[https://nvbugs/5814309][fix] Use NCCL as fallback to avoid crash due to insufficient memory

[hyukn]

Summary by CodeRabbit

• Chores
  • Updated fallback strategy selection for distributed communication operations to use a more general approach.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

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

/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.

[coderabbitai]

📝 Walkthrough

Walkthrough

Two files updated to change the All-Reduce default fallback strategy from NCCL_SYMMETRIC to NCCL. Both changes occur in strategy selection logic when lookup table entries are missing or fallback conditions are triggered.

Changes

Cohort / File(s)	Summary
All-Reduce Fallback Strategy Changes cpp/tensorrt_llm/common/customAllReduceUtils.h, cpp/tensorrt_llm/thop/allreduceOp.cpp	Fallback strategy selection changed from NCCL_SYMMETRIC to NCCL in both selectStrategyLookUpTable (out-of-bounds table access) and selectImplementation (fallback condition)

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~5 minutes

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	PR description is incomplete; required sections (Description and Test Coverage) lack substantive details about the changes and testing.	Add explanation of why NCCL fallback avoids memory crashes and specify which tests validate this change.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly specifies a fix that uses NCCL as a fallback strategy to avoid memory-related crashes, which aligns with the code changes shown in the raw summary.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

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

✨ Finishing touches

• 📝 Generate docstrings

---

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.}

[coderabbitai]

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/common/customAllReduceUtils.h (1)

110-118: Update fallback comment to match NCCL behavior.
The comment still says NCCL_SYMMETRIC, but the code now returns NCCL on OOB lookup.

📝 Suggested fix

-    // Check if the entry is out of bounds, otherwise return NCCL_SYMMETRIC as fallback
+    // Check if the entry is out of bounds; return NCCL as fallback

🤖 Fix all issues with AI agents

In `@cpp/tensorrt_llm/thop/allreduceOp.cpp`:
- Around line 1207-1209: The comment inside ifFallbackToNCCL is stale — it
mentions NCCL_SYMMETRIC while callers (e.g., the branch that returns
AllReduceStrategyType::NCCL) now use NCCL; update the comment in the
ifFallbackToNCCL implementation to refer to NCCL (or
AllReduceStrategyType::NCCL) instead of NCCL_SYMMETRIC and ensure the
explanation matches the current behavior and return value used by callers like
the branch checking ifFallbackToNCCL(seq_len, message_size_bytes,
max_workspace_size).

[hyukn]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #33177 [ run ] triggered by Bot. Commit: 91c49b7

[tensorrt-cicd]

PR_Github #33177 [ run ] completed with state SUCCESS. Commit: 91c49b7
/LLM/release-1.2/L0_MergeRequest_PR pipeline #167 completed with status: 'FAILURE'

⚠️ 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

[hyukn]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #33237 [ run ] triggered by Bot. Commit: 91c49b7

[tensorrt-cicd]

PR_Github #33237 [ run ] completed with state ABORTED. Commit: 91c49b7
LLM/release-1.2/L0_MergeRequest_PR #182 (Blue Ocean) completed with status: ABORTED

[ZhanruiSunCh]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #33298 [ run ] triggered by Bot. Commit: 91c49b7

[tensorrt-cicd]

PR_Github #33298 [ run ] completed with state SUCCESS. Commit: 91c49b7
/LLM/release-1.2/L0_MergeRequest_PR pipeline #192 completed with status: 'SUCCESS'

[hyukn]

@nv-lschneider will help to re-enable it again as the solid fallback tactic when more investigation has been done.
Related PR: #10463