[None][infra] Waive 10 failed cases for main in post-merge 2825

[trtllm-agent]

Auto-generated Waive PR

Created by: TensorRT LLM CI Report (requested by @mzweilz)
Target branch: main
Bug(s): 6426834, 6426841, 6426843, 6426844, 6426847, 6426850, 6426852, 6426860, 6426865, 6426868, 6426886, 6426890

Waive entries added

accuracy/test_disaggregated_serving.py::TestDeepSeekV3Lite::test_guided_decoding[llguidance-mtp_nextn=0] SKIP (https://nvbugs/6426865)
accuracy/test_llm_api_pytorch.py::TestDeepSeekR1::test_fp8_blockscale[throughput_mtp_trtllm] SKIP (https://nvbugs/6426868)
accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16[mtp_nextn=2-attention_dp=True-cuda_graph=True-overlap_scheduler=True-torch_compile=False-enable_chunked_prefill=False-v2_kv_cache=True] SKIP (https://nvbugs/6426847)
accuracy/test_llm_api_pytorch.py::TestQwen3_5_397B_A17B::test_nvfp4[adp4_trtllm] SKIP (https://nvbugs/6426850)
examples/test_ad_speculative_decoding.py::test_autodeploy_eagle3_one_model_acceptance_rate[trtllm-torch-cudagraph] SKIP (https://nvbugs/6426841)
perf/test_perf_sanity.py::test_e2e[disagg_upload-e2e-gb200_deepseek-r1-fp4_128k8k_con128_ctx1_pp8_gen1_dep16_eplb0_mtp1_ccb-NIXL] SKIP (https://nvbugs/6426890)
perf/test_perf_sanity.py::test_e2e[disagg_upload-gen_only-gb200_deepseek-v32-fp4_1k1k_con2048_ctx1_dep4_gen1_dep4_eplb0_mtp1_ccb-NIXL] SKIP (https://nvbugs/6426890)
unittest/_torch/attention/sparse/deepseek_v4/test_compressor_kernel.py::test_prefill_varlen[varlen_hd512_overlap] SKIP (https://nvbugs/6426860)
unittest/_torch/ray_orchestrator/multi_gpu/test_llm_update_weights_multi_gpu.py -m "part1" SKIP (https://nvbugs/6426852)
unittest/disaggregated/test_kv_transfer.py::test_transfer_worker_v2[tp4_pp1_to_tp2_pp2] SKIP (https://nvbugs/6426834)

Already waived (skipped)

• accuracy/test_llm_api_pytorch.py::TestQwen3_5_35B_A3B::test_bf16[tp1-CUTLASS]
• accuracy/test_llm_api_pytorch.py::TestQwen3_5_35B_A3B::test_fp8[enable_block_reuse=False]
• accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_nvfp4_4gpus[moe_backend=CUTLASS-mtp_nextn=0-pp4-fp8kv=True-attention_dp=True-cuda_graph=True-overlap_scheduler=True-low_precision_combine=False-torch_compile=False]

---

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

Summary by CodeRabbit

• Tests
  • Updated test skip lists for several additional incompatible or unstable cases across accuracy, PyTorch, examples, performance sanity, and unit test coverage.
  • Added new waivers for a few model, decoding, dtype, and multi-GPU scenarios to keep automated test runs stable.

[mzweilz]

/bot skip --comment "waive only"

[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: 7c040236-7cde-48a8-8569-0a21a37919e6

📥 Commits

Reviewing files that changed from the base of the PR and between 7a56db9 and e7a995a.

📒 Files selected for processing (1)

• tests/integration/test_lists/waives.txt

---

📝 Walkthrough

Walkthrough

This PR adds ten new SKIP waive entries to tests/integration/test_lists/waives.txt, covering DeepSeek and Qwen3 accuracy variants, a speculative decoding example test, GB200 perf-sanity disaggregated upload configurations, a DeepSeek v4 compressor kernel case, a multi_gpu update-weights case, and a kv-transfer worker mapping.

Changes

Test Waiver Additions

Layer / File(s)	Summary
Accuracy test waives for DeepSeek/Qwen3 tests/integration/test_lists/waives.txt	Adds SKIP entries for TestDeepSeekV3Lite::test_guided_decoding, TestDeepSeekR1::test_fp8_blockscale, TestDeepSeekV3Lite::test_bfloat16 (mtp_nextn=2), and TestQwen3_5_397B_A17B::test_nvfp4 variants.
Examples and perf sanity waives tests/integration/test_lists/waives.txt	Adds a speculative decoding acceptance-rate example waive and two GB200 DeepSeek R1/V32 FP4 disaggregated upload (NIXL) perf-sanity waives.
Unittest waives tests/integration/test_lists/waives.txt	Adds a DeepSeek v4 compressor kernel prefill varlen waive, a multi_gpu update-weights "part1" waive, and a kv-transfer v2 worker tp4_pp1_to_tp2_pp2 waive.

Estimated code review effort: 1 (Trivial) | ~3 minutes

Possibly related PRs

• NVIDIA/TensorRT-LLM#15535: Both PRs add SKIP entries to the same waives.txt for overlapping DeepSeek/disaggregated and Qwen3/FP8 block-scale test variants.
• NVIDIA/TensorRT-LLM#15825: Both PRs add similar entries covering perf sanity disaggregated upload and DeepSeek v4 compressor kernel prefill cases.
• NVIDIA/TensorRT-LLM#16027: Both PRs add new SKIP waive entries to the same file for accuracy/examples and perf-sanity disaggregated upload cases.

Suggested reviewers: mzweilz, dongxuy04, pcastonguay, yuxianq

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	It includes a summary, bugs, and waived entries, but omits the required Description, Test Coverage, and PR Checklist sections.	Add the template sections with a short issue/solution description, relevant test coverage, and PR checklist completion.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title follows the required [None][infra] format and clearly summarizes the main change.
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 #58109 [ skip ] triggered by Bot. Commit: e7a995a Link to invocation

[tensorrt-cicd]

PR_Github #58109 [ skip ] completed with state SUCCESS. Commit: e7a995a
Skipping testing for commit e7a995a

Link to invocation