[None][opt] Add batch wait timeout in fetching requests #6923
Conversation
📝 WalkthroughWalkthroughAdds a batch_wait_timeout config (default 0) exposed via TorchLlmArgs and PyTorchConfig, propagated into PyExecutor, and used by ExecutorRequestQueue to optionally wait up to the timeout to accumulate requests into larger batches. ExecutorRequestQueue constructor signature updated. Changes
Sequence Diagram(s)sequenceDiagram
participant Client
participant PyExecutor
participant ExecReqQueue
participant RequestQueue
Client->>PyExecutor: submit requests
PyExecutor->>ExecReqQueue: construct(..., batch_wait_timeout, ...)
loop fetch batch
ExecReqQueue->>RequestQueue: get(timeout=initial)
alt batch_wait_timeout == 0
ExecReqQueue-->>PyExecutor: return items immediately
else batch_wait_timeout > 0
ExecReqQueue->>RequestQueue: get(timeout=remaining) [loop until timeout or max_batch_size]
ExecReqQueue-->>PyExecutor: return accumulated items
end
end
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes Possibly related PRs
Suggested reviewers
Tip 🔌 Remote MCP (Model Context Protocol) integration is now available!Pro plan users can now connect to remote MCP servers from the Integrations page. Connect with popular remote MCPs such as Notion and Linear to add more context to your reviews and chats. 📜 Recent review detailsConfiguration used: .coderabbit.yaml 💡 Knowledge Base configuration:
You can enable these sources in your CodeRabbit configuration. 📒 Files selected for processing (3)
🚧 Files skipped from review as they are similar to previous changes (1)
🧰 Additional context used📓 Path-based instructions (2)**/*.py📄 CodeRabbit Inference Engine (CODING_GUIDELINES.md)
Files:
**/*.{cpp,cxx,cc,cu,h,hpp,hxx,hh,cuh,py}📄 CodeRabbit Inference Engine (CODING_GUIDELINES.md)
Files:
🔇 Additional comments (8)
✨ 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
|
![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
🔭 Outside diff range comments (1)
tensorrt_llm/_torch/pyexecutor/executor_request_queue.py (1)
44-62: Missing assignment: self.max_batch_size not set; will raise AttributeError at runtime.You use self.max_batch_size in _get_from_request_queue, but it isn’t initialized in init. This will crash when batch_wait_timeout > 0.
Apply this diff to store the constructor argument:
def __init__(self, dist: Distributed, enable_attention_dp: bool, max_batch_size: int, max_beam_width: int, max_num_active_requests: int, enable_iter_perf_stats: bool, batch_wait_timeout: float, is_disaggregated: bool): self.dist = dist self.request_queue: queue.Queue[RequestQueueItem] = queue.Queue() self.waiting_queue: deque[RequestQueueItem] = deque() self.canceled_req_ids = [] self.enable_attention_dp = enable_attention_dp self.max_beam_width = max_beam_width self.max_num_active_requests = max_num_active_requests + self.max_batch_size = max_batch_size self.is_disaggregated = is_disaggregated self.enqueue_lock = threading.Lock() self.next_request_id = max_batch_size self.enable_iter_perf_stats = enable_iter_perf_stats self.start_times = {} self.active = True self.batch_wait_timeout = batch_wait_timeout
🧹 Nitpick comments (2)
tensorrt_llm/_torch/pyexecutor/config.py (1)
53-54: Add a brief docstring for the new knob (optional).The addition looks correct. Consider documenting units (seconds) and semantics (0 = no wait) inline to match llm_args’ description.
tensorrt_llm/_torch/pyexecutor/executor_request_queue.py (1)
71-110: Batching loop semantics look good; consider minor edge-case guard (optional).The “accumulate up to max_batch_size or timeout” logic is sound. Once self.max_batch_size is set (see above), this will behave as intended. Optionally, clamp negative batch_wait_timeout to 0 at init, or rely on the llm_args validator (preferred).
