[https://nvbugs/6079901][fix] Avoid divide-by-zero in KVCacheTransfer… by farazkh80 · Pull Request #13618 · NVIDIA/TensorRT-LLM

farazkh80 · 2026-04-29T18:57:49Z

…Manager::computeBlockTransferBytes for empty pools

The recent kv-cache statistics monitoring commit (580c7a6, "[TRTLLM-11421][feat] Support better kv cache statistics monitoring") derives bytesPerElement as getSizeInBytes()/getSize() inside KVCacheTransferManager::computeBlockTransferBytes. Pools whose primaryPtr is non-null but has zero elements (e.g. layers that contribute no KV cache in NAS-style heterogeneous models such as Llama-3_3-Nemotron-Super-49B-v1) trigger a SIGFPE / integer divide-by-zero on the first onboard transfer that touches such a pool.

Skip pools with zero elements: they legitimately contribute zero transfer bytes, so the stats are unchanged and the divide-by-zero is avoided.

Repro before fix: post-merge L0 stage RTXPro6000D-PyTorch-Post-Merge-1 (test_e2e.py::test_ptp_quickstart_advanced[Nemotron-Super-49B-v1-FP8-...]) crashes with "Floating point exception (8) / Integer divide-by-zero (1)" inside computeBlockTransferBytes during KVCacheManager::addSequenceBatch. After fix the test passes locally on RTX Pro 6000 (SM120) in 73s.

Summary by CodeRabbit

Bug Fixes
- Resolved a crash risk in KV cache transfer operations when encountering empty cache pools, enhancing system reliability and preventing unexpected failures during batch processing.

Description

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
Update tava architecture diagram if there is a significant design change in PR.
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

To see a list of available CI bot commands, please comment /bot help.

coderabbitai · 2026-04-29T19:00:14Z

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 1d33ead3-302f-437f-949c-b386a51ee822

📥 Commits

Reviewing files that changed from the base of the PR and between 3b7af1c and b948b0e.

📒 Files selected for processing (1)

cpp/tensorrt_llm/batch_manager/kvCacheTransferManager.cpp

📝 Walkthrough

Walkthrough

The computeBlockTransferBytes function now includes a guard against empty KV cache pools by checking pool size before computing bytes per element, preventing division-by-zero errors when processing cache transfer operations.

Changes

Cohort / File(s)	Summary
KV Cache Division-by-Zero Guard `cpp/tensorrt_llm/batch_manager/kvCacheTransferManager.cpp`	Added null-check guard in `computeBlockTransferBytes` to skip processing when KV cache pool size is zero, preventing divide-by-zero in `bytesPerElement` calculation while preserving existing partial-vs-full copy byte accounting logic.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~8 minutes

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly identifies the NVBugs ticket, the fix type, and the specific issue being addressed (divide-by-zero in KVCacheTransfer), directly matching the changeset.
Description check	✅ Passed	The PR description thoroughly explains the issue (divide-by-zero crash in heterogeneous models), the solution (skip zero-element pools), includes reproduction details, and confirms the fix was tested. However, the Description and Test Coverage sections of the template remain unfilled.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

Create PR with unit tests

_{Review rate limit: 9/10 reviews remaining, refill in 6 minutes.}

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

farazkh80 · 2026-04-29T19:53:30Z

/bot run --disable-fail-fast

tensorrt-cicd · 2026-04-29T20:00:09Z

PR_Github #46207 [ run ] triggered by Bot. Commit: b948b0e Link to invocation

tensorrt-cicd · 2026-04-30T04:41:08Z

PR_Github #46207 [ run ] completed with state SUCCESS. Commit: b948b0e
/LLM/main/L0_MergeRequest_PR pipeline #36320 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

Please check the failed tests and fix your PR
If you cannot view the failures, ask the CI triggerer to share details
Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

farazkh80 · 2026-04-30T15:48:08Z

/bot help

github-actions · 2026-04-30T15:48:20Z

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) --high-priority]

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. Supports wildcard * for pattern matching (e.g., "*PerfSanity*" matches all stages containing PerfSanity). Examples: "A10-PyTorch-1, xxx", "PerfSanity". 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. Supports wildcard * for pattern matching. Examples: --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx", --extra-stage "Post-Merge".

--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.

--high-priority (OPTIONAL) : Run the pipeline with high priority. This option is restricted to authorized users only and will route the job to a high-priority queue.

