out servers is all based on
nonvolatile memory express (NVMe) devices that are connected over PCIe directly. More
storage expansion drawers can be connected to the system by using Serial Attached SCSI
(SAS) connections.

Table 3-15 lists the low profile SAS adapters that are supported within the Power S1022s and
Power S1022 server models.

Table 3-15 Low profile SAS adapters that are supported in the S1022s and S1022
Feature CCIN Description Operating Order type
code system
support

EJ0M 57B4 PCIe3 LP RAID SAS Adapter Quad-Port 6 AIX, Linux, IBM i Both
Gb x8 (through VIOS)

EJ11 57B4 PCIe3 LP SAS Tape/DVD Adapter AIX, Linux, IBM i


Quad-port 6 Gb x8 (through VIOS)

Chapter 3. Available features and options 123


Table 3-16 list the full-height SAS adapters that are supported within the Power S1014 and
Power S1024 server models, and within the PCIe expansion drawer (EMX0) that is connected
to any of the Power10 processor-based scale-out server models.

Table 3-16 Full-height SAS adapters supported in the S1014, S1024, and PCIe expansion drawers
Feature CCIN Description Operating Order type
code system
support

EJ0J 57B4 PCIe3 RAID SAS Adapter Quad-Port 6 Gb AIX, Linux, IBM i Both
x8 (through VIOS)

EJ10 57B4 PCIe3 SAS Tape/DVD Adapter Quad-port 6 AIX, Linux, IBM i
Gb x8 (through VIOS)

EJ14 57B1 PCIe3 12 GB Cache RAID PLUS SAS AIX, Linux, IBM i
Adapter Quad-port 6 Gb x8 (through VIOS)

3.4.4 USB adapters


Universal Serial Bus (USB) adapters are available to support the connection of devices, such
as external DVD drives to the Power10 processor-based scale-out server models.

Table 3-17 lists the low profile USB adapter that is supported within the Power S1022s and
Power S1022 server models.

Table 3-17 Low profile USB adapter that is supported in the S1022s and S1022
Feature CCIN Description Operating system support Order type
code

EC6J 590F PCIe2 LP 2-Port USB 3.0 Adapter AIX, Linux, IBM i (through VIOS) Both

Table 3-18 lists the full-height USB adapter that is supported within the Power S1014 and
Power S1024 server models, and within the PCIe expansion drawer (EMX0) connected to any
of the Power10 processor-based scale-out server models.

Table 3-18 Full-height USB adapter supported in the S1014, S1024, and PCIe expansion drawers
Feature CCIN Description Operating system support Order type
code

EC6K 590F PCIe2 LP 2-Port USB 3.0 Adapter AIX, Linux, IBM i (through VIOS) Both

3.4.5 Cryptographic coprocessor adapters


Two different Cryptographic coprocessors or accelerators are supported by the Power10
processor-based scale-out server models, both of which are full-height adapters. These
adapters work with the IBM Common Cryptographic Architecture (CCA) to deliver
acceleration to cryptographic workloads.

For more information about the cryptographic coprocessors, the available associated
software, and the available CCA, see this IBM Security® web page.

124 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
PCIe Gen3 cryptographic coprocessor 4767
Secure-key adapter provides cryptographic coprocessor and accelerator functions in a single
PCIe card. The adapter is suited to applications that require high-speed, security-sensitive,
RSA acceleration; cryptographic operations for data encryption and digital signing; secure
management, and the use of cryptographic keys or custom cryptographic applications.

It provides secure storage of cryptographic keys in a tamper-resistant hardware security


module that meets FIPS 140-2 Level 4 security requirements. The adapter is a PCIe Gen3 x4
full-height, half-length card. The adapter runs in dedicated mode only (no PowerVM
virtualization).

This adapter is available only in full-height form factor, and is available in two variations with
two different Feature Codes:
򐂰 #EJ32 does not include a Blind Swap Cassette (BSC) and can be installed only within the
chassis of a Power S1014 or Power S1024 server.
򐂰 #EJ33 includes a Blind Swap Cassette housing, and can be installed only in a PCIe Gen3
I/O expansion drawer enclosure. This option is supported only for the Power S1022s and
Power S1022 server models.

PCIe Gen3 cryptographic coprocessor 4769


The 4769 PCIe Cryptographic Coprocessor features a PCIe local bus-compatible interface.
The coprocessor holds a security-enabled subsystem module and batteries for backup power.

The hardened encapsulated subsystem contains redundant IBM PowerPC® 476 processors,
custom symmetric key and hashing engines to perform AES, DES, TDES, SHA-1 and SHA-
2, MD5 and HMAC, and public key cryptographic algorithm support for RSA and Elliptic Curve
Cryptography.

Other hardware support includes a secure real-time clock, a hardware random number
generator, and a prime number generator. It also contains a separate service processor that
is used to manage self-test and firmware updates. The secure module is protected by a
tamper responding design that protects against various system attacks.

It includes acceleration for: AES; DES; Triple DES; HMAC; CMAC; MD5; multiple SHA
hashing methods; modular-exponentiation hardware, such as RSA and ECC; and full-duplex
direct memory access (DMA) communications.

A security-enabled code-loading arrangement allows control program and application


program loading and refreshes after coprocessor installation in your server. IBM offers an
embedded subsystem control program and a cryptographic application programming
interface (API) that implements the IBM Common Cryptographic Architecture.

The IBM 4769 is verified by NIST at FIPS 140-2 Level 4, the highest level of certification that
is achievable as of this writing for commercial cryptographic devices.

This adapter is available only in full-height form factor, and is available in two variations with
two different Feature Codes:
򐂰 #EJ35 does not include a Blind Swap Cassette (BSC) and can be installed only within the
chassis of a Power S1014 or Power S1024 server.
򐂰 #EJ37 includes a Blind Swap Cassette housing, and can be installed only in a PCIe Gen3
I/O expansion drawer enclosure. This option is supported only for the Power S1022s and
Power S1022 server models.

Chapter 3. Available features and options 125


Table 3-19 lists the cryptographic coprocessor and accelerator adapters that are supported in
the Power10 processor based scale-out servers.

Table 3-19 Cryptographic adapters supported in the Power S1014, S1024, and PCIe expansion drawer
Feature CCIN Description Operating Order type
code system
support

EJ32 4767 PCIe3 Crypto Coprocessor no BSC 4767 AIX, Linux, IBM i Both
(S1014 or S1024 chassis only) Direct onlya

EJ33 4767 PCIe3 Crypto Coprocessor BSC-Gen3 AIX, Linux, IBM i


4767 (PCIe expansion drawer only) Direct onlya

EJ35 C0AF PCIe3 Crypto Coprocessor no BSC 4769 AIX, Linux, IBM i
(S1014 or S1024 chassis only) Direct only

EJ37 C0AF PCIe3 Crypto Coprocessor BSC-Gen3 AIX, Linux, IBM i


4769 (PCIe expansion drawer only) Direct only
a. PowerVM virtualization is not supported for this adapter.

3.5 Internal storage


NVMe solid-state devices (SSD) are used for internal storage in the Power S1014, S1022s,
S1022, and S1024 servers. The Power S1014 and S1024 servers support up to 16 NVMe
drives and the Power S1022s and S1022 servers support up to 8 NVMe drives.

General PCIe slots (C10/C8 and C11) support NVMe just a bunch of flash (JBOF) adapters
and are cabled to the NVMe backplane. Each NVMe JBOF card contains a 52-lane PCIe
Gen4 switch. The connected NVMe devices are individually addressable, and can be
allocated individually to LPARs that are running on the system.

Table 3-20 lists the available internal storage options.

Table 3-20 Internal storage summary


Power S1022s and Power S1014 Power S1024
S1022

NVMe 8-device N/A 1-2 1-2


backplane without Up to 16 devices Up to 16 devices
RDX

NVMe 8-device N/A 1-2 1-2


backplane with RDX Up to 16 devices Up to 16 devices

NVMe 4-device 1-2 N/A N/A


backplane Up to 8 devices

NVMe U.2 7 mm 800 GB


devices (Max 4)

NVMe U.2 15 mm 0.8, 1.6, 3.2, and 6.4 TB


devices

Concurrently Yes
Maintainable NVMe

RDX (optional) No Yes Yes

126 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
3.5.1 S1022s and S1022 Backplane
The Storage backplane with four NVMe U.2 drive slots (#EJ1X) is the base storage
backplane. The internal NVMe is attached to the processor by using a plug-in PCIe NVMe
JBOF card that is included with each storage backplane.

Up to 2 NVMe JBOF cards can be populated in the Power S1022s and S1022 servers with a
1:1 correspondence between the card and the storage backplane. Each JBOF card contains
four connectors that are cabled to connectors on a single 4-device backplane, with each cable
containing signals for two NVMe devices. Only two cables are installed to support a total of
four devices per backplane.

The NVMe JBOF card and storage backplane connection is shown in Figure 3-1.

Storage
Backplane
NVMe JBOF Card

Cables
2 x4 U.2 drive slot

Gen4 U.2 drive slot


52-lane
2 x4 U.2 drive slot
Switch
U.2 drive slot

G4 x16

PCIe G4 x16 slot

Figure 3-1 NVMe JBOF card and storage backplane connection

The NVMe JBOF card is treated as a regular cable card, with the similar EEH support as a
planar switch. The card is not concurrently maintainable because of the cabling that is
required to the NVMe backplane.

3.5.2 S1014 and S1024 Backplane


The Storage backplane with eight NVMe U.2 device slots (#EJ1Y) is the base storage
backplane. The internal NVMe is attached to the processor by using a plug-in PCIe NVMe
JBOF card that ship with each storage backplane.

Up to two NVMe JBOF cards can be populated in the Power S1014 and S1024 servers with a
1:1 correspondence between the card and the storage backplane. Each JBOF card contains
four connectors that are cabled to four connectors on a single 8-device backplane, with each
cable containing signals for two NVMe devices.

Chapter 3. Available features and options 127


The NVMe JBOF card and storage backplane connection is shown in Table 3-2.

Storage
Backplane
NVMe JBOF Card
Cables

2 x4 U.2 drive slot


U.2 drive slot

2 x4 U.2 drive slot


Gen4
U.2 drive slot
52-lane
Switch 2 x4 U.2 drive slot
Cables
U.2 drive slot

2 x4 U.2 drive slot


G4 x16
U.2 drive slot

PCIe G4 x16 slot

Figure 3-2 NVMe JBOF card and storage backplane connection

The NVMe JBOF card is treated as a regular cable card, with the similar EEH support as a
planar switch. The card is not concurrently maintainable because of the cabling that is
required to the NVMe backplane.

3.5.3 NVMe JBOF to backplane cabling


Three PCIe slots support NVMe JBOF cards: C8, C10, and C11. Each of these PCIe slots
can be cabled only to one single NVMe Backplane location. C8 and C10 are mutually
exclusive, such that if a JBOF card is plugged in C10, a JBOF card cannot be in C8. Also, if a
JBOF card is plugged in C8, a JBOF card cannot be in C10.

PCIe slots C8 and C10 can be cabled only to NVMe backplane P1 and PCIe slot C11 can be
cabled only to NVMe backplane P2. A JBOF card never can be plugged in a lower numbered
slot than an OpenCAPI adapter.

Table 3-21 lists the NVMe JBOF card slots that are cabled to NVMe backplanes under various
configurations.

Table 3-21 NVMe JBOF to backplane mappings

NVMe backplane

P1 (Left) P2 (Middle)

Default JBOF card locationa C10 C11

JBOF card location when x16 adapter is C8 C11


required in C10

JBOF card location when OpenCAPI is in C10b N/A C11


a. JBOF cards are plugged in x16 slot first by default.
b. JBOF card is not allowed in lower slot than OpenCAPI.

128 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Each connector on the JBOF card cables to the corresponding connector on the backplane:
򐂰 C0 provides signaling for NVMe drives 0 and 1
򐂰 C1 provides signaling for drives 2 and 3
򐂰 C2 provides signaling for drives 4 and 5
򐂰 C3 provides signaling for drives 6 and 7

In the Power S1022s and S1022 servers, only C1 and C2 are connected. The other
connectors on the JBOF and backplane are left unconnected.

Figure 3-3 shows the connector numbering on the NVMe JBOF card on the left and the
NVMe backplane on the right.

T0 = NVMe C0/C1
T1 = NVMe C2/C3
T3 = NVMe C4/C5
T4 = NVMe C6/C7

Figure 3-3 Connector locations for JBOF card and NVMe backplane

The following methods are used to reduce the likelihood of mis-cabling:


򐂰 Cable lengths allow cabling only from a specific PCIe slot to a specific backplane.
򐂰 Labeling connectors 0 - 3 on the NVMe JBOF card and the backplane indicates how to
install the cables.
򐂰 Four bits in each cable are used by the hypervisor for topology checking to confirm
whether the JBOF card is cabled to the correct backplane and that each of the four
connectors on the JBOF card is cabled to the corresponding connectors on the backplane.

3.5.4 NVMe support


This section provides information about the available PCIe based NVMe storage devices for
the Power S1014, S1022s, S1022, and S1024 servers. These servers are based on the same
PCIe form factor as other PCIe adapters, and are connected to the PCIe slots in the rear of
each server. These servers are different from the U.2 form factor NVMe devices that are
installed in the front of the system and are connected by using the JBOF adapters and
storage backplanes.

For more information about the U.2 form factor NVMe storage devices, see 3.8, “Disk and
media features” on page 134.

Chapter 3. Available features and options 129


Table 3-22 lists the PCIe based NVMe storage devices that are available for the Power
S1022s and S1022 servers.

Table 3-22 PCIe-based NVMe storage devices for the Power S1022s and S1022 servers
Feature Code CCIN Description Minimum Maximum Operating
system support

EC5C 58FD PCIe3 x8 LP 3.2 TB NVMe Flash adapter for 0 9 򐂰 IBM ia


AIX and Linux 򐂰 Linux
򐂰 AIX
EC5E 58FE PCIe3 x8 LP 6.4 TB NVMe Flash adapter for 0 9
AIX and Linux

EC5G 58FC PCIe3 x8 LP 1.6 TB NVMe Flash Adapter for 0 9


AIX and Linux

EC7A 594A PCIe4 LP 1.6 TB NVMe Flash Adapter x8 for 0 9


AIX and Linux

EC7C 594B PCIe4 LP 3.2 TB NVMe Flash Adapter x8 for 0 9


AIX and Linux

EC7E 594C PCIe4 LP 6.4 TB NVMe Flash Adapter x8 for 0 9


AIX and Linux
a. IBM i supported through VIOS.

Table 3-23 lists the PCIe-based NVMe storage devices that are available for the Power S1014
server.

