KEMBAR78
fix: Fix MOE benchmark to rotate buffers to prevent L2 cache reuse by djns99 · Pull Request #4135 · NVIDIA/TensorRT-LLM · GitHub
Skip to content

Conversation

@djns99
Copy link
Collaborator

@djns99 djns99 commented May 7, 2025

Description

This PR fixes some issues in the MOE microbenchmarks:

  • Adds buffer rotation to prevent L2 cache reuse
  • Fixes the seed for random distributions so consecutive iterations are comparable

Test Coverage

GitHub Bot Help

/bot [-h] ['run', 'kill', 'skip', 'reuse-pipeline'] ...

Provide a user friendly way for developers to interact with a Jenkins server.

Run /bot [-h|--help] to print this help message.

See details below for each supported subcommand.

run [--disable-fail-fast --skip-test --stage-list "A10-1, xxx" --gpu-type "A30, H100_PCIe" --add-multi-gpu-test --only-multi-gpu-test --disable-multi-gpu-test --post-merge --extra-stage "H100_PCIe-[Post-Merge]-1, xxx"]

Launch build/test pipelines. All previously running jobs will be killed.

--disable-fail-fast (OPTIONAL) : Disable fail fast on build/tests/infra failures.

--skip-test (OPTIONAL) : Skip all test stages, but still run build stages, package stages and sanity check stages. Note: Does NOT update GitHub check status.

--stage-list "A10-1, xxx" (OPTIONAL) : Only run the specified test stages. Examples: "A10-1, xxx". Note: Does NOT update GitHub check status.

--gpu-type "A30, H100_PCIe" (OPTIONAL) : Only run the test stages on the specified GPU types. Examples: "A30, H100_PCIe". Note: Does NOT update GitHub check status.

--only-multi-gpu-test (OPTIONAL) : Only run the multi-GPU tests. Note: Does NOT update GitHub check status.

--disable-multi-gpu-test (OPTIONAL) : Disable the multi-GPU tests. Note: Does NOT update GitHub check status.

--add-multi-gpu-test (OPTIONAL) : Force run the multi-GPU tests. Will also run L0 pre-merge pipeline.

--post-merge (OPTIONAL) : Run the L0 post-merge pipeline instead of the ordinary L0 pre-merge pipeline.

--extra-stage "H100_PCIe-[Post-Merge]-1, xxx" (OPTIONAL) : Run the ordinary L0 pre-merge pipeline and specified test stages. Examples: --extra-stage "H100_PCIe-[Post-Merge]-1, xxx".

kill

kill

Kill all running builds associated with pull request.

skip

skip --comment COMMENT

Skip testing for latest commit on pull request. --comment "Reason for skipping build/test" is required. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.

reuse-pipeline

reuse-pipeline

Reuse a previous pipeline to validate current commit. This action will also kill all currently running builds associated with the pull request. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.

@djns99
Copy link
Collaborator Author

djns99 commented May 7, 2025

/bot run

@tensorrt-cicd
Copy link
Collaborator

PR_Github #4431 [ run ] triggered by Bot

@tensorrt-cicd
Copy link
Collaborator

PR_Github #4431 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #3191 completed with status: 'FAILURE'

@djns99 djns99 force-pushed the user/djns99/moe_router_fixes branch from a3dc0f8 to 363cbb6 Compare May 9, 2025 05:06
@djns99
Copy link
Collaborator Author

djns99 commented May 9, 2025

/bot run

@tensorrt-cicd
Copy link
Collaborator

PR_Github #4652 [ run ] triggered by Bot

@tensorrt-cicd
Copy link
Collaborator

PR_Github #4652 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #3355 completed with status: 'FAILURE'

@djns99 djns99 force-pushed the user/djns99/moe_router_fixes branch from 363cbb6 to 181ca58 Compare May 19, 2025 23:44
@djns99
Copy link
Collaborator Author

djns99 commented May 19, 2025

/bot run

@tensorrt-cicd
Copy link
Collaborator

PR_Github #5786 [ run ] triggered by Bot

@tensorrt-cicd
Copy link
Collaborator

PR_Github #5786 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #4236 completed with status: 'SUCCESS'

@djns99 djns99 requested a review from jinyangyuan-nvidia May 21, 2025 21:22
@djns99 djns99 force-pushed the user/djns99/moe_router_fixes branch from 181ca58 to ef3424b Compare May 21, 2025 21:22
@djns99
Copy link
Collaborator Author

djns99 commented May 21, 2025

/bot run

@tensorrt-cicd
Copy link
Collaborator

PR_Github #6060 [ run ] triggered by Bot

@tensorrt-cicd
Copy link
Collaborator

PR_Github #6060 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #4432 completed with status: 'FAILURE'

@djns99 djns99 force-pushed the user/djns99/moe_router_fixes branch from ef3424b to bd04586 Compare May 22, 2025 00:52
@djns99
Copy link
Collaborator Author

djns99 commented May 22, 2025

/bot run

@tensorrt-cicd
Copy link
Collaborator

PR_Github #6073 [ run ] triggered by Bot

@tensorrt-cicd
Copy link
Collaborator

PR_Github #6073 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #4442 completed with status: 'FAILURE'

@djns99 djns99 force-pushed the user/djns99/moe_router_fixes branch 2 times, most recently from 03454da to 4c9d838 Compare June 25, 2025 03:07
Signed-off-by: Daniel Stokes <40156487+djns99@users.noreply.github.com>
@djns99 djns99 force-pushed the user/djns99/moe_router_fixes branch from 4c9d838 to 1cc7657 Compare July 14, 2025 22:29
@djns99
Copy link
Collaborator Author

djns99 commented Jul 14, 2025

/bot run

@tensorrt-cicd
Copy link
Collaborator

PR_Github #11852 [ run ] triggered by Bot

@tensorrt-cicd
Copy link
Collaborator

PR_Github #11852 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #8783 completed with status: 'SUCCESS'

@djns99 djns99 merged commit dd2491f into NVIDIA:main Jul 15, 2025
3 checks passed
evezhier pushed a commit to evezhier/TensorRT-LLM that referenced this pull request Jul 16, 2025
…VIDIA#4135)

Signed-off-by: Daniel Stokes <40156487+djns99@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants