[https://nvbugs/5410391][bug] Support to share device buffers in attention meta #6557
Conversation
📝 WalkthroughWalkthroughAccepts reusable CPU/GPU buffers into attention metadata, exposes runtime buffers, refactors buffer allocation/reuse in TRT LLM attention, updates model engine to accumulate/pass those buffers into CUDA-graph metadata creation, adds a safe destructor guard and a comment tweak around CUDA cache clearing. Changes
Sequence Diagram(s)sequenceDiagram
participant Engine as PyTorchModelEngine
participant Store as Engine.cuda_graph_meta_buffers
participant Meta as AttentionMetadata (Trtllm / FlashInfer)
Engine->>Store: read stored buffers (may be empty)
Engine->>Meta: create_cuda_graph_metadata(max_batch, ..., buffers=Store)
Note right of Meta #D6EAF8: post_init_with_buffers allocates/reuses buffers\nand shapes host/device views
Meta-->>Engine: returns metadata
Engine->>Meta: get_runtime_buffers()
Meta-->>Engine: runtime_buffers (dict of tensors)
Engine->>Store: _update_attn_meta_buffers(runtime_buffers)
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 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. ✨ 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 |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 7
🧹 Nitpick comments (2)
tensorrt_llm/_torch/attention_backend/trtllm.py (2)
635-635: Remove unnecessary blank line.While adding blank lines for readability can be helpful, this particular blank line doesn't improve code organization as it separates related initialization code within the same method.
- self.host_kv_cache_block_offsets = torch.empty_like(
718-855: Consider renaming method to follow Python conventions.The method name
__post_init_with_buffers__uses double underscores which typically indicate Python special methods or name mangling. Since this is a custom initialization method, consider using a single underscore prefix like_post_init_with_buffersor a regular public method name likeinitialize_with_buffers.
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (5)
tensorrt_llm/_torch/attention_backend/interface.py(2 hunks)tensorrt_llm/_torch/attention_backend/trtllm.py(2 hunks)tensorrt_llm/_torch/pyexecutor/_util.py(1 hunks)tensorrt_llm/_torch/pyexecutor/cuda_graph_runner.py(1 hunks)tensorrt_llm/_torch/pyexecutor/model_engine.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/attention_backend/trtllm.pytensorrt_llm/_torch/pyexecutor/_util.pytensorrt_llm/_torch/attention_backend/interface.pytensorrt_llm/_torch/pyexecutor/cuda_graph_runner.pytensorrt_llm/_torch/pyexecutor/model_engine.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/attention_backend/trtllm.pytensorrt_llm/_torch/pyexecutor/_util.pytensorrt_llm/_torch/attention_backend/interface.pytensorrt_llm/_torch/pyexecutor/cuda_graph_runner.pytensorrt_llm/_torch/pyexecutor/model_engine.py
🧠 Learnings (2)
📓 Common learnings
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.
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.
Learnt from: CR
PR: NVIDIA/TensorRT-LLM#0
File: CODING_GUIDELINES.md:0-0
Timestamp: 2025-08-01T07:34:42.734Z
Learning: Applies to **/*.py : The code developed for TensorRT-LLM should conform to Python 3.8+.
Learnt from: yiqingy0
PR: NVIDIA/TensorRT-LLM#5198
File: jenkins/mergeWaiveList.py:0-0
Timestamp: 2025-07-22T08:33:49.109Z
Learning: In the TensorRT-LLM waive list merging system, removed lines are always located at the end of the merge waive lists, which is why the mergeWaiveList.py script uses reverse traversal - it's an optimization for this specific domain constraint.
Learnt from: CR
PR: NVIDIA/TensorRT-LLM#0
File: CODING_GUIDELINES.md:0-0
Timestamp: 2025-08-01T07:34:42.734Z
Learning: Applies to **/*.{cpp,h,hpp,cc,cxx,cu,py} : 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.
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.
