[TRTLLM-6650][fix] Enhance CUDA graph + Beam search to correctly handle padding #6665
Conversation
📝 WalkthroughWalkthroughThe changes introduce a Changes
Sequence Diagram(s)sequenceDiagram
participant Test as Test Suite
participant Engine as PyTorchModelEngine
participant Runner as DecodingCUDAGraphRunner
Test->>Engine: Prepare batch with num_prompts and max_beam_width
Engine->>Runner: Initialize with batch_size, max_beam_width
Runner->>Runner: Allocate input_ids, position_ids with scaled shapes
Engine->>Engine: Prepare inputs, handle dummy requests
Engine->>Runner: Execute CUDA graph with prepared tensors
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes Possibly related PRs
Suggested reviewers
Note ⚡️ Unit Test Generation is now available in beta!Learn more here, or try it out under "Finishing Touches" below. 📜 Recent review detailsConfiguration used: .coderabbit.yaml 📒 Files selected for processing (3)
🚧 Files skipped from review as they are similar to previous changes (3)
⏰ 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)
✨ 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 comments)
Other keywords and placeholders
Documentation and Community
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (3)
tensorrt_llm/_torch/pyexecutor/cuda_graph_runner.py(2 hunks)tensorrt_llm/_torch/pyexecutor/model_engine.py(4 hunks)tests/unittest/_torch/test_beam_search.py(3 hunks)
🧰 Additional context used
📓 Path-based instructions (2)
**/*.py
📄 CodeRabbit Inference Engine (CODING_GUIDELINES.md)
**/*.py: The code developed for TensorRT-LLM should conform to Python 3.8+.
Indent Python code with 4 spaces. Do not use tabs.
Always maintain the namespace when importing in Python, even if only one class or function from a module is used.
Python filenames should use snake_case (e.g., some_file.py).
Python classes should use PascalCase (e.g., class SomeClass).
Python functions and methods should use snake_case (e.g., def my_awesome_function():).
Python local variables should use snake_case. Prefix k for variable names that start with a number (e.g., k_99th_percentile = ...).
Python global variables should use upper snake_case and prefix G (e.g., G_MY_GLOBAL = ...).
Python constants should use upper snake_case (e.g., MY_CONSTANT = ...).
Avoid shadowing variables declared in an outer scope in Python.
Initialize all externally visible members of a class in the constructor in Python.
For interfaces that may be used outside a file, prefer docstrings over comments in Python.
Comments in Python should be reserved for code within a function, or interfaces that are local to a file.
Use Google style docstrings for classes and functions in Python, which can be parsed by Sphinx.
Attributes and variables in Python can be documented inline; attribute docstrings will be rendered under the docstring for the class.
Avoid using reflection in Python when functionality can be easily achieved without it.
When using try-except blocks in Python, limit the except to the smallest set of errors possible.
When using try-except blocks to handle multiple possible variable types in Python, keep the body of the try as small as possible, using the else block to implement the logic.
Files:
tensorrt_llm/_torch/pyexecutor/cuda_graph_runner.pytensorrt_llm/_torch/pyexecutor/model_engine.pytests/unittest/_torch/test_beam_search.py
**/*.{cpp,h,hpp,cc,cxx,cu,py}
📄 CodeRabbit Inference Engine (CODING_GUIDELINES.md)
All TensorRT-LLM Open Source Software code should contain an NVIDIA copyright header that includes the current year. This includes .cpp, .h, .cu, .py, and any other source files which are compiled or interpreted.
Files:
tensorrt_llm/_torch/pyexecutor/cuda_graph_runner.pytensorrt_llm/_torch/pyexecutor/model_engine.pytests/unittest/_torch/test_beam_search.py
🧠 Learnings (3)
📚 Learning: in tensorrt-llm's multimodal processing pipeline, shared tensor recovery using `from_shared_tensor()...
Learnt from: yechank-nvidia
PR: NVIDIA/TensorRT-LLM#6254
File: tensorrt_llm/_torch/pyexecutor/model_engine.py:1201-1204
Timestamp: 2025-07-22T09:22:14.726Z
Learning: In TensorRT-LLM's multimodal processing pipeline, shared tensor recovery using `from_shared_tensor()` is only needed during the context phase. Generation requests reuse the already-recovered tensor data and only need to call `strip_for_generation()` to remove unnecessary multimodal data while preserving the recovered tensors. This avoids redundant tensor recovery operations during generation.