Table 3-23 PCIe based NVMe storage adapters for the Power S1014 server
Feature Code CCIN Description Minimum Maximum Operating
system support

EC5B 58FC PCIe3 x8 1.6 TB NVMe Flash Adapter for 0 4 򐂰 IBM ia


AIX/Linux 򐂰 AIX
򐂰 Linux
EC5D 58FD PCIe3 x8 3.2 TB NVMe Flash Adapter for 0 4
AIX/Linux

EC5F 58FE PCIe3 x8 6.4 TB NVMe Flash Adapter for 0 4


AIX/Linux

EC6V 58FC PCIe3 x8 1.6 TB NVMe Flash Adapter for 0 4 IBM ib


IBM i

EC6X 58FD PCIe3 x8 3.2 TB NVMe Flash Adapter for 0 4


IBM i

EC6Z 58FE PCIe3 x8 6.4 TB NVMe Flash Adapter for 0 4


IBM i

EC7B 594A PCIe4 1.6 TB NVMe Flash Adapter x8 for 0 4 򐂰 IBM i


AIX/Linux 򐂰 AIX
򐂰 Linux
EC7D 594B PCIe4 3.2 TB NVMe Flash Adapter x8 for 0 4
AIX/Linux

EC7F 594C PCIe4 6.4 TB NVMe Flash Adapter x8 for 0 4


AIX/Linux

EC7K 594A PCIe4 1.6 TB NVMe Flash Adapter x8 for 0 4 IBM ib


IBM i

130 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Feature Code CCIN Description Minimum Maximum Operating
system support

EC7M 594B PCIe4 3.2 TB NVMe Flash Adapter x8 for 0 4 IBM ib


IBM i

EC7P 594C PCIe4 6.4 TB NVMe Flash Adapter x8 for 0 4


IBM i
a. IBM i supported through VIOS.
b. IBM i 7.5, or later, IBM i 7.4 TR6.

Table 3-24 lists the PCIe-based NVMe storage devices that are available for the Power S1024
server.

Table 3-24 PCIe based NVMe storage devices for the Power S1024 server
Feature code CCIN Description Min Max Operating
system support

EC5B 58FC PCIe3 x8 1.6 TB NVMe Flash 0 9 򐂰 IBM ia


Adapter for AIX/Linux 򐂰 AIX
򐂰 Linux
EC5D 58FD PCIe3 x8 3.2 TB NVMe Flash 0 9
Adapter for AIX/Linux

EC5F 58FE PCIe3 x8 6.4 TB NVMe Flash 0 9


Adapter for AIX/Linux

EC6V 58FC PCIe3 x8 1.6 TB NVMe Flash 0 9 IBM ib


Adapter for IBM i

EC6X 58FD PCIe3 x8 3.2 TB NVMe Flash 0 9


Adapter for IBM i

EC6Z 58FE PCIe3 x8 6.4 TB NVMe Flash 0 9


Adapter for IBM i

EC7B 594A PCIe4 1.6 TB NVMe Flash Adapter 0 9 򐂰 IBM ia


x8 for AIX/Linux 򐂰 AIX
򐂰 Linux
EC7D 594B PCIe4 3.2 TB NVMe Flash Adapter 0 9
x8 for AIX/Linux

EC7F 594C PCIe4 6.4 TB NVMe Flash Adapter 0 9


x8 for AIX/Linux

EC7K 594A PCIe4 1.6 TB NVMe Flash Adapter 0 9 IBM ib


x8 for IBM i

EC7M 594B PCIe4 3.2 TB NVMe Flash Adapter 0 9


x8 for IBM i

EC7P 594C PCIe4 6.4 TB NVMe Flash Adapter 0 9


x8 for IBM i
a. IBM i supported through VIOS.
b. IBM i 7.5, or later, IBM i 7.4 TR6.

Chapter 3. Available features and options 131


3.5.5 RAID support
Data protection functions for NVMe storage devices in the system unit are provided by
operating system mirroring. Typically, NVMe devices support mirroring (RAID1), with devices
plugged in multiples of two. Operating systems can successfully mirror by using an odd
number of NVMe devices that use “namespaces” on the NVMe devices and mirroring the
name spaces.

Several protection options are available for hard disk drives (HDDs) or SSDs that are in
disk-only I/O drawers. Although protecting drives is always preferred, AIX and Linux users can
choose to leave a few or all drives unprotected at their own risk. IBM supports these
configurations.

Supported RAID functions


The following supported PCIe based SAS adapters (see 3.4.3, “SAS adapters” on page 123)
provide hardware support for RAID 0, 5, 6, and 10:
򐂰 RAID 0 provides striping for performance, but does not offer any fault tolerance.
The failure of a single drive results in the loss of all data on the array. This version of RAID
increases I/O bandwidth by simultaneously accessing multiple data paths.
򐂰 RAID 5 uses block-level data striping with distributed parity.
RAID 5 stripes both data and parity information across three or more drives. Fault
tolerance is maintained by ensuring that the parity information for any block of data is
placed on a drive that is separate from the ones that are used to store the data. This
version of RAID provides data resiliency if a single drive fails in a RAID 5 array.
򐂰 RAID 6 uses block-level data striping with dual distributed parity.
RAID 6 is the same as RAID 5, except that it uses a second level of independently
calculated and distributed parity information for more fault tolerance. A RAID 6
configuration requires N+2 drives to accommodate the added parity data, which makes it
less cost-effective than RAID 5 for equivalent storage capacity.
This version of RAID provides data resiliency if one or two drives fail in a RAID 6 array.
When you work with large capacity disks, RAID 6 enables you to sustain data parity during
the rebuild process.
򐂰 RAID 10 is a striped set of mirrored arrays.
RAID 10 is a combination of RAID 0 and RAID 1. A RAID 0 stripe set of the data is created
across a two-disk array for performance benefits. A duplicate of the first stripe set is then
mirrored on another two-disk array for fault tolerance.
This version of RAID provides data resiliency if a single drive fails, and it can provide
resiliency for multiple drive failures.

3.6 External SAS ports


No external SAS ports are available in the Power S1014, S1022s, S1022, and S1024 servers.

132 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
3.7 Media drawers
The IBM System Storage 7226 Model 1U3 Multi-Media Enclosure can accommodate up to
two LTO tape drives, two RDX removable disk drive docking stations, or up to four DVD-RAM
drives. The 7226 offers SAS, USB, and FC electronic interface drive options for attachment to
the Power S1014, S1022s, S1022, and S1024 servers.

The 7226-1U3 multi-media expansion enclosure is a 1U rack-mountable dual bay enclosure


with storage device options of LTO7, 8, and 9 tape drives with SAS or Fibre Channel interface.
The 7226 also offers DVD-RAM SAS and USB drive features and RDX 3.0 docking station
options. Up to two devices (or four DVD-RAM drives) can be installed in any combination in
the 7226 enclosure.

The 7226 offers the following drive features:


򐂰 RDX 3.0 Removable Disk Drive Docking Station (#EU03)
򐂰 DVD Sled with DVD-RAM SATA/SAS Drive (#1420)
򐂰 DVD Sled with DVD-RAM USB Drive (#5762)
򐂰 Half-High LTO Ultrium 7 SAS Tape Drive (#8441)
򐂰 Half-High LTO Ultrium 7 FC Tape Drive (#8446)
򐂰 Half-High LTO Ultrium 8 SAS Tape Drive (#8541)
򐂰 Half-High LTO Ultrium 8 FC Tape Drive (#8546)
򐂰 Half-High LTO Ultrium 9 SAS Tape Drive (#8641)
򐂰 Half-High LTO Ultrium 9 FC Tape Drive (#8646)

For more information about the 7226-1U3 multi-media expansion enclosure and supported
options, see 3.10.4, “Useful rack additions” on page 153.

3.7.1 External DVD drives


The Stand-alone USB DVD drive (#EUA5) is an optional, stand-alone external USB-DVD
device. It requires high current at 5V and must use the front USB 3.0 port on the Power
S1014, S1022s, S1022, and S1024 servers. This device includes a USB cable, which
provides the data path and power to this drive.

3.7.2 RDX removable disk drives


The RDX USB External Docking Station (#EUA4) accommodates RDX removable disk
cartridges of any capacity. The disk is in a protective rugged cartridge enclosure that plugs
into the docking station. The docking station holds one removable rugged disk drive or
cartridge at a time. The rugged removable disk cartridge and docking station can be used in a
similar way to a tape drive.

The RDX USB External Docking Station attaches to a Power server by way of a USB cable,
which carries data and control information. It is not powered by the USB port on the Power
server or Power server USB adapter, but has a separate electrical power cord.

Physically, the docking station is a stand-alone enclosure that is approximately 2.0 x 7.0 x
4.25 inches and can sit on a shelf or on top of equipment in a rack.

Chapter 3. Available features and options 133


Various disk drives are available, as listed in Table 3-25.

Table 3-25 RDX removable disk drives


Feature code Description

1107 USB 500 GB Removable Disk Drive

EU01 1TB Removable Disk Drive Cartridge

EU08a RDX 320 GB Removable Disk Drive

EU15a 1.5TB Removable Disk Drive Cartridge

EU2T 2TB Removable Disk Drive Cartridge


a. Supported only. The feature can be migrated from existing systems only.

3.8 Disk and media features


NVMe SSDs are used for internal storage in the Power S1014, S1022s, S1022, and S1024
servers. The Power S1014 and S1024 servers support up to 16 NVMe drives and the Power
S1022s and S1022 servers support up to eight NVMe drives.

General PCIe slots (C10/C8, C11) support NVMe JBOF cards that are cabled to an NVMe
backplane. NVMe JBOF cards contain a 52-lane PCIe Gen4 switch.

The Power S1014 and S1024 servers also support an optional internal RDX drive that is
attached by way of the USB controller.

Table 3-26 lists the available internal storage options that can be installed in the Power S1014
and S1024 servers.

Table 3-26 Internal storage options in the Power S1014 and S1024 servers
Feature code Description Maximum

EJ1Ya Storage backplane with eight NVMe U.2 drive slots 2

EUA0 RDX USB Internal Docking Station 1


a. Each backplane ships with 1 NVMe JBOF card that plugs into a PCIe slot.

The Power S1014 and S1024 servers with two storage backplanes and RDX drive are shown
in Figure 3-4 on page 135.

134 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Figure 3-4 The Power S1014 and S1024 servers with two storage backplanes and RDX drive

Table 3-27 lists the available U.2 form factor NVMe drive Feature Codes for the Power S1014
and S1024 servers. These codes are different from the PCIe based NVMe storage devices
that can be installed in the PCIe slots in the rear of the server. For more information about the
available PCIe-based NVMe adapters, see 3.5.4, “NVMe support” on page 129.

Table 3-27 U.2 form factor NVMe device features in the Power S1014 and S1024 servers
Feature code CCIN Description Minimum Maximum Operating system
support

EC5V 59BA Enterprise 6.4 TB SSD PCIe4 NVMe 0 16 AIX, IBM ia, and Linux
U.2 module for AIX/Linux

EC5W 59BA Enterprise 6.4 TB SSD PCIe4 NVMe 0 16 IBM ib


U.2 module for IBM i

EC5X 59B7 Mainstream 800 GB SSD PCIe3 NVMe 0 4 AIX and Linux
U.2 module for AIX/Linux

EC7T 59B7 800 GB Mainstream NVMe U.2 SSD 4k 0 4


for AIX/Linux

EKF2 5B53 Enterprise 800 GB SSD PCIe4 NVMe 0 16 IBM i


U.2 module for IBM i

EKF3 5B52 Enterprise 1.6 TB SSD PCIe4 NVMe 0 16 AIX, IBM ia , and Linux
U.2 module for AIX/Linux

EKF4 5B52 Enterprise 1.6 TB SSD PCIe4 NVMe 0 16 IBM i


U.2 module for IBM i

EKF5 5B51 Enterprise 3.2 TB SSD PCIe4 NVMe 0 16 AIX, IBM ia , and Linux
U.2 module for AIX/Linux

EKF6 5B51 Enterprise 3.2 TB SSD PCIe4 NVMe 0 16 IBM i


U.2 module for IBM i

Chapter 3. Available features and options 135


Feature code CCIN Description Minimum Maximum Operating system
support

EKF7 5B50 Enterprise 6.4 TB SSD PCIe4 NVMe 0 16 AIX, IBM ia , and Linux
U.2 module for AIX/Linux

EKF8 5B50 Enterprise 6.4 TB SSD PCIe4 NVMe 0 16 IBM ia


U.2 module for IBM i

ES1E 59B8 Enterprise 1.6 TB SSD PCIe4 NVMe 0 16 AIX, IBM ia , and Linux
U.2 module for AIX/Linux

ES1F 59B8 Enterprise 1.6 TB SSD PCIe4 NVMe 0 16 AIX and IBM ib
U.2 module for IBM i

ES1G 59B9 Enterprise 3.2 TB SSD PCIe4 NVMe 0 16 AIX, IBM ia , and Linux
U.2 module for AIX/Linux

ES1H 59B9 Enterprise 3.2 TB SSD PCIe4 NVMe 0 16 IBM ib


U.2 module for IBM i

ES1K 5947 Enterprise 800 GB SSD PCIe4 NVMe 0 16 IBM ib


U.2 module for IBM i

ES3A 5B53 Enterprise 800 GB SSD PCIe4 NVMe 0 16 IBM i


U.2 module for IBM i

ES3B 5B34 Enterprise 1.6 TB SSD PCIe4 NVMe 0 16 AIX, IBM ia , and Linux
U.2 module for AIX/Linux

ES3C 5B52 Enterprise 1.6 TB SSD PCIe4 NVMe 0 16 IBM i


U.2 module for IBM i

ES3D 5B51 Enterprise 3.2 TB SSD PCIe4 NVMe 0 16 AIX, IBM ia , and Linux
U.2 module for AIX/Linux

ES3E 5B51 Enterprise 3.2 TB SSD PCIe4 NVMe 0 16 IBM i


U.2 module for IBM i

ES3F 5B50 Enterprise 6.4 TB SSD PCIe4 NVMe 0 16 AIX, IBM ia , and Linux
U.2 module for AIX/Linux

ES3G 5B50 Enterprise 6.4 TB SSD PCIe4 NVMe 0 16 IBM ib


U.2 module for IBM i
a. IBM i supported through VIOS.
b. IBM i 7.5, or later, IBM i 7.4 TR6, or later.

Table 3-28 lists the available internal storage option that can be installed in the Power S1022s
and S1022 servers.

Table 3-28 Internal storage option in the Power S1022s and S1022 servers
Feature code Description Maximum
a
EJ1X Storage backplane with four NVMe U.2 drive slots 2
a. Each backplane ships with 1 NVMe JBOF card that plugs into a PCIe slot.

136 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Table 3-29 lists the available U.2 form factor NVMe drive Feature Codes for the
Power S1022s and S1022 servers. These codes are different from the PCIe based NVMe
storage devices that can be installed in the PCIe slots in the rear of the server. For more
information about the available PCIe-based NVMe adapters, see 3.5.4, “NVMe support” on
page 129.

Table 3-29 U.2 form factor NVMe device features in the Power S1022s and S1022 servers
Feature code CCIN Description Minimum Maximum Operating system
support

EC5V 59BA Enterprise 6.4 TB SSD PCIe4 NVMe 0 8 򐂰 AIX


U.2 module for AIX/Linux 򐂰 IBM ia
򐂰 Linux

EC5X 59B7 Mainstream 800 GB SSD PCIe3 0 4 AIX and Linux


NVMe U.2 module for AIX/Linux

EC7T 59B7 800 GB Mainstream NVMe U.2 SSD 0 4


4k for AIX/Linux

EKF3 5B52 Enterprise 1.6 TB SSD PCIe4 NVMe 0 8 򐂰 AIX


U.2 module for AIX/Linux 򐂰 IBM ia
򐂰 Linux
EKF5 5B51 Enterprise 3.2 TB SSD PCIe4 NVMe 0 8
U.2 module for AIX/Linux

EKF7 5B50 Enterprise 6.4 TB SSD PCIe4 NVMe 0 8


U.2 module for AIX/Linux

ES1E 59B8 Enterprise 1.6 TB SSD PCIe4 NVMe 0 8


U.2 module for AIX/Linux

ES1G 59B9 Enterprise 3.2 TB SSD PCIe4 NVMe 0 8


U.2 module for AIX/Linux

ES3B 5B34 Enterprise 1.6 TB SSD PCIe4 NVMe 0 8


U.2 module for AIX/Linux

ES3D 5B51 Enterprise 3.2 TB SSD PCIe4 NVMe 0 8


U.2 module for AIX/Linux

ES3F 5B50 Enterprise 6.4 TB SSD PCIe4 NVMe 0 8


U.2 module for AIX/Linux
a. IBM i supported through VIOS.

