[TRTLLM-6286] [feat] Update CUTLASS to 4.2 and enable SM103 group gemm #7832
Conversation
Signed-off-by: Xiwen Yu <[email protected]>
Signed-off-by: Xiwen Yu <[email protected]>
Signed-off-by: Xiwen Yu <[email protected]>
|
/bot run |
|
PR_Github #19126 [ run ] triggered by Bot |
📝 WalkthroughWalkthroughUpdates the Cutlass submodule pointer, re-enables the SM103 (Blackwell) MOE-GEMM path via conditional compilation and adjusts its dispatch/logging, and removes a runtime patching step from the kernel-generation Python script in favor of direct imports. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
participant Caller
participant Dispatch as moe_gemm_template_dispatch_tma_ws.h
Caller->>Dispatch: dispatch_moe_gemm_tma_ws(...)
alt Build without COMPILE_BLACKWELL_SM103_TMA_GROUPED_GEMMS and arch=SM103
Note right of Dispatch: SM103 path enabled at compile-time
Dispatch->>Dispatch: call SM103-specific grouped GEMM (void)
Dispatch-->>Caller: return
else Other architectures or guard defined
Dispatch->>Dispatch: route to non-SM103 implementation
Dispatch-->>Caller: return
end
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes Pre-merge checks and finishing touches❌ Failed checks (2 warnings)
✅ Passed checks (1 passed)
✨ Finishing touches
🧪 Generate unit tests
Tip 👮 Agentic pre-merge checks are now available in preview!Pro plan users can now enable pre-merge checks in their settings to enforce checklists before merging PRs.
Please see the documentation for more information. Example: reviews:
pre_merge_checks:
custom_checks:
- name: "Undocumented Breaking Changes"
mode: "warning"
instructions: |
Pass/fail criteria: All breaking changes to public APIs, CLI flags, environment variables, configuration keys, database schemas, or HTTP/GraphQL endpoints must be documented in the "Breaking Change" section of the PR description and in CHANGELOG.md. Exclude purely internal or private changes (e.g., code not exported from package entry points or explicitly marked as internal).Please share your feedback with us on this Discord post. 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. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (4)
3rdparty/cutlass (2)
1-1: Prefer aligning the bump with a released tag for traceability.If possible, bump to the exact v4.2.x tag (or note the upstream tag in the PR description) to simplify bisects and downstream audits.
1-1: CI/build hygiene after submodule bump.Recommend invalidating build caches (ccache/CMake/Bazel), bumping container/toolchain tags if needed, and running full CUTLASS‑consuming test suites (including SM103 variants) to avoid ABI/ODR mismatches.
cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_template_dispatch_tma_ws.h (2)
141-141: Prefer#if !defined(...)over#ifndefper project guidelines.Minor preprocessor style tweak for the new SM103 gate.
-#ifndef COMPILE_BLACKWELL_SM103_TMA_GROUPED_GEMMS +#if !defined(COMPILE_BLACKWELL_SM103_TMA_GROUPED_GEMMS) ... -#endif +#endifAlso applies to: 157-157
149-151: Capitalize “Blackwell” in user-facing log.Tiny text polish for consistency with other messages.
- TLLM_LOG_WARNING( - "For best performance please recompile with support for blackwell by " - "passing 103-real as an arch to build_wheel.py."); + TLLM_LOG_WARNING( + "For best performance please recompile with support for Blackwell by " + "passing 103-real as an arch to build_wheel.py.");
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (3)
3rdparty/cutlass(1 hunks)cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_template_dispatch_tma_ws.h(1 hunks)cpp/tensorrt_llm/kernels/cutlass_kernels/python/generate_kernels.py(0 hunks)
💤 Files with no reviewable changes (1)
- cpp/tensorrt_llm/kernels/cutlass_kernels/python/generate_kernels.py
🧰 Additional context used
📓 Path-based instructions (7)
**/*.{h,hpp,hh,hxx,cpp,cxx,cc,cu,cuh}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
**/*.{h,hpp,hh,hxx,cpp,cxx,cc,cu,cuh}: Namespace closing braces must include a trailing comment with the namespace name (e.g., '} // namespace foo').
Prefer const or constexpr variables over #define for constants.
Declare variables that are not modified after initialization as const.
Avoid magic literals in code; except for 0, nullptr, true, false. Use named constants for comparisons and logic.
Use Allman brace style for formatting.
Place the semicolon of an empty for/while loop on a new line.
Bodies of switch/while/do-while/for must be compound statements (brace-delimited), and if/else must always be followed by brace-delimited statements.
Type names (e.g., classes) must be CamelCase starting with an uppercase letter (e.g., FooBar).
Local variables, methods, and namespaces use lowerCamelCase (e.g., localFooBar).
Non-magic-number global variables that are non-static and not in an anonymous namespace must be lowerCamelCase prefixed with 'g' (e.g., gDontUseGlobalFoos).
Non-magic-number globals that are static or in an anonymous namespace use lowerCamelCase prefixed with 's' (e.g., sMutableStaticGlobal).
Locally visible static variables use lowerCamelCase with 's' prefix (e.g., static std::once_flag sFlag).
Private/protected member variables use 'm' prefix with CamelCase (e.g., mNbFooValues). Public members may omit, but 'm' is encouraged for clarity.
Constants (enums, global constants, static constants, and function-scope magic/literal constants) use uppercase SNAKE_CASE with 'k' prefix (e.g., kDIGIT_NUM).
Function-scope constants that are not magic numbers or literals are named like non-constant variables (e.g., bool const pass = a && b).
If macros are necessary, name them in UPPER_SNAKE_CASE (e.g., FOO_VERSION) and prefer constants over #define.
