Matrix-Multiply Assist Best Practices Guide

Table of contents

1. Front cover
2. Notices
   1. Trademarks
3. Preface
   1. Authors
   2. Now you can become a published author, too!
   3. Comments welcome
   4. Stay connected to IBM Redbooks
4. Chapter 1. Matrix multiplication
   1. 1.1 Matrix-multiply operation
   2. 1.2 Vector outer product operation
   3. 1.3 Introduction to Vector Scalar Extension
   4. 1.4 Simple VSX code example for a vector outer product
5. Chapter 2. Matrix-Multiply Assist Architecture
   1. 2.1 Data types
   2. 2.2 Data layout in accumulators and VSRs
   3. 2.3 Instructions
      1. 2.3.1 Accumulator operation instructions
      2. 2.3.2 Outer product instructions
      3. 2.3.3 Advanced feature: Lane masking
6. Chapter 3. Programming with Matrix-Multiply Assist
   1. 3.1 Single-precision GEMM using MMA
   2. 3.2 Double-precision GEMM using MMA and one accumulator
   3. 3.3 Mixed and lower precision matrix multiplication with MMA
      1. 3.3.1 Source matrix reordering with Int8 as example
7. Chapter 4. Advanced programming concepts
   1. 4.1 Multiple accumulators SGEMM for load value reuse
   2. 4.2 Multiple accumulators DGEMM for load value reuse
   3. 4.3 SGEMM performance with advanced cache-blocking
8. Chapter 5. Matrix-Multiply Assist programming with compiler built-ins
   1. 5.1 Simple MMA SGEMM example using built-ins
9. Appendix A. List of Matrix-Multiply Assist compiler built-ins
10. Appendix B. List of Matrix-Multiply Assist instructions in Power ISA v3.1
    1. Related publications
11. Back cover