[https://nvbugs/6479837][fix] Fix OOM of Qwen3_5_35B on a single a100

[JadoTu]

Dev Engineer Review

• Reduced BF16 test resource usage by lowering the KV-cache memory fraction from 0.8 to 0.75 and limiting CUDA graph and LLM batch sizes from 256 to 128.
• The settings are consistent across the test configuration and LLM invocation, addressing the single-A100 OOM without changing public APIs.
• Removed the obsolete A100 waiver for TestQwen3_5_35B_A3B::test_bf16[tp1-CUTLASS]; the waiver entry format and scope remain appropriate.

QA Engineer Review

• Modified TestQwen3_5_35B_A3B.test_bf16.
• The test is covered by the corresponding test_lists/waives.txt entry being removed, allowing CI coverage to run.
• Verdict: sufficient.

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)

• If PR introduces API changes, an appropriate PR label is added - either api-compatible or api-breaking. For api-breaking, include BREAKING in the PR title.

• 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.

[JadoTu]

/bot run --disable-fail-fast

[coderabbitai]

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: 164d0266-cd13-43de-a538-045294fbd2a4

📥 Commits

Reviewing files that changed from the base of the PR and between 1dd8b97 and 3d5c967.

📒 Files selected for processing (2)

• tests/integration/defs/accuracy/test_llm_api_pytorch.py
• tests/integration/test_lists/waives.txt

💤 Files with no reviewable changes (1)

• tests/integration/test_lists/waives.txt

---

Walkthrough

The Qwen3.5-35B-A3B BF16 accuracy test lowers GPU memory and batch-size settings, while its matching A100 skip waiver is removed.

Changes

Qwen BF16 test enablement

Layer / File(s)	Summary
Adjust BF16 capacity and remove waiver tests/integration/defs/accuracy/test_llm_api_pytorch.py, tests/integration/test_lists/waives.txt	The test lowers the free GPU memory fraction and batch-size limits to 128, and the corresponding A100 waiver entry is removed.

Estimated code review effort: 1 (Trivial) | ~5 minutes

Possibly related PRs

• NVIDIA/TensorRT-LLM#16493: Removes a waiver for the same Qwen BF16 test configuration.
• NVIDIA/TensorRT-LLM#15511: Removes related test_llm_api_pytorch skip entries from the waiver file.
• NVIDIA/TensorRT-LLM#16348: Removes an accuracy-test waiver for another Qwen configuration.

Suggested reviewers: qijune

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	The description is missing the required Description and Test Coverage sections and only includes a checklist.	Add short Description and Test Coverage sections explaining the issue, fix, and relevant tests; keep the checklist if desired.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title is concise and clearly summarizes the main fix: resolving Qwen3_5_35B OOM on a single A100.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
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.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[tensorrt-cicd]

PR_Github #61849 [ run ] triggered by Bot. Commit: 3d5c967 Link to invocation

[tensorrt-cicd]

PR_Github #61849 [ run ] completed with state SUCCESS. Commit: 3d5c967
/LLM/main/L0_MergeRequest_PR pipeline #50044 completed with status: 'SUCCESS'

CI Report

Link to invocation