Skip to content

[MLIR] enhance get_scalar op, only requires dyn_leaf_cnt = 1#398

Merged
coderfeli merged 2 commits intomainfrom
pr/enhance-get-scalar
Apr 14, 2026
Merged

[MLIR] enhance get_scalar op, only requires dyn_leaf_cnt = 1#398
coderfeli merged 2 commits intomainfrom
pr/enhance-get-scalar

Conversation

@sjfeng1999
Copy link
Copy Markdown
Collaborator

Motivation

Technical Details

Test Plan

Test Result

Submission Checklist

Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Enhances Fly dialect get_scalar handling so it can accept nested singleton IntTuple shapes (e.g. ((((42)))) / ((((?))))) rather than requiring the tuple attr to be a direct leaf, and adds coverage in the layout-lowering MLIR tests.

Changes:

  • Update GetScalarOp return type inference to unwrap rank-1 nesting until a leaf is reached.
  • Update fly-layout-lowering’s GetScalarOp rewrite to perform the same unwrapping.
  • Add MLIR tests for nested singleton static and dynamic get_scalar.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

File Description
tests/mlir/Transforms/layout_lowering.mlir Adds FileCheck tests validating nested singleton get_scalar is eliminated to constants/SSA values.
lib/Dialect/Fly/Transforms/LayoutLowering.cpp Adjusts GetScalarOp lowering to unwrap nested singleton tuples before extracting the leaf.
lib/Dialect/Fly/IR/FlyOps.cpp Adjusts GetScalarOp type inference to accept nested singleton tuples and improve the diagnostic for non-scalars.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Comment thread lib/Dialect/Fly/Transforms/LayoutLowering.cpp
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
@coderfeli coderfeli merged commit c8dbdaf into main Apr 14, 2026
9 checks passed
@coderfeli coderfeli deleted the pr/enhance-get-scalar branch April 14, 2026 05:21
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