Fused two elementwise kernels for k_nope and k_pe concat#14862
Merged
HaiShaw merged 3 commits intosgl-project:mainfrom Dec 15, 2025
Merged
Fused two elementwise kernels for k_nope and k_pe concat#14862HaiShaw merged 3 commits intosgl-project:mainfrom
HaiShaw merged 3 commits intosgl-project:mainfrom
Conversation
Contributor
|
Warning You have reached your daily quota limit. Please wait up to 24 hours and I will start processing your requests again! |
HaiShaw
reviewed
Dec 11, 2025
| attn_dtype = k_nope.dtype | ||
| k = k_nope.new_empty(*k_shape, dtype=attn_dtype) | ||
| concat_and_cast_mha_k_triton(k, k_nope, k_pe) | ||
| elif _is_hip and self.current_attention_backend == "aiter": |
Collaborator
There was a problem hiding this comment.
Please confirm _is_hip or _is_gfx95_supported
Collaborator
Author
There was a problem hiding this comment.
_is_hip is enough, this optimization can go in any ROCm platform.
Collaborator
|
/tag-and-rerun-ci |
HaiShaw
approved these changes
Dec 11, 2025
Liwansi
added a commit
to iforgetmyname/sglang
that referenced
this pull request
Dec 15, 2025
…n_eagle3_npu * 'main' of https://github.com/sgl-project/sglang: (89 commits) [model-gateway] Remove legacy RouterMetrics and Rename SmgMetrics to Metrics and smg_labels to metrics_labels (sgl-project#15160) [diffusion] fix: fix video model sp when resolution is not specified (sgl-project#15047) [diffusion] fix: fix pytorch non-writable array warning (sgl-project#15017) [diffusion] fix: cache dit with parallel (sgl-project#15163) chore: change npu pr-test a2 runner (sgl-project#15152) [Feature] Fuse mrope all in 1 kernel (sgl-project#14906) Fix num running requests (load) wrong cleared for ongoing requests (sgl-project#15116) Fused two elementwise kernels for k_nope and k_pe concat (sgl-project#14862) fix: adding date and fixing release name issue (sgl-project#15174) [CPU] Add Gemma3RMSNorm kernel in sgl-kernel and add ut (sgl-project#9324) feature: PR wheel (sgl-project#15170) [diffusion] model: support mutli-image input and qwen-image-edit-2509 (sgl-project#15005) fix CompressedTensorsW8A8Int8 min_capability (sgl-project#13914) Tiny improve summary text in `bench_one_batch_server.py` (sgl-project#15158) [model-gateway] add mcp and discovery metrics (sgl-project#15156) fix: move ci-bot (sgl-project#15154) Fix import warnings (sgl-project#15144) ci: adding errors to Github summary (sgl-project#14778) [model-gateway] Add streaming metrics for harmony gRPC router (sgl-project#15147) [model-gateway] upgrade axum and axum server (sgl-project#15146) ... # Conflicts: # python/sglang/srt/server_args.py
tonyluj
pushed a commit
to openanolis/sglang
that referenced
this pull request
Dec 17, 2025
…#14862) 1 TC failure to check, but irrelevant to this code change
YChange01
pushed a commit
to YChange01/sglang
that referenced
this pull request
Jan 13, 2026
…#14862) 1 TC failure to check, but irrelevant to this code change
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.
Motivation
Reduce time cost in concat k_nope and k_pe before doing MHA attention
Modifications
Use the triton kernel to replace the naive torch operations
Accuracy Tests
Benchmarking and Profiling
before fusing two elementwise kernels
This PR
The time cost can be reduced from (169 us + 104 us) to (128us)
Checklist