[TRTLLM-7280][test] Add beam search CudaGraph + Overlap Scheduler tests #7326
Conversation
Signed-off-by: FredricZ-2007 <[email protected]>
Signed-off-by: FredricZ-2007 <[email protected]>
Signed-off-by: FredricZ-2007 <[email protected]>
Signed-off-by: FredricZ-2007 <[email protected]>
📝 WalkthroughWalkthroughAdds beam-search coverage for meta-llama/Llama-3.1-8B-Instruct: four new CNN/DailyMail accuracy entries for beam_width=2 (CUDA-graph and overlap-scheduler, with/without FP8), a new TestBeamSearch test class with four tests, and corresponding QA test-list entries. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
participant T as TestBeamSearch
participant L as LLM API
participant S as Sampler (BeamSearch)
participant O as Overlap Scheduler
participant G as CUDA Graph Runtime
Note over T: configure max_beam_width=2, SamplingParams(beam_search=true)
T->>L: load model & kv cache config
T->>L: run CnnDailymail evaluation
L->>S: request token generation (beam_search)
alt CUDA graph enabled
S->>G: capture / launch CUDA graph (beam_width=2)
else CUDA graph disabled
S->>S: regular kernel launches
end
opt Overlap scheduler enabled
S->>O: schedule overlap of compute and comms
end
S-->>L: return beams & scores
L-->>T: outputs for accuracy check
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes Possibly related PRs
Suggested labels
Suggested reviewers
✨ 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. 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
SupportNeed help? Create a ticket on our support page for assistance with any issues or questions. CodeRabbit Commands (Invoked using PR/Issue comments)Type Other keywords and placeholders
Status, Documentation and Community
|
|
/bot run --skip-test |
|
PR_Github #16780 [ run ] triggered by Bot |
![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 (2)
tests/integration/defs/accuracy/test_llm_api_pytorch.py (2)
2871-2873: Trim CUDA graph capture sizes to reduce warmup/capture time.Max batch size is 2; capturing 4 and 8 is unnecessary overhead here.
- cuda_graph_config=CudaGraphConfig( - batch_sizes=[1, 2, 4, 8], enable_padding=enable_padding), + cuda_graph_config=CudaGraphConfig( + batch_sizes=[1, 2], enable_padding=enable_padding),
2854-2855: KV cache sizing: consider using fraction-based sizing for portability.Hard-coding max_tokens=10000 can be risky on smaller GPUs. Other tests use free_gpu_memory_fraction to adapt.
- kv_cache_config = KvCacheConfig(max_tokens=10000) + kv_cache_config = KvCacheConfig(free_gpu_memory_fraction=0.8)
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
💡 Knowledge Base configuration:
- MCP integration is disabled by default for public repositories
- Jira integration is disabled by default for public repositories
- Linear integration is disabled by default for public repositories
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (4)
tests/integration/defs/accuracy/references/cnn_dailymail.yaml(1 hunks)tests/integration/defs/accuracy/test_llm_api_pytorch.py(1 hunks)tests/integration/test_lists/qa/llm_function_full.txt(1 hunks)tests/integration/test_lists/qa/llm_function_sanity.txt(1 hunks)
🧰 Additional context used
📓 Path-based instructions (2)
**/*.py
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
**/*.py: Code must target Python 3.8+
Indent Python code with 4 spaces; do not use tabs
Preserve module namespaces when importing; import modules/packages and access members via the module (e.g., from package.subpackage import foo; foo.SomeClass())
Python file names should be snake_case
Python class names should be PascalCase
Python functions/methods and local variables should be snake_case; variables beginning with a number should be prefixed with k_ (e.g., k_99th_percentile)
Global variables should be UPPER_SNAKE_CASE prefixed with G_ (e.g., G_MY_GLOBAL); constants should be UPPER_SNAKE_CASE
Avoid shadowing variables from outer scopes; initialize all externally visible members in init
Prefer docstrings for interfaces used outside a file; comments should be reserved for in-function or file-local interfaces
Use Google-style docstrings for classes and functions; attributes and variables may be documented inline with trailing string literals
Avoid reflection when simpler, explicit code suffices (e.g., avoid dict(**locals()) patterns)
In try/except, catch the narrowest exceptions possible
For duck-typing patterns, keep the try body minimal and move logic to else to avoid masking unrelated failures
Files:
tests/integration/defs/accuracy/test_llm_api_pytorch.py
**/*.{c,cc,cpp,cxx,h,hh,hpp,hxx,cu,cuh,py}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
Prepend the NVIDIA copyright header (current year) to all source files (.cpp, .h, .cu, .py, etc.)
