[TRTLLM-7989][infra] Bundle UCX and NIXL libs in the TRTLLM python package #7766
Conversation
|
/bot run --stage-list "Build-Docker-Images" |
|
PR_Github #18777 [ run ] triggered by Bot |
📝 WalkthroughWalkthroughAdds optional NIXL discovery in CMake when UCX is enabled, introduces default NIXL_ROOT and non-fatal failure in FindNIXL, simplifies NIXL utils CMake gating, moves UCX/NIXL/etcd installs earlier in Docker flow, switches from LD_LIBRARY_PATH to patchelf RPATH patching, vendors UCX/NIXL into wheels, updates package_data, and parameterizes tests. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
participant Dev as build_wheel.py
participant Sys as System Paths
participant PELF as patchelf
participant Pkg as Wheel Package (libs)
rect rgb(245,245,255)
note over Dev,Pkg: UCX wrapper handling
Dev->>Pkg: Copy libtensorrt_llm_ucx_wrapper.so
Dev->>PELF: set RPATH of UCX wrapper to $ORIGIN/ucx/
alt /usr/local/ucx exists
Dev->>Pkg: Clear libs/ucx/*
Dev->>Sys: Read /usr/local/ucx/lib/*.so*
Dev->>Pkg: Vendor into libs/ucx/
end
end
rect rgb(245,255,245)
note over Dev,Pkg: NIXL wrapper handling
Dev->>Pkg: Copy libtensorrt_llm_nixl_wrapper.so
Dev->>PELF: set RPATH of NIXL wrapper to $ORIGIN/nixl/
alt /opt/nvidia/nvda_nixl exists
Dev->>Pkg: Clear libs/nixl/*
Dev->>Sys: Read nvda_nixl/lib/x86_64-linux-gnu or aarch64-linux-gnu
Dev->>Pkg: Vendor into libs/nixl/
end
end
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Pre-merge checks and finishing touches❌ Failed checks (2 warnings)
✅ Passed checks (1 passed)
✨ Finishing touches
🧪 Generate unit tests
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 (2)
cpp/cmake/modules/FindNIXL.cmake (1)
69-77: Imported target misconfigured: IMPORTED_LOCATION must be a single file.You’re assigning three libs to IMPORTED_LOCATION. Link the helper libs via INTERFACE_LINK_LIBRARIES instead.
if(NIXL_FOUND) if(NOT TARGET NIXL::nixl) add_library(NIXL::nixl SHARED IMPORTED) - set_target_properties( - NIXL::nixl - PROPERTIES INTERFACE_INCLUDE_DIRECTORIES ${NIXL_INCLUDE_DIR} - IMPORTED_LOCATION ${NIXL_LIBRARY} ${NIXL_BUILD_LIBRARY} - ${SERDES_LIBRARY}) + set_target_properties(NIXL::nixl PROPERTIES + INTERFACE_INCLUDE_DIRECTORIES "${NIXL_INCLUDE_DIR}" + IMPORTED_LOCATION "${NIXL_LIBRARY}") + # Link companion libs through interface + target_link_libraries(NIXL::nixl INTERFACE + "${NIXL_BUILD_LIBRARY}" "${SERDES_LIBRARY}" + ${UCX_BACKEND_LIBRARY} ${UCX_UTILS_LIBRARY}) endif() else()setup.py (1)
205-210: Fix Python version compatibility mismatch (PEP 604 unions vs python_requires).
str | Nonerequires Python 3.10+, butpython_requires=">=3.7"will break source installs on 3.7–3.9. Either raisepython_requiresto >=3.10 or switch toOptional[str].Apply the following updates (outside this hunk):
- from typing import List + from typing import List, Optional ... - precompiled: str | None = os.getenv("TRTLLM_USE_PRECOMPILED") - precompiled_location: str | None = os.getenv("TRTLLM_PRECOMPILED_LOCATION") + precompiled: Optional[str] = os.getenv("TRTLLM_USE_PRECOMPILED") + precompiled_location: Optional[str] = os.getenv("TRTLLM_PRECOMPILED_LOCATION")Alternatively (if 3.10+ is intended), update:
- python_requires=">=3.7, <4") + python_requires=">=3.10, <4")Also applies to: 263-263
🧹 Nitpick comments (9)
tests/unittest/test_pip_install.py (1)
58-60: Make installs deterministic and reduce layers by updating/combining apt installs.Add apt-get update and combine related packages with --no-install-recommends to avoid transient failures and bloat.
