[RemoveLayoutConversions]: Reduce loop carried values #4915
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Tiotto, Ettore <[email protected]>
Signed-off-by: Tiotto, Ettore <[email protected]>
There was a problem hiding this comment.
Pull Request Overview
This PR implements an optimization to reduce loop carried values in the RemoveLayoutConversions pass by reusing equivalent loop results with layout conversion operations instead of carrying redundant values through loops.
- Adds logic to identify and replace redundant loop arguments with equivalent ones plus layout conversions
- Handles LoadOp and StoreOp operations that use these loop results
- Reduces memory usage and improves loop efficiency by eliminating unnecessary loop-carried tensor pointer values
Reviewed Changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.
| File | Description |
|---|---|
| RemoveLayoutConversions.cpp | Implements the core optimization logic to reduce loop carried values by replacing redundant arguments with conversions |
| backward_combine_dpas_dot_layout.mlir | Adds test case to verify the optimization works correctly for tensor pointer operations in loops |
Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.
third_party/intel/lib/TritonIntelGPUTransforms/RemoveLayoutConversions.cpp
Show resolved
Hide resolved
third_party/intel/lib/TritonIntelGPUTransforms/RemoveLayoutConversions.cpp
Show resolved
Hide resolved
third_party/intel/lib/TritonIntelGPUTransforms/RemoveLayoutConversions.cpp
Outdated
Show resolved
Hide resolved
etiotto
changed the title
Signed-off-by: Tiotto, Ettore <[email protected]>
Comment on lines
+1275
to
+1278
| if (!isa<BlockArgument>(pair.first)) | ||
| continue; | ||
|
|
||
| auto arg = cast<BlockArgument>(pair.first); |
There was a problem hiding this comment.
Suggested change
| if (!isa<BlockArgument>(pair.first)) | |
| continue; | |
| auto arg = cast<BlockArgument>(pair.first); | |
| auto arg = dyn_cast<BlockArgument>(pair.first); | |
| if (!arg) | |
| continue; | |
Comment on lines
+1295
to
+1296
| for (OpOperand &use : loopRes.getUses()) { | ||
| Operation *user = use.getOwner(); |
There was a problem hiding this comment.
?
Suggested change
| for (OpOperand &use : loopRes.getUses()) { | |
| Operation *user = use.getOwner(); | |
| for (Operation *user : loopRes.getUsers()) { |
Comment on lines
+381
to
+383
| %24 = arith.truncf %23#0 : tensor<64x256xf32, #dpas> to tensor<64x256xf16, #dpas> | ||
| %27 = tt.make_tensor_ptr %arg2, [%c0_i64, %c0_i64], [%c0_i64, %c1_i64], [%c0_i32, %c0_i32] {order = array<i32: 1, 0>} : <tensor<64x256xf16, #dpas>> | ||
| tt.store %27, %24 {boundaryCheck = array<i32: 0, 1>} : !tt.ptr<tensor<64x256xf16, #dpas>> |
There was a problem hiding this comment.
looks like these 3 lines can be removed?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR implements an optimization to reduce loop carried values in the RemoveLayoutConversions pass by reusing equivalent loop results with layout conversion operations instead of carrying redundant values through loops.
Fixes issue #4901