[None][test] Waive 3 failed cases for main in QA CI

[trtllm-agent]

Auto-generated Waive PR

Created by: TensorRT LLM CI Report (requested by qa@nvidia.com)
Target branch: main
Bug(s): 6474894, 6546909, 6562582, 6563551

Waive entries added

full:H100/accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_bf16_4gpu[dep4] SKIP (https://nvbugs/6563551)
full:H20/accuracy/test_llm_api_pytorch_multimodal.py::TestQwen3_5_27B_VL::test_auto_dtype SKIP (https://nvbugs/6562582)
full:H20/accuracy/test_llm_api_pytorch.py::TestLlama3_1_8BInstruct::test_fp8_4gpus[tp4-fp8kv=True-attn_backend=FLASHINFER-torch_compile=True] SKIP (https://nvbugs/6546909)

Already waived (skipped)

• full:B200/accuracy/test_llm_api_pytorch.py::TestNemotronV3Ultra::test_nvfp4_4gpus_block_reuse[TEP4]

---

This PR was auto-generated by TensorRT LLM CI Report. Please review the waive entries before merging.

Dev Engineer Review

• Added three QA CI waiver entries in tests/integration/test_lists/waives.txt.
• Entries cover the H100 Qwen3 Next BF16 test and the H20 Qwen3.5 VL automatic-dtype and Llama 3.1 FP8 4-GPU tests.
• Each entry includes its associated NV bug reference.
• The change adds no public API or code changes.
• Review should confirm that the test paths, formatting, and bug references match existing waiver conventions.

QA Engineer Review

• Modified tests/integration/test_lists/waives.txt.
• Added three skipped QA CI entries.
• No test-db/ or qa/ files were modified.
• CBTS coverage data is unavailable.
• Verdict: needs follow-up.

[xinhe-nv]

/bot skip --comment "skip failed tests"

[coderabbitai]

Walkthrough

Three skipped integration-test entries were added to tests/integration/test_lists/waives.txt. The entries cover one H100 test and two H20 tests. Each entry references an NV bug.

Changes

Integration test waivers

Layer / File(s)	Summary
Add GPU-specific waiver entries tests/integration/test_lists/waives.txt	Added one H100 waiver for the Qwen3 Next BF16 test and two H20 waivers for Llama 3.1 FP8 and Qwen3.5 multimodal tests. Each waiver references an NV bug.

Estimated code review effort: 1 (Trivial) | ~2 minutes

Possibly related PRs

• NVIDIA/TensorRT-LLM#17217: Updates the same integration-test waiver list.
• NVIDIA/TensorRT-LLM#17252: Adds skipped entries to the same waiver list.
• NVIDIA/TensorRT-LLM#17301: Modifies NV-bug-linked skip entries in the same file.

Suggested reviewers: bowenfu, brnguyen2, crazydemo

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly identifies the addition of three failed-case waivers for main-branch QA CI.
Description check	✅ Passed	The description explains the automated waiver purpose, target branch, bug references, and all three affected test entries.
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.}

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@tests/integration/test_lists/waives.txt`:
- Line 258: Add a repository-supported NVIDIA copyright header including 2026 to
tests/integration/test_lists/waives.txt without introducing a raw non-empty line
that verify_waive_list treats as a waiver; if no supported format exists, update
scripts/check_test_list.py::verify_waive_list and its parser to recognize and
skip the header while preserving waiver validation.

🪄 Autofix

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 1546557c-342f-4873-805f-36f5f499a14e

📥 Commits

Reviewing files that changed from the base of the PR and between 1310036 and cf76582.

📒 Files selected for processing (1)

• tests/integration/test_lists/waives.txt

[tensorrt-cicd]

PR_Github #64163 [ skip ] triggered by Bot. Commit: cf76582 Link to invocation

[tensorrt-cicd]

PR_Github #64163 [ skip ] completed with state SUCCESS. Commit: cf76582
Skipping testing for commit cf76582

Link to invocation

[brnguyen2]

Approving — the comments below are optional touch-ups, not blockers.

All three test IDs resolve to real tests and param IDs on disk, no duplicates, sorted correctly — mergeable.

One thing to check before merge: the entry citing https://nvbugs/6546909 — that bug was filed against a different test/parameterization on a different GPU than the H20 4-GPU FLASHINFER case being waived here. The failures may share a root cause, but as filed the bug won't be re-verified against this config, so the waive risks being orphaned (or removed prematurely) when the bug closes. Worth adding this case to that bug or filing one that covers it.