Skip to content

[MLIR][Affine] Fix signature of mlir::affine::permuteLoops #111100

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Nov 6, 2024
Merged

[MLIR][Affine] Fix signature of mlir::affine::permuteLoops #111100

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion mlir/include/mlir/Dialect/Affine/LoopUtils.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -136,7 +136,7 @@ bool isValidLoopInterchangePermutation(ArrayRef<AffineForOp> loops,
/// to inner. Returns the position in `inputNest` of the AffineForOp that
/// becomes the new outermost loop of this nest. This method always succeeds,
/// asserts out on invalid input / specifications.
unsigned permuteLoops(MutableArrayRef<AffineForOp> inputNest,
unsigned permuteLoops(ArrayRef<AffineForOp> inputNest,
ArrayRef<unsigned> permMap);

// Sinks all sequential loops to the innermost levels (while preserving
Expand Down
8 changes: 4 additions & 4 deletions mlir/lib/Dialect/Affine/Utils/LoopUtils.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1379,7 +1379,7 @@ mlir::affine::isPerfectlyNested(ArrayRef<AffineForOp> loops) {

// input[i] should move from position i -> permMap[i]. Returns the position in
// `input` that becomes the new outermost loop.
unsigned mlir::affine::permuteLoops(MutableArrayRef<AffineForOp> input,
unsigned mlir::affine::permuteLoops(ArrayRef<AffineForOp> input,
ArrayRef<unsigned> permMap) {
assert(input.size() == permMap.size() && "invalid permutation map size");
// Check whether the permutation spec is valid. This is a small vector - we'll
Expand All @@ -1406,8 +1406,8 @@ unsigned mlir::affine::permuteLoops(MutableArrayRef<AffineForOp> input,
// Move the innermost loop body to the loop that would be the innermost in the
// permuted nest (only if the innermost loop is going to change).
if (permMap.back() != input.size() - 1) {
auto *destBody = input[invPermMap.back().second].getBody();
auto *srcBody = input.back().getBody();
Block *destBody = ((AffineForOp)input[invPermMap.back().second]).getBody();
Copy link
Member

@MacDue MacDue Oct 13, 2024

Choose a reason for hiding this comment

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

Why add the C style casts?

Copy link
Member

@MacDue MacDue Oct 13, 2024

Choose a reason for hiding this comment

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

I feel like MutableArrayRef is probably correct given it's mutating the AffineForOps... 🤔

Copy link
Member

@ftynse ftynse Oct 28, 2024

Choose a reason for hiding this comment

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

It doesn't mutate the value in the array, so ArrayRef is fine. Not a fan of C-style casts, but okay.

Copy link
Contributor Author

@bondhugula bondhugula Nov 6, 2024 •
edited
Loading

Choose a reason for hiding this comment

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

The C-style cast is to work around const'ness friction in MLIR - MLIR IR types don't use consts but ArrayRef and similar LLVM data structures containing them will give you const versions of those IR types.

There's no AffineForOp mutation happening here.

Block *srcBody = ((AffineForOp)input.back()).getBody();
destBody->getOperations().splice(destBody->begin(),
srcBody->getOperations(), srcBody->begin(),
std::prev(srcBody->end()));
Expand Down Expand Up @@ -1437,7 +1437,7 @@ unsigned mlir::affine::permuteLoops(MutableArrayRef<AffineForOp> input,
continue;

// Move input[i] to its surrounding loop in the transformed nest.
auto *destBody = input[parentPosInInput].getBody();
auto *destBody = ((AffineForOp)input[parentPosInInput]).getBody();
destBody->getOperations().splice(destBody->begin(),
input[i]->getBlock()->getOperations(),
Block::iterator(input[i]));
Expand Down