[https://nvbugs/5636986][fix] Fix DeepGemmMoe get_buffer calls

[VALLIS-NERIA]

Summary by CodeRabbit

• Refactor
  • Updated workspace memory allocation for mixture of experts operations to use tensor shape descriptors instead of flat buffer sizes.

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.

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 the stage-list parameter 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.md
and the scripts/test_to_stage_mapping.py helper.

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.

[VALLIS-NERIA]

/bot run

[coderabbitai]

📝 Walkthrough

Walkthrough

Modified workspace tensor specifications in the Fused MOE deep GEMM operation by replacing scalar buffer size calculations with multi-dimensional shape descriptors (lists) for three workspace buffers: workspace_0, workspace_1, and workspace_sf.

Changes

Cohort / File(s)	Summary
MOE workspace buffer shape conversion tensorrt_llm/_torch/modules/fused_moe/ops/moe_op_deepgemm.py	Updated _get_deepgemm_workspace function to specify workspace tensor buffers using multi-dimensional shape lists instead of flat scalar sizes: workspace_0, workspace_1, and workspace_sf each transitioned from scalar products to list-based shape descriptors maintaining dimensional structure

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

• Verify that the memory allocation and buffer management logic downstream can properly consume list-based shape descriptors instead of scalar sizes
• Confirm mathematical equivalence: ensure the product of dimensions in each list equals the original scalar calculations
• Check for any other references to these workspace tensors that may assume scalar size specifications

Pre-merge checks and finishing touches

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	The PR description is incomplete. Both the 'Description' and 'Test Coverage' sections are empty placeholders, with no substantive explanation of the issue, solution, or test coverage provided.	Fill in the Description section explaining what bug is being fixed and why, and complete the Test Coverage section with relevant test cases.

✅ Passed checks (1 passed)

Check name	Status	Explanation
Title check	✅ Passed	The PR title clearly identifies the main change: fixing DeepGemmMoe get_buffer calls, with proper format including NVBugs ticket and type designation.

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

---

📜 Recent review details

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between b9e5315 and d8fec4e.

📒 Files selected for processing (1)

• tensorrt_llm/_torch/modules/fused_moe/ops/moe_op_deepgemm.py (1 hunks)

🧰 Additional context used

📓 Path-based instructions (3)

**/*.{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:

• tensorrt_llm/_torch/modules/fused_moe/ops/moe_op_deepgemm.py

**/*.py

📄 CodeRabbit inference engine (CODING_GUIDELINES.md)

**/*.py: Python code must target Python 3.8+.
Indent Python code with 4 spaces; do not use tabs.
Maintain module namespace when importing; prefer 'from package.subpackage import foo' then 'foo.SomeClass()' instead of importing the class directly.
Python filenames should be snake_case (e.g., some_file.py).
Python classes use PascalCase names.
Functions and methods use snake_case names.
Local variables use snake_case; prefix 'k' for variables that start with a number (e.g., k_99th_percentile).
Global variables use upper SNAKE_CASE prefixed with 'G' (e.g., G_MY_GLOBAL).
Constants use upper SNAKE_CASE (e.g., MY_CONSTANT).
Avoid shadowing variables from an outer scope.
Initialize all externally visible members of a class in the constructor.
Prefer docstrings for interfaces that may be used outside a file; comments for in-function or file-local interfaces.
Use Google-style docstrings for classes and functions (Sphinx-parsable).
Document attributes and variables inline so they render under the class/function docstring.
Avoid reflection when a simpler, explicit approach suffices (e.g., avoid dict(**locals()) patterns).
In try/except, catch the most specific exceptions possible.
For duck-typing try/except, keep the try body minimal and use else for the main logic.

Files:

• tensorrt_llm/_torch/modules/fused_moe/ops/moe_op_deepgemm.py

**/*.{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:

• tensorrt_llm/_torch/modules/fused_moe/ops/moe_op_deepgemm.py

🧠 Learnings (6)

📓 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: 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.

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`.

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.

