gemm-optimization

Official

Optimize GEMM kernels on AMD CDNA with FlyDSL.

AuthorROCm
Version1.0.0
Installs0

System Documentation

What problem does it solve?

FlyDSL GEMM optimization provides a comprehensive guide to maximizing performance of GEMM kernels on AMD CDNA GPUs by detailing tiling strategies, memory layouts, and MFMA scheduling patterns.

Core Features & Use Cases

  • Comprehensive tiling strategies, LDS ping-pong double buffering, XOR bank-conflict mitigation, A/B prefetch pipelines, two-stage software pipelining, hot_loop_scheduler MFMA scheduling, epilogue options (Direct Store vs CShuffle), TFLOPS and bandwidth analysis, ATT trace bottleneck diagnosis, and optimization recipes.
  • Real-world scenario: configure tile sizes, enable 2-stage LDS, compare direct store vs CShuffle, profile with ATT traces, and iteratively optimize kernels.

Quick Start

Follow the guide to implement GEMM optimizations for an AMD CDNA GPU using FlyDSL, starting with selecting tile_m, tile_n, and tile_k and enabling 2-stage LDS where appropriate.

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: gemm-optimization
Download link: https://github.com/ROCm/FlyDSL/archive/main.zip#gemm-optimization

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.