The Stand-alone USB DVD drive (#EUA5) is an optional, stand-alone external USB-DVD
device. This device includes a USB cable. The cable provides the data path and power to this
drive.

SAS backplane is not supported on the Power S1014, S1022s, S1022, and S1024 servers.
SAS drives can be placed only in IBM EXP24SX SAS Storage Enclosures, which are
connected to system units by using a serial-attached SCSI (SAS) ports in PCIe based SAS
adapters.

For more information about the available SAS adapters, see 3.4.3, “SAS adapters” on
page 123.

Chapter 3. Available features and options 137


3.9 External IO subsystems
If more PCIe slots beyond the system node slots are required, the Power S1014, S1022s,
S1022, and S1024 servers support adding I/O expansion drawers.

If you need more directly connected storage capacity than is available within the internal
NVMe storage device bays, you can attach external disk subsystems to the Power S1014,
S1022s, S1022, and S1024 servers:
򐂰 EXP24SX SAS Storage Enclosures
򐂰 IBM System Storage

3.9.1 PCIe Gen3 I/O expansion drawer


This 19-inch, 4U (4 EIA) enclosure provides PCIe Gen3 slots outside of the system unit. It has
two module bays. One 6-slot fan-out Module (#EMXH) can be placed in each module bay.
Two 6-slot modules provide a total of 12 PCIe Gen3 slots. Each fan-out module is connected
to a PCIe3 Optical Cable adapter that is installed in the system unit over an active optical
CXP cable (AOC) pair or CXP copper cable pair.

The PCIe Gen3 I/O Expansion Drawer has two redundant, hot-plug power supplies. Each
power supply has its own separately ordered power cord. The two power cords plug into a
power supply conduit that connects to the power supply. The single-phase AC power supply is
rated at 1030 W and can use 100 - 120 V or 200 - 240 V. If 100 - 120 V is used, the maximum
is 950 W. It is a best practice that the power supply connects to a power distribution unit
(PDU) in the rack. IBM Power PDUs are designed for a 200 - 240 V electrical source.

A blind swap cassette (BSC) is used to house the full-height adapters that are installed in
these slots. The BSC is the same BSC that is used with previous generation 12X attached I/O
drawers (#5802, #5803, #5877, and #5873). The drawer includes a full set of BSCs, even if
the BSCs are empty.

Concurrent repair, and adding or removing PCIe adapters, is done by HMC-guided menus or
by operating system support utilities.

Figure 3-5 shows a PCIe Gen3 I/O expansion drawer.

Figure 3-5 PCIe Gen3 I/O expansion drawer

138 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Figure 3-6 shows the back view of the PCIe Gen3 I/O expansion drawer.

PCIe Optical Interface to CEC PCIe Optical Interface to CEC

Fan-out Module Fan-out Module


6 PCIe G3 Slots 6 PCIe G3 Slots
4 x8 & 2 x16 4 x8 & 2 x16 Dual Power Cords

Figure 3-6 Rear view of the PCIe Gen3 I/O expansion drawer

Chapter 3. Available features and options 139


I/O drawers and usable PCI slots
Figure 3-7 shows the rear view of the PCIe Gen3 I/O expansion drawer that is equipped with
two PCIe3 6-slot fan-out modules with the location codes for the PCIe adapter slots.

P1-C1 P1-C6 P2-C3 P2-C4


P1-C2 P1-C5 P2-C2 P2-C5
P1-C3 P1-C4 P2-C1 P2-C6

Figure 3-7 Rear view of a PCIe Gen3 I/O expansion drawer with PCIe slots location codes

Table 3-30 lists the PCI slots in the PCIe Gen3 I/O expansion drawer that is equipped with two
PCIe3 6-slot fan-out modules.

Table 3-30 PCIe slot locations for the PCIe Gen3 I/O expansion drawer with two fan-out modules
Slot Location code Description

Slot 1 P1-C1 PCIe3, x16

Slot 2 P1-C2 PCIe3, x8

Slot 3 P1-C3 PCIe3, x8

Slot 4 P1-C4 PCIe3, x16

Slot 5 P1-C5 PCIe3, x8

Slot 6 P1-C6 PCIe3, x8

Slot 7 P2-C1 PCIe3, x16

Slot 8 P2-C2 PCIe3, x8

Slot 9 P2-C3 PCIe3, x8

Slot 10 P2-C4 PCIe3, x16

Slot 11 P2-C5 PCIe3, x8

Slot 12 P2-C6 PCIe3, x8

140 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Consider the following points about Table 3-30 on page 140:
򐂰 All slots support full-length, full-height adapters or short (LP) adapters with a full-height tail
stock in single-wide, Gen3, BSC.
򐂰 Slots C1 and C4 in each PCIe3 6-slot fan-out module are x16 PCIe3 buses, and slots C2,
C3, C5, and C6 are x8 PCIe buses.
򐂰 All slots support enhanced error handling (EEH).
򐂰 All PCIe slots are hot-swappable and support concurrent maintenance.

Table 3-31 lists the maximum number of I/O drawers that are supported and the total number
of PCI slots that are available to the server.

Table 3-31 Maximum number of I/O drawers that are supported and total number of PCI slots
Server Maximum number of Maximum number of Maximum PCIe
I/O exp drawers I/O fan-out modules slots

Power S1014 (1-socket)a 1b 1 10

Power S1022s (1-socket) 1b 1 10

Power S1022s (2-socket) 2 4 30

Power S1022 (1-socket) 1b 1 10

Power S1022 (2-socket) 2 4 30

Power S1024 (1-socket) 1b 1 10

Power S1024 (2-socket) 2 4 30


a. The PCIe expansion drawer (#EMX0) cannot be used with the four-core configuration Power
S1014 server.
b. The empty PCIe module bay must be populated by a filler module.

PCIe3 x16 to CXP Converter Adapter


The PCIe3 x16 to CXP Converter adapter provides two ports for the attachment of two
expansion drawer cables. One adapter supports the attachment of one PCIe3 6-slot fanout
module in an EMX0 PCIe Gen3 I/O expansion drawer.

Table 3-32 lists the available converter adapter that can be installed in the Power S1022s and
S1022 servers.

Table 3-32 Available converter adapter in the Power S1022s and S1022
Feature code Slot priorities Maximum Slot priorities Maximum
(one processor) number of (two number of
adapters processors) adapters
supported supported

EJ24a 10 1 3, 0, 4, 10 4
a. single-wide, low-profile

Chapter 3. Available features and options 141


Table 3-33 lists the available converter adapter that can be installed in the Power S1014 and
S1024 servers.

Table 3-33 Available converter adapter in the Power S1014 and S1024
Feature code Slot priorities Maximum Slot priorities Maximum
(one processor) number of (two number of
adapters processors) adapters
supported supported

EJ24a 10 1 3, 0, 4, 10 4
a. single-wide, full-height

The PCIe3 x16 to CXP Converter Adapter (#EJ24) is shown in Figure 3-8.

T1

T2

Figure 3-8 PCIe3 x16 to CXP Converter Adapter (#EJ24)

PCIe Gen3 I/O expansion drawer optical cabling


I/O drawers are connected to the adapters in the system node through data transfer cables:
򐂰 3 m Active Optical Cable Pair for PCIe3 Expansion Drawer (#ECCX)
򐂰 10 m Active Optical Cable Pair for PCIe3 Expansion Drawer (#ECCY)
򐂰 3 m Copper CXP Cable Pair for PCIe3 Expansion Drawer (#ECCS)

Although these cables are not redundant, the loss of one cable reduces the I/O bandwidth
(that is, the number of lanes that are available to the I/O module) by 50%.

Note: Consider the following points:


򐂰 Use the 3 m cables for intra-rack installations. Use the 10 m cables for inter-rack
installations.
򐂰 You cannot mix copper and optical cables on the same PCIe Gen3 I/O drawer. Both
fan-out modules use copper cables or both use optical cables.

142 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
A minimum of one PCIe3 x16 to CXP Converter adapter for PCIe3 Expansion Drawer is
required to connect to the PCIe3 6-slot fan-out module in the I/O expansion drawer. The
fan-out module has two CXP ports. The top CXP port of the fan-out module is cabled to the
top CXP port of the PCIe3 x16 to CXP Converter adapter. The bottom CXP port of the fan-out
module is cabled to the bottom CXP port of the same PCIe3 x16 to CXP Converter adapter.

To set up the cabling correctly, complete the following steps:


1. Connect an optical cable or copper CXP cable to connector T1 on the PCIe3 x16 to CXP
Converter adapter in your server.
2. Connect the other end of the optical cable or copper CXP cable to connector T1 on one of
the PCIe3 6-slot fan-out modules in your expansion drawer.
3. Connect another cable to connector T2 on the PCIe3 x16 to CXP Converter adapter in
your server.
4. Connect the other end of the cable to connector T2 on the PCIe3 6-slot fan-out module in
your expansion drawer.
5. Repeat steps 1 - 4 for the other PCIe3 6-slot fan-out module in the expansion drawer, if
required.

Drawer connections: Each fan-out module in a PCIe3 Expansion Drawer can be


connected only to a single PCIe3 x16 to CXP Converter adapter for PCIe3 Expansion
Drawer.

Figure 3-9 shows the connector locations for the PCIe Gen3 I/O Expansion Drawer.

Figure 3-9 Connector locations for the PCIe Gen3 I/O expansion drawer

Chapter 3. Available features and options 143


Figure 3-10 shows the typical optical cable connections.

Server I/O Expansion Drawer

Optical cable adapter Fanout module


T1 T1
Optical cables
T2 T2

Optical cable adapter Fanout module


T1 T1
Optical cables
T2 T2

Figure 3-10 Typical optical cable connections

PCIe Gen3 I/O expansion drawer system power control network cabling
No system power control network (SPCN) is used to control and monitor the status of power
and cooling within the I/O drawer. SPCN capabilities are integrated into the optical cables.

3.9.2 EXP24SX SAS Storage Enclosure


The EXP24SX SAS storage enclosure (#ESLS) is the only direct attached storage device
(DASD) drawer that is available for the Power S1014, S1022s, S1022, and S1024 servers to
provide non-NVMe storage device capacity.

The EXP24SX drawer is a storage expansion enclosure with 24 2.5-inch SFF SAS drive bays.
It supports up to 24 hot-plug HDDs or SSDs in only 2 EIA rack units (2U) of space in a 19-inch
rack. The EXP24SX SFF bays use SFF Gen2 (SFF-2) carriers or trays.

The EXP24SX drawers feature the following high-reliability design points:


򐂰 SAS bays that support hot-swap
򐂰 Redundant and hot-plug power and fan assemblies
򐂰 Dual power cords
򐂰 Redundant and hot-plug Enclosure Services Managers (ESMs)
򐂰 Redundant data paths to all drives
򐂰 LED indicators on drives, bays, ESMs, and power supplies that support problem
identification
򐂰 Through the SAS adapters and controllers, drives can be protected with RAID and
mirroring and hot-spare capability

144 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Figure 3-11 shows the EXP24SX drawer.

Figure 3-11 EXP24SX drawer

With AIX/Linux/VIOS, the EXP24SX can be ordered as configured with four sets of 6 bays
(mode 4), two sets of 12 bays (mode 2), or one set of 24 bays (mode 1). With IBM i, one set of
24 bays (mode 1) is supported. It is possible to change the mode setting in the field by using
software commands along with a documented procedure.

Note: Consider the following points:


򐂰 For the EXP24SX drawer, a maximum of 24 2.5-inch SSDs or 2.5-inch HDDs are
supported in the 24 SAS bays. HDDs and SSDs cannot be mixed in the same mode-1
drawer. HDDs and SSDs can be mixed in a mode-2 or mode-4 drawer, but they cannot
be mixed within a logical split of the drawer. For example, in a mode-2 drawer with two
sets of 12 bays, one set can hold SSDs and one set can hold HDDs, but SSDs and
HDDs cannot be mixed in the same set of 12 bays.
򐂰 When changing modes, a skilled, technically qualified person must follow the special
documented procedures. Improperly changing modes can destroy RAID sets, which
prevents access to data, or might allow partitions to access data from another partition.

Chapter 3. Available features and options 145


The front view of the #ESLS storage enclosure with mode group and drive locations is shown
in Figure 3-12.

1 2 3 4 5 6 7 8 9 0
1 1
1 2
1 3
1 4
1 5
1 6
1 7
1 8
1 9
1 0
2 1
2 2
2 3
2 4
2
-D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1 -D
1
P P P P P P P P P P P P P P P P P P P P P P P P

MODE 1 GROUP 1 24 disks x1

MODE 2 GROUP 1 GROUP 2 12 disks x2

MODE 4 GROUP 1 GROUP 2 GROUP 3 GROUP 4 6 disks x4

Figure 3-12 Front view of the ESLS storage enclosure with mode groups and drive locations

Four mini-SAS HD ports on the EXP24SX are attached to PCIe Gen3 SAS adapters. The
following PCIe3 SAS adapters support the EXP24SX:
򐂰 PCIe3 RAID SAS Adapter Quad-port 6 Gb x8 (#EJ0J)
򐂰 PCIe3 12 GB Cache RAID Plus SAS Adapter Quad-port 6 Gb x8 (#EJ14)
򐂰 PCIe3 LP RAID SAS Adapter Quad-port 6 Gb x8 (#EJ0M)

The attachment between the EXP24SX drawer and the PCIe Gen 3 SAS adapter is through
SAS YO12 or X12 cables. The PCIe Gen 3 SAS adapters support 6 Gb throughput. The
EXP24SX drawer can support up to 12 Gb throughput if future SAS adapters support that
capability.

The following cable options are available:


򐂰 X12 cable: 3-meter copper (#ECDJ), 4.5-meter optical (#ECDK), 10-meter optical
(#ECDL)
򐂰 YO12 cables: 1.5-meter copper (feature ECDT), 3-meter copper (#ECDU)
򐂰 1 M 100 GbE Optical Cable QSFP28 (AOC) (#EB5K)
򐂰 1.5 M 100 GbE Optical Cable QSFP28 (AOC) (#EB5L)
򐂰 2 M 100 GbE Optical Cable QSFP28 (AOC) (#EB5M)
򐂰 3 M 100 GbE Optical Cable QSFP28 (AOC) (#EB5R)
򐂰 5 M 100 GbE Optical Cable QSFP28 (AOC) (#EB5S)
򐂰 10 M 100 GbE Optical Cable QSFP28 (AOC) (#EB5T)
򐂰 15 M 100 GbE Optical Cable QSFP28 (AOC) (#EB5U)

146 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
򐂰 20 M 100 GbE Optical Cable QSFP28 (AOC) (#EB5V)
򐂰 30 M 100 GbE Optical Cable QSFP28 (AOC) (#EB5W)
򐂰 50 M 100 GbE Optical Cable QSFP28 (AOC) (#EB5X)

Six SAS connectors are at the rear of the EXP24SX drawers to which SAS adapters or
controllers are attached. They are labeled T1 T2, and T3; two T1s, two T2s, and two T3s
connectors. Consider the following points:
򐂰 In mode 1, two or four of the six ports are used. Two T2 ports are used for a single SAS
adapter, and two T2 and two T3 ports are used with a paired set of two adapters or a dual
adapters configuration.
򐂰 In mode 2 or mode 4, four ports are used, two T2s and two T3 connectors to access all the
SAS bays.
򐂰 The T1 connectors are not used.