Learnt from: djns99
Repo: NVIDIA/TensorRT-LLM PR: 7104
File: cpp/tensorrt_llm/kernels/cutlass_kernels/include/moe_gemm_kernels.h:141-145
Timestamp: 2025-08-21T02:41:10.565Z
Learning: In TensorRT-LLM MOE GEMM kernels (cpp/tensorrt_llm/kernels/cutlass_kernels/include/moe_gemm_kernels.h), the stride_act and stride_weight pointers in TmaWarpSpecializedGroupedGemmInput are intentionally declared as void* rather than typed pointers because the actual stride type is determined at runtime based on factors like the swap_ab flag and layout decisions. This runtime type determination makes compile-time type safety impossible, so void* is the correct approach.

📚 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:

• tensorrt_llm/_torch/modules/fused_moe/ops/moe_op_deepgemm.py

📚 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:

• tensorrt_llm/_torch/modules/fused_moe/ops/moe_op_deepgemm.py

📚 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:

• tensorrt_llm/_torch/modules/fused_moe/ops/moe_op_deepgemm.py

📚 Learning: 2025-08-14T23:23:27.449Z

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.

Applied to files:

• tensorrt_llm/_torch/modules/fused_moe/ops/moe_op_deepgemm.py

📚 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:

• tensorrt_llm/_torch/modules/fused_moe/ops/moe_op_deepgemm.py

🧬 Code graph analysis (1)

tensorrt_llm/_torch/modules/fused_moe/ops/moe_op_deepgemm.py (1)

tensorrt_llm/_torch/memory_buffer_utils.py (1)

• get_buffer (52-113)

⏰ 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 (3)

tensorrt_llm/_torch/modules/fused_moe/ops/moe_op_deepgemm.py (3)

91-95: LGTM! Correct fix for workspace_0 buffer allocation.

The change properly passes a shape list [expert_size_per_partition, m_max, fp8_dim] to get_buffer, matching its expected signature. The use of fp8_dim = max(hidden_size, intermediate_size) ensures the buffer is sized correctly for both downstream usages at lines 218-220 (with hidden_size) and 249-251 (with intermediate_size).

---

98-105: LGTM! Correct fix for workspace_1 buffer allocation.

The shape list correctly uses max(intermediate_size * 2, hidden_size) to accommodate both downstream usages: the h1 intermediate tensor at line 235-236 (size intermediate_size * 2) and the h3 output tensor at line 268-269 (size hidden_size).

---

108-112: LGTM! Correct fix for workspace_sf buffer allocation.

The shape list properly sizes the scaling factor workspace. The allocation uses scale_k_padded computed from fp8_dim = max(hidden_size, intermediate_size) at line 84, which correctly accommodates both downstream usages where scale_k_padded is recalculated based on hidden_size (line 215-216) and intermediate_size (line 253-254).

---

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.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[tensorrt-cicd]

PR_Github #23620 [ run ] triggered by Bot. Commit: d8fec4e

[tensorrt-cicd]

PR_Github #23620 [ run ] completed with state SUCCESS. Commit: d8fec4e
/LLM/main/L0_MergeRequest_PR pipeline #17773 completed with status: 'FAILURE'

[VALLIS-NERIA]

/bot run

[tensorrt-cicd]

PR_Github #23714 [ run ] triggered by Bot. Commit: d8fec4e

[tensorrt-cicd]

PR_Github #23714 [ run ] completed with state SUCCESS. Commit: d8fec4e
/LLM/main/L0_MergeRequest_PR pipeline #17846 completed with status: 'SUCCESS'
Pipeline passed with automatic retried tests. Check the rerun report for details.

[HuiGao-NV]

LGTM

[VALLIS-NERIA]

/bot reuse-pipeline

[tensorrt-cicd]

PR_Github #23799 [ reuse-pipeline ] triggered by Bot. Commit: 58208de

[tensorrt-cicd]

PR_Github #23799 [ reuse-pipeline ] completed with state SUCCESS. Commit: 58208de
Reusing PR_Github #23714 for commit 58208de