[https://nvbugs/6668807][fix] Remove invalid GLM KV cache assertion

[yizhang-nv]

Dev Engineer Review

• Removed the invalid GLM-5.2 KV cache manager version assertion.
• Removed the NVBug 6668807 waiver.
• The changes are limited to the intended test and waiver entries.
• The waiver file remains consistent with the expected format.

QA Engineer Review

• Modified TestGLM52::test_nvfp4 in tests/integration/defs/accuracy/test_llm_api_pytorch.py.
• Removed its skip entry from tests/integration/test_lists/waives.txt.
• The test remains covered for the DGX B200 8-GPU PyTorch pre-merge stage.
• Verdict: sufficient.

Description

Remove the invalid frontend assertion for GLM-5.2's KV cache manager version. The frontend llm.args intentionally retains the "auto" sentinel, while model loading resolves GLM's model preference to KV cache manager V2 inside the worker.

Remove the corresponding NVBug 6668807 waiver so the GLM-5.2 NVFP4 accuracy test runs again.

Test Coverage

• Pre-commit hooks passed.
• accuracy/test_llm_api_pytorch.py::TestGLM52::test_nvfp4[tp_size=8-ep_size=8] will be exercised by the DGX B200 8-GPU PyTorch pre-merge stage.

PR Checklist

• 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.

[yizhang-nv]

/bot run --disable-fail-fast --stage-list "DGX_B200-8_GPUs-PyTorch-1, DGX_B200-8_GPUs-PyTorch-2, DGX_B200-8_GPUs-PyTorch-3, DGX_B200-8_GPUs-PyTorch-4"

[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: 87d8bb71-bb77-4df2-ae21-479524cef6b3

📥 Commits

Reviewing files that changed from the base of the PR and between 5767bed and 68c9934.

📒 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 (2)

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

Included review availability: Your plan provides up to 12 included reviews per hour; 11 remain after this review.

---

Walkthrough

The GLM-5.2 NVFP4 integration test no longer requires KV-cache manager version 2. Its skip entry was also removed. NVFP4 validation and GSM8K evaluation remain unchanged.

Changes

GLM-5.2 NVFP4 test

Layer / File(s)	Summary
Remove KV-cache manager requirement tests/integration/defs/accuracy/test_llm_api_pytorch.py, tests/integration/test_lists/waives.txt	The test no longer asserts use_kv_cache_manager_v2. The corresponding skip entry was removed. NVFP4 quantization validation and GSM8K evaluation remain unchanged.

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

Merge Risk: ⚪ Minimal · up to 68c99

This change removes an invalid GLM-5.2 frontend assertion and its test waiver so the accuracy test can run again; no actionable merge-blocking risk remains after normal checks and review.

Suggested reviewers: brnguyen2

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly identifies NVBug 6668807, uses the valid fix type, and describes removal of the invalid GLM KV cache assertion.
Description check	✅ Passed	The description explains the issue and solution, identifies the removed waiver, lists relevant test coverage, and includes the completed checklist.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff. Analyzed 0 functions across 0…
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.

Full details: Docstring Coverage

Explanation

No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff. Analyzed 0 functions across 0 files.

✨ 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 #69986 [ run ] triggered by Bot. Commit: 68c9934 Link to invocation

[tensorrt-cicd]

PR_Github #69986 [ run ] completed with state FAILURE. Commit: 68c9934
/LLM/main/L0_MergeRequest_PR pipeline #57266 (Partly Tested) completed with status: 'UNSTABLE'

CI Report

⚠️ Multi-GPU Label Required:
Multi-GPU tests require the ci: full pre-merge approved label on this PR. Ask a member of NVIDIA/trt-llm-ci-approvers to add the label, then re-trigger CI with the same bot command (no rebase needed).

⚠️ 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

Link to invocation