[TRTLLM-10126][feat] Increase topk upper limit to 22 for NVLinkOneSid… #10229
Conversation
|
/bot run |
📝 WalkthroughWalkthroughExpanded MoE communication kernel support to handle TOP_K = 22 by adding a specialized vectorized reduction path in the CUDA kernel, updating the maximum top_k constant from 16 to 22, and extending test coverage to validate the new capability. Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes Pre-merge checks and finishing touches❌ Failed checks (2 warnings)
✅ Passed checks (1 passed)
✨ Finishing touches
🧪 Generate unit tests (beta)
📜 Recent review detailsConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro 📒 Files selected for processing (3)
🧰 Additional context used📓 Path-based instructions (4)**/*.{cpp,h,cu,cuh}📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
Files:
**/*.{cpp,h,cu,cuh,py}📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
Files:
**/*.h📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
Files:
**/*.py📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
Files:
🧠 Learnings (11)📓 Common learnings📚 Learning: 2025-08-09T20:57:04.084ZApplied to files:
📚 Learning: 2025-08-19T03:35:20.866ZApplied to files:
📚 Learning: 2025-09-23T15:12:38.312ZApplied to files:
📚 Learning: 2025-08-08T22:03:40.707ZApplied to files:
📚 Learning: 2025-09-23T14:58:05.372ZApplied to files:
📚 Learning: 2025-09-02T13:42:44.885ZApplied to files:
📚 Learning: 2025-08-21T02:39:12.009ZApplied to files:
📚 Learning: 2025-08-20T07:43:36.447ZApplied to files:
📚 Learning: 2025-09-23T15:01:00.070ZApplied to files:
📚 Learning: 2025-08-14T21:04:50.248ZApplied to files:
⏰ 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). (2)
🔇 Additional comments (4)
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: 0
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.cu (1)
48-97: Critical: Missing case for top_k = 22 in SWITCH_TOP_K macro.The
SWITCH_TOP_Kmacro does not include a case fortop_k = 22, which means any call withtop_k = 22will hit the default case at Line 95 and fail with "Unsupported top_k". This prevents the new TOP_K = 22 kernel specialization (lines 657-718) from being reached at runtime, even though the implementation is present.🔎 Proposed fix: Add case 22 to the macro
#define SWITCH_TOP_K(top_k, TOP_K, ...) \ switch (top_k) \ { \ + case 22: \ + { \ + constexpr int TOP_K = 22; \ + __VA_ARGS__; \ + break; \ + } \ case 16: \ { \ constexpr int TOP_K = 16; \ __VA_ARGS__; \ break; \ } \
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (3)
cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.cucpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.htests/unittest/_torch/multi_gpu/test_moe_a2a.py
🧰 Additional context used
📓 Path-based instructions (4)
**/*.{cpp,h,cu,cuh}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
**/*.{cpp,h,cu,cuh}: Closing braces of namespaces should have a comment saying the namespace it closes:} // namespace foo
Preferconstorconstexprvariables over#definewhenever possible, as the latter are not visible to the compiler
A variable that is not modified after its initialization should be declared asconst
For naming of constants in C++, follow the naming section conventions
Except0(only used in comparison for checking signness/existence/emptiness) andnullptr,true,false, all other literals should only be used for variable initialization in C++
Use the Allman indentation style in C++
Put the semicolon for an emptyfororwhileloop in a new line in C++
The statement forming the body of aswitch,while,do .. whileorforstatement shall be a compound statement (use brace-delimited statements) in C++
If and else should always be followed by brace-delimited statements, even if empty or a single statement in C++
C++ filenames should use camel case with first letter lowercase:thisIsASubDirandthisIsAFilename.cpp
All files involved in the compilation of a compilation target (.exe/.so) must have filenames that are case-insensitive unique in C++
All types (including class names) in C++ should use camel case with uppercase first letter:FooBarClass
Local variables, methods and namespaces in C++ should use camel case with first letter lowercase:localFooBar
Non-magic-number global variables that are non-static and not defined in anonymous namespace in C++ should use camel case prefixed by a lower case 'g':gDontUseGlobalFoos
Non-magic-number global variables that are static or defined in an anonymous namespace in C++ should use camel case prefixed by a lower case 's':sMutableStaticGlobal
Locally visible static variables in C++ should use camel case with lowercase prefix 's' as the first letter:static std::once_flag sFlag;
Public, private and protected class member variables in C++ should use camel case prefi...