Figure 3-13 shows the connector locations for the EXP24SX storage enclosure.

MODE 1 MODE 2 MODE 4


1 1 1
tsr
T3 2
T3 T3 2
T3 2 T3 o s d
3 3 T3 3 P le re
i
d b u
4 4 4 a q
T2 T2 T2 T2 T2
re C e
5 5
T2
5 a X R
6 6 6
h
S
7 7 7
8 8 8
9 9 9
10 10 10
T1 T1 T1 T1 T1 T1
11 11 11
12 12 12
13 13 13
14 14 14
15 15 15
16 16 16
17 17 17
18 18 18
19 19 19
20 20 20
21 21 21
22 22 22
23 23 23
24 24 24

ESM1 ESM2 ESM1 ESM2 ESM1 ESM2

Figure 3-13 Rear view of the EXP24SX with location codes and different split modes

For more information about SAS cabling and cabling configurations, see this IBM
Documentation web page.

Chapter 3. Available features and options 147


3.9.3 IBM Storage
With IBM Storage, every bit of your organization’s data is available to you in the most secure
and insightful way possible, so that you can take informed action on it quickly and
consistently.

For more information about the various offerings, see Data Storage Solutions.

The following sections highlight some of the offerings.

IBM Elastic Storage System


IBM Elastic Storage® System is a modern implementation of software-defined storage (SDS).
The IBM Elastic Storage System 3500 and IBM Elastic Storage System 5000 make it easier
for you to deploy fast, highly scalable storage for AI and big data.

With the low latency and high-performance NVMe storage technology and up to 8 YB global
file system and global data services of IBM Spectrum® Scale, the IBM Elastic Storage
System 3500 and 5000 nodes can grow to multi-yottabyte configurations. They also can be
integrated into a federated global storage system.

For more information, see IBM Elastic Storage System.

IBM FlashSystem: Flash data storage


The IBM FlashSystem family is a portfolio of cloud-enabled storage systems that can be
easily deployed and quickly scaled to help optimize storage configurations, streamline issue
resolution, and reduce storage costs.

IBM FlashSystem is built with IBM Spectrum Virtualize software to help deploy sophisticated
hybrid cloud storage solutions, accelerate infrastructure modernization, address
cybersecurity needs, and maximize value by using the power of AI. New IBM FlashSystem
models deliver the performance to facilitate cyber security without compromising production
workloads.

For more information, see IBM FlashSystem.

IBM DS8000 Storage system


IBM DS8900F is the next generation of enterprise data systems, which are built with
advanced Power9 processor technology. Designed for data-intensive and mission-critical
workloads, DS8900F adds next-level performance, data protection, resiliency, and availability
across your hybrid cloud solutions. It accomplishes this goal by using ultra-low latency, better
than 99.99999 (seven nines) availability, transparent cloud tiering, and advanced data
protection against malware and ransomware.

For more information, see IBM DS8000 Storage system.

148 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
3.10 System racks
Except for the S1014 tower model IBM Power S1014, the S1022s, S1022, and S1024 servers
fit into a standard 19-inch rack. These server models are all certified and tested in the IBM
Enterprise racks (7965-S42, 7014-T42, or 7014-T00). Customers can choose to place the
server in other racks if they are confident that those racks have the strength, rigidity, depth,
and hole pattern characteristics that are needed. Contact IBM Support to determine whether
other racks are suitable.

Order information: Only the IBM Enterprise 42U slim rack (7965-S42) is available and
supported for factory integration and installation of the server. The other Enterprise racks
(7014-T42 and 7014-T00) are supported only for installation into existing racks. Multiple
servers can be installed into a single IBM Enterprise rack in the factory or field.

If a system is installed in a rack or cabinet that is not from IBM, ensure that the rack meets the
requirements that are described in 3.10.5, “Original equipment manufacturer racks” on
page 155.

Responsibility: The customer is responsible for ensuring the installation of the server in
the preferred rack or cabinet results in a configuration that is stable, serviceable, and safe.
It also must be compatible with the drawer requirements for power, cooling, cable
management, weight, and rail security.

3.10.1 IBM Enterprise 42U Slim Rack 7965-S42


The 2.0-meter (79-inch) Model 7965-S42 is compatible with past and present IBM Power
servers and provides an excellent 19-inch rack enclosure for your data center. Its 600 mm
(23.6 in.) width combined with its 1100 mm (43.3 in.) depth plus its 42 EIA unit enclosure
capacity provides great footprint efficiency for your systems. It can be placed easily on
standard 24-inch floor tiles.

The 7965-S42 rack includes space for up to four PDUs in side pockets. Extra PDUs beyond
four are mounted horizontally and each uses 1U of rack space.

The Enterprise Slim Rack comes with options for the installed front door:
򐂰 Basic Black/Flat (#ECRM)
򐂰 High-End appearance (#ECRF)
򐂰 OEM Black (#ECRE)

All options include perforated steel, which provides ventilation, physical security, and visibility
of indicator lights in the installed equipment within. All options also include a lock and
mechanism that is identical to the lock on the rear doors.

Only one front door must be included for each rack ordered. The basic door (#ECRM) and
OEM door (#ECRE) can be hinged on the left or right side.

Orientation: #ECRF must not be flipped because the IBM logo is upside down.

Chapter 3. Available features and options 149


At the rear of the rack, a perforated steel rear door (#ECRG) or a Rear Door Heat Exchanger
(RDHX) can be installed. The only supported Rear Door Heat Exchanger is IBM machine
type 1164-95X, which can remove up to 30,000 watts (102,000 BTU) of heat per hour by
using chilled water. The no-charge Feature Code #ECR2 is included with the Enterprise Slim
Rack as an indicator when the Rear Door Heat Exchanger is ordered.

The basic door (#ECRG) can be hinged on the left or right side, and includes a lock and
mechanism identical to the lock on the front door. The basic rear door (#ECRG) or the RDHX
indicator (#ECR2) must be included with the order of a new Enterprise Slim Rack.

Because of the depth of the S1022s and S1022 server models, the 5-inch rear rack extension
(#ECRK) is required for the Enterprise Slim Rack to accommodate these systems. This
extension expands the space available for cable management and allows the rear door to
close safely.

3.10.2 AC power distribution units


The IBM Enterprise Slim Rack can be ordered with a selection of different Power Distribution
Units (PDUs) that deliver power to the individual servers from an incoming AC power source.
The choice and number of PDUs to install in the rack depends on the incoming power supply
available on site, the number of systems being powered, and the power consumption of those
systems.

Rack-integrated system orders require at least two PDU devices be installed in the rack to
support independent connection of redundant power supplies in the server.

Standard Power Distribution Unit


The standard PDU (#7188) mounts in a 19-inch rack and provides 12 C13 power outlets. The
PDU has six 16 A circuit breakers, with two power outlets per circuit breaker.

To connect to the standard PDU, and system units and expansion units must use a power
cord with a C14 plug to connect to #7188. One of the following power cords must be used to
distribute power from a wall outlet to the #7188 PDU: #6489, #6491, #6492, #6653, #6654,
#6655, #6656, #6657, #6658, or #6667.

High-function Power Distribution Unit


High-function PDUs provide more electrical power per PDU and offer better “PDU footprint”
efficiency. In addition, they are intelligent PDUs that provide insight to power usage by
receptacle and remote power on and off capability for easier support by individual receptacle.

The following high-function PDUs are orderable as #ECJJ, #ECJL, #ECJN, and #ECJQ:
򐂰 High Function 9xC19 PDU plus (#ECJJ)
This intelligent, switched 200 - 240 volt AC PDU includes nine C19 receptacles on the
front of the PDU and three C13 receptacles on the rear of the PDU. The PDU is mounted
on the rear of the rack, which makes the nine C19 receptacles easily accessible.
򐂰 High Function 9xC19 PDU plus 3-Phase (#ECJL)
This intelligent, switched 208-volt 3-phase AC PDU includes nine C19 receptacles on the
front of the PDU and three C13 receptacles on the rear of the PDU. The PDU is mounted
on the rear of the rack, which makes the nine C19 receptacles easily accessible.

150 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
򐂰 High Function 12xC13 PDU plus (#ECJN)
This intelligent, switched 200 - 240 volt AC PDU includes 12 C13 receptacles on the front
of the PDU. The PDU is mounted on the rear of the rack, which makes the 12 C13
receptacles easily accessible.
򐂰 High Function 12xC13 PDU plus 3-Phase (#ECJQ)
This intelligent, switched 208-volt 3-phase AC PDU includes 12 C13 receptacles on the
front of the PDU. The PDU is mounted on the rear of the rack, which makes the 12 C13
receptacles easily accessible.

Table 3-34 lists the Feature Codes for the high-function PDUs.

Table 3-34 High-function PDUs available with IBM Enterprise Slim Rack (7965-S42)
PDUs 1-phase or 3-phase 3-phase 208 V depending on
depending on country wiring country wiring standards
standards

Nine C19 receptacles ECJJ ECJL

Twelve C13 receptacles ECJN ECJQ

The PDU receives power through a UTG0247 power-line connector. Each PDU requires one
PDU-to-wall power cord. Various power cord features are available for various countries and
applications by varying the PDU-to-wall power cord, which must be ordered separately.

Each power cord provides the unique design characteristics for the specific power
requirements. To match new power requirements and save previous investments, these
power cords can be requested with an initial order of the rack, or with a later upgrade of the
rack features.

Table 3-35 lists the available PDU-to-wall power cord options for the PDU features, which
must be ordered separately.

Table 3-35 PDU-to-wall power cord options for the PDU features
Feature Wall plug Rated voltage Phase Rated amperage Geography
code (V AC)

6489 IEC 309, 3P+N+G, 230 3 32 amps/phase EMEA


32 A

6491 IEC 309, P+N+G, 230 1 63 amps


63 A

6492 IEC 309, 2P+G, 200 - 208, 240 1 48 amps US, Canada, LA,
60 A and Japan

6653 IEC 309, 3P+N+G, 230 3 16 amps/phase Internationally


16 A available

6654 NEMA L6-30 200 - 208, 240 1 24 amps US, Canada, LA,
and Japan
6655 RS 3750DP 200 - 208, 240 1 24 amps
(watertight)

6656 IEC 309,P+N+G, 230 1 24 amps EMEA


32 A

6657 PDL 230-240 1 32 amps Australia, New


Zealand

Chapter 3. Available features and options 151


Feature Wall plug Rated voltage Phase Rated amperage Geography
code (V AC)

6658 Korean plug 220 1 30 amps North and South


Korea

6667 PDL 380-415 3 32 amps Australia, New


Zealand

Notes: Ensure that a suitable power cord feature is configured to support the power that is
being supplied. Based on the power cord that is used, the PDU can supply 4.8 - 19.2 kVA.
The power of all the drawers that are plugged into the PDU must not exceed the power
cord limitation.

For maximum availability, a preferred approach is to connect power cords from the same
system to two separate PDUs in the rack, and to connect each PDU to independent power
sources.

For more information about power requirements of and the power cord for the 7965-S42 rack,
see this IBM Documentation web page.

PDU installation
The IBM Enterprise Slim Rack includes four side mount pockets to allow for the vertical
installation of PDUs. This configuration frees up more of the horizontal space in the rack for
the installation of systems and other equipment. Up to four PDU devices can be installed
vertically in each rack, so any other PDU devices must be installed horizontally. When PDUs
are mounted horizontally in a rack, they each use 1 EIA (1U) of rack space.

Note: When a new IBM Power server is factory installed in an IBM rack that also includes a
PCIe expansion drawer, all of the PDUs for that rack are installed horizontally by default.
This configuration allows for extra space in the sides of the rack to enhance cable
management.

3.10.3 Rack-mounting rules


Consider the following primary rules when you mount the system into a rack:
򐂰 The system can be placed at any location in the rack. For rack stability, start filling the rack
from the bottom.
򐂰 Any remaining space in the rack can be used to install other systems or peripheral
devices. Ensure that the maximum permissible weight of the rack is not exceeded and the
installation rules for these devices are followed.
򐂰 Before placing the system into the service position, follow the rack manufacturer’s safety
instructions regarding rack stability.

152 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
3.10.4 Useful rack additions
This section highlights several rack addition solutions for IBM Power rack-based systems.

IBM System Storage 7226 Model 1U3 Multi-Media Enclosure


The IBM System Storage 7226 Model 1U3 Multi-Media Enclosure can accommodate up to
two tape drives, two RDX removable disk drive docking stations, or up to four DVD-RAM
drives.

The IBM System Storage 7226 Multi-Media Enclosure supports LTO Ultrium and DAT160
Tape technology, DVD-RAM, and RDX removable storage requirements on the following IBM
systems:
򐂰 IBM POWER6 processor-based systems
򐂰 IBM POWER7 processor-based systems
򐂰 IBM POWER8 processor-based systems
򐂰 IBM POWER9 processor-based systems
򐂰 IBM POWER10 processor-based systems

The IBM System Storage 7226 Multi-Media Enclosure offers the drive feature options that are
listed in Table 3-36.

Table 3-36 Supported drive features for the 7226-1U3


Feature code Description

1420 DVD-RAM SATA/SAS optical drive

1422 Extra DVD-RAM SATA/SAS optical drive for #1420

5762 DVD-RAM USB optical drive

5757 Extra DVD-RAM USB optical drive for #5762

8441 LTO Ultrium 7 half height SAS tape drive

8446 LTO Ultrium 7 half height Fibre Channel tape drive

8541 LTO Ultrium 8 half height SAS tape drive

8546 LTO Ultrium 8 half height Fibre Channel tape drive

8641 LTO Ultrium 9 half height SAS tape drive

8646 LTO Ultrium 9 half height Fibre Channel tape drive

EU03 RDX 3.0 Removable Disk Docking Station

The following options are available:


򐂰 The LTO Ultrium 7 drive offers a data rate of up to 300 MBps with compression. It also
provides read/write compatibility with Ultrium 7 and Ultrium 6 media formats, and
read-only compatibility with Ultrium 5 media formats. By using data compression, an
LTO-7 cartridge can store up to 15 TB of data.
򐂰 The LTO Ultrium 8 drive offers a data rate of up to 750 MBps with compression. It also
provides read/write compatibility with Ultrium 8 and Ultrium 7 media formats. It is not read
or write compatible with other Ultrium media formats. By using data compression, an
LTO-8 cartridge can store up to 30 TB of data.

