[AutoDeploy] Merge feat/ad_2025_06_13 feature branch

[lucaslie]

mass integration of https://github.com/nv-auto-deploy/TensorRT-LLM/tree/feat/ad_2025_06_13

see individual PRs for details

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 [--disable-fail-fast --skip-test --stage-list "A10-1, xxx" --gpu-type "A30, H100_PCIe" --add-multi-gpu-test --only-multi-gpu-test --disable-multi-gpu-test --post-merge --extra-stage "H100_PCIe-[Post-Merge]-1, xxx"]

Launch build/test pipelines. All previously running jobs will be killed.

--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 "A10-1, xxx" (OPTIONAL) : Only run the specified test stages. Examples: "A10-1, 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.

--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. Will also run 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-[Post-Merge]-1, xxx" (OPTIONAL) : Run the ordinary L0 pre-merge pipeline and specified test stages. Examples: --extra-stage "H100_PCIe-[Post-Merge]-1, xxx".

For guidance on mapping tests to stage names, see docs/source/reference/ci-overview.md.

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.

[lucaslie]

/bot run

[Copilot]

Pull Request Overview

This PR consolidates custom operator usage under the torch.ops.auto_deploy namespace across the codebase and extends multigpu sharding tests with a new GQA‐aware block.

• Refactored all Torch operator calls (rope, attention, quant, moe, dist, etc.) to use torch.ops.auto_deploy.*.
• Updated unit tests and transformation logic to reference the new auto_deploy namespace.
• Added GQA_Block in multigpu graph sharding tests with head‐dimension–aware sharding support.

Reviewed Changes

Copilot reviewed 50 out of 50 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
tests/unittest/_torch/auto_deploy/unit/singlegpu/transformations/library/test_rope_transformation.py	Replaced torch.ops.rope.* with torch.ops.auto_deploy.* for RoPE variants.
tests/unittest/_torch/auto_deploy/unit/singlegpu/transformations/library/test_quantization.py	Updated quantization operators to torch.ops.auto_deploy.torch_quant_fp*.
tests/unittest/_torch/auto_deploy/unit/multigpu/transformations/library/test_graph_sharding.py	Introduced GQA_Block and adapted sharding checks to use auto_deploy ops.
tensorrt_llm/_torch/auto_deploy/transformations/library/sharding.py	Enhanced _insert_sharded_matmul with min_local_shape and switched to auto_deploy distributed ops.
tensorrt_llm/_torch/auto_deploy/custom_ops/*	Registered and renamed all custom ops under the auto_deploy:: namespace.

Comments suppressed due to low confidence (4)

tests/unittest/_torch/auto_deploy/unit/multigpu/transformations/library/test_graph_sharding.py:18

• [nitpick] Class names should follow PEP8 CamelCase conventions; consider renaming GQA_Block to GQABlock.

class GQA_Block(nn.Module):

tensorrt_llm/_torch/auto_deploy/shim/demollm.py:322

• TorchSampler is not imported in this module; add the appropriate import (e.g., from tensorrt_llm._torch.pyexecutor.sampler import TorchSampler) to avoid undefined symbol errors.

        # set request id if necessary

tensorrt_llm/_torch/auto_deploy/transformations/library/sharding.py:74

• The docstring should be updated to document the new min_local_shape parameter, explaining how it constrains tensor splitting based on head dimension.

def _insert_sharded_matmul(

tests/unittest/_torch/auto_deploy/unit/multigpu/transformations/library/test_graph_sharding.py:83

• Add a dedicated test case for GQA_Block to verify that verify_local_weight_sizes correctly validates head-dimension sharding constraints under GQA scenarios.

    batch_size = 4

[tensorrt-cicd]

PR_Github #9764 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #9764 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #7198 completed with status: 'SUCCESS'