ROCm 7.13 nightly adds strix halo optimizations
Summary
AMD's ROCm 7.13 tech preview adds optimizations for Strix Halo (Ryzen AI Max 300) and open-sources the ROCprof Trace Decoder.
Similar Articles
strix-llama new branch: prompt processing +21–54% over Vulkan and up to 2.7× over stock ROCm kernel at 128k, same outputs quality
rdna35-perf branch of strix-llama.cpp 为 Strix Halo (Radeon 8060S) 带来新的 RDNA3.5 flash attention 与 GEMM 内核优化,prompt processing 比 Vulkan 快 21–54%,在 128k 长上下文下比原版 ROCm 最快达 2.7 倍,且输出质量与原版 ROCm 基本一致。
@no_stp_on_snek: Spent the last few days helping @Italianclownz speed up ROCmFPX, his llama.cpp fork, on a borrowed Strix Halo (Ryzen AI…
Developer helped speed up ROCmFPX, a llama.cpp fork, on AMD Strix Halo (Ryzen AI Max+ 395), achieving major prefill and decode speedups with 16 merged PRs and improved quantization accuracy.
Strix Halo ROCm + MTP Notes (May 2026)
Technical benchmark comparing ROCm and Vulkan backends for LLM inference on Strix Halo hardware after MTP merged into llama.cpp, revealing ROCm suffers severe performance drops at full context while Vulkan remains stable.
AMD ROCm 7.14 "TheRock" tech preview tagged for latest AMD GPU compute stack
AMD has tagged the ROCm 7.14 'TheRock' tech preview, bringing AI training enhancements, performance improvements up to 16% for select AI workloads like Comfy UI, and ongoing Windows support for the open-source GPU compute stack.
Big News for AMD / Strix Halo+ Owners
The NPU on AMD Strix Halo devices is now usable for AI inference, enabling hybrid mode that combines NPU and iGPU for faster prompt processing. Tools like Lemonade and AMD's ROCm software make this possible.