Skip to content

[None][fix] Bump DeepGEMM to nv_dev HEAD for SM120 attention support#16217

Open
lucifer1004 wants to merge 1 commit into
NVIDIA:mainfrom
lucifer1004:fix/deepgemm-sm120-pin
Open

[None][fix] Bump DeepGEMM to nv_dev HEAD for SM120 attention support#16217
lucifer1004 wants to merge 1 commit into
NVIDIA:mainfrom
lucifer1004:fix/deepgemm-sm120-pin

Conversation

@lucifer1004

@lucifer1004 lucifer1004 commented Jul 10, 2026

Copy link
Copy Markdown
Collaborator

The pinned DeepGEMM commit predates the SM120 implementations of the DSA/DSv4 indexer attention kernels (sm120_fp8_mqa_logits / sm120_fp8_paged_mqa_logits), so on SM120/SM121 the sparse-attention indexer fails with an unsupported-architecture error even though current DeepGEMM nv_dev supports the arch.

Bump the 3rdparty fetch pin to nv_dev a6b593d2826719dcf4892609af7b84ee23aaf32a, which carries the SM120 kernels. Validated as part of DeepSeek-V4 bring-up on RTX PRO 6000 Blackwell (the indexer dispatches and the model passes gsm8k gating at tp2/tp4).

🤖 Generated with Claude Code

Summary by CodeRabbit

  • Chores
    • Updated an underlying third-party component to a newer revision.
    • No changes to exported functionality or user-facing behavior.

The pinned DeepGEMM (245dc5d, 2026-06-04) predates the SM120 support
merged on nv_dev (NVIDIA#324, 2026-06-24): its attention APIs branch on
arch_major 9/10 only and DeepSeek-V4 serving on SM120 dies at engine
warmup with 'Unsupported architecture' from
get_paged_mqa_logits_metadata. The current nv_dev head carries
arch_major 12 branches across the attention interface (fp8/fp4 MQA
logits, paged logits metadata, SM120 block_kv sizing) plus the
indexer_n_heads=16 support, enabling the DSv4 sparse indexer on
SM120 without any TensorRT-LLM-side workaround.

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Signed-off-by: Zihua Wu <13583761+lucifer1004@users.noreply.github.com>
@lucifer1004 lucifer1004 requested a review from a team as a code owner July 10, 2026 04:53
@coderabbitai

coderabbitai Bot commented Jul 10, 2026

Copy link
Copy Markdown
Contributor

Review Change Stack

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: 41f831a8-c9aa-448a-a143-388a967236b5

📥 Commits

Reviewing files that changed from the base of the PR and between 6929451 and 3b2f760.

📒 Files selected for processing (1)
  • 3rdparty/fetch_content.json

📝 Walkthrough

Walkthrough

The DeepGEMM dependency pin in 3rdparty/fetch_content.json is updated from commit 245dc5d6a5fe344c61505fe71011d203141d4479 to a6b593d2826719dcf4892609af7b84ee23aaf32a.

Changes

DeepGEMM Dependency Update

Layer / File(s) Summary
Update DeepGEMM dependency pin
3rdparty/fetch_content.json
The deepgemm git_tag is updated to the new commit hash.

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

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name Status Explanation Resolution
Description check ⚠️ Warning The description explains the change, but it omits the required template sections for Description, Test Coverage, and PR Checklist. Add the required template headings and fill in concise Description, Test Coverage, and PR Checklist sections.
✅ Passed checks (4 passed)
Check name Status Explanation
Title check ✅ Passed The title matches the change and follows the required [None][fix] format.
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.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant