[None][chore] Cleanup KV cache manager

[Funatiq]

Summary by CodeRabbit

• Breaking Changes
  • Removed public APIs from KV cache block management interfaces, including block allocation, shared block replacement, and offset configuration methods. Applications directly using these interfaces will require updates to reflect the simplified API.

Description

• Remove unused functions.

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.

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 the stage-list parameter 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.md
and the scripts/test_to_stage_mapping.py helper.

kill

kill

Kill all running builds associated with pull request.

skip

skip --comment COMMENT

Skip 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-pipeline

Reuse 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.

[Funatiq]

There are some functions that are only used in tests:

• BlockManagerTestWindowSizeToShare
  • BlockManager::calculateWindowSizeToShare
• runPartialCopyTest
  • BlockManager::offloadBlock
  • BlockManager::onboardBlock
  • BlockManager::verifyQueueIntegrity

@netanel-haber @thorjohnsen can these also be removed? I find it a bit concerning that they are not used in the standard workflow.

[Funatiq]

/bot run

[tensorrt-cicd]

PR_Github #21475 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #21475 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #16214 completed with status: 'FAILURE'

[Funatiq]

/bot run

[tensorrt-cicd]

PR_Github #21570 [ run ] triggered by Bot

[tensorrt-cicd]

PR_Github #21570 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #16281 completed with status: 'SUCCESS'

[coderabbitai]

📝 Walkthrough

Walkthrough

Public API methods are removed from the KV cache manager classes. Specifically, KVCacheBlock::isShared(), WindowBlockManager::replaceSharedBlock(), and BlockManager::allocateBlock(), replaceSharedBlock(), setOffsets(), and updateCacheBlockOffsetsAtIdx() methods are deleted from both the header and implementation files, indicating a consolidation of block management interfaces.

Changes

Cohort / File(s)

Summary

KV Cache Manager API Removals cpp/include/tensorrt_llm/batch_manager/kvCacheManager.h, cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp

Removed public methods: KVCacheBlock::isShared(), WindowBlockManager::replaceSharedBlock(GenerationRequest&, SizeType32), BlockManager::allocateBlock(GenerationRequest&, SizeType32), BlockManager::replaceSharedBlock(GenerationRequest&, SizeType32, SizeType32), BlockManager::setOffsets(kernels::KVCacheIndex*, nvinfer1::Dims const&, SizeType32, SizeType32, KVCacheBlock::IdType, SizeType32), and BlockManager::updateCacheBlockOffsetsAtIdx(GenerationRequest&, SizeType32, SizeType32). Removed helper getAllSequenceBlocks(). Eliminates block sharing queries, replacement operations, and offset management interfaces.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~5 minutes

Pre-merge checks and finishing touches

❌ Failed checks (2 warnings)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 13.33% which is insufficient. The required threshold is 80.00%.	You can run @coderabbitai generate docstrings to improve docstring coverage.
Description Check	⚠️ Warning	The pull request description does not follow the repository’s template: it omits a properly formatted title, provides only a one-line description without context or rationale, leaves the Test Coverage section empty, and does not clearly list which files or functions were changed. As a result, readers and automated tools cannot verify that the PR meets the project’s documentation standards.	Please add a title using the “[JIRA/NVBugs/GitHub/None][type] Summary” format, expand the Description section to explain what functions are being removed and why, populate the Test Coverage section with relevant test information or confirmations that no tests are required, and ensure the PR Checklist accurately reflects any actions taken.

✅ Passed checks (1 passed)

Check name	Status	Explanation
Title Check	✅ Passed	The title “[None][chore] Cleanup KV cache manager” succinctly follows the repository convention, clearly indicates a chore, and accurately summarizes the primary change of removing unused functionality from the KV cache manager.

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

---

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.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[netanel-haber]

@Funatiq if calculate etc. is only used in tests, then it is effectively dead code and can be removed AFAIC

[Funatiq]

@eopXD Is there any reason why we don't use calculateWindowSizeToShare anymore? I saw you removed it here.

[eopXD]

@eopXD Is there any reason why we don't use calculateWindowSizeToShare anymore? I saw you removed it here.

calculateWindowSizeToShare sets the memory pool ratio to the theoretical optimial.

During my development of the KV cache, my intention was to first align the allocation behavior between the original implementation, which is to treat all window size as full attention, allocating equal proportions to the memory pools.

Have you benchmarked if this theoretical optimal works better than giving a bit more space to the SWA memory pool? I am asking this because when the pool is tight, a context request will take up much space and batch size throughput will fluctuate down. So I am unsure if we are better off giving a rather sub-optimal allocation toward the pools.

Given this, I only added option to control the allocation ratio and did not go further to recover the theoretical optimal setting. I believe we can derive to the best setting through experiment and if we really need a better estimation by default I will help support it.

[eopXD]

Thank you for the clean up!!