Fix ci breakage for missing protection to is_exclusive

[yanqinz2]

Before submitting

• I agree to license this contribution under the terms of LICENSE.txt.
• I ran pre-commit run and committed any formatting changes.
• I added GitHub labels: one cat-*, one or more mod-*, and one orig-* (see label list).

Affected area

Summary

Add protection flag to is_exclusive to avoid ci breakage on Blackwell

Why

Related issues

API and compatibility impact

Testing

Summary by CodeRabbit

• New Features

  • Added support for advanced tensor-memory allocation, deallocation, and block-scaled matrix operations across supported GEMM and grouped GEMM workloads.
  • Improved compatibility across GPU architectures and software versions through guarded operation handling.

• Bug Fixes

  • Refined architecture-specific test coverage and gating for supported block-scale configurations.

• Tests

  • Added validation ensuring newer operation parameters are handled only where supported.

[coderabbitai]

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: aea4bcb6-d6a7-498d-9894-edd09830abdf

📥 Commits

Reviewing files that changed from the base of the PR and between 0cc2d62 and 94b022f.

📒 Files selected for processing (23)

• python/cudnn/gemm/frost/kernel_templates/_tile_helpers.py
• python/cudnn/gemm/frost/kernel_templates/sm100_block_scale_matmul_1ctamma.py
• python/cudnn/gemm/frost/kernel_templates/sm100_block_scale_matmul_1ctamma_static.py
• python/cudnn/gemm/frost/kernel_templates/sm100_block_scale_matmul_2ctamma.py
• python/cudnn/gemm/frost/kernel_templates/sm100_block_scale_matmul_2ctamma_static.py
• python/cudnn/gemm/frost/kernel_templates/sm100_matmul_1ctamma.py
• python/cudnn/gemm/frost/kernel_templates/sm100_matmul_1ctamma_static.py
• python/cudnn/gemm/frost/kernel_templates/sm100_matmul_2ctamma.py
• python/cudnn/gemm/frost/kernel_templates/sm100_matmul_2ctamma_static.py
• python/cudnn/gemm/frost/kernel_templates/sm100_matmul_mainloop_1ctamma.py
• python/cudnn/gemm/frost/kernel_templates/sm100_matmul_mainloop_2ctamma.py
• python/cudnn/gemm/frost/kernel_templates/sm100_moe_grouped_block_scale_matmul_fwd_1ctamma.py
• python/cudnn/gemm/frost/kernel_templates/sm100_moe_grouped_block_scale_matmul_fwd_2ctamma.py
• python/cudnn/gemm/frost/kernel_templates/sm100_moe_grouped_matmul_fwd_1ctamma.py
• python/cudnn/gemm/frost/kernel_templates/sm100_moe_grouped_matmul_fwd_2ctamma.py
• python/cudnn/gemm/frost/kernel_templates/sm103_block_scale_matmul_1ctamma.py
• python/cudnn/gemm/frost/kernel_templates/sm103_block_scale_matmul_2ctamma.py
• python/cudnn/gemm/frost/kernel_templates/sm107_block_scale_matmul_1ctamma.py
• python/cudnn/gemm/frost/kernel_templates/sm107_block_scale_matmul_2ctamma.py
• python/cudnn/gemm/frost/kernel_templates/sm107_moe_grouped_block_scale_matmul_fwd_1ctamma.py
• python/cudnn/gemm/frost/kernel_templates/sm107_moe_grouped_block_scale_matmul_fwd_2ctamma.py
• test/python/gemm/frost/test_block_scale_matmul.py
• test/python/gemm/frost/test_matmul.py

---

📝 Walkthrough

Walkthrough

The change adds shared TCGEN05 wrappers for allocation, deallocation, and block-scaled MMA. Frost GEMM templates now use these wrappers instead of direct NVVM operations. Tests add architecture gating and AST validation for version-specific keyword forwarding.

Changes

TCGEN05 helper migration

Layer / File(s)	Summary
Shared TCGEN05 wrappers python/cudnn/gemm/frost/kernel_templates/_tile_helpers.py	Adds wrappers for allocation, deallocation, and block-scaled MMA. Optional arguments are forwarded conditionally.
Block-scaled kernel migration python/cudnn/gemm/frost/kernel_templates/sm100_block_scale_matmul_*, python/cudnn/gemm/frost/kernel_templates/sm103_block_scale_matmul_*, python/cudnn/gemm/frost/kernel_templates/sm107_block_scale_matmul_*, python/cudnn/gemm/frost/kernel_templates/sm100_moe_grouped_block_scale_matmul_fwd_*, python/cudnn/gemm/frost/kernel_templates/sm107_moe_grouped_block_scale_matmul_fwd_*	Routes TMEM allocation, block-scaled MMA, and CTA cleanup through _tile_helpers.
Standard kernel migration python/cudnn/gemm/frost/kernel_templates/sm100_matmul_*, python/cudnn/gemm/frost/kernel_templates/sm100_matmul_mainloop_*, python/cudnn/gemm/frost/kernel_templates/sm100_moe_grouped_matmul_fwd_*	Routes TMEM allocation and deallocation through shared wrappers in single-CTA and two-CTA paths.
Architecture and wrapper validation test/python/gemm/frost/test_block_scale_matmul.py, test/python/gemm/frost/test_matmul.py	Updates SM107 GPU gating and adds AST tests for guarded keyword forwarding and direct NVVM call removal.

Estimated code review effort: 3 (Moderate) | ~20 minutes

Merge Risk: ⚪ Minimal · up to 94b02

This change adds guarded keyword handling across matmul templates to preserve CI and compatibility behavior, with limited expected production impact. Merge readiness is minimal; only a small follow-up remains to confirm the two new tests carry the required level marker.

Possibly related PRs

• NVIDIA/cudnn-frontend#593: Introduces related _tile_helpers abstractions and kernel call-site changes.

Suggested labels: cat-bug, orig-nv-eng, mod-cutedsl, mod-frost

Suggested reviewers: anerudhan

🚥 Pre-merge checks | ✅ 3 | ❌ 2

❌ Failed checks (2 warnings)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	The description includes the required headings, but only Summary is completed; affected area, rationale, compatibility, testing, and checklist items are missing.	Complete the affected area, Why, API and compatibility impact, and Testing sections, and confirm the required checklist items and labels.
Docstring Coverage	⚠️ Warning	Docstring coverage is 29.17% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (3 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly identifies the CI fix for missing is_exclusive protection, which matches the primary change.
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.}