📚 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/_util.py
🪛 Ruff (0.12.2)
tensorrt_llm/_torch/pyexecutor/model_engine.py
980-980: Undefined name spec_max_draft_tokens
(F821)
⏰ 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 (4)
tensorrt_llm/_torch/pyexecutor/_util.py (1)
243-246: Good improvement to memory measurement accuracy.Moving
torch.cuda.empty_cache()after the memory measurements ensures accurate GPU memory usage data before clearing the cache. The added comment clearly explains the reasoning.tensorrt_llm/_torch/pyexecutor/model_engine.py (3)
428-429: LGTM!The initialization of buffer tracking dictionaries is appropriate for the buffer sharing feature.
927-940: LGTM!The buffer update logic correctly tracks and maintains the maximum-sized buffers across CUDA graph invocations.
982-996: get_runtime_buffers implementation verifiedI’ve confirmed that
get_runtime_buffersis defined in theTrtllmAttentionMetadataclass (tensorrt_llm/_torch/attention_backend/trtllm.py:694). Bothattn_metadataand any metadata produced viacreate_cuda_graph_metadatasupport this call—no changes needed.
|
PR_Github #13807 [ run ] triggered by Bot |
|
PR_Github #13807 [ run ] completed with state |
|
/bot run |
|
PR_Github #13814 [ run ] triggered by Bot |
|
PR_Github #13814 [ run ] completed with state |
|
/bot run |
|
PR_Github #13910 [ run ] triggered by Bot |
|
PR_Github #13910 [ run ] completed with state |
|
/bot run |
|
PR_Github #13931 [ run ] triggered by Bot |
|
PR_Github #13931 [ run ] completed with state |
|
/bot run |
|
PR_Github #13975 [ run ] triggered by Bot |
|
PR_Github #13975 [ run ] completed with state |
|
/bot run |
|
PR_Github #14002 [ run ] triggered by Bot |
|
PR_Github #14002 [ run ] completed with state |
|
B200 cases meet "RuntimeError: CUDA error: CUDA-capable device(s) is/are busy or unavailable". |
|
/bot run |
|
PR_Github #14047 [ run ] triggered by Bot |
|
PR_Github #14047 [ run ] completed with state |
HuiGao-NV
force-pushed
the
attnmeta_share_buffer
branch
from
77b7289 to
bf5e322
Compare
|
/bot reuse-pipeline |
Signed-off-by: Hui Gao <[email protected]>
Signed-off-by: Hui Gao <[email protected]>
Signed-off-by: Hui Gao <[email protected]>
Signed-off-by: Hui Gao <[email protected]>
Signed-off-by: Hui Gao <[email protected]>
Signed-off-by: Hui Gao <[email protected]>
HuiGao-NV
force-pushed
the
attnmeta_share_buffer
branch
from
bf5e322 to
1e92f00
Compare
|
/bot reuse-pipeline |
|
PR_Github #16117 [ reuse-pipeline ] triggered by Bot |
|
PR_Github #16118 [ reuse-pipeline ] triggered by Bot |
|
PR_Github #16117 [ reuse-pipeline ] completed with state |
|
PR_Github #16118 [ reuse-pipeline ] completed with state |
…ntion meta (#6557) Signed-off-by: Hui Gao <[email protected]>
…ntion meta (NVIDIA#6557) Signed-off-by: Hui Gao <[email protected]> Signed-off-by: Wangshanshan <[email protected]>
…ntion meta (NVIDIA#6557) Signed-off-by: Hui Gao <[email protected]> Signed-off-by: Wangshanshan <[email protected]>
…ntion meta (NVIDIA#6557) Signed-off-by: Hui Gao <[email protected]> Signed-off-by: Wangshanshan <[email protected]>
…ntion meta (NVIDIA#6557) Signed-off-by: Hui Gao <[email protected]> Signed-off-by: Wangshanshan <[email protected]>
…ntion meta (NVIDIA#6557) Signed-off-by: Hui Gao <[email protected]> Signed-off-by: Wangshanshan <[email protected]>
6 participants
Summary by CodeRabbit
New Features
Improvements
Bug Fixes
Description
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.
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.