Refactor: move DeepEP from Docker images to wheel building

[yuantailing]

[TRTLLM-6011] Improvement of DeepEP integration flow to avoid docker upgrade

Description

Move DeepEP from Docker images to wheel building.

The DeepEP library porting is expected to update fast, so it's too heavy to upgrade the Docker image for each DeepEP upgradation.

This PR added the following files into .whl, which makes the wheel self-contained.

tensorrt_llm
├── deep_ep
│   ├── LICENSE
│   ├── __init__.py
│   ├── buffer.py
│   └── utils.py
├── deep_ep_cpp_tllm.cpython-312-x86_64-linux-gnu.so
├── deep_ep_cpp_tllm.pyi
└── libs
    └── nvshmem
        ├── License.txt
        ├── nvshmem_bootstrap_uid.so.3
        └── nvshmem_transport_ibgda.so.103

Test Coverage

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 [--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.

[yuantailing]

/bot run --stage-list "DGX_H100-4_GPUs-PyTorch-DeepSeek-1"

[tensorrt-cicd]

PR_Github #10095 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #10095 [ run ] completed with state FAILURE
/LLM/main/L0_MergeRequest_PR pipeline #7453 (Partly Tested) completed with status: 'FAILURE'

[yuantailing]

/bot run --stage-list "DGX_H100-4_GPUs-PyTorch-DeepSeek-1"

[tensorrt-cicd]

PR_Github #10269 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #10269 [ run ] completed with state FAILURE
/LLM/main/L0_MergeRequest_PR pipeline #7589 (Partly Tested) completed with status: 'FAILURE'

[yuantailing]

/bot run --stage-list "DGX_H100-4_GPUs-PyTorch-DeepSeek-1"

[tensorrt-cicd]

PR_Github #10343 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #10343 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #7648 (Partly Tested) completed with status: 'FAILURE'

[yuantailing]

/bot run --stage-list "DGX_H100-4_GPUs-PyTorch-DeepSeek-1"

[tensorrt-cicd]

PR_Github #10388 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #10388 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #7680 (Partly Tested) completed with status: 'SUCCESS'
Pipeline passed with automatic retried tests. Check the rerun report for details.

[yuantailing]

/bot run --stage-list "Build-Docker-Images"

[tensorrt-cicd]

PR_Github #10429 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #10429 [ run ] completed with state FAILURE
/LLM/main/L0_MergeRequest_PR pipeline #7711 (Partly Tested) completed with status: 'FAILURE'

[yuantailing]

/bot run --stage-list "Build-Docker-Images"

[tensorrt-cicd]

PR_Github #10434 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #10434 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #7716 (Partly Tested) completed with status: 'SUCCESS'

[yuantailing]

/bot run --disable-fail-fast --add-multi-gpu-test

[tensorrt-cicd]

PR_Github #10506 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #11102 [ run ] triggered by Bot

[yuantailing]

/bot run --disable-fail-fast --add-multi-gpu-test

[tensorrt-cicd]

PR_Github #11103 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #11102 [ run ] completed with state ABORTED

[yuantailing]

/bot run --disable-fail-fast --add-multi-gpu-test

[tensorrt-cicd]

PR_Github #11108 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #11103 [ run ] completed with state ABORTED

[yuantailing]

/bot run --disable-fail-fast --add-multi-gpu-test

[tensorrt-cicd]

PR_Github #11110 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #11108 [ run ] completed with state ABORTED

[tensorrt-cicd]

PR_Github #11110 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #8216 completed with status: 'SUCCESS'

[yuantailing]

/bot skip --comment "Only comment changes after PR_Github #11110 [ run ] completed with state SUCCESS"

[tensorrt-cicd]

PR_Github #11155 [ skip ] triggered by Bot

[tensorrt-cicd]

PR_Github #11155 [ skip ] completed with state SUCCESS
Skipping testing for commit ca3cdcf

[yuantailing]

Known issue: if the number of CUDA architecutes that ≥9.0 is an even number, the following error occurs:

-- deep_ep DEEP_EP_CUDA_ARCHITECTURES: 90-real;100-real
-- deep_ep DEEP_EP_URL: https://github.com/deepseek-ai/DeepEP/archive/c381dadf43a85062f6a8947592017ee513abc70b.tar.gz
CMake Error at tensorrt_llm/deep_ep/CMakeLists.txt:179 (set_target_properties):
  set_target_properties called with incorrect number of arguments.

The root cause is ${DEEP_EP_CUDA_ARCHITECTURES} is not quoted. The CI passed because there are 1 or 3 archs on CI.

This problem will be fixed by #5816 .