-subprocess.check_call("apt-get -y install libzmq3-dev", shell=True) -subprocess.check_call("apt-get -y install python3-pip", shell=True) +subprocess.check_call( + "apt-get update && apt-get -y install --no-install-recommends libzmq3-dev python3-pip", + shell=True)Note: Ruff S602/S607 applies broadly in this file; fine to defer a broader refactor.
docker/common/install_nixl.sh (1)
41-42: Guard patchelf usage.Script assumes patchelf is preinstalled; fail fast with a clear message to avoid silent runtime link issues.
-cd builddir && ninja install - -find /opt/nvidia/nvda_nixl/lib/${ARCH_NAME} -type f -name "*.so*" -exec patchelf --set-rpath '$ORIGIN:$ORIGIN/plugins:$ORIGIN/ucx:$ORIGIN/../' {} \; +cd builddir && ninja install + +# Ensure patchelf is available +if ! command -v patchelf >/dev/null 2>&1; then + echo "ERROR: patchelf not found. Please install patchelf in the base image before running install_nixl.sh." + exit 1 +fi + +find /opt/nvidia/nvda_nixl/lib/${ARCH_NAME} -type f -name "*.so*" \ + -exec patchelf --set-rpath '$ORIGIN:$ORIGIN/plugins:$ORIGIN/ucx:$ORIGIN/../' {} \;docker/common/install_ucx.sh (1)
30-31: Also guard patchelf; mirror NIXL script behavior.Avoid opaque failures when the tool is missing in some bases.
-rm -rf ucx # Remove UCX source to save space - -find ${UCX_INSTALL_PATH}/lib -type f -name "*.so*" -exec patchelf --set-rpath '$ORIGIN:$ORIGIN/plugins:$ORIGIN/ucx:$ORIGIN/../' {} \; +rm -rf ucx # Remove UCX source to save space + +# Ensure patchelf is available +if ! command -v patchelf >/dev/null 2>&1; then + echo "ERROR: patchelf not found. Please install patchelf in the base image before running install_ucx.sh." + exit 1 +fi + +find ${UCX_INSTALL_PATH}/lib -type f -name "*.so*" \ + -exec patchelf --set-rpath '$ORIGIN:$ORIGIN/plugins:$ORIGIN/ucx:$ORIGIN/../' {} \;docker/Dockerfile.multi (1)
92-103: Install patchelf before invoking UCX/NIXL installers.Both scripts now require patchelf; install it once here with a simple distro fallback.
# Install UCX first COPY docker/common/install_ucx.sh install_ucx.sh -RUN bash ./install_ucx.sh && rm install_ucx.sh +RUN (command -v patchelf >/dev/null 2>&1) || \ + ( (command -v apt-get >/dev/null && apt-get update && apt-get install -y patchelf) || \ + (command -v dnf >/dev/null && dnf install -y patchelf) || \ + (command -v yum >/dev/null && yum install -y patchelf) || \ + (echo "ERROR: patchelf not found and package manager unknown" && exit 1) ) +RUN bash ./install_ucx.sh && rm install_ucx.sh # Install NIXL COPY docker/common/install_nixl.sh install_nixl.sh RUN bash ./install_nixl.sh && rm install_nixl.shcpp/cmake/modules/FindNIXL.cmake (2)
26-35: Defaulting NIXL_ROOT: confirm FORCE semantics are desired.Using CACHE ... FORCE will overwrite existing cache entries if NIXL_ROOT evaluates falsey. If users pass an empty string to force auto-discovery, this will override it. Consider dropping FORCE or using set(... CACHE ... FORCE) only when first configure.
62-67: Duplicate find_package_handle_standard_args call; remove the second.There’s no install step between the two calls; keeping both is redundant and noisy.
