Cognitive Computing Featuring the IBM Power System AC922

A draft IBM Redpaper publication

Updated 17 September 2019

cover image

IBM Form #: REDP-5555-00
(55 pages)

More options

Rate and comment

Authors: Scott Vetter, Ivaylo Bozhinov, Boran Lee, Gustavo Santos


This IBM® Redpaper™ publication, covers the advantages of using IBM Power System AC922 for cognitive solutions, and how it can enhance client's busineses.

In order to optimize the hardware and software, IBM works in partnering with NVIDIA, Mellanox,, SQream, Kinetica and other prominent companies, to design the Power AC922 server, specifically enhanced the Cognitive era.

Most outstanding hardware features, as NVIDIA NVLink 2.0 and PCIe 4.0, are described in this publication to show all the advantages that clients can take in comparison with our competitors.

We also include a brief description about what is cognitive computing, and how IBM use Watson™ Machine Learning Cognitive Solutions to bring more value to the client's ecosystem. Additionally we display performance charts that brings all the advantages of using Power AC922 against x86 competitors. In the last chapter we describe the most remarkable use cases where IBM solve real problems using cognitive solutions.

This IBM Redpaper publication, is aimed at IT technical audiences, especially decision-making levels that need a full look at the benefits and improvements that an IBM Cognitive Solution can offer. It also brings valuable information to data science professionals, so they can plan their modeling needs and finally provides information to the infrastructure support group in charge of maintain the solution.

Table of contents

Chapter 1. Introduction to cognitive computing
Chapter 2. IBM Power System AC922 for cognitive computing
Chapter 3. Cognitive solutions
Chapter 4. Use cases

These pages are Web versions of IBM Redbooks- and Redpapers-in-progress. They are published here for those who need the information now and may contain spelling, layout and grammatical errors. This material has not been submitted to any formal IBM test and is published AS IS. It has not been the subject of rigorous review. Your feedback is welcomed to improve the usefulness of the material to others.

Follow IBM Redbooks

Follow IBM Redbooks