[https://nvbugs/5821415][test] update rtx6k test list

[xinhe-nv]

@coderabbitai summary

Description

• remove all llama tests from test list.
• sort test list, remove dup lines.

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.

[xinhe-nv]

/bot run --stage-list ""

[coderabbitai]

📝 Walkthrough

Walkthrough

This PR updates the test selection list for RTX6k QA environments. It expands the DeepSeekV3Lite bfloat16 parameter matrices with new cuda_graph and attention_dp configurations, adjusts existing parameter combinations for different mtp_nextn values, and modifies the selection of downstream model variant tests (nvfp4, guided decoding) and cross-suite accuracy/e2e test entries.

Changes

RTX6k Test Matrix Selection

Layer / File(s)	Summary
DeepSeekV3Lite bfloat16 parameter matrices tests/integration/test_lists/qa/llm_function_rtx6k.txt	Expands mtp_nextn=0 matrix with cuda_graph=True and attention_dp=True configurations (with overlap_scheduler=False across torch_compile variants). Adjusts mtp_nextn=2 parameter combinations for cuda_graph, overlap_scheduler, torch_compile, and v2_kv_cache.
DeepSeekV3Lite variant and downstream test selection tests/integration/test_lists/qa/llm_function_rtx6k.txt	Updates enabled DeepSeekV3Lite variant test identifiers (bfloat16_4gpus_online_eplb, guided decoding variants, nvfp4, nvfp4_4gpus_online_eplb, batch-waiting scenarios) with entries newly added, removed, or reordered.
Cross-model accuracy and e2e test selection tests/integration/test_lists/qa/llm_function_rtx6k.txt	Modifies selection and order of test entries across GPTOSS, model accuracy tests, encode/multimodal, e2e, and disaggregated test suites.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Possibly related PRs

• NVIDIA/TensorRT-LLM#14866: Both PRs adjust handling of TestDeepSeekV3Lite::test_bfloat16 parameterizations involving mtp_nextn and torch_compile (main PR via test list selection, related PR via test skip list).

Suggested reviewers

• StanleySun639

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
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.
Title check	✅ Passed	The title accurately describes the main change: updating the RTX6K test list file (llm_function_rtx6k.txt) by removing llama tests and reorganizing entries.
Description check	✅ Passed	The description provides specific details about the changes: removing all llama tests and sorting/deduplicating the test list, though the Test Coverage section remains unfilled.

_{✏️ 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.}

[coderabbitai]

Caution

Some comments are outside the diff and can’t be posted inline due to platform limitations.

⚠️ Outside diff range comments (1)

tests/integration/test_lists/qa/llm_function_rtx6k.txt (1)

22-33: ⚠️ Potential issue | 🟠 Major | ⚡ Quick win

Missing attention_dp=True + cuda_graph=True + overlap_scheduler=False coverage for mtp_nextn=0.

Coverage looks incomplete for the intended DeepSeekV3Lite bfloat16 matrix expansion. Please add these entries to tests/integration/test_lists/qa/llm_function_rtx6k.txt:

• accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16[mtp_nextn=0-attention_dp=True-cuda_graph=True-overlap_scheduler=False-torch_compile=False-enable_chunked_prefill=False-v2_kv_cache=False]
• accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16[mtp_nextn=0-attention_dp=True-cuda_graph=True-overlap_scheduler=False-torch_compile=False-enable_chunked_prefill=False-v2_kv_cache=True]
• accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16[mtp_nextn=0-attention_dp=True-cuda_graph=True-overlap_scheduler=False-torch_compile=True-enable_chunked_prefill=False-v2_kv_cache=False]
• accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16[mtp_nextn=0-attention_dp=True-cuda_graph=True-overlap_scheduler=False-torch_compile=True-enable_chunked_prefill=False-v2_kv_cache=True]

Coverage status: insufficient for this matrix slice in this PR cohort.

As per coding guidelines, for files under tests/** feedback should be actionable and explicitly state whether coverage is sufficient or insufficient.

🤖 Prompt for 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.

In `@tests/integration/test_lists/qa/llm_function_rtx6k.txt` around lines 22 - 33,
The test list is missing four parameterized cases for TestDeepSeekV3Lite
bfloat16 with mtp_nextn=0 and attention_dp=True + cuda_graph=True +
overlap_scheduler=False; edit
tests/integration/test_lists/qa/llm_function_rtx6k.txt and add the four entries
for accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16 with the
combinations torch_compile={False,True} and v2_kv_cache={False,True} (as listed
in the review comment) so the matrix slice
(attention_dp=True,cuda_graph=True,overlap_scheduler=False,mtp_nextn=0) is fully
covered.

🤖 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.

Outside diff comments:
In `@tests/integration/test_lists/qa/llm_function_rtx6k.txt`:
- Around line 22-33: The test list is missing four parameterized cases for
TestDeepSeekV3Lite bfloat16 with mtp_nextn=0 and attention_dp=True +
cuda_graph=True + overlap_scheduler=False; edit
tests/integration/test_lists/qa/llm_function_rtx6k.txt and add the four entries
for accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16 with the
combinations torch_compile={False,True} and v2_kv_cache={False,True} (as listed
in the review comment) so the matrix slice
(attention_dp=True,cuda_graph=True,overlap_scheduler=False,mtp_nextn=0) is fully
covered.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 6636a6ec-60ff-4044-a23c-b8a9042c2bc5

📥 Commits

Reviewing files that changed from the base of the PR and between e1212ad and 5d3fbcb.

📒 Files selected for processing (1)

• tests/integration/test_lists/qa/llm_function_rtx6k.txt

[xinhe-nv]

/bot run --stage-list ""

[tensorrt-cicd]

PR_Github #51946 [ run ] triggered by Bot. Commit: 06941d7 Link to invocation

[tensorrt-cicd]

PR_Github #51948 [ run ] triggered by Bot. Commit: 06941d7 Link to invocation

[tensorrt-cicd]

PR_Github #51948 [ run ] completed with state SUCCESS. Commit: 06941d7
/LLM/main/L0_MergeRequest_PR pipeline #41298 (Partly Tested) completed with status: 'SUCCESS'

CI Report

Link to invocation

[xinhe-nv]

/bot reuse-pipeline

[tensorrt-cicd]

PR_Github #51969 [ reuse-pipeline ] triggered by Bot. Commit: 64d5740 Link to invocation

[tensorrt-cicd]

PR_Github #51969 [ reuse-pipeline ] completed with state SUCCESS. Commit: 64d5740
Reusing PR_Github #51948 (Partly Tested) for commit 64d5740

Link to invocation