-# Re-attempt to find NIXL after installation -find_package_handle_standard_args( - NIXL - FOUND_VAR NIXL_FOUND - REQUIRED_VARS NIXL_INCLUDE_DIR NIXL_LIBRARY NIXL_BUILD_LIBRARY SERDES_LIBRARY)scripts/build_wheel.py (2)
753-760: Ensure patchelf exists (fail fast) and consider vendoring UCX libs RPATH too.Guard the call so wheel builds fail with a clear error if patchelf is missing.
- build_run( - f'patchelf --set-rpath \'$ORIGIN/ucx/\' {lib_dir / "libtensorrt_llm_ucx_wrapper.so"}' - ) + build_run( + f'(command -v patchelf >/dev/null 2>&1) || ' + f'(echo "ERROR: patchelf not found; required to set RPATH on UCX wrapper" && exit 1); ' + f'patchelf --set-rpath \'$ORIGIN/ucx/\' {lib_dir / "libtensorrt_llm_ucx_wrapper.so"}' + )Optionally, for robustness outside the Docker build (where UCX/NIXL libs may not have been pre-patched), also iterate over copied libs under libs/ucx and set their RPATH to $ORIGIN:$ORIGIN/../ as done in the Docker install scripts.
769-778: Same guard for NIXL wrapper; also patch vendored NIXL libs if present.Mirror the UCX change and optionally patch libs/nixl/* to ensure self-contained wheels.
- build_run( - f'patchelf --set-rpath \'$ORIGIN/nixl/\' {lib_dir / "libtensorrt_llm_nixl_wrapper.so"}' - ) + build_run( + f'(command -v patchelf >/dev/null 2>&1) || ' + f'(echo "ERROR: patchelf not found; required to set RPATH on NIXL wrapper" && exit 1); ' + f'patchelf --set-rpath \'$ORIGIN/nixl/\' {lib_dir / "libtensorrt_llm_nixl_wrapper.so"}' + )tests/unittest/others/test_kv_cache_transceiver.py (1)
80-82: Skip gracefully when a backend isn’t available.In some environments only one backend may be built/packaged; make this test auto-skip instead of failing hard.
Apply this diff:
- cache_transceiver_config = trtllm.CacheTransceiverConfig( - backend=backend, max_tokens_in_buffer=512) + try: + cache_transceiver_config = trtllm.CacheTransceiverConfig( + backend=backend, max_tokens_in_buffer=512) + except Exception as e: + pytest.skip(f"{backend} backend not available: {e}")Optionally also guard transceiver construction (outside this hunk):
# around lines creating kv_cache_transceiver_ctx/gen try: kv_cache_transceiver_ctx = create_kv_cache_transceiver( mapping, dist, kv_cache_manager_ctx, attention_type, cache_transceiver_config) kv_cache_transceiver_gen = create_kv_cache_transceiver( mapping, dist, kv_cache_manager_gen, attention_type, cache_transceiver_config) except Exception as e: pytest.skip(f"{backend} backend not available during creation: {e}")
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (11)
cpp/CMakeLists.txt(1 hunks)cpp/cmake/modules/FindNIXL.cmake(2 hunks)cpp/tensorrt_llm/executor/cache_transmission/nixl_utils/CMakeLists.txt(0 hunks)docker/Dockerfile.multi(1 hunks)docker/common/install_nixl.sh(1 hunks)docker/common/install_ucx.sh(1 hunks)scripts/build_wheel.py(2 hunks)setup.py(1 hunks)tests/integration/test_lists/test-db/l0_sanity_check.yml(1 hunks)tests/unittest/others/test_kv_cache_transceiver.py(1 hunks)tests/unittest/test_pip_install.py(1 hunks)
💤 Files with no reviewable changes (1)
- cpp/tensorrt_llm/executor/cache_transmission/nixl_utils/CMakeLists.txt
🧰 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:
tests/unittest/test_pip_install.pytests/unittest/others/test_kv_cache_transceiver.pyscripts/build_wheel.pysetup.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:
tests/unittest/test_pip_install.pytests/unittest/others/test_kv_cache_transceiver.pyscripts/build_wheel.pysetup.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:
tests/unittest/test_pip_install.pytests/unittest/others/test_kv_cache_transceiver.pyscripts/build_wheel.pysetup.py
🧠 Learnings (7)
📓 Common learnings
Learnt from: venkywonka
PR: NVIDIA/TensorRT-LLM#6029
File: .github/pull_request_template.md:45-53
Timestamp: 2025-08-27T17:50:13.264Z
Learning: For PR templates in TensorRT-LLM, avoid suggesting changes that would increase developer overhead, such as converting plain bullets to mandatory checkboxes. The team prefers guidance-style bullets that don't require explicit interaction to reduce friction in the PR creation process.
