[TRTLLM-5812][feat] support FP8 row-wise dense GEMM in torch flow #5615
Conversation
DylanChen-NV changed the title DylanChen-NV changed the title 
juney-nvidia requested review from Naveassaf and Tracin and removed request for dongxuy04 and liji-nv DylanChen-NV force-pushed the row_wise_torch_flow branch from 20840d9 to 41579d6 Compare | /bot run |
| PR_Github #10461 [ run ] triggered by Bot |
| PR_Github #10461 [ run ] completed with state |
| /bot run |
| PR_Github #10464 [ run ] triggered by Bot |
| PR_Github #10464 [ run ] completed with state |
| /bot run |
| PR_Github #10473 [ run ] triggered by Bot |
| PR_Github #10473 [ run ] completed with state |
| /bot run |
| PR_Github #10503 [ run ] triggered by Bot |
| PR_Github #10503 [ run ] completed with state |
| /bot run |
| PR_Github #10512 [ run ] triggered by Bot |
| PR_Github #10512 [ run ] completed with state |
DylanChen-NV force-pushed the row_wise_torch_flow branch from e1a100c to f6230e9 Compare | /bot run |
| PR_Github #10562 [ run ] triggered by Bot |
| PR_Github #10562 [ run ] completed with state |
tomeras91 left a comment
tomeras91 left a comment There was a problem hiding this comment.
Overall LGTM. Added a few nit comments.
What's more important: Can you edit the PR description, remove the irrelevant info from the description template and add important information on what is added in this PR?
@tomeras91 Thanks for the comments. I’ve implemented all the suggested changes. Please feel free to approve or let me know if further changes are needed. |
DylanChen-NV force-pushed the row_wise_torch_flow branch from 11461a1 to d4950c7 Compare | /bot run |
| PR_Github #11086 [ run ] triggered by Bot |
| PR_Github #11086 [ run ] completed with state |
DylanChen-NV force-pushed the row_wise_torch_flow branch from d4950c7 to cea6589 Compare | /bot reuse-pipeline |
| PR_Github #11131 [ reuse-pipeline ] triggered by Bot |
| PR_Github #11131 [ reuse-pipeline ] completed with state |
Signed-off-by: Dylan Chen <191843203+DylanChen-NV@users.noreply.github.com>
Signed-off-by: Dylan Chen <191843203+DylanChen-NV@users.noreply.github.com>
Signed-off-by: Dylan Chen <191843203+DylanChen-NV@users.noreply.github.com>
Signed-off-by: Dylan Chen <191843203+DylanChen-NV@users.noreply.github.com>
Signed-off-by: Dylan Chen <191843203+DylanChen-NV@users.noreply.github.com>
Signed-off-by: Dylan Chen <191843203+DylanChen-NV@users.noreply.github.com>
DylanChen-NV force-pushed the row_wise_torch_flow branch from cea6589 to 819c9b8 Compare | /bot reuse-pipeline |
| PR_Github #11133 [ reuse-pipeline ] triggered by Bot |
| PR_Github #11133 [ reuse-pipeline ] completed with state |
…IDIA#5615) Signed-off-by: Dylan Chen <191843203+DylanChen-NV@users.noreply.github.com> Signed-off-by: Yuxin <yuxinz@nvidia.com>
5 participants
[TRTLLM-5812][feat] support FP8 row-wise dense GEMM in torch flow
Description
Adapter existing FP8 row-wise dense GEMM kernels (sm89/sm90/sm120) to torch workflow.
Support FP8 row-wise GEMM as a TunableRunner for better performance.
Test Coverage
tests/unittest/_torch/thop/test_fp8_rowwise_linear.py: Test the output correctness of row-wise torch op
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.
Details
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".For guidance on mapping tests to stage names, see
docs/source/reference/ci-overview.md.kill
killKill all running builds associated with pull request.
skip
skip --comment COMMENTSkip 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-pipelineReuse 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.