[AMD][TRITON_KERNELS] Improve matmul ogs config on RDNA#931
Merged
leeliu103 merged 1 commit intorelease/internal/3.6.xfrom Mar 17, 2026
Merged
[AMD][TRITON_KERNELS] Improve matmul ogs config on RDNA#931leeliu103 merged 1 commit intorelease/internal/3.6.xfrom
leeliu103 merged 1 commit intorelease/internal/3.6.xfrom
Conversation
The default configuration for matmul_ogs resulted in significant register spilling on RDNA when the batch size exceeded 512. With this optimization, we observe a notable improvement in GPT-OSS end-to-end performance. gpt-oss-20B on 1x GPU | Device | Optimization | Throughput (reqs/sec) | Improvement | | ----------- | --------------------| --------------------- | ----------- | | Navi31-48GB | Baseline | 3.84 | 100% | | Navi31-48GB | Config Optimization | 6.18 | 161% | | Navi48-32GB | Baseline | 8.58 | 100% | | Navi48-32GB | Config Optimization | 10.07 | 117% | 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
leeliu103
force-pushed
the
backport-rdna-matmul-config-3.6
branch
from
e76303e to
fd25401
Compare
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The default configuration for matmul_ogs resulted in significant register spilling on RDNA when the batch size exceeded 512.
With this optimization, we observe a notable improvement in GPT-OSS end-to-end performance.
gpt-oss-20B on 1x GPU