[https://nvbugs/5680911][fix] Remove @cache decorator from enable_worker_single_process_for_tp1() to fix env var staleness#10098
[https://nvbugs/5680911][fix] Remove @cache decorator from enable_worker_single_process_for_tp1() to fix env var staleness#10098zheyuf wants to merge 45 commits intoNVIDIA:mainfrom
Conversation
Signed-off-by: Zheyu Fu <zheyuf@NVIDIA.com>
zheyuf
force-pushed
the
fix_spec_gate
branch
from
8922ca8 to
e7d542a
Compare
|
/bot run --disable-fail-fast --extra-stage "H100_PCIe-PyTorch-4, B200_PCIe-PyTorch-1" |
|
PR_Github #28862 [ run ] triggered by Bot. Commit: |
|
PR_Github #28862 [ run ] completed with state
|
|
/bot run --disable-fail-fast --extra-stage "H100_PCIe-PyTorch-4, B200_PCIe-PyTorch-1" |
|
PR_Github #28863 [ run ] triggered by Bot. Commit: |
|
PR_Github #28863 [ run ] completed with state
|
|
/bot run --disable-fail-fast |
|
PR_Github #28899 [ run ] triggered by Bot. Commit: |
|
PR_Github #28899 [ run ] completed with state
|
…st_spec_gate. Signed-off-by: Zheyu Fu <zheyuf@NVIDIA.com>
zheyuf
force-pushed
the
fix_spec_gate
branch
from
b48acb4 to
5ab0d1e
Compare
Signed-off-by: Zheyu Fu <zheyuf@nvidia.com>
|
/bot run --disable-fail-fast |
|
PR_Github #29066 [ run ] triggered by Bot. Commit: |
|
PR_Github #29066 [ run ] completed with state
|
Signed-off-by: Zheyu Fu <zheyuf@NVIDIA.com>
Signed-off-by: Zheyu Fu <zheyuf@NVIDIA.com>
Signed-off-by: Zheyu Fu <zheyuf@nvidia.com>
Signed-off-by: Zheyu Fu <zheyuf@NVIDIA.com>
|
/bot run --disable-fail-fast |
zheyuf
changed the title
|
PR_Github #29199 [ run ] triggered by Bot. Commit: |
zheyuf
changed the title
zheyuf
changed the title
|
PR_Github #29199 [ run ] completed with state
|
|
/bot run --disable-fail-fast |
|
PR_Github #29220 [ run ] triggered by Bot. Commit: |
|
PR_Github #31822 [ run ] triggered by Bot. Commit: |
|
PR_Github #31822 [ run ] completed with state
|
|
/bot run |
|
PR_Github #31859 [ run ] triggered by Bot. Commit: |
|
PR_Github #31859 [ run ] completed with state
|
|
/bot run |
|
PR_Github #32014 [ run ] triggered by Bot. Commit: |
|
PR_Github #32014 [ run ] completed with state
|
|
/bot run |
|
PR_Github #32026 [ run ] triggered by Bot. Commit: |
|
PR_Github #32026 [ run ] completed with state
|
Signed-off-by: Zheyu Fu <zheyuf@nvidia.com>
|
/bot run |
|
PR_Github #32176 [ run ] triggered by Bot. Commit: |
|
PR_Github #32176 [ run ] completed with state
|
|
/bot run |
|
PR_Github #32188 [ run ] triggered by Bot. Commit: |
|
This PR is intended to fix bugs for release/1.2 but is currently targeting main. Since the cutoff between 1.2 and 1.3 has already passed, I will close this and open a new PR to submit these changes directly to release/1.2. |
|
PR_Github #32188 [ run ] completed with state
|
3 participants
Summary by CodeRabbit
Tests
Chores
✏️ Tip: You can customize this high-level summary in your review settings.
Description
This PR contains a one-line change that requires review from trt-llm-llmapi-devs.
Problem
The @cache decorator on enable_worker_single_process_for_tp1() causes the function to return a stale cached result instead of re-reading os.environ on each call. This breaks directory-level CI test runs where different unit tests need different process modes.
Root Cause
When CI runs directory tests like:
python -m pytest unittest/_torch/speculative -k not eagle3The first test
(test_draft_len_schedule.py::test_correctness_across_batch_sizes[ngram-schedule0])runs without settingTLLM_WORKER_USE_SINGLE_PROCESS, soenable_worker_single_process_for_tp1()caches False.Later tests like
test_dynamic_spec_decode.py::test_dynamic_spec_decode[True]set the env var to True viamonkeypatch.setenv(), butenable_worker_single_process_for_tp1()always returns the cached False instead of reading the updated env var.Impact
This creates a particularly insidious bug: test_dynamic_spec_decode requires single-process mode. When the test silently runs in multi-process mode due to the stale cache, the mock function inside the unit test has no effect—the test passes without actually validating the intended behavior. In CI directory runs, it becomes a no-op that doesn't catch regressions.
Fix
Remove the @cache decorator. The performance impact is negligible since enable_worker_single_process_for_tp1() is only called once per LLM initialization (not in the inference hot path), and os.environ.get() is a simple dictionary lookup.
NVBugs Fixed in this PR
Additional Changes
Waived test_dynamic_spec_decode (https://nvbugspro.nvidia.com/bug/5758449), which is buggy: Due to the @cache decorator, this test has been silently running in multi-process mode, making it never raise alarms in CI. Now that the decorator is removed and the test would run correctly in single-process mode, it needs to be fixed to actually work as intended.
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
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.
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.