Skip to content

[https://nvbugs/6481375][test] Unwaive passing DSV3-Lite tests - #17038

Merged
lfr-0531 merged 2 commits into
NVIDIA:mainfrom
lfr-0531:user/fanrongl/fix-dsv3-lite-nvbugs-6481375-6400067
Aug 4, 2026
Merged

[https://nvbugs/6481375][test] Unwaive passing DSV3-Lite tests#17038
lfr-0531 merged 2 commits into
NVIDIA:mainfrom
lfr-0531:user/fanrongl/fix-dsv3-lite-nvbugs-6481375-6400067

Conversation

@lfr-0531

@lfr-0531 lfr-0531 commented Jul 30, 2026 •
edited by coderabbitai Bot
Loading

Copy link
Copy Markdown
Collaborator

Description

Unwaive DSV3-Lite tests tracked by NVBug 6481375 and NVBug 6400067 after the underlying failures were fixed.

  • Re-enable disaggregated gen-first coverage.
  • Re-enable BF16 MTP coverage with both legacy and V2 KV cache.

Test Coverage

  • DSV3-Lite disaggregated gen-first test: passed.
  • DSV3-Lite BF16 MTP tests with both KV cache modes: passed.
  • Related KV cache V2 unit tests and pre-commit checks: passed.

Dev Engineer Review

  • Removed three obsolete skip entries from tests/integration/test_lists/waives.txt.
  • The changes target one TestDeepSeekV3Lite::test_gen_first case and two GB200 TestDeepSeekV3Lite::test_bfloat16 cases.
  • The change has no API, performance, or error-handling impact.
  • The waiver scope is limited to the intended DSV3-Lite tests.
  • The waiver-file format and bug-reference consistency are preserved.
  • No duplicate or unrelated waiver changes were identified.

QA Engineer Review

  • No test-db/ or qa/ files were modified.
  • Removed one gen-first waiver and two GB200 BF16 waivers from tests/integration/test_lists/waives.txt.
  • The related tests are re-enabled for CI coverage.
  • Listed tests, KV cache V2 unit tests, pre-commit checks, and CI pipelines passed.
  • Verdict: sufficient.

lfr-0531 added 2 commits July 30, 2026 02:17
@lfr-0531
[https://nvbugs/6481375][test] Unwaive DSV3-Lite gen-first test
51b451a 
Signed-off-by: Fanrong Li <23290157+lfr-0531@users.noreply.github.com>
@lfr-0531
[https://nvbugs/6400067][test] Unwaive DSV3-Lite GB200 tests
2f7c87d 
Signed-off-by: Fanrong Li <23290157+lfr-0531@users.noreply.github.com>
@lfr-0531

lfr-0531 commented Jul 30, 2026

Copy link
Copy Markdown
Collaborator Author

/bot run --disable-fail-fast

@tensorrt-cicd

tensorrt-cicd commented Jul 30, 2026

Copy link
Copy Markdown
Collaborator

PR_Github #62659 [ run ] triggered by Bot. Commit: 2f7c87d Link to invocation

@tensorrt-cicd

tensorrt-cicd commented Jul 30, 2026

Copy link
Copy Markdown
Collaborator

PR_Github #62659 [ run ] completed with state SUCCESS. Commit: 2f7c87d
/LLM/main/L0_MergeRequest_PR pipeline #50799 completed with status: 'SUCCESS'

CI Report

Link to invocation

@lfr-0531
lfr-0531 marked this pull request as ready for review August 3, 2026 13:23
@lfr-0531
lfr-0531 requested review from a team as code owners August 3, 2026 13:23
@coderabbitai

coderabbitai Bot commented Aug 3, 2026

Copy link
Copy Markdown
Contributor

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: d8c96ef3-35de-4fd4-a1a5-816af98e87a1

📥 Commits

Reviewing files that changed from the base of the PR and between d0543dc and 2f7c87d.

📒 Files selected for processing (1)
  • tests/integration/test_lists/waives.txt
💤 Files with no reviewable changes (1)
  • tests/integration/test_lists/waives.txt

Walkthrough

The waiver list removes three skipped TestDeepSeekV3Lite entries: one test_gen_first case and two GB200 test_bfloat16 cases.