📜 Review details
Configuration used: .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 (5)
tensorrt_llm/_torch/pyexecutor/config.py(1 hunks)tensorrt_llm/_torch/pyexecutor/executor_request_queue.py(4 hunks)tensorrt_llm/_torch/pyexecutor/py_executor.py(2 hunks)tensorrt_llm/llmapi/llm_args.py(3 hunks)tests/unittest/api_stability/references/llm.yaml(1 hunks)
🧰 Additional context used
📓 Path-based instructions (2)
**/*.py
📄 CodeRabbit Inference Engine (CODING_GUIDELINES.md)
**/*.py: Python code must target Python 3.8+
Python indentation: 4 spaces, no tabs
Maintain module namespace in imports (from package.subpackage import foo; then use foo.SomeClass())
Python file names use snake_case
Python class names use PascalCase
Python functions/methods and local variables use snake_case; variables starting with a number get k_ prefix (e.g., k_99th_percentile)
Global variables use G_ prefixed UPPER_SNAKE_CASE (e.g., G_MY_GLOBAL)
Constants use UPPER_SNAKE_CASE in Python
Avoid shadowing variables from outer scopes in Python
Initialize all externally visible members of a Python class in init
Prefer docstrings for interfaces used outside a file; comments for local code
Use Google-style docstrings for classes and functions (Sphinx-parsable)
Document attributes/variables inline with short docstrings
Avoid reflection when simple alternatives exist (e.g., prefer explicit parameters over dict(**locals()))
In try/except, catch the narrowest exceptions possible
For duck-typing with try/except, keep try body minimal and put logic in else
Files:
tensorrt_llm/_torch/pyexecutor/config.pytensorrt_llm/_torch/pyexecutor/executor_request_queue.pytensorrt_llm/_torch/pyexecutor/py_executor.pytensorrt_llm/llmapi/llm_args.py
**/*.{cpp,cxx,cc,cu,h,hpp,hxx,hh,cuh,py}
📄 CodeRabbit Inference Engine (CODING_GUIDELINES.md)
Prepend NVIDIA copyright header (current year) to all source files
Files:
tensorrt_llm/_torch/pyexecutor/config.pytensorrt_llm/_torch/pyexecutor/executor_request_queue.pytensorrt_llm/_torch/pyexecutor/py_executor.pytensorrt_llm/llmapi/llm_args.py
🧬 Code Graph Analysis (1)
tensorrt_llm/llmapi/llm_args.py (1)
tensorrt_llm/builder.py (1)
default(50-58)
⏰ 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 (3)
tests/unittest/api_stability/references/llm.yaml (1)
126-129: API reference entry added correctly.The new parameter is reflected with the correct type, default, and status. This keeps the public surface consistent with the implementation.
tensorrt_llm/_torch/pyexecutor/py_executor.py (1)
189-189: Propagation of batch_wait_timeout into the executor queue: LGTM
- Reads pytorch_backend_config.batch_wait_timeout and forwards it to ExecutorRequestQueue via keyword arg. Good placement and ordering.
Also applies to: 240-242
tensorrt_llm/llmapi/llm_args.py (1)
2398-2400: Forwarding into PyTorchConfig: LGTMThe new knob is correctly included in get_pytorch_backend_config so it flows into backend config and the executor.
Shunkangz
force-pushed
the
fetch_time_out
branch
from
463dcec to
0ffafbb
Compare
|
/bot run |
|
PR_Github #15387 [ run ] triggered by Bot |
|
PR_Github #15387 [ run ] completed with state |
Shunkangz
force-pushed
the
fetch_time_out
branch
from
0ffafbb to
77e2592
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #15425 [ run ] triggered by Bot |
|
PR_Github #15425 [ run ] completed with state |
Shunkangz
force-pushed
the
fetch_time_out
branch
from
77e2592 to
d9d1734
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #15569 [ run ] triggered by Bot |
|
PR_Github #15569 [ run ] completed with state |
Signed-off-by: Shunkang <[email protected]>
Signed-off-by: Shunkang <[email protected]>
Signed-off-by: Shunkang <[email protected]>
Signed-off-by: Shunkang <[email protected]>
Co-authored-by: pcastonguay <[email protected]> Signed-off-by: Shunkangz <[email protected]>
Shunkangz
force-pushed
the
fetch_time_out
branch
from
bb33483 to
d9b7963
Compare
|
/bot reuse-pipeline |
|
PR_Github #15742 [ reuse-pipeline ] triggered by Bot |
|
PR_Github #15742 [ reuse-pipeline ] completed with state |
6 participants
Summary by CodeRabbit
New Features
Tests
Description
In this PR, I add a new argument
batch_wait_timeout. If set this argument with value larger than 0, the PyExecutor will fetch the requests from request queue first. If the number of requests are larger thanmax_batch_size, the fetch return function will return directly. Otherwise, the fetching function will wait forbatch_wait_timeoutto fetch as more requests as possible.With this optimization, we can try to put as many requests as possible in one forward step calculation.
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.