📚 Learning: 2025-07-28T17:06:08.621Z
Learnt from: moraxu
PR: NVIDIA/TensorRT-LLM#6303
File: tests/integration/test_lists/qa/examples_test_list.txt:494-494
Timestamp: 2025-07-28T17:06:08.621Z
Learning: In TensorRT-LLM testing, it's common to have both CLI flow tests (test_cli_flow.py) and PyTorch API tests (test_llm_api_pytorch.py) for the same model. These serve different purposes: CLI flow tests validate the traditional command-line workflow, while PyTorch API tests validate the newer LLM API backend. Both are legitimate and should coexist.
Applied to files:
tests/integration/test_lists/test-db/l0_sanity_check.yml
📚 Learning: 2025-09-09T09:40:45.658Z
Learnt from: fredricz-20070104
PR: NVIDIA/TensorRT-LLM#7645
File: tests/integration/test_lists/qa/llm_function_core.txt:648-648
Timestamp: 2025-09-09T09:40:45.658Z
Learning: In TensorRT-LLM test lists, it's common and intentional for the same test to appear in multiple test list files when they serve different purposes (e.g., llm_function_core.txt for comprehensive core functionality testing and llm_function_core_sanity.txt for quick sanity checks). This duplication allows tests to be run in different testing contexts.
Applied to files:
tests/integration/test_lists/test-db/l0_sanity_check.yml
📚 Learning: 2025-08-18T09:08:07.687Z
Learnt from: tongyuantongyu
PR: NVIDIA/TensorRT-LLM#6984
File: cpp/tensorrt_llm/CMakeLists.txt:297-299
Timestamp: 2025-08-18T09:08:07.687Z
Learning: In the TensorRT-LLM project, artifacts are manually copied rather than installed via `cmake --install`, so INSTALL_RPATH properties are not needed - only BUILD_RPATH affects the final artifacts.
Applied to files:
scripts/build_wheel.py
📚 Learning: 2025-08-06T13:58:07.506Z
Learnt from: galagam
PR: NVIDIA/TensorRT-LLM#6487
File: tests/unittest/_torch/auto_deploy/unit/singlegpu/test_ad_trtllm_bench.py:1-12
Timestamp: 2025-08-06T13:58:07.506Z
Learning: In TensorRT-LLM, test files (files under tests/ directories) do not require NVIDIA copyright headers, unlike production source code files. Test files typically start directly with imports, docstrings, or code.
Applied to files:
setup.py
📚 Learning: 2025-09-16T09:30:09.686Z
Learnt from: tongyuantongyu
PR: NVIDIA/TensorRT-LLM#7763
File: cpp/tensorrt_llm/CMakeLists.txt:297-301
Timestamp: 2025-09-16T09:30:09.686Z
Learning: In the TensorRT-LLM project, NCCL libraries are loaded earlier by PyTorch libraries or the bindings library, so the main shared library doesn't need NCCL paths in its RPATH - the libraries will already be available in the process address space when needed.
Applied to files:
setup.py
📚 Learning: 2025-08-01T15:14:45.673Z
Learnt from: yibinl-nvidia
PR: NVIDIA/TensorRT-LLM#6506
File: examples/models/core/mixtral/requirements.txt:3-3
Timestamp: 2025-08-01T15:14:45.673Z
Learning: In TensorRT-LLM, examples directory can have different dependency versions than the root requirements.txt file. Version conflicts between root and examples dependencies are acceptable because examples are designed to be standalone and self-contained.
