[TRTLLM-7831][feat] Support block wise FP8 in wide ep #7423
Conversation
|
This PR depends on #7422 |
xxi-nv
force-pushed
the
supportFP8BlockWideEp_release
branch
from
2d4604d to
d3074c1
Compare
|
Caution Review failedFailed to post review comments. Configuration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro 💡 Knowledge Base configuration:
You can enable these sources in your CodeRabbit configuration. 📒 Files selected for processing (3)
🧰 Additional context used📓 Path-based instructions (2)**/*.{cpp,cc,cxx,cu,h,hpp,hh,hxx,cuh,py}📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
Files:
**/*.py📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
Files:
🧠 Learnings (1)📚 Learning: 2025-08-14T06:36:40.701ZApplied to files:
🧬 Code graph analysis (3)tensorrt_llm/_torch/modules/fused_moe/moe_backend.py (4)
tests/unittest/_torch/modules/test_fused_moe.py (7)
tensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.py (2)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
📝 WalkthroughWalkthroughAdds a pluggable MoE backend abstraction with Cutlass and DeepGemm implementations, integrates lazy backend selection and invocation into WideEPMoE, updates quantization method gating by SM version, and introduces a multi-GPU FP8 blockwise Wide MoE unit test comparing outputs against a DeepGemm reference. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
participant U as Caller
participant W as WideEPMoE
participant S as MoEBackendSelection
participant C as MoECutlassBackend
participant D as MoEDeepGemmBackend
U->>W: forward(...)
alt Backend not initialized
W->>S: select_backend(self)
alt SM==100 and has_deepseek_fp8_block_scales
S-->>W: MoEDeepGemmBackend
W->>D: cache instance as moe_backend_impl
else
S-->>W: MoECutlassBackend
W->>C: cache instance as moe_backend_impl
end
end
W->>W: prepare inputs/weights/quant_scales
opt Tuning/finalize
W->>+moe_backend_impl: finalize_tactic(tuner_input, ...)
moe_backend_impl-->>-W: tactics ready / no-op
end
W->>+moe_backend_impl: run_moe(input, tokens, weights, scales, ...)
moe_backend_impl-->>-W: final_hidden_states
W-->>U: output
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Possibly related PRs
Suggested reviewers
✨ Finishing Touches
🧪 Generate unit tests
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
SupportNeed help? Create a ticket on our support page for assistance with any issues or questions. CodeRabbit Commands (Invoked using PR/Issue comments)Type Other keywords and placeholders
Status, Documentation and Community
|
|
/bot run |
xxi-nv
requested review from
kaiyux and
QiJune
and removed request for
mikeiovine
kaiyux
force-pushed
the
supportFP8BlockWideEp_release
branch
from
abfad92 to
17b76e3
Compare
|
/bot run |
|
PR_Github #17258 [ run ] triggered by Bot |
|
PR_Github #17258 [ run ] completed with state |
xxi-nv
force-pushed
the
supportFP8BlockWideEp_release
branch
from
17b76e3 to
029204e
Compare
|
/bot run |
|
PR_Github #17475 [ run ] triggered by Bot |
|
PR_Github #17475 [ run ] completed with state |
|
/bot run |
|
PR_Github #17513 [ run ] triggered by Bot |
|
PR_Github #17513 [ run ] completed with state |
|
PR_Github #17576 [ run ] completed with state |
kaiyux
force-pushed
the
supportFP8BlockWideEp_release
branch
from
029204e to
29ebc85
Compare
|
/bot run --disable-fail-fast --add-multi-gpu-test |
|
PR_Github #17698 [ run ] triggered by Bot |
|
/bot run |
|
PR_Github #17722 [ run ] triggered by Bot |
|
PR_Github #17722 [ run ] completed with state |
|
/bot run --disable-fail-fast --add-multi-gpu-test |
|
PR_Github #17729 [ run ] triggered by Bot |
|
PR_Github #17698 [ run ] completed with state |
|
/bot run --disable-fail-fast --add-multi-gpu-test |
|
PR_Github #17751 [ run ] triggered by Bot |
|
PR_Github #17729 [ run ] completed with state |
|
PR_Github #17751 [ run ] completed with state |
Signed-off-by: xxi <[email protected]> modified: tensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.py new file: tensorrt_llm/_torch/modules/fused_moe/moe_backend.py modified: tests/unittest/_torch/modules/test_fused_moe.py modified: tensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.py new file: tensorrt_llm/_torch/modules/fused_moe/moe_backend.py modified: tests/unittest/_torch/modules/test_fused_moe.py