Files:
cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.cucpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.h
**/*.{cpp,h,cu,cuh,py}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
All TensorRT-LLM Open Source Software code should contain an NVIDIA copyright header that includes the year of its latest meaningful modification
Files:
cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.cucpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.htests/unittest/_torch/multi_gpu/test_moe_a2a.py
**/*.h
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
**/*.h: Use a preprocessor guard in C++ header files with the formatTRTLLM_<FILENAME>_Hderived from the filename in all caps
The preprocessor guard name in C++ must have prefixTRTLLM_followed by the filename, all in caps. Only use the file name, not directory names
Do not use prefix with underscore in C++ preprocessor guard symbols as such symbols are reserved in C++ standard for compilers or implementation
Do not use trailing underscore in C++ preprocessor guard symbols (unlike Google C++ guideline)
Files:
cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.h
**/*.py
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
**/*.py: Code developed for TensorRT-LLM should conform to Python 3.8+
Indent Python code with 4 spaces. Do not use tabs
Always maintain the namespace when importing in Python, even if only one class or function from a module is used
Python files should use snake_case naming:some_file.py
Python classes should use PascalCase naming:class SomeClass
Python functions and methods should use snake_case naming:def my_awesome_function():
Python local variables should use snake_case naming:my_variable = ...
Python variable names that start with a number should be prefixed with 'k':k_99th_percentile = ...
Python global variables should use upper snake_case with prefix 'G':G_MY_GLOBAL = ...
Python constants should use upper snake_case naming:MY_CONSTANT = ...
Avoid shadowing variables declared in an outer scope in Python
Initialize all externally visible members of a Python class in the constructor
For Python interfaces that may be used outside a file, prefer docstrings over comments
Python comments should be reserved for code within a function, or interfaces that are local to a file
Use Google style docstrings in Python for classes and functions, which can be parsed by Sphinx
Python attributes and variables can be documented inline with type and description
Avoid using reflection in Python when functionality can be easily achieved without reflection
When using try-except blocks in Python, limit the except to the smallest set of errors possible
When using try-except blocks in Python to handle multiple possible variable types (duck-typing), keep the body of the try as small as possible, using the else block for logic
Files:
tests/unittest/_torch/multi_gpu/test_moe_a2a.py
🧠 Learnings (11)
📓 Common learnings
Learnt from: djns99
Repo: NVIDIA/TensorRT-LLM PR: 6915
File: cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_kernels.cu:4010-4012
Timestamp: 2025-08-14T23:23:27.449Z
Learning: For MOE (Mixture of Experts) code reviews in TensorRT-LLM, avoid repeatedly suggesting finalize fusion validation checks and safety assertions. The user djns99 has indicated these suggestions are repetitive and unwanted across multiple MOE-related changes.
Learnt from: ChristinaZ
Repo: NVIDIA/TensorRT-LLM PR: 7068
File: cpp/tensorrt_llm/kernels/moeTopKFuncs.cuh:169-172
Timestamp: 2025-08-20T07:43:36.447Z
Learning: In TensorRT-LLM MOE kernels, when processing up to 128 experts across 32 threads, each thread handles at most 4 experts (N < 5 constraint), where N represents candidates per thread rather than total system capacity.
Learnt from: djns99
Repo: NVIDIA/TensorRT-LLM PR: 7104
File: cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_kernels.cu:1475-1480
Timestamp: 2025-08-21T02:39:12.009Z
Learning: The min latency mode functionality in TensorRT-LLM MOE kernels (cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_kernels.cu) is deprecated and no longer being maintained/updated, as confirmed by djns99. Bug reports and optimization suggestions for the computeStridesTmaWarpSpecializedLowLatencyKernel and related min latency code paths should be deprioritized.