Applied to files:
setup.py
🧬 Code graph analysis (1)
tests/unittest/others/test_kv_cache_transceiver.py (1)
tensorrt_llm/mapping.py (1)
Mapping(32-513)
🪛 Ruff (0.12.2)
tests/unittest/test_pip_install.py
58-58: subprocess call with shell=True seems safe, but may be changed in the future; consider rewriting without shell
(S602)
58-58: Starting a process with a partial executable path
(S607)
⏰ 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 (6)
docker/Dockerfile.multi (1)
92-103: Reordering UCX/NIXL/etcd into devel stage: LGTM.This reduces duplication in TRITON stages and ensures libs are available for wheel packaging.
cpp/CMakeLists.txt (1)
561-561: Optional NIXL discovery under ENABLE_UCX: LGTM.Keeps NIXL non-fatal while allowing downstream CMake to gate features by NIXL_FOUND.
tests/unittest/others/test_kv_cache_transceiver.py (2)
67-71: Backend parametrization looks good.Nice addition; ids ordering aligns with integration test nodeids.
73-73: Signature update wired correctly.The new backend arg is plumbed through and used below.
tests/integration/test_lists/test-db/l0_sanity_check.yml (1)
35-36: Sanity matrix extension is fine, but rely on test-level skip for missing backends.Given env variance, the unit test’s skip guard will prevent false reds in CI if NIXL/UCX is absent.
If the runner matches exact nodeids, please confirm the collected id is:
others/test_kv_cache_transceiver.py::test_kv_cache_transceiver_single_process[NIXL-mha-ctx_fp16_gen_fp16]
and likewise for UCX, after a localpytest --collect-onlyrun.setup.py (1)
106-109: Including NIXL/UCX assets in wheels: OK.Globs match the vendoring flow; licenses under libs/* will be bundled.
|
/bot run --stage-list "Build-Docker-Images" |
|
PR_Github #18807 [ run ] triggered by Bot |
|
PR_Github #18777 [ run ] completed with state |
|
/bot run |
|
PR_Github #18820 [ run ] triggered by Bot |
|
PR_Github #18807 [ run ] completed with state |
|
PR_Github #18820 [ run ] completed with state |
bo-nv
force-pushed
the
user/bo/rel-1.2-nixl-infra
branch
from
4b6cd94 to
6bdf9ba
Compare
|
/bot run --stage-list "Build-Docker-Images" |
|
PR_Github #18847 [ run ] triggered by Bot |
|
/bot run --stage-list "Build-Docker-Images" |
|
PR_Github #18862 [ run ] triggered by Bot |
|
PR_Github #18847 [ run ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #18895 [ run ] triggered by Bot |
|
PR_Github #18862 [ run ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #18929 [ run ] triggered by Bot |
|
PR_Github #18895 [ run ] completed with state |
|
PR_Github #18929 [ run ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #19285 [ run ] triggered by Bot |
|
PR_Github #19279 [ run ] completed with state |
Signed-off-by: Bo Deng <deemod@nvidia.com>
|
/bot run --disable-fail-fast |
|
PR_Github #19304 [ run ] triggered by Bot |
|
PR_Github #19285 [ run ] completed with state |
|
PR_Github #19304 [ run ] completed with state |
|
/bot run --skip-test --extra-stage "A10-PackageSanityCheck-PY310-UB2204-CU12" |
|
PR_Github #19421 [ run ] triggered by Bot |
|
/bot run --disable-fail-fast |
|
PR_Github #19421 [ run ] completed with state |
|
PR_Github #19424 [ run ] triggered by Bot |
|
PR_Github #19424 [ run ] completed with state |
|
/bot run |
|
PR_Github #19463 [ run ] triggered by Bot |
|
PR_Github #19463 [ run ] completed with state |
bo-nv
changed the title
|
@Shixiaowei02 @zeroepoch please help review, thanks! |
bo-nv
changed the title
bo-nv
changed the title
bo-nv
changed the title
bo-nv
changed the title
bo-nv
changed the title
…ckage (NVIDIA#7766) Signed-off-by: Bo Deng <deemod@nvidia.com>
…ckage (NVIDIA#7766) Signed-off-by: Bo Deng <deemod@nvidia.com>
…ckage (NVIDIA#7766) Signed-off-by: Bo Deng <deemod@nvidia.com>
3 participants
Summary by CodeRabbit
New Features
Chores
Tests
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
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.