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.

farazkh80 · 2026-05-06T13:50:36Z

/bot run --disable-fail-fast

tensorrt-cicd · 2026-05-06T13:57:04Z

PR_Github #47012 [ run ] triggered by Bot. Commit: b948b0e Link to invocation

tensorrt-cicd · 2026-05-06T22:43:43Z

PR_Github #47012 [ run ] completed with state SUCCESS. Commit: b948b0e
/LLM/main/L0_MergeRequest_PR pipeline #36987 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

Please check the failed tests and fix your PR
If you cannot view the failures, ask the CI triggerer to share details
Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

farazkh80 · 2026-05-13T15:54:14Z

/bot run

tensorrt-cicd · 2026-05-13T15:59:54Z

PR_Github #48210 [ run ] triggered by Bot. Commit: a40cfce Link to invocation

tensorrt-cicd · 2026-05-13T17:48:28Z

PR_Github #48210 [ run ] completed with state SUCCESS. Commit: a40cfce
/LLM/main/L0_MergeRequest_PR pipeline #38028 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

Please check the failed tests and fix your PR
If you cannot view the failures, ask the CI triggerer to share details
Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

…Manager::computeBlockTransferBytes for empty pools The recent kv-cache statistics monitoring commit (580c7a6, "[TRTLLM-11421][feat] Support better kv cache statistics monitoring") derives bytesPerElement as getSizeInBytes()/getSize() inside KVCacheTransferManager::computeBlockTransferBytes. Pools whose primaryPtr is non-null but has zero elements (e.g. layers that contribute no KV cache in NAS-style heterogeneous models such as Llama-3_3-Nemotron-Super-49B-v1) trigger a SIGFPE / integer divide-by-zero on the first onboard transfer that touches such a pool. Skip pools with zero elements: they legitimately contribute zero transfer bytes, so the stats are unchanged and the divide-by-zero is avoided. Repro before fix: post-merge L0 stage RTXPro6000D-PyTorch-Post-Merge-1 (test_e2e.py::test_ptp_quickstart_advanced[Nemotron-Super-49B-v1-FP8-...]) crashes with "Floating point exception (8) / Integer divide-by-zero (1)" inside computeBlockTransferBytes during KVCacheManager::addSequenceBatch. After fix the test passes locally on RTX Pro 6000 (SM120) in 73s. Signed-off-by: Faraz Khoubsirat <[email protected]>

farazkh80 · 2026-05-23T19:37:44Z

/bot run

tensorrt-cicd · 2026-05-23T19:43:55Z

PR_Github #50065 [ run ] triggered by Bot. Commit: 48a9ea9 Link to invocation

tensorrt-cicd · 2026-05-24T04:14:13Z

PR_Github #50065 [ run ] completed with state SUCCESS. Commit: 48a9ea9
/LLM/main/L0_MergeRequest_PR pipeline #39621 completed with status: 'SUCCESS'
Pipeline passed with automatic retried tests. Check the rerun report for details.

CI Report

Link to invocation

NVIDIA#13618) Signed-off-by: Faraz Khoubsirat <[email protected]>

farazkh80 requested a review from a team as a code owner April 29, 2026 18:57

github-actions Bot assigned farazkh80 Apr 29, 2026

farazkh80 changed the title ~~[nvbugs/6079901][fix] Avoid divide-by-zero in KVCacheTransfer…~~ [https://nvbugs/6079901][fix] Avoid divide-by-zero in KVCacheTransfer… Apr 29, 2026

nvpohanh approved these changes Apr 30, 2026

View reviewed changes

farazkh80 force-pushed the fix/nvbugs-6079901-kv-cache-transfer-divide-by-zero branch from b948b0e to a40cfce Compare May 13, 2026 15:53

farazkh80 force-pushed the fix/nvbugs-6079901-kv-cache-transfer-divide-by-zero branch from a40cfce to 3dc73e4 Compare May 20, 2026 15:07

farazkh80 force-pushed the fix/nvbugs-6079901-kv-cache-transfer-divide-by-zero branch from 3dc73e4 to 48a9ea9 Compare May 23, 2026 19:37

farazkh80 merged commit c02475e into NVIDIA:main May 24, 2026
7 checks passed

Conversation

farazkh80 commented Apr 29, 2026 • edited by coderabbitai Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary by CodeRabbit

Description

Test Coverage

PR Checklist

GitHub Bot Help

Uh oh!

coderabbitai Bot commented Apr 29, 2026

Walkthrough

Changes

Estimated code review effort

❌ Failed checks (1 warning)

Uh oh!

farazkh80 commented Apr 29, 2026

Uh oh!

tensorrt-cicd commented Apr 29, 2026

Uh oh!

tensorrt-cicd commented Apr 30, 2026

Uh oh!

farazkh80 commented Apr 30, 2026

Uh oh!

github-actions Bot commented Apr 30, 2026

GitHub Bot Help

kill

skip

reuse-pipeline

Uh oh!

farazkh80 commented May 6, 2026

Uh oh!

tensorrt-cicd commented May 6, 2026

Uh oh!

tensorrt-cicd commented May 6, 2026

Uh oh!

farazkh80 commented May 13, 2026

Uh oh!

tensorrt-cicd commented May 13, 2026

Uh oh!

tensorrt-cicd commented May 13, 2026

Uh oh!

farazkh80 commented May 23, 2026

Uh oh!

tensorrt-cicd commented May 23, 2026

Uh oh!

tensorrt-cicd commented May 24, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

farazkh80 commented Apr 29, 2026 •

edited by coderabbitai Bot

Loading