[TRTLLM-6898][feat] Add Cute DSL nvfp4 linear op

[limin2021]

Summary by CodeRabbit

• New Features
  • Added an optional NVFP4 block‑scaled GEMM path accelerated by CuTe DSL for Blackwell GPUs, with autotuning and BF16 outputs. Configurable via a new Linear constructor flag; existing behavior remains the default when disabled. Automatically gated by Python 3.12+ and runtime availability.
• Tests
  • Introduced unit and performance tests validating correctness and benchmarking the new NVFP4 CuTe DSL path across multiple shapes.
• Chores
  • Added a new dependency for Python 3.12+ to enable the CuTe/Cutlass DSL integration.

Description

[TRTLLM-6898][feat]Add Cute DSL nvfp4 linear op-step1
(1) install nvidia-cutlass-dsl 4.1.0 pkg. (require python >= 3.12)
(2) add nvfp4_bs_gemm into trtllm linear op.
(3) I copy the nvfp4 cute_dsl kernel from cutlass dsl examples, and add scalar parameter support to adapt to trtllm's linear op's interface.

Test Coverage

op UT:
pytest -s -o log_cli=true tests/unittest/_torch/thop/parallel/test_fp4_linear.py -k "test_fp4_linear"

model UT:
pytest -s -o log_cli=true "tests/integration/defs/accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_nvfp4"

Perf

to add perf data here.

TODO;
add pdl, and other optimizations for this kernel.

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 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.

[coderabbitai]

📝 Walkthrough

Walkthrough

Adds a new Blackwell SM100 CuTe-DSL persistent block-scaled GEMM kernel and utilities, integrates a Python 3.12+ gated FP4/FP8 GEMM path into Torch custom ops, wires it into Linear via a new flag and scalar alpha handling, updates dependencies, and introduces unit/perf tests for the new path.

Changes

Cohort / File(s)	Summary
Dependencies requirements.txt	Adds dependency nvidia-cutlass-dsl==4.1.0 for Python >= 3.12.
CuTe DSL Blackwell kernel and utils tensorrt_llm/_torch/custom_ops/cute_dsl_kernels/blackwell/dense_blockscaled_gemm_persistent.py, tensorrt_llm/_torch/custom_ops/cute_dsl_kernels/blackwell/utils.py	Introduces persistent block-scaled GEMM kernel for SM100, wrapper, run harness, scale-factor converter, and runtime pointer utilities (_Pointer, make_ptr).
Torch custom ops integration tensorrt_llm/_torch/custom_ops/torch_custom_ops.py	Adds Python 3.12+ gated CuTe-DSL NVFP4 GEMM path, tuner-backed runner, public op cute_dsl_nvfp4_gemm_blackwell, and fake-op registrations.
Linear module wiring tensorrt_llm/_torch/modules/linear.py	Adds flag use_cute_dsl_nvfp4_blockscaling_mm; routes NVFP4 path to new op using scalar_alpha; populates scalar_alpha in weight-loading paths.
Tests tests/unittest/_torch/thop/parallel/test_fp4_linear.py	Adds NVFP4 CuTe-DSL functional and perf tests (Python 3.12 gated), including padding helper and scripted perf entry.

Sequence Diagram(s)

sequenceDiagram
  autonumber
  actor User
  participant Linear as Linear (NVFP4)
  participant Ops as torch_custom_ops
  participant Tuner as AutoTuner/TunableRunner
  participant Kernel as KernelWrapper
  participant GPU as Device Kernel

  User->>Linear: forward(input)
  alt use_cute_dsl_nvfp4_blockscaling_mm
    Linear->>Ops: cute_dsl_nvfp4_gemm_blackwell(input, weight, scales, alpha_s)
    Ops->>Ops: Check Python>=3.12 & HAS_CUTLASS_DSL
    alt Available
      Ops->>Tuner: get_valid_tactics()
      Tuner->>Kernel: compile/select tactic
      Kernel->>GPU: launch(m, n, k, pointers, epilogue)
      GPU-->>Kernel: result
      Kernel-->>Ops: BF16 tensor
      Ops-->>Linear: BF16 tensor
    else Unavailable
      Ops-->>Linear: raise RuntimeError
    end
  else legacy path
    Linear->>Ops: nvfp4_gemm(..., alpha)
    Ops-->>Linear: result
  end
  Linear-->>User: output

Loading

sequenceDiagram
  autonumber
  participant Ops as torch_custom_ops (init)
  participant Env as Runtime

  Ops->>Env: import cutlass DSL (Python>=3.12)
  alt Success
    Env-->>Ops: loaded
    Ops-->>Ops: HAS_CUTLASS_DSL=True
  else Failure
    Env-->>Ops: not available
    Ops-->>Ops: HAS_CUTLASS_DSL=False
  end

Loading

Estimated code review effort

🎯 5 (Critical) | ⏱️ ~120 minutes

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.

• Built-in checks – Quickly apply ready-made checks to enforce title conventions, require pull request descriptions that follow templates, validate linked issues for compliance, and more.
• Custom agentic checks – Define your own rules using CodeRabbit’s advanced agentic capabilities to enforce organization-specific policies and workflows. For example, you can instruct CodeRabbit’s agent to verify that API documentation is updated whenever API schema files are modified in a PR. Note: Upto 5 custom checks are currently allowed during the preview period. Pricing for this feature will be announced in a few weeks.

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.

Pre-merge checks (2 passed, 1 warning)

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 34.55% which is insufficient. The required threshold is 80.00%.	You can run @coderabbitai generate docstrings to improve docstring coverage.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title Check	✅ Passed	The title succinctly follows the repository convention by including the Jira ticket, change type, and a clear summary of the primary feature (“Add Cute DSL nvfp4 linear op”), accurately reflecting the main deliverable of the pull request.
Description Check	✅ Passed	The pull request description provides the mandatory sections—Description, Test Coverage, and PR Checklist—and gives clear details on the feature added, how it was tested with both unit and model tests, and the items required by the project’s template.

✨ Finishing Touches

• 📝 Generate Docstrings

🧪 Generate unit tests

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

---

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.}

[limin2021]

/bot run

[limin2021]

/bot run

[tensorrt-cicd]

PR_Github #18160 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #18161 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #18160 [ run ] completed with state ABORTED

[limin2021]

/bot run

[tensorrt-cicd]

PR_Github #18169 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #18161 [ run ] completed with state ABORTED

[tensorrt-cicd]

PR_Github #18169 [ run ] completed with state FAILURE
/LLM/main/L0_MergeRequest_PR pipeline #13615 completed with status: 'FAILURE'

[limin2021]

/bot run

[limin2021]

/bot run

[tensorrt-cicd]

PR_Github #18604 [ run ] triggered by Bot

[limin2021]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #18606 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #18604 [ run ] completed with state ABORTED
LLM/main/L0_MergeRequest_PR #13966 (Blue Ocean) completed with status: ABORTED

[limin2021]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #18626 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #18606 [ run ] completed with state ABORTED
LLM/main/L0_MergeRequest_PR #13967 (Blue Ocean) completed with status: ABORTED

[tensorrt-cicd]

PR_Github #18626 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #13984 completed with status: 'FAILURE'

[limin2021]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #18655 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #18655 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #14012 completed with status: 'SUCCESS'
Pipeline passed with automatic retried tests. Check the rerun report for details.