Chapter 3. Available features and options 153


򐂰 The LTO Ultrium 9 drive offers a data rate of up to 1000 MBps with compression. It also
provides read/write compatibility with Ultrium 9 and Ultrium 8 media formats. It is not read
or write compatible with other Ultrium media formats. By using data compression, an
LTO-9 cartridge can store up to 45 TB of data.
򐂰 DVD-RAM: The 9.4 GB SAS Slim Optical Drive with a SATA/SAS or USB interface option
is compatible with most standard DVD disks.
򐂰 RDX removable disk drives: The RDX USB docking station is compatible with most RDX
removable disk drive cartridges when it is used in the same operating system. The 7226
offers the following RDX removable drive capacity options:
– 500 GB (#1107)
– 1.0 TB (#EU01)
– 2.0 TB (#EU2T)

Removable RDX drives are in a rugged cartridge that inserts in to an RDX removable (USB)
disk docking station (#EU03). RDX drives are compatible with docking stations, which are
installed internally in Power8, Power9, and Power10 processor-based servers (where
applicable) or the IBM System Storage 7226 Multi-Media Enclosure (7226-1U3).

Figure 3-14 shows the IBM System Storage 7226 Multi-Media Enclosure with a single RDX
docking station and two DVD-RAM devices installed.

Figure 3-14 IBM System Storage 7226 Multi-Media Enclosure

The IBM System Storage 7226 Multi-Media Enclosure offers a customer-replaceable unit
(CRU) maintenance service to help make the installation or replacement of new drives
efficient. Other 7226 components also are designed for CRU maintenance.

The IBM System Storage 7226 Multi-Media Enclosure is compatible with most Power8,
Power9, and Power10 processor-based systems that offer current level AIX, IBM i, and Linux
operating systems.

Unsupported: IBM i does not support 7226 USB devices.

For a complete list of host software versions and release levels that support the IBM System
Storage 7226 Multi-Media Enclosure, see System Storage Interoperation Center (SSIC).

154 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Flat panel display options
The IBM 7316 Model TF5 is a rack-mountable flat panel console kit that also can be
configured with the tray pulled forward and the monitor folded up, which provides full viewing
and keying capability for the HMC operator.

The Model TF5 is a follow-on product to the Model TF4 and offers the following features:
򐂰 A slim, sleek, and lightweight monitor design that occupies only 1U (1.75 in.) in a 19-inch
standard rack
򐂰 A 18.5-inch (409.8 mm x 230.4 mm) flat panel TFT monitor with truly accurate images and
virtually no distortion
򐂰 The ability to mount the IBM Travel Keyboard in the 7316-TF5 rack keyboard tray

3.10.5 Original equipment manufacturer racks


The system can be installed in a suitable OEM rack if that the rack conforms to the EIA-310-D
standard for 19-inch racks. This standard is published by the Electrical Industries Alliance. For
more information, see this IBM Documentation web page.

The IBM Documentation provides the general rack specifications, including the following
information:
򐂰 The rack or cabinet must meet the EIA Standard EIA-310-D for 19-inch racks that was
published 24 August 1992. The EIA-310-D standard specifies internal dimensions; for
example, the width of the rack opening (width of the chassis), the width of the module
mounting flanges, and the mounting hole spacing.
򐂰 The front rack opening must be a minimum of 450 mm (17.72 in.) wide, and the
rail-mounting holes must be 465 mm +/- 1.6 mm (18.3 in. +/- 0.06 in.) apart on center
(horizontal width between vertical columns of holes on the two front-mounting flanges and
on the two rear-mounting flanges).
Figure 3-15 is a top view showing the rack specification dimensions.

Figure 3-15 Rack specifications (top-down view)

Chapter 3. Available features and options 155


򐂰 The vertical distance between mounting holes must consist of sets of three holes that are
spaced (from bottom to top) 15.9 mm (0.625 in.), 15.9 mm (0.625 in.), and 12.7 mm
(0.5 in.) on center, which makes each three-hole set of vertical hole spacing 44.45 mm
(1.75 in.) apart on center.
Figure 3-16 shows the vertical distances between the mounting holes.

Figure 3-16 Vertical distances between mounting holes

򐂰 The following rack hole sizes are supported for racks where IBM hardware is mounted:
– 7.1 mm (0.28 in.) plus or minus 0.1 mm (round)
– 9.5 mm (0.37 in.) plus or minus 0.1 mm (square)

The rack or cabinet must support an average load of 20 kg (44 lb.) of product weight per EIA
unit. For example, a four EIA drawer has a maximum drawer weight of 80 kg (176 lb.).

156 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
4

Chapter 4. Enterprise solutions


In this chapter, we describe the major pillars that can help enterprises achieve their business
goals and the reasons why Power10 processor-based scale-out servers provide a significant
contribution to that end.

This chapter includes the following topics:


򐂰 4.1, “PowerVM virtualization” on page 158
򐂰 4.2, “IBM PowerVC overview” on page 167
򐂰 4.3, “Digital transformation and IT modernization” on page 169
򐂰 4.4, “Protect trust from core to cloud” on page 174
򐂰 4.5, “Running artificial intelligence where operational data is created, processed, and
stored” on page 180

© Copyright IBM Corp. 2022. 157


4.1 PowerVM virtualization
The PowerVM platform is the family of technologies, capabilities, and offerings that delivers
industry-leading virtualization for enterprises. It is the umbrella branding term for the Power
processor-based server virtualization products, including the following examples:
򐂰 IBM Power Hypervisor
򐂰 Logical partitioning
򐂰 IBM Micro-Partitioning®
򐂰 VIOS
򐂰 Live Partition Mobility (LPM)

PowerVM is a combination of hardware enablement and software.

Note: PowerVM Enterprise Edition license entitlement is included with each Power10
processor-based, scale-out server. PowerVM Enterprise Edition is available as a hardware
feature (#5228) and supports up to 20 partitions per core, VIOS, and multiple shared
processor pools (SPPs), and offers LPM.

4.1.1 IBM Power Hypervisor


Power processor-based servers are combined with PowerVM technology and offer the
following key capabilities that can help to consolidate and simplify IT environments:
򐂰 Improve server usage and share I/O resources to reduce the total cost of ownership
(TCO) and better use IT assets.
򐂰 Improve business responsiveness and operational speed by dynamically reallocating
resources to applications as needed to better match changing business needs or handle
unexpected changes in demand.
򐂰 Simplify IT infrastructure management by making workloads independent of hardware
resources so that business-driven policies can be used to deliver resources that are based
on time, cost, and service-level requirements.

Combined with features in the Power10 processor-based scale-out servers, the IBM Power
Hypervisor delivers functions that enable other system technologies, including the following
examples:
򐂰 Logical partitioning (LPAR)
򐂰 Virtualized processors
򐂰 IEEE virtual local area network (VLAN)-compatible virtual switches
򐂰 Virtual SCSI adapters
򐂰 Virtual Fibre Channel adapters
򐂰 Virtual consoles

The Power Hypervisor is a basic component of the system’s firmware and offers the following
functions:
򐂰 Provides an abstraction between the physical hardware resources and the LPARs that use
them.
򐂰 Enforces partition integrity by providing a security layer between LPARs.
򐂰 Controls the dispatch of virtual processors to physical processors.
򐂰 Saves and restores all processor state information during a logical processor context
switch.

158 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
򐂰 Controls hardware I/O interrupt management facilities for LPARs.
򐂰 Provides VLAN channels between LPARs that help reduce the need for physical Ethernet
adapters for inter-partition communication.
򐂰 Monitors the enterprise baseboard management controller (eBMC) or the flexible service
processor (FSP) of the system and performs a reset or reload if it detects the loss of one
of the eBMC or FSP controllers, and notifies the operating system if the problem is not
corrected.

The Power Hypervisor is always active, regardless of the system configuration or whether it is
connected to the managed console. It requires memory to support the resource assignment
of the LPARs on the server.

The amount of memory that is required by the Power Hypervisor firmware varies according to
the following memory usage factors:
򐂰 For hardware page tables (HPTs)
򐂰 To support I/O devices
򐂰 For virtualization

Memory usage for hardware page tables


Each partition on the system includes its own HPT that contributes to hypervisor memory
usage. The HPT is used by the operating system to translate from effective addresses to
physical real addresses in the hardware. This translation allows multiple operating systems to
run simultaneously in their own logical address space.

Whenever a virtual processor for a partition is dispatched on a physical processor, the


hypervisor indicates to the hardware the location of the partition HPT that can be used when
translating addresses.

The amount of memory for the HPT is based on the maximum memory size of the partition
and the HPT ratio. The default HPT ratio is 1/128th (for AIX, Virtual I/O Server [VIOS], and
Linux partitions) of the maximum memory size of the partition. AIX, VIOS, and Linux use
larger page sizes (16 and 64 KB) instead of the use of 4 KB pages.

The use of larger page sizes reduces the overall number of pages that must be tracked;
therefore, the overall size of the HPT can be reduced. For example, the HPT is 2 GB for an
AIX partition with a maximum memory size of 256 GB.

When defining a partition, the maximum memory size that is specified is based on the amount
of memory that can be dynamically added to the dynamic partition (DLPAR) without changing
the configuration and restarting the partition.

In addition to setting the maximum memory size, the HPT ratio can be configured. The
hpt_ratio parameter for the chsyscfg Hardware Management Console (HMC) command can
be issued to define the HPT ratio that is used for a partition profile. The following values are
valid:
򐂰 1:32
򐂰 1:64
򐂰 1:128
򐂰 1:256
򐂰 1:512

Chapter 4. Enterprise solutions 159


Specifying a smaller absolute ratio (1/512 is the smallest value) decreases the overall
memory that is assigned to the HPT. Testing is required when changing the HPT ratio
because a smaller HPT might incur more CPU consumption because the operating system
might need to reload the entries in the HPT more frequently. Most customers choose to use
the IBM provided default values for the HPT ratios.

Memory usage for I/O devices


In support of I/O operations, the hypervisor maintains structures that are called the
translation control entities (TCEs), which provide an information path between I/O devices
and partitions.

The TCEs also provide the address of the I/O buffer, which is an indication of read versus
write requests, and other I/O-related attributes. Many TCEs are used per I/O device, so
multiple requests can be active simultaneously to the same physical device. To provide better
affinity, the TCEs are spread across multiple processor chips or drawers to improve
performance while accessing the TCEs.

For physical I/O devices, the base amount of space for the TCEs is defined by the hypervisor
that is based on the number of I/O devices that are supported. A system that supports
high-speed adapters also can be configured to allocate more memory to improve I/O
performance. Linux is the only operating system that uses these extra TCEs so that the
memory can be freed for use by partitions if the system uses only AIX or IBM i operating
systems.

Memory usage for virtualization features


Virtualization requires more memory to be allocated by the Power Hypervisor for hardware
statesave areas and various virtualization technologies. For example, on Power10
processor-based systems, each processor core supports up to eight simultaneous
multithreading (SMT) threads of execution, and each thread contains over 80 different
registers.

The Power Hypervisor must set aside save areas for the register contents for the maximum
number of virtual processors that are configured. The greater the number of physical
hardware devices, the greater the number of virtual devices, the greater the amount of
virtualization, and the more hypervisor memory is required.

For efficient memory consumption, wanted and maximum values for various attributes
(processors, memory, and virtual adapters) must be based on business needs, and not set to
values that are significantly higher than requirements.

Predicting memory that is used by the Power Hypervisor


The IBM System Planning Tool (SPT) can be used to estimate the amount of hypervisor
memory that is required for a specific server configuration. After the SPT executable file is
downloaded and installed, you can define a configuration by selecting the correct hardware
platform and the installed processors and memory, and defining partitions and partition
attributes. SPT can estimate the amount of memory that is assigned to the hypervisor, which
assists you when you change a configuration or deploy new servers.

The Power Hypervisor provides the following types of virtual I/O adapters:
򐂰 Virtual SCSI
The Power Hypervisor provides a virtual SCSI mechanism for the virtualization of storage
devices. The storage virtualization is accomplished by using two paired adapters: a virtual
SCSI server adapter and a virtual SCSI customer adapter.

160 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
򐂰 Virtual Ethernet
The Power Hypervisor provides a virtual Ethernet switch function that allows partitions fast
and secure communication on the same server without any need for physical
interconnection or connectivity outside of the server if a Layer 2 bridge to a physical
Ethernet adapter is set in one VIOS partition, also known as Shared Ethernet Adapter
(SEA).
򐂰 Virtual Fibre Channel
A virtual Fibre Channel adapter is a virtual adapter that provides customer LPARs with a
Fibre Channel connection to a storage area network through the VIOS partition. The VIOS
partition provides the connection between the virtual Fibre Channel adapters on the VIOS
partition and the physical Fibre Channel adapters on the managed system.
򐂰 Virtual (tty) console
Each partition must have access to a system console. Tasks, such as operating system
installation, network setup, and various problem analysis activities, require a dedicated
system console. The Power Hypervisor provides the virtual console by using a virtual tty
and a set of hypervisor calls to operate on them. Virtual tty does not require the purchase
of any other features or software, such as the PowerVM Edition features.

Logical partitions
LPARs and the use of virtualization increase the usage of system resources while adding a
level of configuration possibilities.

Logical partitioning is the ability to make a server run as though it were two or more
independent servers. When you logically partition a server, you divide the resources on the
server into subsets, called LPARs. You can install software on an LPAR, and the LPAR runs
as an independent logical server with the resources that you allocated to the LPAR.

LPARs are also referred to in some documentation as virtual machines (VMs), which make
them appear to be similar to what other hypervisors offer. However, LPARs provide a higher
level of security and isolation and other features.

Processors, memory, and I/O devices can be assigned to LPARs. AIX, IBM i, Linux, and VIOS
can run on LPARs. VIOS provides virtual I/O resources to other LPARs with general-purpose
operating systems.

LPARs share a few system attributes, such as the system serial number, system model, and
processor FCs. All other system attributes can vary from one LPAR to another.

Micro-Partitioning
When you use the Micro-Partitioning technology, you can allocate fractions of processors to
an LPAR. An LPAR that uses fractions of processors is also known as a shared processor
partition or micropartition. Micropartitions run over a set of processors that is called a shared
processor pool (SPP), and virtual processors are used to enable the operating system to
manage the fractions of processing power that are assigned to the LPAR.

From an operating system perspective, a virtual processor cannot be distinguished from a


physical processor, unless the operating system is enhanced to determine the difference.
Physical processors are abstracted into virtual processors that are available to partitions.

Chapter 4. Enterprise solutions 161


On the Power10 processor-based server, a partition can be defined with a processor capacity
as small as 0.05 processing units. This number represents 0.05 of a physical core. Each
physical core can be shared by up to 20 shared processor partitions, and the partition’s
entitlement can be incremented fractionally by as little as 0.05 of the processor.

The shared processor partitions are dispatched and time-sliced on the physical processors
under the control of the Power Hypervisor. The shared processor partitions are created and
managed by the HMC.

Note: Although the Power10 processor-based scale-out server supports up to 20 shared


processor partitions per core, the real limit depends on application workload demands that
are used on the server.

Processing mode
When you create an LPAR, you can assign entire processors for dedicated use, or you can
assign partial processing units from an SPP. This setting defines the processing mode of the
LPAR.

Dedicated mode
In dedicated mode, physical processors are assigned as a whole to partitions. The SMT
feature in the Power10 processor core allows the core to run instructions from one, two, four,
or eight independent software threads simultaneously.

Shared dedicated mode


On Power10 processor-based servers, you can configure dedicated partitions to become
processor donors for idle processors that they own, which allows for the donation of spare
CPU cycles from dedicated processor partitions to an SPP.

