Skip to content

Repository files navigation

RDNA AI Optimization

Reproducible inference and training optimization for consumer AMD Radeon RDNA GPUs — what runs, how it compares to stock, and whether to use upstream, an image, or a patch.

Start here

  1. Hardware compatibility
  2. Performance dashboard
  3. Upstream / image / patch status
  4. Quick start

Scope

RDNA4 (gfx1200/1201) primary · vLLM / llama.cpp / DiffSynth / ComfyUI · LLM·VLM·video · GEMM / attention / MoE kernels · RCCL / QuickReduce.

Results

Topic Platform Highlight Detail
Qwen3.6-27B-FP8 · vLLM TP=2 2×R9700 · graph · MTP on Live+ TG128 decode 74.1 vs Community 64.3 / Stock 46.1 · 256k @ c≈2 (FP8-QKV + UA) PERFORMANCE

More models will land as additional rows here; full tables stay in each topic’s PERFORMANCE.md.

Boundaries

  • Docker is user space only; host GPU must pass rocminfo / rocm-smi.
  • Results do not generalize across VRAM, GPU count, or PCIe/NUMA/P2P without revalidation.
  • working/ is untracked; only reviewed artifacts enter the release tree (promotion rules).

Map

Path Purpose
docs/ Getting started, methodology, hardware, upstream
frameworks/ · kernels/ · comm/ Released topics
working/ Optimization workbench

License

LICENSE. Upstream subtrees keep their own licenses.

About

No description, website, or topics provided.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages