[TRTLLM-6549][fix] add kv cache time output back #7798
Conversation
zhengd-nv
requested review from
Shixiaowei02,
Tabrizian,
chuangz0 and
pcastonguay
|
/bot run |
|
PR_Github #18952 [ run ] triggered by Bot |
📝 WalkthroughWalkthroughAdds environment-gated per-transfer CSV measurement output to sender and receiver paths in dataTransceiver. Updates TransferSession constructor to accept an optional LlmRequest pointer and a boolean flag. Implements per-rank send/recv CSV files. Introduces thread-safe file writes on receiver. Adds an integration test validating CSV outputs and basic cross-checks. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
actor Env as Environment
participant Sender as CacheSender::Impl
participant Receiver as CacheReceiver::Impl
participant TS as TransferSession
participant FS as Filesystem
rect rgb(235, 245, 255)
note over Env,Sender: Configure output path via env var
Env->>Sender: getTransferOutputPath("send")
Env->>Receiver: getTransferOutputPath("recv")
end
par Send Path
Sender->>TS: construct(..., nullptr, hasOutputPath)
alt hasOutputPath
Sender->>FS: open rank_<rank>_send.csv (append)
end
note over TS: Perform transfer (send)
Sender->>TS: release session
alt hasOutputPath
TS-->>Sender: exportMeasure(header+rows, isSend=true)
Sender->>FS: write CSV rows
end
and Receive Path
Receiver->>TS: construct(..., &llmRequest, hasOutputPath)
alt hasOutputPath
Receiver->>FS: open rank_<rank>_recv.csv (append) [mutex]
end
note over TS: Perform transfer (recv)
Receiver->>TS: unformat complete
alt hasOutputPath
TS-->>Receiver: exportMeasure(header+rows, isSend=false)
Receiver->>FS: write CSV rows [mutex]
end
end
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes ✨ Finishing touches
🧪 Generate unit tests
Tip 👮 Agentic pre-merge checks are now available in preview!Pro plan users can now enable pre-merge checks in their settings to enforce checklists before merging PRs.
Please see the documentation for more information. Example: reviews:
pre_merge_checks:
custom_checks:
- name: "Undocumented Breaking Changes"
mode: "warning"
instructions: |
Pass/fail criteria: All breaking changes to public APIs, CLI flags, environment variables, configuration keys, database schemas, or HTTP/GraphQL endpoints must be documented in the "Breaking Change" section of the PR description and in CHANGELOG.md. Exclude purely internal or private changes (e.g., code not exported from package entry points or explicitly marked as internal).Please share your feedback with us on this Discord post. 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 Pre-merge checks❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
|
![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
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
cpp/tensorrt_llm/batch_manager/dataTransceiver.cpp (1)
30-34: Missing headers for std::filesystem and std::ofstream.
std::filesystemandstd::ofstreamare used but<filesystem>and<fstream>are not included, which can break some toolchains.#include <future> #include <map> #include <memory> #include <unordered_map> +#include <filesystem> +#include <fstream>
🧹 Nitpick comments (6)
tests/integration/defs/disaggregated/test_disaggregated.py (3)
589-591: Silence linter for unused fixture arg or drop it.
disaggregated_test_rootis unused (Ruff ARG001). Either prefix with_or add a noqa.Apply one of:
-def test_disaggregated_kv_cache_time_output(disaggregated_test_root, llm_venv, +def test_disaggregated_kv_cache_time_output(_disaggregated_test_root, llm_venv,or
-def test_disaggregated_kv_cache_time_output(disaggregated_test_root, llm_venv, +def test_disaggregated_kv_cache_time_output(disaggregated_test_root, llm_venv, # noqa: ARG001
608-610: Assumption on rank ↔ role mapping may be brittle.Hard-coding
rank_0_send.csvandrank_1_recv.csvties the test to current role placement. Consider globbing for*_send.csvand*_recv.csvto reduce flakiness across launcher/config changes.Example:
- send_file = os.path.join(output_path, "rank_0_send.csv") - recv_file = os.path.join(output_path, "rank_1_recv.csv") + import glob + send_file = sorted(glob.glob(os.path.join(output_path, "*_send.csv")))[0] + recv_file = sorted(glob.glob(os.path.join(output_path, "*_recv.csv")))[0]
615-621: CSV parsing robustness.Strip whitespace before float conversion and skip blank lines to avoid sporadic failures on trailing newlines.
- sample = lines[1].split(',') + sample = lines[1].strip().split(',')And inside the recv loop:
- for line in lines: - sample_recv = line.split(',') + for line in lines: + if not line.strip(): + continue + sample_recv = line.strip().split(',')cpp/tensorrt_llm/batch_manager/dataTransceiver.cpp (3)
159-173: getTransferOutputPath: directory creation error handling.
fs::create_directories(path)may throw on permission issues. If you want resiliency, catch and log, then disable output. Otherwise, this is fine.
300-310: Open measurement file in append mode to avoid truncation.Current open mode truncates on first open; appending preserves prior runs and matches header-guard logic using
tellp().- mMeasuresFile.open(outputPath); + mMeasuresFile.open(outputPath, std::ios::out | std::ios::app);
615-630: Receiver: append mode and minimal critical section.
- Use append mode to prevent truncation.
- Optional: open outside the hot path once; current lazy-open is fine, but keep the locked section minimal.
- if (!mMeasuresFile.is_open()) - { - auto outputPath = getTransferOutputPath("recv"); - mMeasuresFile.open(outputPath); + if (!mMeasuresFile.is_open()) + { + auto outputPath = getTransferOutputPath("recv"); + mMeasuresFile.open(outputPath, std::ios::out | std::ios::app); TLLM_CHECK_WITH_INFO( mMeasuresFile.is_open(), "Failed to open transfer output file: %s", outputPath.string().c_str()); } session.exportMeasure(mMeasuresFile, false);
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (2)
cpp/tensorrt_llm/batch_manager/dataTransceiver.cpp(7 hunks)tests/integration/defs/disaggregated/test_disaggregated.py(1 hunks)
🧰 Additional context used
📓 Path-based instructions (6)
**/*.{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/integration/defs/disaggregated/test_disaggregated.pycpp/tensorrt_llm/batch_manager/dataTransceiver.cpp
**/*.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/integration/defs/disaggregated/test_disaggregated.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/integration/defs/disaggregated/test_disaggregated.pycpp/tensorrt_llm/batch_manager/dataTransceiver.cpp
**/*.{h,hpp,hh,hxx,cpp,cxx,cc,cu,cuh}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
**/*.{h,hpp,hh,hxx,cpp,cxx,cc,cu,cuh}: Namespace closing braces must include a trailing comment with the namespace name (e.g., '} // namespace foo').
Prefer const or constexpr variables over #define for constants.
Declare variables that are not modified after initialization as const.
Avoid magic literals in code; except for 0, nullptr, true, false. Use named constants for comparisons and logic.
Use Allman brace style for formatting.
Place the semicolon of an empty for/while loop on a new line.
Bodies of switch/while/do-while/for must be compound statements (brace-delimited), and if/else must always be followed by brace-delimited statements.
Type names (e.g., classes) must be CamelCase starting with an uppercase letter (e.g., FooBar).
Local variables, methods, and namespaces use lowerCamelCase (e.g., localFooBar).
Non-magic-number global variables that are non-static and not in an anonymous namespace must be lowerCamelCase prefixed with 'g' (e.g., gDontUseGlobalFoos).
Non-magic-number globals that are static or in an anonymous namespace use lowerCamelCase prefixed with 's' (e.g., sMutableStaticGlobal).
Locally visible static variables use lowerCamelCase with 's' prefix (e.g., static std::once_flag sFlag).
Private/protected member variables use 'm' prefix with CamelCase (e.g., mNbFooValues). Public members may omit, but 'm' is encouraged for clarity.
Constants (enums, global constants, static constants, and function-scope magic/literal constants) use uppercase SNAKE_CASE with 'k' prefix (e.g., kDIGIT_NUM).
Function-scope constants that are not magic numbers or literals are named like non-constant variables (e.g., bool const pass = a && b).
If macros are necessary, name them in UPPER_SNAKE_CASE (e.g., FOO_VERSION) and prefer constants over #define.
Use LLVM clang-format; wrap lines at a maximum of 120 columns; use '// clang-format off/on' sparingly with justification.
Use smart pointers for heap allocations; prefer unique_ptr for sole ownership, shared_ptr for shared...
Files:
cpp/tensorrt_llm/batch_manager/dataTransceiver.cpp
**/*.{cpp,cxx,cc,cu,h,hpp,hh,hxx,cuh}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
C++ filenames should be lowerCamelCase (first letter lowercase) and must be case-insensitive unique within a compilation target.
Files:
cpp/tensorrt_llm/batch_manager/dataTransceiver.cpp
**/*.{h,hpp,hh,hxx,cpp,cxx,cc}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
**/*.{h,hpp,hh,hxx,cpp,cxx,cc}: Prefer anonymous namespaces over 'static' for internal linkage of functions.
All templates (class/function/member/static) must be instantiated at least once; non-POD classes should have private data members.