The dedicated partition maintains absolute priority for dedicated CPU cycles. Enabling this
feature can help increase system usage without compromising the computing power for
critical workloads in a dedicated processor mode LPAR.

Shared mode
In shared mode, LPARs use virtual processors to access fractions of physical processors.
Shared partitions can define any number of virtual processors (the maximum number is 20
times the number of processing units that are assigned to the partition).

The Power Hypervisor dispatches virtual processors to physical processors according to the
partition’s processing units entitlement. One processing unit represents one physical
processor’s processing capacity. All partitions receive a total CPU time equal to their
processing unit’s entitlement.

The logical processors are defined on top of virtual processors. Therefore, even with a virtual
processor, the concept of a logical processor exists, and the number of logical processors
depends on whether SMT is turned on or off.

4.1.2 Multiple shared processor pools


Multiple SPPs are supported on Power10 processor-based servers. This capability allows a
system administrator to create a set of micropartitions with the purpose of controlling the
processor capacity that can be used from the physical SPP.

162 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Micropartitions are created and then identified as members of the default processor pool or a
user-defined SPP. The virtual processors that exist within the set of micropartitions are
monitored by the Power Hypervisor. Processor capacity is managed according to
user-defined attributes.

If the Power server is under heavy load, each micropartition within an SPP is assured of its
processor entitlement, plus any capacity that might be allocated from the reserved pool
capacity if the micropartition is uncapped.

If specific micropartitions in an SPP do not use their processing capacity entitlement, the
unused capacity is ceded, and other uncapped micropartitions within the same SPP can use
the extra capacity according to their uncapped weighting. In this way, the entitled pool
capacity of an SPP is distributed to the set of micropartitions within that SPP.

All Power servers that support the multiple SPP capability have a minimum of one (the
default) SPP and up to a maximum of 64 SPPs.

This capability helps customers reduce the TCO significantly when the costs of software or
database licenses depend on the number of assigned processor-cores.

4.1.3 Virtual I/O Server


The VIOS is part of PowerVM. It is the specific appliance that allows the sharing of physical
resources among LPARs to allow more efficient usage (for example, consolidation). In this
case, the VIOS owns the physical I/O resources (SCSI, Fibre Channel, network adapters, or
optical devices) and allows client partitions to share access to them, which minimizes and
optimizes the number of physical adapters in the system.

The VIOS eliminates the requirement that every partition owns a dedicated network adapter,
disk adapter, and disk drive. The VIOS supports OpenSSH for secure remote logins. It also
provides a firewall for limiting access by ports, network services, and IP addresses.

Figure 4-1 shows an overview of a VIOS configuration.

Figure 4-1 Architectural view of the VIOS

It is a preferred practice to run dual VIO servers per physical server to allow for redundancy of
all I/O paths for client LPARs.

Chapter 4. Enterprise solutions 163


Shared Ethernet Adapter
A SEA can be used to connect a physical Ethernet network to a virtual Ethernet network. The
SEA provides this access by connecting the Power Hypervisor VLANs to the VLANs that are
on the external switches.

Because the SEA processes packets at Layer 2, the original MAC address and VLAN tags of
the packet are visible to other systems on the physical network. IEEE 802.1 VLAN tagging is
supported.

By using the SEA, several customer partitions can share one physical adapter. You also can
connect internal and external VLANs by using a physical adapter. The SEA service can be
hosted only in the VIOS (not in a general-purpose AIX or Linux partition) and acts as a
Layer 2 network bridge to securely transport network traffic between virtual Ethernet
networks (internal) and one or more (Etherchannel) physical network adapters (external).
These virtual Ethernet network adapters are defined by the Power Hypervisor on the VIOS.

Virtual SCSI
Virtual SCSI is used to view a virtualized implementation of the SCSI protocol. Virtual SCSI is
based on a customer/server relationship. The VIOS LPAR owns the physical I/O resources
and acts as a server or in SCSI terms, a target device. The client LPARs access the virtual
SCSI backing storage devices that are provided by the VIOS as clients.

The virtual I/O adapters (a virtual SCSI server adapter and a virtual SCSI client adapter) are
configured by using an HMC. The virtual SCSI server (target) adapter is responsible for
running any SCSI commands that it receives, and is owned by the VIOS partition. The virtual
SCSI client adapter allows a client partition to access physical SCSI and SAN-attached
devices and LUNs that are mapped to be used by the client partitions. The provisioning of
virtual disk resources is provided by the VIOS.

N_Port ID Virtualization
N_Port ID Virtualization (NPIV) is a technology that allows multiple LPARs to access one or
more external physical storage devices through the same physical Fibre Channel adapter.
This adapter is attached to a VIOS partition that acts only as a pass-through that manages
the data transfer through the Power Hypervisor.

Each partition features one or more virtual Fibre Channel adapters, each with their own pair
of unique worldwide port names. This configuration enables you to connect each partition to
independent physical storage on a SAN. Unlike virtual SCSI, only the client partitions see the
disk.

For more information and requirements for NPIV, see IBM PowerVM Virtualization Managing
and Monitoring, SG24-7590.

4.1.4 Live Partition Mobility


LPM allows you to move a running LPAR from one system to another without disruption.
Inactive partition mobility allows you to move a powered-off LPAR from one system to another
one.

LPM provides systems management flexibility and improves system availability by avoiding
the following situations:
򐂰 Planned outages for hardware upgrade or firmware maintenance.
򐂰 Unplanned downtime. With preventive failure management, if a server indicates a potential
failure, you can move its LPARs to another server before the failure occurs.

164 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
For more information and requirements for LPM, see IBM PowerVM Live Partition Mobility,
SG24-7460.

HMCV10.1.1020.0 and VIOS 3.1.3.21 or later provide the following enhancements to the
LPM feature:
򐂰 Automatically choose fastest network for LPM memory transfer
򐂰 Allow LPM when a virtual optical device is assigned to a partition

4.1.5 Active Memory Mirroring


Active Memory Mirroring (AMM) for Hypervisor is available as an option (#EM8G) to enhance
resilience by mirroring critical memory that is used by the PowerVM hypervisor so that it can
continue operating if a memory failure occurs.

A portion of available memory can be proactively partitioned such that a duplicate set can be
used upon noncorrectable memory errors. This feature can be implemented at the granularity
of DDIMMs or logical memory blocks.

4.1.6 Remote Restart


Remote Restart is a high availability option for partitions. If an error occurs that causes a
server outage, a partition that is configured for Remote Restart can be restarted on a different
physical server. At times, it might take longer to start the server, in which case the Remote
Restart function can be used for faster reprovisioning of the partition. Typically, this process
can be done faster than restarting the server that stopped and then restarting the partitions.

The Remote Restart function relies on technology that is similar to LPM where a partition is
configured with storage on a SAN that is shared (accessible) by the server that hosts the
partition.

HMC V10R1 provides an enhancement to the Remote Restart Feature that enables remote
restart when a virtual optical device is assigned to a partition.

4.1.7 POWER processor modes


Although they are not virtualization features, the POWER processor modes are described
here because they affect various virtualization features.

On Power servers, partitions can be configured to run in several modes, including the
following modes:
򐂰 POWER8
This native mode for Power8 processors implements Version 2.07 of the IBM Power ISA.
For more information, see this IBM Documentation web page.
򐂰 POWER9
This native mode for Power9 processors implements Version 3.0 of the IBM Power ISA.
For more information, see this IBM Documentation web page.
򐂰 Power10
This native mode for Power10 processors implements Version 3.1 of the IBM Power ISA.
For more information, see this IBM Documentation web page.

Chapter 4. Enterprise solutions 165


Figure 4-2 shows the available processor modes on a Power10 processor-based scale-out
server.

Figure 4-2 Processor modes

Processor compatibility mode is important when LPM migration is planned between different
generations of server. An LPAR that might be migrated to a machine that is managed by a
processor from another generation must be activated in a specific compatibility mode.

Note: Migrating an LPAR from a POWER7 processor-based server to a Power10


processor-based scale-out server by using LPM is not supported; however, the following
steps can be completed to accomplish this task:
1. Migrate LPAR from POWER7 processor-based server to Power8 or Power9
processor-based server by using LPM.
2. Migrate then the LPAR from the Power8 or Power9 processor-based server to a
Power10 processor-based scale-out server.

The operating system that is running on the POWER7 processor-based server must be
supported on Power10 processor-based scale-out server, or must be upgraded to a
supported level before starting the above steps.

166 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
4.1.8 Single Root I/O virtualization
Single Root I/O Virtualization (SR-IOV) is an extension to the Peripheral Component
Interconnect Express (PCIe) specification that allows multiple operating systems to
simultaneously share a PCIe adapter with little or no runtime involvement from a hypervisor or
other virtualization intermediary.

SR-IOV is a PCI standard architecture that enables PCIe adapters to become self-virtualizing.
It enables adapter consolidation through sharing, much like logical partitioning enables server
consolidation. With an adapter capable of SR-IOV, you can assign virtual slices of a single
physical adapter to multiple partitions through logical ports;, which is done without a VIOS.

For more information about the virtualization features, see the following publications:
򐂰 IBM PowerVM Best Practices, SG24-8062
򐂰 IBM PowerVM Virtualization Introduction and Configuration, SG24-7940
򐂰 IBM PowerVM Virtualization Managing and Monitoring, SG24-7590
򐂰 IBM Power Systems SR-IOV: Technical Overview and Introduction, REDP-5065

4.2 IBM PowerVC overview


IBM Power Virtualization Center (PowerVC) is an advanced virtualization and cloud
management offering that provides simplified virtualization management and cloud
deployments for IBM AIX, IBM i, and Linux VMs, including Red Hat OpenShift (CoreOS) that
is running on IBM Power. It is built on the Open Source OpenStack project to ensure
consistent APIs and interfaces with other cloud management tools. PowerVC is designed to
improve administrator productivity and to simplify the cloud management of VMs on Power
servers.

By using PowerVC, the following tasks can be performed:


򐂰 Create VMs and resize the VMs CPU and memory.
򐂰 Attach disk volumes or other networks to those VMs.
򐂰 Import VMs and volumes so that they can be managed by IBM PowerVC.
򐂰 Deploy new VMs with storage and network from an image in a few minutes.
򐂰 Monitor the use of resources in your environment.
򐂰 Take snapshots of a VM or clone a VM.
򐂰 Migrate VMs while they are running (live migration between physical servers).
򐂰 Automated Remote restart VMs if a server failure occurs.
򐂰 Automatically balance cloud workloads by using the Dynamic Resource Optimizer (DRO).
򐂰 Use advanced storage technologies, such as VDisk mirroring, IBM HyperSwap, and IBM
Global mirror.
򐂰 Put a server into maintenance mode with automatic distribution of LPARs to other server
and back using LPM.
򐂰 Create a private cloud with different projects or tenants that are independent from each
other but use the same resources.
򐂰 Create a self-service portal with an approval workflow.
򐂰 Meter resource usage as basis for cost allocation.

Chapter 4. Enterprise solutions 167


The use of PowerVC management tools in a Power environment includes the following
benefits:
򐂰 Improve resource usage to reduce capital expense and power consumption.
򐂰 Increase agility and execution to respond quickly to changing business requirements.
򐂰 Increase IT productivity and responsiveness.
򐂰 Simplify Power virtualization management.
򐂰 Accelerate repeatable, error-free virtualization deployments.

IBM PowerVC can manage AIX, IBM i, and Linux-based VMs that are running under
PowerVM virtualization and are connected to an HMC or by using NovaLink. As of this writing,
the release supports the scale-out and the enterprise Power servers that are built on IBM
Power8, IBM Power9, and Power10.

Note: The Power S1014, S1022s, S1022, and S1024 servers are supported by PowerVC
2.0.3 or later. More fix packs might be required. For more information, see this IBM
Support Fix Central web page.

4.2.1 IBM PowerVC functions and advantages


When more than 70% of IT budgets are spent on operations and maintenance, IT customers
expect vendors to focus their new development efforts to reduce IT costs and foster
innovation within IT departments.

IBM PowerVC gives IBM Power customers the following advantages:


򐂰 It is deeply integrated with IBM Power hardware and software.
򐂰 It provides virtualization management tools.
򐂰 It eases the integration of servers that are managed by PowerVM in automated IT
environments, such as clouds.
򐂰 It is a building block of IBM Infrastructure as a Service (IaaS), based on Power.
򐂰 PowerVC integrates with other cloud management tools, such as Ansible, Terraform, or
Red Hat OpenShift, and can be integrated into orchestration tools, such as VMware
vRealize, or the SAP Landscape Management (LaMa).
򐂰 PowerVC also provides an easy exchange of VM images between private and
public clouds by using the integration of Cloud Object Storage into PowerVC.

IBM PowerVC is an addition to the PowerVM set of enterprise virtualization technologies that
provide virtualization management. It is based on open standards and integrates server
management with storage and network management.

Because IBM PowerVC is based on the OpenStack initiative, Power can be managed by tools
that are compatible with OpenStack standards. When a system is controlled by IBM
PowerVC, it can be managed in one of three ways:
򐂰 By a system administrator that uses the IBM PowerVC graphical user interface (GUI)
򐂰 By a system administrator that uses scripts that contain the IBM PowerVC
Representational State Transfer (REST) APIs
򐂰 By higher-level tools that call IBM PowerVC by using standard OpenStack API

The following PowerVC offerings are available:


򐂰 IBM PowerVC
򐂰 IBM PowerVC for Private Cloud

168 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
The PowerVC for Private Cloud edition adds a self-service portal with which users can deploy
and manage their own LPARs and workloads, and offers further cloud management functions.
These functions include more project level metering, approval flows, and notification
capabilities.

PowerVC provides a systems management product that enterprise customers require to


effectively manage the advanced features that are offered by IBM premium hardware. It
reduces resource use and manages workloads for performance and availability.

For more information about PowerVC, see IBM PowerVC Version 2.0 Introduction and
Configuration, SG24-8477.

4.3 Digital transformation and IT modernization


Digital transformation is the process of the use of digital technologies to create business
processes, or modify processes. It also means changing the culture and experiences to meet
the changing needs of the business and the market. This reimagining of business in the digital
age is at the heart of digital transformation.

In the IBM vision, digital transformation takes a customer-centric and digital-centric approach
to all aspects of business: from business models to customer experiences, processes and
operations. It uses artificial intelligence, automation, hybrid cloud, and other digital
technologies to use data and drive intelligent workflows, enable faster and smarter decision
making, and a real-time response to market disruptions. Ultimately, it changes customer
expectations and creates business opportunities.

To date, one of the main technologies that is recognized as an enabler to digital


transformation and as the path to application modernization, is the Red Hat OpenShift
Container Platform. Associating OpenShift Container Platform with the IBM Power
processor-based platform realizes several advantages, which are described next.

4.3.1 Application and services modernization


Modernizing established and cloud native applications and services by using containers that
are deployed by way of a platform that is designed with Kubernetes orchestration at its core is
arguably the most forward-looking strategy from which any business benefits. Unsurprisingly,
it is also at the heart of the IBM and Red Hat strategy for the hybrid multicloud reality of the
digital landscape of tomorrow.

Red Hat OpenShift Container Platfom is a container orchestration platform that is based on
Kubernetes that helps develop containerized applications with open source technology that is
ready for the Enterprise. Red Hat OpenShift Container Platfom facilitates management and
deployments in hybrid and multicloud environments by using full-stack automated operations.

