Apply black 26.3.1 formatting to glu_hadamard_quant and glu python files

[Anerudhan]

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

FE OSS kernels or CuTeDSL

Summary

Formatting-only change: run black --line-length 160 (black 26.3.1, the version the format CI now uses) over the Python tree. Six files under python/cudnn/gemm/cutedsl/grouped/ (glu_hadamard_quant/* and glu/_blockscaled_api.py) get reformatted; no code changes.

Why

The style-check CI job (analysis:clang-format, which also runs black --check) fails on current develop because these files were merged before the CI's black version was bumped to 26.3.1, which formats them differently. Every open PR/MR now inherits this failure. Applying the new formatter output unblocks the check.

Related issues

None.

API and compatibility impact

None — formatting only; black verifies AST equivalence.

Testing

find test/ python/ -regex '.*\.py$' | xargs black --check --line-length 160
# All done! 599 files would be left unchanged.
find ./ -maxdepth 1 -regex '.*\.py$' | xargs black --check --line-length 160
# All done! 1 file would be left unchanged.

(black 26.3.1, matching CI.)

🤖 Generated with Claude Code

Summary by CodeRabbit

• Refactor
  • Reorganized and standardized internal GEMM, GLU, Hadamard, and NVFP4 quantization code.
  • Preserved existing Rubin kernel behavior, validation, scheduling, and quantization workflows.
• Style
  • Improved formatting, spacing, multiline expressions, and error-message presentation.
  • Simplified redundant syntax without changing results or functionality.

[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: 86241d20-b808-45b3-938f-154d05ef9979

📥 Commits

Reviewing files that changed from the base of the PR and between 9b2058b and 0604b17.

📒 Files selected for processing (6)

• python/cudnn/gemm/cutedsl/grouped/glu/_blockscaled_api.py
• python/cudnn/gemm/cutedsl/grouped/glu_hadamard_quant/api.py
• python/cudnn/gemm/cutedsl/grouped/glu_hadamard_quant/moe_blockscaled_grouped_gemm_glu_hadamard_quant.py
• python/cudnn/gemm/cutedsl/grouped/glu_hadamard_quant/moe_blockscaled_grouped_gemm_glu_hadamard_quant_rubin.py
• python/cudnn/gemm/cutedsl/grouped/glu_hadamard_quant/quant_utils.py
• python/cudnn/gemm/cutedsl/grouped/glu_hadamard_quant/rht_utils.py

Included review availability: Your plan provides up to 12 included reviews per hour; 10 remain after this review.

---

📝 Walkthrough

Walkthrough

Changes

GLU Hadamard quantization

Layer / File(s)	Summary
Validation and helper expression cleanup python/cudnn/gemm/cutedsl/grouped/glu/_blockscaled_api.py, python/cudnn/gemm/cutedsl/grouped/glu_hadamard_quant/...	Existing validation, sizing, range, transform, and scaling expressions are reformatted or condensed.
Rubin B-reuse GEMM setup python/cudnn/gemm/cutedsl/grouped/glu_hadamard_quant/moe_blockscaled_grouped_gemm_glu_hadamard_quant_rubin.py	B-reuse validation, MMA setup, scale broadcast layout, output partitioning, and split GEMM calls retain their existing behavior.
Rubin activation epilogue python/cudnn/gemm/cutedsl/grouped/glu_hadamard_quant/moe_blockscaled_grouped_gemm_glu_hadamard_quant_rubin.py	The ACT epilogue now uses an explicit loop for B-keep and B-reuse halves.
Rubin RHT and NVFP4 epilogue python/cudnn/gemm/cutedsl/grouped/glu_hadamard_quant/moe_blockscaled_grouped_gemm_glu_hadamard_quant_rubin.py	The quantized epilogue now uses an explicit loop for FWHT, NVFP4 stores, scale-row writes, and synchronization.

Estimated code review effort: 2 (Simple) | ~10 minutes

Merge Risk: ⚪ Minimal · up to 0604b

This change only reformats existing Python code to match the repository’s Black version and does not alter behavior or compatibility. No actionable merge-blocking risk remains beyond normal checks and review.

Possibly related PRs

• NVIDIA/cudnn-frontend#637: Reformats code in the same GLU Hadamard quantization modules.
• NVIDIA/cudnn-frontend#645: Reformats the same SiTU-GLU situ_beta1 specialization code.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly and concisely identifies the Black 26.3.1 formatting change and affected Python files.
Description check	✅ Passed	The description completes all required sections and provides the scope, rationale, compatibility impact, and exact formatting checks.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
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.}