Enable shuffled KV cache layout for MiniMax vLLM#1199
Open
jiacao-amd wants to merge 1 commit intoSemiAnalysisAI:mainfrom
Open
Enable shuffled KV cache layout for MiniMax vLLM#1199jiacao-amd wants to merge 1 commit intoSemiAnalysisAI:mainfrom
jiacao-amd wants to merge 1 commit intoSemiAnalysisAI:mainfrom
Conversation
![claude[bot]](https://avatars.githubusercontent.com/in/1236702?s=60&v=4){kind=small}
cquil11
approved these changes
Apr 27, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
VLLM_ROCM_SHUFFLE_KV_CACHE_LAYOUT=1in the MiniMax-M2.5 FP8 MI355X vLLM benchmark.Why
vLLM's ROCm AITER attention backend expects the shuffled KV cache layout for its fast attention path. Without this environment variable, the benchmark can run with the default KV cache layout and miss the
intended AITER attention kernels, leaving MI355X MiniMax-M2.5 FP8 throughput below the optimized path.
This script already enables
VLLM_ROCM_USE_AITER=1and launches vLLM with--attention-backend ROCM_AITER_FA; settingVLLM_ROCM_SHUFFLE_KV_CACHE_LAYOUT=1makes the KV cache layout match that backend.Testing
bash -n benchmarks/single_node/minimaxm2.5_fp8_mi355x.sh