Files:
cpp/tensorrt_llm/batch_manager/dataTransceiver.cpp
🧠 Learnings (1)
📚 Learning: 2025-08-06T08:18:28.669Z
Learnt from: zhengd-nv
PR: NVIDIA/TensorRT-LLM#6633
File: cpp/tensorrt_llm/batch_manager/dataTransceiverImpl.cpp:145-155
Timestamp: 2025-08-06T08:18:28.669Z
Learning: In cpp/tensorrt_llm/batch_manager/dataTransceiverImpl.cpp, the existing `mMtxForMap` mutex in DataSenderImpl is sufficient to synchronize measurement file operations in the `release` method, as all file operations occur within the same critical section that protects the `mRequestToSession` map access.
Applied to files:
cpp/tensorrt_llm/batch_manager/dataTransceiver.cpp
🧬 Code graph analysis (2)
tests/integration/defs/disaggregated/test_disaggregated.py (1)
tests/integration/defs/conftest.py (4)
disaggregated_test_root(2381-2386)llm_venv(715-732)disaggregated_example_root(274-279)llama_model_root(974-1049)
cpp/tensorrt_llm/batch_manager/dataTransceiver.cpp (1)
cpp/tensorrt_llm/common/envUtils.cpp (2)
getEnvKVCacheTransferOutputPath(395-399)getEnvKVCacheTransferOutputPath(395-395)
🪛 Ruff (0.12.2)
tests/integration/defs/disaggregated/test_disaggregated.py
589-589: Unused function argument: disaggregated_test_root
(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 (3)
cpp/tensorrt_llm/batch_manager/dataTransceiver.cpp (3)
295-312: Sender: file-write synchronization looks sufficient.File writes are under
mMtxForMap; aligns with prior guidance that this lock guards file ops inrelease(). No further action.
359-362: Constructor callsites updated correctly.Passing
nullptrfor sender and&llmRequestfor receiver matches usage inexportMeasure(). Boolean flag derived from env is clear.Please confirm
TransferSession7‑arg ctor is defined accordingly and all other callsites were updated.
874-876: Receiver-side locking is appropriate.Per-instance mutex for
mMeasuresFileavoids interleaved writes under concurrent receives.
zhengd-nv
changed the title
|
PR_Github #18952 [ run ] completed with state |
|
/bot run |
|
PR_Github #18988 [ run ] triggered by Bot |
|
PR_Github #18988 [ run ] completed with state |
|
/bot run |
|
PR_Github #19067 [ run ] triggered by Bot |
|
PR_Github #19067 [ run ] completed with state |
|
/bot run |
|
PR_Github #19116 [ run ] triggered by Bot |
|
PR_Github #19116 [ run ] completed with state |
|
/bot run |
|
PR_Github #19154 [ run ] triggered by Bot |
|
PR_Github #19154 [ run ] completed with state |
Signed-off-by: zhengd-nv <[email protected]>
zhengd-nv
force-pushed
the
fix-kv-time-output
branch
from
218ed2e to
754a0db
Compare
|
/bot run |
|
PR_Github #19247 [ run ] triggered by Bot |
|
PR_Github #19247 [ run ] completed with state |
|
/bot run |
|
PR_Github #19338 [ run ] triggered by Bot |
|
/bot run |
|
PR_Github #19459 [ run ] triggered by Bot |
|
PR_Github #19459 [ run ] completed with state |
|
/bot run |
|
PR_Github #19504 [ run ] triggered by Bot |
|
PR_Github #19504 [ run ] completed with state |
|
/bot run |
|
PR_Github #19555 [ run ] triggered by Bot |
|
/bot run |
|
PR_Github #19598 [ run ] triggered by Bot |
|
PR_Github #19555 [ run ] completed with state |
|
/bot run |
|
PR_Github #19630 [ run ] triggered by Bot |
|
PR_Github #19598 [ run ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #19637 [ run ] triggered by Bot |
|
PR_Github #19630 [ run ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #19641 [ run ] triggered by Bot |
|
PR_Github #19637 [ run ] completed with state |
|
PR_Github #19641 [ run ] completed with state |
Signed-off-by: zhengd-nv <[email protected]>
4 participants
Summary by CodeRabbit
New Features
Tests
Description
The logic of KV cache time output (aka. KV cache measure in #6633 ) are accidently removed in #7659 . This PR adds it back with additional test to avoid being removed again.
Test Coverage
test_disaggregated_kv_cache_time_outputintests/integration/defs/disaggregated/test_disaggregated.pyPR 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.