Changes

Cohort / File(s) Summary
DeepSeekV3Lite waiver cleanup
tests/integration/test_lists/waives.txt		 Removes three skipped test entries from the integration waiver list.

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

Possibly related PRs

Suggested reviewers: brnguyen2, qijune, xinhe-nv

🚥 Pre-merge checks | ✅ 5
✅ Passed checks (5 passed)
Check name Status Explanation
Title check ✅ Passed The title follows the required ticket and type format and clearly states that passing DSV3-Lite tests were unwaived.
Description check ✅ Passed The description explains the issue, solution, affected tests, and test coverage; the omitted checklist does not prevent a complete review.
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.
✨ Finishing Touches
🧪 Generate unit tests (beta)
  • Create PR with unit tests

Comment @coderabbitai help to get the list of available commands.

This was referenced Aug 3, 2026
Merged
Open
@BowenFu

BowenFu commented Aug 4, 2026

Copy link
Copy Markdown

The test_gen_first[noadp-mtp0] removal is covered — build 50799 is green on 2f7c87d and you're the engineer on 6481375 with a can't-repro on main.

The two full:GB200/...test_bfloat16[...torch_compile=True-enable_chunked_prefill=True...] lines aren't. full: means post-merge-only on GB200, and 50799 ran DGX_H100 / DGX_B200 -cbts stages (8 tests) — no GB200 stage. 6400067 is also still Dev - Open - To fix / P1, unlike 6481375.

Per that bug's own instructions, could you run /bot run --extra-stage "<GB200 post-merge stage>" on this head so those two are exercised before they land? Happy to approve once that's green.

@xinhe-nv

xinhe-nv commented Aug 4, 2026

Copy link
Copy Markdown
Collaborator

/bot run --only-qa-verify test accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16[mtp_nextn=2-attention_dp=True-cuda_graph=True-overlap_scheduler=True-torch_compile=True-enable_chunked_prefill=True-v2_kv_cache=False]

@tensorrt-cicd

tensorrt-cicd commented Aug 4, 2026

Copy link
Copy Markdown
Collaborator

PR_Github #63623 [ run ] triggered by Bot. Commit: 2f7c87d Link to invocation

@tensorrt-cicd

tensorrt-cicd commented Aug 4, 2026

Copy link
Copy Markdown
Collaborator

PR_Github #63623 [ run ] completed with state SUCCESS. Commit: 2f7c87d
LLM_FUNCTION_AUTO_V2C #350 completed with status: 'SUCCESS'
QA verify test: accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16[mtp_nextn=2-attention_dp=True-cuda_graph=True-overlap_scheduler=True-torch_compile=True-enable_chunked_prefill=True-v2_kv_cache=False] (NVBug 6481375, branch user/fanrongl/fix-dsv3-lite-nvbugs-6481375-6400067, fork lfr-0531, dry_run_close=true)

Link to invocation

@lfr-0531
lfr-0531 merged commit 164a0e3 into NVIDIA:main Aug 4, 2026
19 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@xinhe-nv xinhe-nv xinhe-nv approved these changes
@schetlur-nv schetlur-nv Awaiting requested review from schetlur-nv schetlur-nv is a code owner automatically assigned from NVIDIA/trt-llm-devs
@brnguyen2 brnguyen2 Awaiting requested review from brnguyen2 brnguyen2 is a code owner automatically assigned from NVIDIA/trt-llm-devs
@yiqingy0 yiqingy0 Awaiting requested review from yiqingy0 yiqingy0 is a code owner automatically assigned from NVIDIA/trt-llm-infra-devs
@dpitman-nvda dpitman-nvda Awaiting requested review from dpitman-nvda dpitman-nvda is a code owner automatically assigned from NVIDIA/trt-llm-infra-devs

Assignees

@lfr-0531 lfr-0531

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

4 participants

@lfr-0531 @tensorrt-cicd @BowenFu @xinhe-nv