Applied to files:
tensorrt_llm/_torch/pyexecutor/model_engine.py
📚 Learning: in tensorrt_llm/executor/worker.py, the lora adapter cache optimization logic that checks `is_adapte...
Learnt from: amitz-nv
PR: NVIDIA/TensorRT-LLM#5616
File: tensorrt_llm/executor/worker.py:375-384
Timestamp: 2025-07-17T09:01:27.402Z
Learning: In tensorrt_llm/executor/worker.py, the LoRA adapter cache optimization logic that checks `is_adapter_in_cpu_cache()` and conditionally passes None for weights/config has a known race condition issue that cannot be solved with simple error handling or verification checks. This is a known limitation that requires a more comprehensive solution.
Applied to files:
tensorrt_llm/_torch/pyexecutor/model_engine.py
📚 Learning: in tensorrt-llm testing, it's common to have both cli flow tests (test_cli_flow.py) and pytorch api ...
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/unittest/_torch/test_beam_search.py
🔇 Additional comments (7)
tensorrt_llm/_torch/pyexecutor/cuda_graph_runner.py (2)
37-37: LGTM! Themax_beam_widthparameter is properly added.The parameter follows Python naming conventions and has a sensible default value of 1 for non-beam search scenarios.
Also applies to: 53-53
60-67: Tensor shapes correctly updated to account for beam width.The shapes for both
input_idsandposition_idsare consistently updated to includemax_beam_widthin the batch dimension calculation, which is necessary for proper beam search support in CUDA graphs.tests/unittest/_torch/test_beam_search.py (1)
66-66: LGTM! Explicit batch sizes properly configured.The explicit batch sizes
[1, 2, 4, 8]are appropriate for testing CUDA graph behavior with various batch configurations, particularly for validating the padding functionality.tensorrt_llm/_torch/pyexecutor/model_engine.py (4)
845-846: Correct removal of beam width multiplication.The batch size calculation no longer needs to multiply by
max_beam_widthsince this is now handled internally by theDecodingCUDAGraphRunnerwhen creating tensor shapes.
979-982: LGTM! CUDA graph runner instantiation properly updated.The instantiation correctly passes both
batch_sizeandmax_beam_widthas separate parameters, aligning with the updated constructor signature incuda_graph_runner.py.
1376-1379: Good defensive programming to prevent None access errors.Correctly skips adding sequence slots for CUDA graph dummy requests, which prevents potential access errors when these slots are used for cache indirection operations.
1508-1520: Cache indirection correctly updated to handle filtered sequence slots.The logic properly uses the filtered
gen_request_seq_slotslist (which excludes dummy requests) for cache indirection operations, ensuring correct beam search behavior without accessing invalid slots.
stnie
force-pushed
the
develop/beam_search_features/cuda_graph_padding
branch
from
a92b1bd to
5a88215
Compare
- Added max_beam_width parameter to DecodingCUDAGraphRunner for correct input sizing. - Updated input and position ID tensor shapes to accommodate the new beam width. - Adjusted model engine to utilize the max_beam_width when creating CUDA graph runners. - Adjusted model engine to ignore dummy requests when updating cache indirection - Modified unit tests to include additional cases for varying prompt counts and batch sizes. To test CUDA Graph with padding Signed-off-by: Stefan Niebler <[email protected]>
Signed-off-by: Stefan Niebler <[email protected]>
stnie
force-pushed
the
develop/beam_search_features/cuda_graph_padding
branch
from
5a88215 to
cb4dd58
Compare
|
/bot run |
|
PR_Github #14484 [ run ] triggered by Bot |
|
PR_Github #14484 [ run ] completed with state |
5 participants
Summary by CodeRabbit
New Features
Bug Fixes
Tests
Description
Cuda Graph uses dummy requests to pad batches. These were not correctly handled and lead to errors, when using beam search. This PR introduces changes to enhance readability as well as fix beam search + padding requests:
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 [--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.