Signed-off-by: xxi <[email protected]> modified: tensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.py deleted: tensorrt_llm/_torch/modules/fused_moe/moe_backend.py new file: tensorrt_llm/_torch/modules/fused_moe/ops/__init__.py new file: tensorrt_llm/_torch/modules/fused_moe/ops/moe_op.py new file: tensorrt_llm/_torch/modules/fused_moe/ops/moe_op_cutlass.py new file: tensorrt_llm/_torch/modules/fused_moe/ops/moe_op_deepgemm.py
Head branch was pushed to by a user without write access
xxi-nv
force-pushed
the
supportFP8BlockWideEp_release
branch
from
ad0f413 to
d41f571
Compare
|
/bot run --disable-fail-fast --add-multi-gpu-test |
|
PR_Github #17945 [ run ] triggered by Bot |
|
PR_Github #17945 [ run ] completed with state |
|
/bot run --disable-fail-fast --add-multi-gpu-test |
|
bot run --disable-fail-fast --add-multi-gpu-test |
|
/bot run --disable-fail-fast --add-multi-gpu-test |
|
PR_Github #17961 [ run ] triggered by Bot |
|
PR_Github #17961 [ run ] completed with state |
Summary by CodeRabbit
New Features
Tests
Description
Test Coverage
PR Checklist
Please review the following before submitting your PR:
PR description clearly explains what and why. If using CodeRabbit's summary, please make sure it makes sense.
PR Follows TRT-LLM CODING GUIDELINES to the best of your knowledge.
Test cases are provided for new code paths (see test instructions)
Any new dependencies have been scanned for license and vulnerabilities
CODEOWNERS updated if ownership changes
Documentation updated as needed
The reviewers assigned automatically/manually are appropriate for the PR.
Please check this after reviewing the above items as appropriate for this PR.
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 [--reuse-test (optional)pipeline-id --disable-fail-fast --skip-test --stage-list "A10-PyTorch-1, xxx" --gpu-type "A30, H100_PCIe" --test-backend "pytorch, cpp" --add-multi-gpu-test --only-multi-gpu-test --disable-multi-gpu-test --post-merge --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx" --detailed-log --debug(experimental)]Launch build/test pipelines. All previously running jobs will be killed.
--reuse-test (optional)pipeline-id(OPTIONAL) : Allow the new pipeline to reuse build artifacts and skip successful test stages from a specified pipeline or the last pipeline if no pipeline-id is indicated. If the Git commit ID has changed, this option will be always ignored. The DEFAULT behavior of the bot is to reuse build artifacts and successful test results from the last pipeline.--disable-reuse-test(OPTIONAL) : Explicitly prevent the pipeline from reusing build artifacts and skipping successful test stages from a previous pipeline. Ensure that all builds and tests are run regardless of previous successes.--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-PyTorch-1, xxx"(OPTIONAL) : Only run the specified test stages. Examples: "A10-PyTorch-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.--test-backend "pytorch, cpp"(OPTIONAL) : Skip test stages which don't match the specified backends. Only support [pytorch, cpp, tensorrt, triton]. Examples: "pytorch, cpp" (does not run test stages with tensorrt or triton backend). Note: Does NOT update GitHub pipeline 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 in addition to running 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-TensorRT-Post-Merge-1, xxx"(OPTIONAL) : Run the ordinary L0 pre-merge pipeline and specified test stages. Examples: --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx".--detailed-log(OPTIONAL) : Enable flushing out all logs to the Jenkins console. This will significantly increase the log volume and may slow down the job.--debug(OPTIONAL) : Experimental feature. Enable access to the CI container for debugging purpose. Note: Specify exactly one stage in thestage-listparameter to access the appropriate container environment. Note: Does NOT update GitHub check status.For guidance on mapping tests to stage names, see
docs/source/reference/ci-overview.mdand the
scripts/test_to_stage_mapping.pyhelper.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.