Containers are key elements of the IT transformation journey toward modernization.


Containers are software executable units in which the application code is packaged (along
with its libraries and dependencies) in common and consistent ways so that it can be run
anywhere on desktop computers and any type of server or on the cloud.

To do this, containers take advantage of a form of operating system virtualization, in which


operating system functions are used effectively to isolate processes and control the amount
of CPU, memory, and disk to which those processes can access.

Chapter 4. Enterprise solutions 169


Containers are small, fast, and portable because, unlike a VM, they do not need to include a
guest operating system in every instance. Instead, they can instead use the functions and
resources of the host operating system.

Containers first appeared decades ago with releases, such as FreeBSD jails and AIX
Workload Partitions (WPARs). However, most modern developers remember 2013 as the
beginning of the modern container era with the introduction of Docker.

One way to better understand a container is to understand how it differs from a traditional VM.
In traditional virtualization (on-premises and in the cloud), a hypervisor is used to virtualize
the physical hardware. Therefore, each VM contains a guest operating system and a virtual
copy of the hardware that the operating system requires to run, with an application and its
associated libraries and dependencies.

Instead of virtualizing the underlying hardware, containers virtualize the operating system
(usually Linux) so that each individual container includes only the application and its libraries
and dependencies. The absence of the guest operating system is the reason why containers
are so light and therefore, fast and portable.

In addition to AIX WPARs, IBM i projects came from the 1980s. The IBM i team devised an
approach to create a container for objects (that is, programs, databases, security objects, and
so on). This container can be converted into an image that can be transported from a
development environment to a test environment, another system, or the cloud. A significant
difference between this version of containers and the containers that we know today is the
name: on IBM i they are called libraries and a container image is called a backup file.

The IBM Power platform delivers a high container density per core, with multiple CPU threads
to enable higher throughput. By using PowerVM virtualization, cloud native applications also
can be colocated alongside applications that are related to AIX or IBM i worlds. This ability
makes available API connections to business-critical data for higher bandwidth and lower
latency than other technologies.

Only with IBM Power can you have a flexible and efficient use of resources, manage peaks,
and support traditional and modern workloads with the capabilities of capacity on demand or
shared processor pools. Hardware is not just a commodity; rather, it must be carefully
evaluated.

4.3.2 System automation with Ansible


Enterprises can spend large amounts of IT administrative time performing repetitive tasks and
running manual processes. Tasks, such as updating, patching, compliance checks,
provisioning new VMs or LPARs, and ensuring that the correct security updates are in place,
are taking time away from more valuable business activities.

The ability to automate by using Red Hat Ansible returns valuable time to the system
administrators.

The Red Hat Ansible Automation Platform for Power is fully enabled, so enterprises can
automate several tasks within AIX, IBM i, and Linux all the way up to provisioning VMs and
deploying applications. Ansible also can be combined with HMC, PowerVC, and Power Virtual
Server to provision infrastructure anywhere you need, including cloud solutions from other
IBM Business Partners or third-party providers that are based on Power processor-based
servers.

170 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
A first task after the initial installation or set-up of a new LPAR is to ensure that the correct
patches are installed. Also, extra software (whether it is open source software, ISV software,
or perhaps their own enterprise software) must be installed. Ansible features a set of
capabilities to roll out new software, which makes it popular in Continuous
Integration/Continuous Delivery (CI/CD) pipeline environments. Orchestration and integration
of automation with security products represent other ways in which Ansible can be applied
within the data center.

Despite the wide adoption of AIX and IBM i in many different business sectors by different
types of customers, Ansible can help introduce the Power processor-based technology to
customers who believe that AIX and IBM i skills are a rare commodity that is difficult to locate
in the marketplace, but want to take advantage of all the features of the hardware platform.
The Ansible experience is identical across Power or x86 processor-based technology and the
same tools can be used in IBM Cloud and other cloud providers.

AIX and IBM i skilled customers can also benefit from the extreme automation solutions that
are provided by Ansible.

The Power processor-based architecture features unique advantages over commodity server
platforms, such as x86, because the engineering teams that are working on the processor,
system boards, virtualization. and management appliances collaborate closely to ensure an
integrated stack that works seamlessly. This approach is in stark contrast to the multivendor
x86 processor-based technology approach, in which the processor, server, management, and
virtualization must be purchased from different (and sometimes competing) vendors.

The Power stack engineering teams partnered closely to deliver the enterprise server
platform, which results in an IT architecture with industry-leading performance, scalability, and
security (see Figure 4-3).

Figure 4-3 Power stack

Every layer in the Power stack is optimized to make the Power10 processor-based technology
the platform of choice for mission-critical enterprise workloads. This stack includes the
Ansible Automation Platform, which is described next.

Chapter 4. Enterprise solutions 171


Ansible Automation Platform
The Ansible Automation Platform integrates with IBM Power processor-based technology,
which is included in the Certified Integrations section of the Red Hat Ansible website.

Many Ansible collections are available for IBM Power processor-based technology, which (at
the time of this writing) were downloaded more than 25,000 times by customers, are now
included in the Red Hat Ansible Automation Platform. As a result, these modules are covered
by Red Hat’s 24x7 enterprise support team, which collaborates with the respective Power
processor-based technology development teams.

IBM Power in the Ansible ecosystem


A series of Ansible collections is available for the Power processor-based platform. A
collection is a group of modules, Playbooks, and roles. Embracing Ansible in the Power
community, our AIX and IBM i community feature a comprehensive set of modules available.
Some examples are development of tools to manage AIX logical volumes, which put module
interfaces over the installation key command, or managing the AIX init tab entries.

From an IBM i perspective, a pertinent example is the ability to run SQL queries against the
integrated IBM Db2 database that is built into the IBM i platform, manage object authorities,
and so on. All of these modules and playbooks can be combined by an AIX administrator or
IBM i administrator to perform complex tasks rapidly and consistently.

The IBM operating system development teams, alongside community contributors, develop
modules that are sent to an open source community (named Ansible Galaxy). Every
developer can post any object that can be a candidate for a collection in the open Ansible
Galaxy community and possibly certified to be supported by IBM with a subscription to Red
Hat Ansible Automation Platform (see Figure 4-4).

Figure 4-4 Power content in the Ansible ecosystem

Ansible modules for AIX


As part of the broader offering of Ansible Content for IBM Power, the IBM Power - AIX
collection is available from Ansible Galaxy and has community support.

The collection includes modules and sample playbooks that help to automate tasks and is
available at this web page.

172 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Ansible modules for IBM i
Ansible Content for IBM Power - IBM i provides modules, action plug-ins, roles, and sample
playbooks to automate tasks on IBM i workloads, including the following examples:
򐂰 Command execution
򐂰 System and application configuration
򐂰 Work management
򐂰 Fix management
򐂰 Application deployment

For more information about the collection, see this web page.

Ansible modules for HMC


The IBM Power - HMC collection provides modules that can be used to manage
configurations and deployments of HMC systems and the servers that they manage. The
collection content helps to include workloads on Power processor-based platforms as part of
an enterprise automation strategy through the Ansible ecosystem.

For more information about this collection, see this web page.

Ansible modules for VIOS


The IBM Power - VIOS collection provides modules that can be used to manage
configurations and deployments of Power VIOS systems. The collection content helps to
include workloads on Power processor-based platforms as part of an enterprise automation
strategy through the Ansible system.

For more information, see this web page.

Ansible modules for Oracle on AIX


This repository contains a collection that can be used to install and manage Oracle single
instance database 19c on AIX operating system and creates test database on AIX file system
and on Oracle ASM. This collection automates Oracle 19c database installation and creation
steps.

For more information, see this web page.

Ansible modules for Oracle RAC on AIX


This collection provides modules that can be used to install and manage Oracle 19c RAC on
AIX operating system. This collection also automates Oracle 19c RAC installation and
creation steps.

For more information, see this web page.

Ansible modules for SAP on AIX


The Ansible Content for SAP Software on AIX provides roles to automate administrator tasks
for SAP installations on AIX, such as installing the SAP Host Agent, starting and stopping
SAP instances, and upgrading the SAP kernel.

For more information, see this web page.

Chapter 4. Enterprise solutions 173


Ansible modules for SAP on IBM i
The Ansible Content for SAP Software on IBM i provides roles to automate administrator
tasks for SAP installations on IBM i, such as installing the SAP Host Agent, starting and
stopping SAP instances, and upgrading the SAP kernel.

For more information, see this web page.

4.4 Protect trust from core to cloud


The IT industry has long relied on perimeter security strategies to protect its most valuable
assets, such as user data and intellectual property. The use of firewalls and other
network-based tools to inspect and validate users that are entering and leaving the network is
no longer enough as digital transformation and the shift to hybrid cloud infrastructure are
changing the way industries do business.

Many organizations also are adapting their business models, and have thousands of people
that are connecting from home computers that are outside the control of an IT department.
Users, data, and resources are scattered all over the world, which makes it difficult to connect
them quickly and securely. Also, without a traditional local infrastructure for security,
employees’ homes are more vulnerable to compromise, which puts the business at risk.

Many companies are operating with a set of security solutions and tools, even without them
being fully integrated or automated. As a result, security teams spend more time on manual
tasks. They lack the context and information that is needed to effectively reduce the attack
surface of their organization. Rising numbers of data breaches and increasing global
regulations make securing networks difficult.

Applications, users, and devices need fast and secure access to data, so much so that an
entire industry of security tools and architectures was created to protect them.

Although enforcing a data encryption policy is an effective way to minimize the risk of a data
breach that, in turn, minimizes costs, only a few enterprises at the worldwide level have an
encryption strategy that is applied consistently across the entire organization. This is true in
large part because such policies add complexity and cost, and negatively affect performance,
which can mean missed SLAs to the business.

The rapidly evolving cyberthreat landscape also requires focus on cyber-resilience. Persistent
and end-to-end security is the best way to reduce exposure to threats.

Prevention is the best protection


Because prevention is the best protection, Power10 processor-based servers provide
industry-leading isolation, and integrity that help prevent ransomware from being installed.
The following features and implementation help IBM Power customers to protect their
business:
򐂰 Host and firmware secure and trusted boot.
򐂰 Guest operating system secure boot (AIX now, Linux upcoming).
򐂰 Built-in operating system runtime integrity: AIX Trusted Execution and Linux IMA.
򐂰 Most secure multi-tenant environment with orders of magnitude lower Common
Vulnerabilities and Exposures (CVE) versus x86 hypervisors.

174 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
򐂰 Orders of magnitude lower operating system CVEs for AIX and IBM i.
򐂰 Simplified patching with PowerSC.
򐂰 Multi-Factor Authentication with PowerSC.

Early Detection is critical


Integrated security and compliance management with PowerSC makes it more difficult to
misconfigure systems and easier to detect anomalies. Third-party offerings, for example, IBM
Security QRadar®, enhance inherent security with early anomaly detection.

Fast and Efficient Recovery


It is easier to deploy resiliency strategies with IBM PowerHA® and IBM Storage safeguarded
copies, while also using the collaboration with IBM Storage and security services for fast
detection and automated recovery of affected data.

4.4.1 Power10 processor-based, technology-integrated security ecosystem


The IBM Power processor-based platform offers the most secure and reliable servers in their
class. Introducing Power10 processor-based technology in 2021, IBM further extended the
industry-leading security and reliability of the Power processor-based platform, with focus on
protecting applications and data across hybrid cloud environments.

Also introduced were significant innovations along the following major dimensions:
򐂰 Advanced Data Protection that offers simple to use and efficient capabilities to protect
sensitive data through mechanisms, such as encryption and multi-factor authentication.
򐂰 Platform Security ensures that the server is hardened against tampering, continuously
protecting its integrity, and ensuring strong isolation among multi-tenant workloads.
Without strong platform security, all other system security measures are at risk.
򐂰 Security Innovation for Modern Threats provides the ability to stay ahead of new types of
cybersecurity threats by using emerging technologies.
򐂰 Integrated Security Management addresses the key challenge of ensuring correct
configuration of the many security features across the stack, monitoring them, and
reacting if unexpected changes are detected.

The Power10 processor-based servers are enhanced to simplify and integrate security
management across the stack, which reduces the likelihood of administrator errors.

In the Power10 processor-based scale-out servers, all data is protected by a greatly simplified
end-to-end encryption that extends across the hybrid cloud without detectable performance
impact and prepares for future cyberthreats.

Power10 processor-core technology features built-in security integration:


򐂰 Stay ahead of current and future data threats with better cryptographic performance and
support for quantum-safe cryptography and fully homomorphic encryption (FHE).
򐂰 Enhance the security of applications with more hardened defense against return-oriented
programming (ROP) attacks.
򐂰 Simplified single-interface hybrid cloud security management without any required setup.
򐂰 Protect your applications and data with the most secure VM isolation in the industry with
orders of magnitude lower CVEs than hypervisors related to x86 processor-based servers.

Chapter 4. Enterprise solutions 175


Also, workloads on the Power10 processor-based scale-out servers benefit from
cryptographic algorithm acceleration, which allows algorithms, such as Advanced Encryption
Standard (AES), SHA2, and SHA3 to run significantly faster than Power9 processor-based
servers on a per core basis. This performance acceleration allows features, such as AIX
Logical Volume Encryption, to be enabled with low-performance overhead.

4.4.2 Cryptographic engines and transparent memory encryption


Power10 processor technology is engineered to achieve significantly faster encryption
performance in comparison to IBM Power9 processor-based servers. Power10
processor-based scale-out servers are updated for today’s most demanding encryption
standards and anticipated future cryptographic standards, such as post-quantum and FHE,
and brings new enhancements to container security.

Transparent memory encryption is designed to simplify the deployment of encryption, and


support end-to-end security without affecting performance by using hardware features for a
seamless user experience. The protection that is introduced in all layers of an infrastructure is
shown in Figure 4-5.

Figure 4-5 Protect data in memory with transparent memory encryption

4.4.3 Quantum-safe cryptography support


To be prepared for the Quantum era, the Power S1014, S1022s, S1022, and S1024 are built
to efficiently support future cryptography, such as Quantum-safe cryptography and FHE. The
software libraries for these solutions are optimized for the Power10 processor-chip instruction
set architecture (ISA) and are to be available in the respective open source communities.

Quantum-safe cryptography refers to the efforts to identify algorithms that are resistant to
attacks by classical and quantum computers in preparation for the time when large-scale
quantum computers are built.

176 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Homomorphic encryption refers to encryption techniques that permit systems to perform
computations on encrypted data without decrypting the data first. The software libraries for
these solutions are optimized for the Power processor-chip ISA.

4.4.4 IBM PCIe3 Crypto Coprocessor BSC-Gen3 4769


IBM PCIe3 Crypto Coprocessor BSC-Gen3 4769 (#EJ35 or #EJ37) provides hardware
security protection for the most sensitive secrets. As described in 3.4.5, “Cryptographic
coprocessor adapters” on page 124, this adapter supports IBM AIX, IBM i without the use of
the VIO server, or supported distribution of Linux operating systems.

It provides a comprehensive set of cryptographic functions, including the common AES,


TDES, RSA, and ECC functions for data confidentiality and data integrity support. In addition,
CCA features extensive functions for key management and many functions of special interest
to the banking and finance industry.

