[None][test] add GLM nvfp4 stress test

[xinhe-nv]

Summary by CodeRabbit

• Tests
  • Added new stress testing scenario for GLM-5-NVFP4 model with disaggregated server configuration.
  • Expanded test suite with configurations supporting tensor and pipeline parallelism, enabling comprehensive coverage for distributed deployment scenarios.
  • https://prod.blsm.nvidia.com/swqa-tensorrt-qa-test/view/TRT-LLM-Function-Pipelines/job/DEBUG_LLM_FUNCTION_CLUSTER_TEST/1626/testReport/B200.disaggregated/test_disaggregated/

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)

• 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.

[coderabbitai]

📝 Walkthrough

Walkthrough

Adds a new disaggregated stress-test scenario for GLM-5-NVFP4. A new YAML configuration file defines context and generation servers with TP4/EP4/DP parallelism, FP8 KV cache, DEEPGEMM MoE backend, and CUDA graph settings. The test parametrization and QA test list are updated to register and enable the new scenario.

Changes

GLM-5-NVFP4 Disaggregated Stress Test

Layer / File(s)	Summary
Disaggregated server YAML config tests/integration/defs/disaggregated/test_configs/disagg_config_ctxtp4ep4_gentp4ep4_glm5_nvfp4_dp_tllm.yaml	New YAML defining pytorch backend servers with TP4/PP1/EP4, enable_attention_dp: true, FP8 KV cache (enable_block_reuse: false, free_gpu_memory_fraction: 0.8), DEEPGEMM MoE backend, DEFAULT cache transceiver with max_tokens_in_buffer: 16384, CUDA graph with padding for generation servers, and null CUDA graph config for context servers.
Test parametrization and QA list registration tests/integration/defs/disaggregated/test_disaggregated.py, tests/integration/test_lists/qa/llm_function_stress.txt	Maps glm5_nvfp4_tp4_ep4_dp_stress to the new YAML in the config-selection dict, adds a TestConfig entry with request count, accuracy threshold, cancellation settings, and skip_less_device(4) / skip_pre_blackwell marks, and registers the test case in the QA stress list.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~5 minutes

Possibly related PRs

• NVIDIA/TensorRT-LLM#14278: Extends the same test_disaggregated.py parametrization and adds a similarly structured YAML config and QA list entry for a different model (Qwen3 32B FP8).

Suggested reviewers

• tburt-nv
• pcastonguay

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	The PR description is empty, containing only the template structure without any actual content in the Description or Test Coverage sections.	Add a meaningful description explaining what stress test was added, why (e.g., testing GLM-5-NVFP4 with specific parallelism settings), and which tests provide coverage for these changes.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title '[None][test] add GLM nvfp4 stress test' is concise and directly related to the main change: adding a stress test for GLM with NVFP4 quantization across configuration, test code, and test list files.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
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.

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

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[xinhe-nv]

/bot run --skip-test

[tensorrt-cicd]

PR_Github #117 [ run ] triggered by Bot. Commit: 5dec4e5 Link to invocation

[tensorrt-cicd]

PR_Github #54845 [ run ] triggered by Bot. Commit: 5dec4e5 Link to invocation

[tensorrt-cicd]

PR_Github #117 [ run ] completed with state FAILURE. Commit: 5dec4e5
/LLM/PipelineMonitor/L0_MergeRequest_PR pipeline #91 (Partly Tested) completed with status: 'FAILURE'

CI Report

⚠️ 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

CI Agent Failure Analysis

Link to invocation

[tensorrt-cicd]

PR_Github #54845 [ run ] completed with state SUCCESS. Commit: 5dec4e5
/LLM/main/L0_MergeRequest_PR pipeline #43857 (Partly Tested) completed with status: 'SUCCESS'

CI Report

Link to invocation