[None][test] Enable fused QKV projection in Eagle CI

[peihu-nv]

@coderabbitai summary

Description

The MiniMax-M3 NVFP4 Eagle3 pre-merge test covers MSA, CUDA graphs, overlap scheduling, and both attention-DP modes, but it leaves fuse_qkv_index_projection disabled. This means the fused sparse QKV/index projection path used by deployment configurations is not safeguarded by CI.

Enable fuse_qkv_index_projection in the existing Eagle3 test so both MiniMax-M3 CI configurations exercise this path.

The changed test file also activates AST validation for an existing GB200 MiniMax-M3 waiver. Correct its truncated parametrization ID so it names the active InferenceMAX test exactly.

Test Coverage

• Both exact MiniMax-M3 Eagle3 pre-merge configurations pass on 4x GB200: TP4/EP4, MSA, CUDA graphs, overlap scheduler enabled, InferenceMAX, with attention DP disabled and enabled (2 passed).
• Homebrew Python 3.12 py_compile passes for the changed test file.
• Changed-file pre-commit hooks pass for both modified files, including validate-test-lists.
• git diff --check passes.

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.

[peihu-nv]

/bot run

[tensorrt-cicd]

PR_Github #62589 [ run ] triggered by Bot. Commit: 5ecdc14 Link to invocation

[peihu-nv]

/bot run

[tensorrt-cicd]

PR_Github #62589 [ run ] completed with state FAILURE. Commit: 5ecdc14
/LLM/main/L0_MergeRequest_PR pipeline #50729 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

[peihu-nv]

/bot run

[tensorrt-cicd]

PR_Github #62597 [ run ] triggered by Bot. Commit: adfccbd Link to invocation

[tensorrt-cicd]

PR_Github #62598 [ run ] triggered by Bot. Commit: adfccbd Link to invocation

[tensorrt-cicd]

PR_Github #62597 [ run ] completed with state ABORTED. Commit: adfccbd

Link to invocation

[tensorrt-cicd]

PR_Github #62598 [ run ] completed with state FAILURE. Commit: adfccbd
/LLM/main/L0_MergeRequest_PR pipeline #50738 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

[peihu-nv]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #62612 [ run ] triggered by Bot. Commit: adfccbd Link to invocation

[tensorrt-cicd]

PR_Github #62612 [ run ] completed with state SUCCESS. Commit: adfccbd
/LLM/main/L0_MergeRequest_PR pipeline #50753 completed with status: 'SUCCESS'

CI Report

Link to invocation

[brb-nv]

LGTM.