The coprocessor holds a security-enabled subsystem module and batteries for backup power.
The hardened encapsulated subsystem contains two sets of two 32-bit PowerPC 476FP
reduced-instruction-set-computer (RISC) processors running in lockstep with cross-checking
to detect soft errors in the hardware.

It also contains a separate service processor that is used to manage:


򐂰 Self-test and firmware updates.
򐂰 RAM, flash memory and battery-powered memory.
򐂰 Secure time-of-day.
򐂰 Cryptographic quality random number generator.
򐂰 AES, DES, triple DES, HMAC, CMAC, MD5, and multiple SHA hashing methods.
򐂰 Modular-exponentiation hardware, such as RSA and ECC.
򐂰 Full-duplex direct memory access (DMA) communications.

A security-enabled code-loading arrangement allows control program and application


program loading and refreshes after coprocessor installation in your server.

IBM offers an embedded subsystem control program and a cryptographic API that
implements the IBM Common Cryptographic Architecture (CCA) Support Program that can
be accessed from the internet at no charge to the user.

Feature #EJ35 and #EJ37 are feature codes that represent the same physical card with the
same CCIN of C0AF. Different feature codes are used to indicate whether a blind swap
cassette is used and its type: #EJ35 indicates no blind swap cassette, #EJ37 indicates a
Gen 3 blind swap cassette.

The 4769 PCIe Cryptographic Coprocessor is designed to deliver the following functions:
򐂰 X.509 certificate services support
򐂰 ANSI X9 TR34-2019 key exchange services that use the public key infrastructure (PKI)
򐂰 ECDSA secp256k1
򐂰 CRYSTALS-Dilithium, a quantum-safe algorithm for digital signature generation and
verification
򐂰 Rivest-Shamir-Adleman (RSA) algorithm for digital signature generation and verification
with keys up to 4096 bits

Chapter 4. Enterprise solutions 177


򐂰 High-throughput Secure Hash Algorithm (SHA), MD5 message digest algorithm,
Hash-Based Message Authentication Code (HMAC), Cipher-based Message
Authentication Code (CMAC), Data Encryption Standard (DES), Triple Data Encryption
Standard (Triple DES), and Advanced Encryption Standard (AES)-based encryption for
data integrity assurance and confidentiality, including AES Key Wrap (AESKW) that
conforms to ANSI X9.102.
򐂰 Elliptic-curve cryptography (ECC) for digital signature and key agreement.
򐂰 Support for smart card applications and personal identification number (PIN) processing.
򐂰 Secure time-of-day.
򐂰 Visa Data Secure Platform (DSP) point-to-point encryption (P2PE) with standard Visa
format-preserving encryption (FPE) and format-preserving, Feistel-based Format
Preserving Encryption (FF1, FF2, FF2.1). Format Preserving Counter Mode (FPCM), as
defined in x9.24 Part 2.

4.4.5 IBM PowerSC support


The Power10 processor-based scale-out servers benefit from the integrated security
management capabilities that are offered by IBM PowerSC. This Power software portfolio
manages security and compliance on every Power processor-based platform that is running
the following operating systems:
򐂰 AIX
򐂰 IBM i
򐂰 The supported distributions and versions of Linux

PowerSC is introducing more features to help customers manage security end-to-end across
the stack to stay ahead of various threats. Specifically, PowerSC 2.0 adds support for
Endpoint Detection and Response (EDR), host-based intrusion detection, block listing, and
Linux.

Security features are beneficial only if they can be easily and accurately managed. Power10
processor-based scale-out servers benefit from the integrated security management
capabilities that are offered by IBM PowerSC.

PowerSC is a key part of the Power solution stack. It provides features, such as compliance
automation, to help with various industry standards, real-time file integrity monitoring,
reporting to support security audits, patch management, trusted logging, and more.

By providing all of these capabilities within a clear and modern web-based user interface,
PowerSC simplifies the management of security and compliance significantly.

The PowerSC Multi-Factor Authentication (MFA) capability provides more assurance that only
authorized people access the environments by requiring at least one extra authentication
factor to prove that you are the person you say you are. MFA is included in PowerSC 2.0.

Because stolen or guessed passwords are still one of the most common ways for hackers to
access systems, having an MFA solution in place allows you to prevent a high percentage of
potential breaches.

This step is important on the journey toward implementing a zero trust security posture.

178 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
PowerSC 2.0 also includes Endpoint Detection and Response (EDR), which provides the
following features:
򐂰 Intrusion Detection and Prevention (IDP)
򐂰 Log inspection and analysis
򐂰 Anomaly detection, correlation, and incident response
򐂰 Response triggers
򐂰 Event context and filtering

4.4.6 Secure and Trusted boot


IBM Power servers provide a highly secure server platform. IBM Power10 processor-based
hardware and firmware includes PowerVM features to provide a more secure platform for
on-premises and cloud deployments.

The key PowerVM features include the following examples:


򐂰 A secure initial program load (IPL) process feature that is called Secure Boot allows only
suitably signed firmware components to run on the system processors. Each component
of the firmware stack, including host boot, the Power Hypervisor (PHYP), and partition
firmware (PFW), is signed by the platform manufacturer and verified as part of the IPL
process.
򐂰 A framework to support remote attestation of the system firmware stack through a
hardware trusted platform module (TPM).

The terms Secure Boot and Trusted Boot have specific connotations. The terms are used as
distinct, yet complementary concepts, as described next.

Secure Boot
This feature protects system integrity by using digital signatures to perform a
hardware-protected verification of all firmware components. It also distinguishes between the
host system trust domain and the eBMC or FSP trust domain by controlling service processor
and service interface access to sensitive system memory regions.

Trusted Boot
This feature creates cryptographically strong and protected platform measurements that
prove that specific firmware components ran on the system. You can assess the
measurements by using trusted protocols to determine the state of the system and use that
information to make informed security decisions.

4.4.7 Enhanced CPU: BMC isolation


Separating CPU and service processor trust domains in Power10 processor-based scale-out
servers improves the protection from external attacks.

Power10 technology introduces innovations to address emerging threats, specifically with


extra features and enhancements to defend against application domain vulnerabilities, such
as return-oriented programming (ROP) attacks (a security leverage technique is used by
attackers to run code on a target system). This capability uses a new in-core hardware
architecture that imparts minimal performance overhead (approximately only 1 - 2% for some
sample workloads tested).

Chapter 4. Enterprise solutions 179


In the Power10 processor-based scale-out servers, the eBMC chip is connected to the two
network interface cards by using NCSI (to support the connection to HMCs) and also have a
PCIe x1 connection that connects to the backplane. This connection is used by PowerVM for
partition management traffic, but cannot be used for guest LPAR traffic. A guest LPAR needs
its own physical or virtual network interface PCIe card (or cards) for external connection.

Hardware assist is necessary to avoid tampering with the stack. The Power platform added
four instructions (hashst, hashchk, hashstp, and hashchkp) to handle ROP in the Power ISA
3.1B.

4.5 Running artificial intelligence where operational data is


created, processed, and stored
Maintaining separate platforms for AI and business applications make deploying AI in
production difficult. The following results are possible:
򐂰 Reduced end-to-end availability for applications and data access
򐂰 Increased risk of violating service level agreements because of overhead of sending
operational data and receiving predictions from external AI platforms
򐂰 Increased complexity and cost of managing different environments and external
accelerators

Because AI is set to deploy everywhere, attention is turning from how fast data science teams
can build AImodels to how fast inference can be run against new data with those trained AI
models. Enterprises are asking their engineers and scientists to review new solutions and
new business models where the use of GPUs is no longer fundamental, especially because
this method became more expensive.

To support this shift, the Power10 processor-based server delivers faster business insights by
running AI in place with four Matrix Math Accelerator (MMA) units to accelerate AI in each
Power10 technology-based processor-core. The robust execution capability of the processor
cores with MMA acceleration, enhanced SIMD, and enhanced data bandwidth, provides an
alternative to external accelerators, such as GPUs.

It also reduces the time and cost that is associated with the related device management for
execution of statistical machine learning and inferencing workloads. These features, which
are combined with the possibility to consolidate multiple environments for AI model execution
on a Power10 processor-based server together with other different types of environments,
reduces costs and leads to a greatly simplified solution stack for the deployment of AI
workloads.

Operationalizing AI inferencing directly on a Power10 processor-based server brings AI closer


to data. This ability allows AI to inherit and benefit from the Enterprise Qualities of Service
(QoS): reliability, availability, and security of the Power10 processor-based platform and
support a performance boost. Enterprise business workflows can now readily and
consistently use insights that are built with the support of AI.

The use of data gravity on Power10 processor-cores enables AI to run during a database
operation or concurrently with an application, for example. This feature is key for
time-sensitive use cases. It delivers fresh input data to AI faster and enhances the quality and
speed of insight.

180 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
As no-code application development, pay-for-use model repositories, auto-machine learning,
and AI-enabled application vendors continue to evolve and grow, the corresponding software
products are brought over to the Power10 processor-based platform. Python and code from
major frameworks and tools, such as TensorFlow, PyTorch, and XGBoost, run on the
Power10 processor-based platform without any changes.

Open Neural Network Exchange (ONNX) models can be brought over from x86 or Arm
processor-based servers or other platforms and small-sized VMs or Power Virtual Server
(PowerVS) instances for deployment on Power10 processor-based servers. This Power10
technology gives customers the ability to train models on independent hardware but deploy
on enterprise grade servers.

The Power10 processor-core architecture includes an embedded MMA. This MMA is


extrapolated for a Power10 processor-based server to provide up to 5x faster AI inference for
32-bit floating point (FP32) precision to infuse AI into business applications and drive greater
insights. Up to 10x faster AI inference can be realized by using reduced precision data types,
such as 16-bit brain float (Bfloat16) and 8-bit integer (INT8), when compared with a previous
generation Power9 processor-based server.

The IBM development teams optimized common math libraries so that AI tools benefit from
the acceleration that is provided by the MMA units of the Power10 chip. The benefits of MMA
acceleration can be realized for statistical machine learning and inferencing, which provides a
cost-effective alternative to external accelerators or GPUs.

4.5.1 Train anywhere, deploy on Power10 processor-based server


IBM Power10 processor-based technology addresses challenges through hardware and
software innovation. Machine learning and deep learning models that are trained on any
system or cloud that is based on Power or x86 processor-based servers (even with different
endianness) can be deployed on the Power10 processor-based server and run without any
changes.

Because Power10 cores are equipped with four MMAs for matrix and tensor math,
applications can run models against colocated data without the need for external
accelerators, GPUs, or separate AI platforms. Power10 technology uses the “train anywhere,
deploy here” principle to operationalize AI.

A model can be trained on a public or private cloud and then deployed on a Power server (see
Figure 4-6 on page 182) by using the following procedure:
1. The trained model is registered with its version in the model vault. This vault is a VM or
LPAR with tools, such as IBM Watson® OpenScale, BentoML, or Tensorflow Serving, to
manage the model lifecycle.
2. The model is pushed out to the destination (in this case, a VM or an LPAR that is running
a database with an application). The model might be used by the database or the
application.
3. Transactions that are received by the database and application trigger model execution
and generate predictions or classifications. These predictions also can be stored locally.
For example, these predictions can be the risk or fraud that is associated with the
transaction or product classifications that are to be used by downstream applications.
4. A copy of the model output (prediction or classification) is sent to the model operations
(ModelOps) engine for calculation of drift by comparison with Ground Truth.

Chapter 4. Enterprise solutions 181


5. If the drift exceeds a threshold, the model retrain triggers are generated, and a new model
is trained by using a current data set. Otherwise, a new model is not trained.
6. Retrained models are then taken through steps 1 - 5.

Figure 4-6 Operationalizing AI

By using a combination of software and hardware innovation, the Power10 processor-based


scale-out server models can meet the model performance, response time, and throughput
KPIs of databases and applications that are infused with AI.

182 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Related publications

The publications that are listed in this section are considered particularly suitable for a more
detailed discussion of the topics that are covered in this paper.

IBM Redbooks
The following IBM Redbooks publications provide more information about the topics in this
document. Some publications that are referenced in this list might be available in softcopy
only:
򐂰 IBM Power Private Cloud with Shared Utility Capacity: Featuring Power Enterprise Pools
2.0, SG24-8476
򐂰 SAP HANA Data Management and Performance on IBM Power Systems, REDP-5570
򐂰 IBM PowerAI: Deep Learning Unleashed on IBM Power Systems Servers, SG24-8409
򐂰 IBM Power E1080 Technical Overview and Introduction, REDP-5649
򐂰 IBM Power E1050 Technical Overview and Introduction, REDP-5684
򐂰 IBM Power System AC922 Technical Overview and Introduction, REDP-5494
򐂰 IBM Power System E950: Technical Overview and Introduction, REDP-5509
򐂰 IBM Power System E980: Technical Overview and Introduction, REDP-5510
򐂰 IBM Power System L922 Technical Overview and Introduction, REDP-5496
򐂰 IBM Power System S822LC for High Performance Computing Introduction and Technical
Overview, REDP-5405
򐂰 IBM Power Systems H922 and H924 Technical Overview and Introduction, REDP-5498
򐂰 IBM Power Systems LC921 and LC922: Technical Overview and Introduction,
REDP-5495
򐂰 IBM Power Systems S922, S914, and S924 Technical Overview and Introduction
Featuring PCIe Gen 4 Technology, REDP-5595
򐂰 IBM PowerVM Best Practices, SG24-8062
򐂰 IBM PowerVM Virtualization Introduction and Configuration, SG24-7940
򐂰 IBM PowerVM Virtualization Managing and Monitoring, SG24-7590
򐂰 IBM PowerVC Version 2.0 Introduction and Configuration, SG24-8477

You can search for, view, download, or order these documents and other Redbooks
publications, Redpapers, web docs, drafts, and additional materials, at the following website:
ibm.com/redbooks

© Copyright IBM Corp. 2022. 183


Online resources
The following websites are also relevant as further information sources:
򐂰 IBM Fix Central:
http://www.ibm.com/support/fixcentral/
򐂰 IBM documentation:
https://www.ibm.com/docs/en
򐂰 IBM Power10 documentation:
https://www.ibm.com/docs/en/power10
򐂰 IBM Portal for OpenPOWER - POWER9 Monza Module:
https://www.ibm.com/systems/power/openpower/tgcmDocumentRepository.xhtml?aliasI
d=POWER9_Monza
򐂰 IBM Power:
https://www.ibm.com/it-infrastructure/power
򐂰 IBM Storage:
https://www.ibm.com/it-infrastructure/storage
򐂰 IBM Systems Energy Estimator:
https://ibm.biz/energy-estimator
򐂰 IBM System Planning Tool:
https://www.ibm.com/support/pages/ibm-system-planning-tool-power-processor-base
d-systems-0
򐂰 OpenCAPI:
http://opencapi.org/technical/use-cases/
򐂰 OpenPOWER Foundation:
https://openpowerfoundation.org/
򐂰 Power Systems Capacity on Demand:
http://www.ibm.com/systems/power/hardware/cod/
򐂰 Support for IBM Systems:
https://www.ibm.com/mysupport/s/?language=en_US

Help from IBM


IBM Support and downloads
ibm.com/support

IBM Global Services


ibm.com/services

184 IBM Power S1014, S1022s, S1022, and S1024 Technical Overview and Introduction
Back cover

REDP-5675-00

ISBN 0738460761

Printed in U.S.A.

®
ibm.com/redbooks

You might also like