[https://jirasw.nvidia.com/browse/TRTLLM-4645] support mutliCtasKvMode for high-throughput MLA kernels #5426
Conversation
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
|
/bot run |
|
PR_Github #9662 [ run ] triggered by Bot |
|
PR_Github #9662 [ run ] completed with state |
|
/bot run |
|
PR_Github #9676 [ run ] triggered by Bot |
|
PR_Github #9676 [ run ] completed with state |
PerkzZheng
force-pushed
the
user/perkzz/high-throughput-mla
branch
from
9f5941d to
3698cd4
Compare
|
/bot run --disable-fail-fast |
1 similar comment
|
/bot run --disable-fail-fast |
|
PR_Github #9783 [ run ] triggered by Bot |
… heurisitc Signed-off-by: Perkz Zheng <[email protected]>
Signed-off-by: Perkz Zheng <[email protected]>
Signed-off-by: Perkz Zheng <[email protected]>
|
PR_Github #9783 [ run ] completed with state |
PerkzZheng
force-pushed
the
user/perkzz/high-throughput-mla
branch
from
3698cd4 to
462e339
Compare
|
/bot run |
|
PR_Github #9808 [ run ] triggered by Bot |
qiaoxj07
approved these changes
Jun 25, 2025
|
PR_Github #9808 [ run ] completed with state |
dominicshanshan
pushed a commit
to dominicshanshan/TensorRT-LLM
that referenced
this pull request
Jul 9, 2025
…e for high-throughput MLA kernels (NVIDIA#5426) Signed-off-by: Perkz Zheng <[email protected]>
dominicshanshan
pushed a commit
to dominicshanshan/TensorRT-LLM
that referenced
this pull request
Jul 10, 2025
…e for high-throughput MLA kernels (NVIDIA#5426) Signed-off-by: Perkz Zheng <[email protected]>
dominicshanshan
pushed a commit
to dominicshanshan/TensorRT-LLM
that referenced
this pull request
Jul 10, 2025
…e for high-throughput MLA kernels (NVIDIA#5426) Signed-off-by: Perkz Zheng <[email protected]>
dominicshanshan
pushed a commit
to dominicshanshan/TensorRT-LLM
that referenced
this pull request
Jul 10, 2025
…e for high-throughput MLA kernels (NVIDIA#5426) Signed-off-by: Perkz Zheng <[email protected]>
dominicshanshan
pushed a commit
to dominicshanshan/TensorRT-LLM
that referenced
this pull request
Jul 10, 2025
…e for high-throughput MLA kernels (NVIDIA#5426) Signed-off-by: Perkz Zheng <[email protected]>
dominicshanshan
pushed a commit
to dominicshanshan/TensorRT-LLM
that referenced
this pull request
Jul 11, 2025
…e for high-throughput MLA kernels (NVIDIA#5426) Signed-off-by: Perkz Zheng <[email protected]>
dominicshanshan
pushed a commit
to dominicshanshan/TensorRT-LLM
that referenced
this pull request
Jul 11, 2025
…e for high-throughput MLA kernels (NVIDIA#5426) Signed-off-by: Perkz Zheng <[email protected]>
dominicshanshan
pushed a commit
to dominicshanshan/TensorRT-LLM
that referenced
this pull request
Jul 11, 2025
…e for high-throughput MLA kernels (NVIDIA#5426) Signed-off-by: Perkz Zheng <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
[https://jirasw.nvidia.com/browse/TRTLLM-4645] support mutliCtasKvMode for high-throughput MLA kernels && update the heuristic
The trtllm-gen commit: 086115ea6893c06b351af19e01aa0288f61d6823
This PR adds the multiCtasKvMode + high-throughput MLA kernels, which can have 1.45x speedup in one important large-EP case.
It also removes unused MLA kernels (like fp16) to reduce the binary size.