📚 Learning: 2025-08-09T20:57:04.084Z
Learnt from: sklevtsov-nvidia
Repo: NVIDIA/TensorRT-LLM PR: 3294
File: cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_tma_warp_specialized_input.cu:118-127
Timestamp: 2025-08-09T20:57:04.084Z
Learning: In the CUTLASS MoE finalize fusion implementation (cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_tma_warp_specialized_input.cu), when setting `fused_finalize_epilogue.stride_final_output` with shape `(hidden_size, num_output_tokens, 1)`, the `num_rows_in_final_output` should be set to `num_output_tokens` (not `hidden_size`) because of a swap+transpose operation that maps rows of the output tensor to `hidden_size` and columns to `num_output_tokens`.
Applied to files:
cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.cucpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.h
📚 Learning: 2025-08-19T03:35:20.866Z
Learnt from: djns99
Repo: NVIDIA/TensorRT-LLM PR: 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/communicationKernels/moeAlltoAllKernels.cu
📚 Learning: 2025-09-23T15:12:38.312Z
Learnt from: nv-lschneider
Repo: NVIDIA/TensorRT-LLM PR: 7910
File: cpp/tensorrt_llm/thop/allreduceOp.cpp:352-446
Timestamp: 2025-09-23T15:12:38.312Z
Learning: In TensorRT-LLM NCCL device allreduce implementation (cpp/tensorrt_llm/thop/allreduceOp.cpp), the goto pattern in runNCCLAllReduceDeviceFusion is intentionally used for future extensibility, allowing multiple switch cases to fallback to the default handler. While not aesthetically ideal, this pattern supports adding more fusion cases later that can reuse the same fallback logic.
Applied to files:
cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.cu
📚 Learning: 2025-08-08T22:03:40.707Z
Learnt from: sklevtsov-nvidia
Repo: NVIDIA/TensorRT-LLM PR: 3294
File: cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_kernels.cu:1198-1209
Timestamp: 2025-08-08T22:03:40.707Z
Learning: In the CUTLASS MoE kernels (cpp/tensorrt_llm/cutlass_extensions), when `layout_info.fusion` is set to `TmaWarpSpecializedGroupedGemmInput::EpilogueFusion::FINALIZE`, the `router_scales` parameter must be non-null by design. The fused finalize kernel epilogue does not perform nullptr checks and requires valid router scales to function correctly. This is an implicit contract that callers must satisfy when enabling the FINALIZE fusion mode.
Applied to files:
cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.cu
📚 Learning: 2025-09-23T14:58:05.372Z
Learnt from: nv-lschneider
Repo: NVIDIA/TensorRT-LLM PR: 7910
File: cpp/tensorrt_llm/kernels/nccl_device/config.cu:42-49
Timestamp: 2025-09-23T14:58:05.372Z
Learning: In TensorRT-LLM NCCL device kernels (cpp/tensorrt_llm/kernels/nccl_device/), the token partitioning intentionally uses ceil-like distribution (same token_per_rank for all ranks) to ensure all ranks launch the same number of blocks. This is required for optimal NCCL device API barrier performance, even though it may launch extra blocks for non-existent tokens on later ranks. Runtime bounds checking in the kernel (blockID validation) handles the overshoot cases.
Applied to files:
cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.cu
📚 Learning: 2025-09-02T13:42:44.885Z
Learnt from: pcastonguay
Repo: NVIDIA/TensorRT-LLM PR: 7455
File: tensorrt_llm/_torch/pyexecutor/py_executor.py:1852-1860
Timestamp: 2025-09-02T13:42:44.885Z
Learning: In MPI communication within TensorRT-LLM pipeline parallelism, different communication types (tokens, logits, termination sync) must use disjoint tag namespaces to avoid message routing collisions when using the same source/destination patterns.
Applied to files:
cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.cu
📚 Learning: 2025-08-21T02:39:12.009Z
Learnt from: djns99
Repo: NVIDIA/TensorRT-LLM PR: 7104
File: cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_kernels.cu:1475-1480
Timestamp: 2025-08-21T02:39:12.009Z
Learning: The min latency mode functionality in TensorRT-LLM MOE kernels (cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_kernels.cu) is deprecated and no longer being maintained/updated, as confirmed by djns99. Bug reports and optimization suggestions for the computeStridesTmaWarpSpecializedLowLatencyKernel and related min latency code paths should be deprioritized.
