Skip to content

Fill CUDA Transpose operator opset gap (23 → 25)#27740

Merged
tianleiwu merged 4 commits intomainfrom
copilot/update-transpose-cuda-operator
Apr 6, 2026
Merged

Fill CUDA Transpose operator opset gap (23 → 25)#27740
tianleiwu merged 4 commits intomainfrom
copilot/update-transpose-cuda-operator

Conversation

Copy link
Copy Markdown
Contributor

Copilot AI commented Mar 18, 2026

Description

Extends the CUDA Transpose kernel registration from opset 23 to opset 25.

  • transpose.cc: Cap existing opset 23 kernel to versioned (23, 24), add new non-versioned kernel at opset 25
  • cuda_execution_provider.cc: Update forward declarations and BuildKernelCreateInfo entries to match; add new // Opset 25 section
  • docs/OperatorKernels.md: Update CUDA Transpose entry from 23+ to 25+ with new [23, 24] versioned range

No functional or type constraint changes — the kernel implementation is identical across these opsets.

Motivation and Context

CUDA EP's Transpose registration stopped at opset 23 while the ONNX spec defines it through opset 25. This is one of the P1 gaps tracked in #27729, following the same pattern as #27728.

Limitation

This PR does not add support of new data type for Transpose:

  • int2 (opset 25)
  • float8e8m0 (opset 24)
  • float4e2m1 (opset 23)
  • float8e4m3fn,float8e4m3fnuz, float8e5m2, float8e5m2fnuz, uint4, int4 (opset 21)

Co-authored-by: tianleiwu <30328909+tianleiwu@users.noreply.github.com>
@tianleiwu
Copy link
Copy Markdown
Contributor

/azp run Windows GPU Doc Gen CI Pipeline

@azure-pipelines
Copy link
Copy Markdown

Azure Pipelines successfully started running 1 pipeline(s).

@tianleiwu
Copy link
Copy Markdown
Contributor

@copilot, please update docs/OperatorKernels.md, and merge latest main branch to this branch.

Copilot AI and others added 2 commits March 18, 2026 23:12
Co-authored-by: tianleiwu <30328909+tianleiwu@users.noreply.github.com>
Copy link
Copy Markdown
Contributor Author

Copilot AI commented Mar 18, 2026

@copilot, please update docs/OperatorKernels.md, and merge latest main branch to this branch.

Updated docs/OperatorKernels.md with the new Transpose opset ranges (25+ and [23, 24]) and merged latest main into the branch. See commit e47282a.

Copilot AI requested a review from tianleiwu March 18, 2026 23:13
@tianleiwu tianleiwu marked this pull request as ready for review March 18, 2026 23:47
@tianleiwu
Copy link
Copy Markdown
Contributor

/azp run Linux QNN CI Pipeline, Win_TRT_Minimal_CUDA_Test_CI, Windows ARM64 QNN CI Pipeline, Windows GPU Doc Gen CI Pipeline

@azure-pipelines
Copy link
Copy Markdown

Azure Pipelines successfully started running 4 pipeline(s).

@tianleiwu tianleiwu requested a review from titaiwangms April 2, 2026 07:45
@tianleiwu tianleiwu merged commit cd48875 into main Apr 6, 2026
89 checks passed
@tianleiwu tianleiwu deleted the copilot/update-transpose-cuda-operator branch April 6, 2026 19:58
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.

3 participants