[None][test] fix perf test cases issue of incorrect match

[ruodil]

Summary by CodeRabbit

• Tests
  • Expanded performance testing infrastructure with new multi-GPU test configurations for 8-GPU benchmark scenarios.
  • Reorganized and realigned test entries across different GPU types and hardware configurations for improved consistency.
  • Enhanced test coverage for comprehensive performance benchmarking validation across multiple system configurations.

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)

• 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.

[ruodil]

/bot skip --comment "skip test as just modifying cases"

[coderabbitai]

📝 Walkthrough

Walkthrough

Modifies test configuration to add multi-GPU test blocks for 8-GPU systems (L40S, H100, H20, H200), adds a new test entry for qwen model under RTX-6000 Server section, and removes a duplicate test entry while reintroducing an aligned 8-GPU variant block with compute capability conditions.

Changes

Cohort / File(s)	Summary
GPU Test Configuration tests/integration/test_lists/qa/llm_perf_core.yml	Adds new multi-GPU test block for 8 GPUs with system_gpu_count >= 8 and compute_capability 8.0–9.0 conditions; adds separate qwen3_235b_a22b_fp8 test entry for RTX-6000 Server section; removes existing duplicate qwen3_235b_a22b_fp8 entry; reorganizes 8-GPU block alignment in L40S/H100/H20/H200 section.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~8 minutes

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	The PR description is entirely empty—all template sections including Description, Test Coverage, and PR Checklist are left blank or unchanged from the template.	Provide a detailed description explaining what issue is being fixed, the solution implemented, relevant test cases, and confirm the PR checklist items have been reviewed.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title '[None][test] fix perf test cases issue of incorrect match' clearly identifies a test fix related to perf test case matching issues, which aligns with the changeset that modifies test configuration.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

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

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

---

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

[tensorrt-cicd]

PR_Github #38539 [ skip ] triggered by Bot. Commit: c4905ec Link to invocation

[yufeiwu-nv]

Suggested change

	gt: 8.0
	gte: 9.0

[yufeiwu-nv]

maybe remove 8XL40S because it's low-priority and 8XGPU is SYS topo which lead to unstable

[yufeiwu-nv]

if remove 8XL40S, maybe you can merge 3b into 4

[yufeiwu-nv]

Suggested change

	# 3b: L40S, H100, H20, H200 (8 GPUs)
	# 3b: H100, H20, H200 (8 GPUs)

[yufeiwu-nv]

Suggested change

	# 3b: L40S, H100, H20, H200 (8 GPUs)
	# 3b: H100, H20, H200 (8 GPUs)

[tensorrt-cicd]

PR_Github #38539 [ skip ] completed with state SUCCESS. Commit: c4905ec
Skipping testing for commit c4905ec

Link to invocation