Applied to files:
cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.cu
📚 Learning: 2025-08-20T07:43:36.447Z
Learnt from: ChristinaZ
Repo: NVIDIA/TensorRT-LLM PR: 7068
File: cpp/tensorrt_llm/kernels/moeTopKFuncs.cuh:169-172
Timestamp: 2025-08-20T07:43:36.447Z
Learning: In TensorRT-LLM MOE kernels, when processing up to 128 experts across 32 threads, each thread handles at most 4 experts (N < 5 constraint), where N represents candidates per thread rather than total system capacity.
Applied to files:
cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.cucpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.h
📚 Learning: 2025-09-23T15:01:00.070Z
Learnt from: nv-lschneider
Repo: NVIDIA/TensorRT-LLM PR: 7910
File: cpp/tensorrt_llm/kernels/nccl_device/config.cu:15-17
Timestamp: 2025-09-23T15:01:00.070Z
Learning: In TensorRT-LLM NCCL device kernels, the <sstream> header is not needed as an explicit include in config.cu because it's provided transitively through other headers. Local compilation testing confirms this works without the explicit include.
Applied to files:
cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.cu
📚 Learning: 2025-08-14T21:04:50.248Z
Learnt from: thorjohnsen
Repo: NVIDIA/TensorRT-LLM PR: 6910
File: cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp:0-0
Timestamp: 2025-08-14T21:04:50.248Z
Learning: In KV cache onboarding logic during prefill in cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp, when calculating which blocks fall within the attention window, use getTokensPerBlock() to advance token indices rather than block->getUniqueTokens().size(), because the calculation needs to consider the post-prefill state where blocks will be filled to capacity, not their current token count.
Applied to files:
cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.cu
⏰ 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). (2)
- GitHub Check: Check PR Checklist Resolution
- GitHub Check: Pre-commit Check
🔇 Additional comments (4)
cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.h (1)
29-29: LGTM: Top-k limit increased to 22.The constant update aligns with the PR objective to support higher top-k values for NVLinkOneSided AlltoAll.
cpp/tensorrt_llm/kernels/communicationKernels/moeAlltoAllKernels.cu (2)
657-718: Reduction logic for TOP_K = 22 is correct.The multi-stage reduction correctly handles 22 accumulators: 22 → 11 → 6 → 3 → 1. The odd numbers at each stage (a20 in stage 2, a8 and a16 in stage 3) are properly carried forward and summed in subsequent stages.
719-719: LGTM: Correctly changed to else if.The change from
if constexprtoelse if constexpris correct since the TOP_K == 22 case now precedes it.tests/unittest/_torch/multi_gpu/test_moe_a2a.py (1)
568-568: LGTM: Test coverage for top_k = 22.The test configuration validates the new top_k = 22 capability for the combine path, ensuring both dispatch and combine kernels work correctly with the increased limit.
|
You also need to modify here |
|
PR_Github #29583 [ run ] triggered by Bot. Commit: |
nv-guomingz
force-pushed
the
user/guomingz/enlarge_moe_topk_to_22
branch
from
6756de5 to
c47f2fc
Compare
Thx, updated. |
|
PR_Github #29583 [ run ] completed with state
|
|
/bot run |
|
PR_Github #29617 [ run ] triggered by Bot. Commit: |
|
PR_Github #29617 [ run ] completed with state |
…ed AlltoAll. Signed-off-by: nv-guomingz <[email protected]>
nv-guomingz
force-pushed
the
user/guomingz/enlarge_moe_topk_to_22
branch
from
c47f2fc to
555a43c
Compare
|
/bot run |
|
PR_Github #29683 [ run ] triggered by Bot. Commit: |
|
PR_Github #29683 [ run ] completed with state
|
|
/bot run |
|
PR_Github #29800 [ run ] triggered by Bot. Commit: |
3 participants
…ed AlltoAll.
Summary by CodeRabbit
New Features
Tests
✏️ Tip: You can customize this high-level summary in your review settings.
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
Update tava architecture diagram if there is a significant design change in PR.
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.
Details
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.