Files:
tests/integration/defs/accuracy/test_llm_api_pytorch.py
🧠 Learnings (3)
📓 Common learnings
Learnt from: dcampora
PR: NVIDIA/TensorRT-LLM#6867
File: tensorrt_llm/_torch/pyexecutor/sampler.py:67-72
Timestamp: 2025-08-13T16:20:37.987Z
Learning: In TensorRT-LLM sampler code, performance is prioritized over additional validation checks. The beam_width helper method intentionally returns the first request's beam_width without validating consistency across all requests to avoid performance overhead from iterating through the entire batch.
Learnt from: eopXD
PR: NVIDIA/TensorRT-LLM#6767
File: cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp:0-0
Timestamp: 2025-08-15T06:46:53.813Z
Learning: In the TensorRT-LLM KV cache manager, SWA (Sliding Window Attention) combined with beam search is currently in a broken/non-functional state and is planned for future rework. During preparatory refactoring phases, code related to SWA+beam search may intentionally remain in a non-working state until the broader rework is completed.
📚 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/qa/llm_function_full.txt
📚 Learning: 2025-08-26T09:49:04.956Z
Learnt from: pengbowang-nv
PR: NVIDIA/TensorRT-LLM#7192
File: tests/integration/test_lists/test-db/l0_dgx_b200.yml:56-72
Timestamp: 2025-08-26T09:49:04.956Z
Learning: In TensorRT-LLM test configuration files, the test scheduling system handles wildcard matching with special rules that prevent duplicate test execution even when the same tests appear in multiple yaml files with overlapping GPU wildcards (e.g., "*b200*" and "*gb200*").
Applied to files:
tests/integration/test_lists/qa/llm_function_full.txt
🧬 Code graph analysis (1)
tests/integration/defs/accuracy/test_llm_api_pytorch.py (5)
tests/integration/defs/accuracy/accuracy_core.py (4)
LlmapiAccuracyTestHarness(788-799)CnnDailymail(209-226)evaluate(147-206)evaluate(707-717)tests/integration/defs/conftest.py (2)
llm_models_root(77-83)parametrize_with_ids(1786-1811)tensorrt_llm/llmapi/llm_args.py (2)
KvCacheConfig(946-1077)CudaGraphConfig(108-165)tensorrt_llm/sampling_params.py (1)
SamplingParams(125-486)tensorrt_llm/quantization/mode.py (1)
QuantAlgo(23-46)
⏰ 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 (5)
tests/integration/defs/accuracy/references/cnn_dailymail.yaml (1)
156-165: Beam-search references for Llama-3.1-8B-Instruct look consistent with new tests.Entries match the new extra_acc_spec strings for cuda_graph and overlap_scheduler at beam_width=2, with and without FP8. LGTM.
tests/integration/test_lists/qa/llm_function_sanity.txt (1)
146-155: Sanity list additions align with parametrized IDs.Eight cuda_graph+overlap_scheduler beam-search cases (enable_padding × disable_overlap_scheduler) and two overlap-only cases are correctly enumerated. LGTM.
tests/integration/defs/accuracy/test_llm_api_pytorch.py (2)
2875-2878: Confirm reference tag choice when overlap scheduler is toggled.The cuda_graph tests toggle disable_overlap_scheduler but always compare against extra_acc_spec="type=cuda_graph,beam_width=2". If overlap scheduling can affect outputs, consider adding distinct references (e.g., "type=cuda_graph+overlap_scheduler,...") or fix overlap disable to a single setting. If outputs are identical (expected), keep as-is.
2850-2945: New BeamSearch coverage is well-scoped and maps cleanly to cnn_dailymail references.
- Correct SamplingParams (n=best_of=max_beam_width, use_beam_search=True).
- Proper max_beam_width and batch sizing.
- FP8 variants assert quant algo.
tests/integration/test_lists/qa/llm_function_full.txt (1)
552-561: Full list entries correctly mirror the new TestBeamSearch cases.All 10 combinations are present and named consistently with auto-generated IDs. LGTM.
fredricz-20070104
changed the title
Signed-off-by: FredricZ-2007 <[email protected]>
|
PR_Github #16780 [ run ] completed with state |
Signed-off-by: FredricZ-2007 <[email protected]>
Signed-off-by: FredricZ-2007 <[email protected]>
Signed-off-by: FredricZ-2007 <[email protected]>
Signed-off-by: FredricZ-2007 <[email protected]>
|
/bot reuse-pipeline |
|
PR_Github #16952 [ reuse-pipeline ] triggered by Bot |
|
PR_Github #16952 [ reuse-pipeline ] completed with state |
…ts (NVIDIA#7326) Signed-off-by: FredricZ-2007 <[email protected]>
Add functional test cases(Accuracy) for
Beam Search + Cudagraph + Overlap Scheduler.
Summary by CodeRabbit
New Features
Bug Fixes
Tests
Chores
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.
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.