[TRTLLM-9175][test] ensure sampling is async #9076
Conversation
|
/bot run |
📝 WalkthroughWalkthroughThe pull request introduces test infrastructure enhancements for controlled warmup behavior and CUDA synchronization validation. It refactors test harnesses across sampling tests to use a new parametrized warmup runner, adds data structures for sampling parameter and result tracking, introduces backend instrumentation for FlashInfer testing, and improves timeout diagnostics using faulthandler. Changes
Sequence Diagram(s)sequenceDiagram
participant Test as Test Runner
participant WarmupRunner as _run_test_with_warmup
participant UUT as Unit Under Test
participant Backend as Sampling Backend
participant Checker as Sync/Batch Checker
Test->>WarmupRunner: Provide UUT, params, is_warmup flag
WarmupRunner->>WarmupRunner: Warmup phase (if enabled)
WarmupRunner->>Checker: Setup interception
Note over Checker: Inject batching/sync checks
WarmupRunner->>UUT: Execute with warmup context
UUT->>Backend: Call sampling
Backend->>Checker: Intercept sampling call
Checker->>Checker: Log sampling inputs<br/>(top_k, top_p, temp, probs)
Checker->>Backend: Validate propagation
Backend-->>UUT: Return samples
UUT-->>WarmupRunner: Return result
WarmupRunner->>Checker: Collect logged inputs
WarmupRunner-->>Test: Return wrapped result
Estimated code review effort🎯 4 (Complex) | ⏱️ ~50 minutes
Pre-merge checks and finishing touches❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✨ Finishing touches
🧪 Generate unit tests (beta)
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: 1
🧹 Nitpick comments (1)
tests/unittest/_torch/sampler/test_torch_sampler.py (1)
362-402: Consider documenting the memory allocation strategy.The
_run_test_with_warmupfunction allocates large buffers (default 4GB) during warmup to prevent PyTorch's caching allocator from triggering device synchronization. While this is a sound strategy, consider adding a note in the docstring about when users might need to adjustwarmup_sizes_bytesfor GPUs with limited memory.
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (2)
tests/unittest/_torch/sampler/test_torch_sampler.py(16 hunks)tests/unittest/utils/util.py(4 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:
tests/unittest/_torch/sampler/test_torch_sampler.pytests/unittest/utils/util.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/_torch/sampler/test_torch_sampler.pytests/unittest/utils/util.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/_torch/sampler/test_torch_sampler.pytests/unittest/utils/util.py
🧬 Code graph analysis (1)
tests/unittest/_torch/sampler/test_torch_sampler.py (4)
tests/unittest/utils/util.py (1)
assert_no_cuda_sync(503-537)tensorrt_llm/_torch/pyexecutor/sampler.py (3)
get_generator(660-674)_request_get_sampling_params(252-262)_unbatch_sampling_results(1370-1412)tensorrt_llm/_torch/pyexecutor/scheduler.py (1)
ScheduledRequests(20-41)tensorrt_llm/_torch/pyexecutor/sampling_utils_flashinfer.py (4)
sample_grouped_strategies(555-576)FlashInferGroupedStrategySampler(515-576)_StrategyImpls(46-512)StrategyImpl(47-169)
🪛 Ruff (0.14.4)
tests/unittest/_torch/sampler/test_torch_sampler.py
423-423: Unused function argument: is_warmup
(ARG001)
1278-1278: zip() without an explicit strict= parameter
Add explicit value for parameter strict=
(B905)
2173-2173: Unused function argument: is_warmup
(ARG001)
⏰ 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 (11)
tests/unittest/utils/util.py (4)
15-15: LGTM!The faulthandler import is appropriate for the enhanced timeout diagnostics in
assert_no_cuda_sync.
489-494: LGTM!The signature formatting with trailing commas and explicit keyword-only parameters improves readability and maintainability.
503-534: LGTM!The enhanced CUDA synchronization detection mechanism combines Python's faulthandler with PyTorch's sync debug mode, providing comprehensive coverage for both general and PyTorch-specific synchronization issues. The 2x timeout margin ensures the device sleep outlasts legitimate async operations.
476-476: No breaking changes detected—the change is safe.All instantiations of
DeviceSleepCtluse zero arguments (DeviceSleepCtl()), which is compatible withkw_only=Truebecause the single field_cancellation_requestedhas a default value. No code attempts to use positional arguments.tests/unittest/_torch/sampler/test_torch_sampler.py (7)
15-30: LGTM!The expanded imports support the new test infrastructure with protocols, context managers, and comprehensive type annotations.
423-559: LGTM!The refactoring to use the new
_run_test_with_warmuppattern correctly integrates the warmup and sync checking infrastructure. The unusedis_warmupparameter flagged by static analysis is required by theUutProviderprotocol and can be safely ignored.
1114-1148: LGTM!The integration of synchronization checking into the
_samplemethod is well-designed. Checking only the first iteration prevents queue saturation issues while still validating async behavior.
1240-1400: LGTM!The refactoring of
test_probscorrectly integrates the warmup infrastructure. Using separate request objects for warmup prevents state pollution from attached probability outputs.
1402-1844: LGTM!The new helper methods provide comprehensive test instrumentation. The backend patching strategy and statistical validation using G-tests are appropriate for validating sampling correctness.
1932-2073: LGTM!The refactoring of
test_samplescorrectly integrates the warmup infrastructure while maintaining comprehensive sampling validation through statistical tests.
2172-2266: LGTM!The refactoring of
test_unbatch_sampling_resultscorrectly integrates the warmup and sync checking infrastructure. The unusedis_warmupparameter flagged by static analysis is required by theUutProviderprotocol.
|
PR_Github #24206 [ run ] triggered by Bot. Commit: |
ixlmar
force-pushed
the
feat/async-sampling-tests
branch
from
b389bfa to
31880b6
Compare
|
/bot run |
|
PR_Github #24207 [ run ] triggered by Bot. Commit: |
|
PR_Github #24206 [ run ] completed with state |
Signed-off-by: ixlmar <206748156+ixlmar@users.noreply.github.com>
ixlmar
force-pushed
the
feat/async-sampling-tests
branch
from
31880b6 to
aa376d6
Compare
|
/bot run |
|
PR_Github #24209 [ run ] triggered by Bot. Commit: |
|
PR_Github #24207 [ run ] completed with state |
|
PR_Github #24209 [ run ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #24281 [ run ] triggered by Bot. Commit: |
|
PR_Github #24281 [ run ] completed with state |
Funatiq
left a comment
Funatiq
left a comment
There was a problem hiding this comment.
LGTM. Thanks for breaking down the tests and adding comments.
3 participants
Description
Extends unit tests to check that there are no device syncs in
sample_async.Opportunistic refactor of
TestBatchedSampling.test_samples, as suggested in [TRTLLM-8377][test] unit tests for TorchSampler batched sampling #9012 (review). Consider reviewing using something likegit diff --color=always --color-moved=plain --color-moved-ws=allow-indentation-change HEAD^, since the PR actually only changes ~300 LoC.Test Coverage
PR improves existing tests.
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 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.
Summary by CodeRabbit