Use LLVM clang-format; wrap lines at a maximum of 120 columns; use '// clang-format off/on' sparingly with justification.
Use smart pointers for heap allocations; prefer unique_ptr for sole ownership, shared_ptr for shared...
Files:
cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_template_dispatch_tma_ws.h
**/*.{cpp,cxx,cc,cu,h,hpp,hh,hxx,cuh}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
C++ filenames should be lowerCamelCase (first letter lowercase) and must be case-insensitive unique within a compilation target.
Files:
cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_template_dispatch_tma_ws.h
**/*.{h,hpp,hh,hxx,cpp,cxx,cc,cu,cuh,py}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
Use only spaces, no tabs; indent with 4 spaces.
Files:
cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_template_dispatch_tma_ws.h
**/*.{h,hpp,hh,hxx}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
Document new class interfaces and function prototypes with Doxygen; use //! for single-line and //!< for members.
Files:
cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_template_dispatch_tma_ws.h
**/*.{h,hpp,hh,hxx,cpp,cxx,cc}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
**/*.{h,hpp,hh,hxx,cpp,cxx,cc}: Prefer anonymous namespaces over 'static' for internal linkage of functions.
All templates (class/function/member/static) must be instantiated at least once; non-POD classes should have private data members.
Files:
cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_template_dispatch_tma_ws.h
**/*.{h,hpp,hh,hxx,cuh}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
Use include guards named 'TRTLLM_<FILE_NAME_IN_CAPS_WITH_UNDERSCORES>_H' (no leading or trailing underscore; directory names excluded).
Files:
cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_template_dispatch_tma_ws.h
**/*.{cpp,cxx,cc,h,hpp,hh,hxx,cu,cuh,py}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
Prepend the NVIDIA Apache-2.0 copyright header with current year to the top of all source files (e.g., .cpp, .h, .cu, .py).
Files:
cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_template_dispatch_tma_ws.h
🧠 Learnings (2)
📓 Common learnings
Learnt from: djns99
PR: NVIDIA/TensorRT-LLM#7104
File: cpp/tensorrt_llm/cutlass_extensions/include/cutlass_extensions/epilogue/fusion/sm90_visitor_scatter.hpp:399-417
Timestamp: 2025-08-21T21:48:35.135Z
Learning: CUTLASS extensions in TensorRT-LLM (located under cpp/tensorrt_llm/cutlass_extensions/) are designed to integrate with and extend functionality in the external CUTLASS repository. When analyzing these extensions, their consumers and functionality wiring may exist in the CUTLASS codebase rather than within TensorRT-LLM itself.
📚 Learning: 2025-08-19T03:35:20.866Z
Learnt from: djns99
PR: NVIDIA/TensorRT-LLM#6915
File: cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_kernels.cu:4616-4626
Timestamp: 2025-08-19T03:35:20.866Z
Learning: In the MOE profiler TMA workspace preparation (cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_kernels.cu), the overlapping of TMA WS regions for NONE and FINALIZE variants is deliberate design to save memory space, as confirmed by djns99. The comment "reuse the same pointers to save space" reflects this intentional behavior.
Applied to files:
cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_template_dispatch_tma_ws.h
🧬 Code graph analysis (1)
cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_template_dispatch_tma_ws.h (1)
cpp/tensorrt_llm/cutlass_extensions/include/cutlass_extensions/gemm_configs.h (13)
if(308-311)if(312-315)if(316-319)if(320-323)if(324-327)if(328-331)if(332-335)if(336-339)if(340-343)if(488-495)if(496-499)TileShape(218-487)ClusterShape(267-430)
⏰ 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)
- GitHub Check: Pre-commit Check
🔇 Additional comments (2)
3rdparty/cutlass (1)
1-1: Sanity‑check SM103 Group GEMM enablement + Python cutlass_library import
- SM103 compile guards/flags are present: cpp/tensorrt_llm/kernels/cutlass_kernels/CMakeLists.txt (COMPILE_BLACKWELL_SM103_TMA_GEMMS / COMPILE_BLACKWELL_SM103_TMA_GROUPED_GEMMS), cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_template_dispatch_tma_ws.h, cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/launchers/moe_gemm_tma_ws_launcher.inl, cpp/tensorrt_llm/kernels/cutlass_kernels/fp4_gemm/fp4_gemm_template.h, and cutlass_extensions/include/cutlass_extensions/gemm_configs.h (SM103 aliased to SM100).
- Kernel generator imports cutlass_library: cpp/tensorrt_llm/kernels/cutlass_kernels/python/generate_kernels.py:6 — no python package/module named cutlass_library was found under 3rdparty/cutlass; confirm the submodule exposes this package or that CI/PYTHONPATH installs/points to it.
- Confirm CI builds/tests cover SM103 and that COMPILE_BLACKWELL_SM103_TMA_GROUPED_GEMMS is being set where required.
cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_template_dispatch_tma_ws.h (1)
152-156: Approved — void-dispatch not returned.
Confirmed dispatchMoeGemmFinalDispatchTmaWarpSpecialized(...) is invoked then a standalonereturn;; noreturn dispatchMoeGemmFinalDispatchTmaWarpSpecialized<...>(...)patterns found.
|
PR_Github #19126 [ run ] completed with state |
|
/bot run |
|
PR_Github #19177 [ run ] triggered by Bot |
|
PR_Github #19177 [ run ] completed with state |
cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_template_dispatch_tma_ws.h
Show resolved
Hide resolved
NVIDIA#7832) Signed-off-by: Xiwen Yu <[email protected]>
NVIDIA#7832) Signed-off-by: Xiwen Yu <[email protected]>
NVIDIA#7832) Signed-off-by: Xiwen Yu <[email protected]>
3 participants
Summary by CodeRabbit
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.