-
Notifications
You must be signed in to change notification settings - Fork 780
[DT] Collapse MaterializeEncodingIntoPaddingPass into the generic pass. #22472
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.
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
base: main
Are you sure you want to change the base?
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Pull Request Overview
This PR consolidates encoding materialization by merging the separate padding encoding pass into the main MaterializeDeviceEncodingPass. The changes eliminate the specialized MaterializeEncodingIntoPaddingPass and handle both data tiling and padding encoding resolution through a unified pass with configurable resolver types.
Key changes:
- Removed the separate
MaterializeEncodingIntoPaddingPassand its associatedaddEncodingToPaddingPassesfunction - Extended
MaterializeDeviceEncodingPasswith a newtest-gpu-encoding-resolveroption that supports bothgpu_data_tilingandgpu_paddingresolver kinds - Updated test files to use the unified pass with appropriate resolver configuration
Reviewed Changes
Copilot reviewed 14 out of 14 changed files in this pull request and generated 1 comment.
Show a summary per file
| File | Description |
|---|---|
| Passes.cpp | Removed call to addEncodingToPaddingPasses from the pass pipeline |
| Passes.td | Replaced boolean test-cl-gpu-target option with enum-based test-gpu-encoding-resolver option supporting multiple resolver types |
| Passes.h | Removed declaration of addEncodingToPaddingPasses function |
| MaterializeEncodingIntoPadding.cpp | Deleted entire file containing the specialized padding pass |
| MaterializeEncoding.cpp | Refactored to handle both data tiling and padding resolvers based on the new enum option |
| EncodingUtils.h | Added TestingResolverKind enum defining resolver types for testing |
| materialize_encoding_*.mlir | Updated test RUN lines to use the unified pass with appropriate resolver configuration |
| CMakeLists.txt / BUILD.bazel | Removed MaterializeEncodingIntoPadding.cpp from build files |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
ebf0278 to
4c5c8a7
Compare
|
I need to update few more comments and docs. |
Signed-off-by: hanhanW <[email protected]>
Signed-off-by: hanhanW <[email protected]>
1f2a5ba to
8c765b2
Compare
Signed-off-by: hanhanW <[email protected]>
Signed-off-by: hanhanW <[email protected]>
This PR consolidates encoding materialization by merging the separate padding encoding pass into the main
MaterializeDeviceEncodingPass. The changes eliminate the specializedMaterializeEncodingIntoPaddingPassand handle both data tiling and padding encoding resolution through a unified pass. The pass still prioritizes the pre-configured encoding resolver. If it is not present, it tries testing encoding resolver.Key changes:
MaterializeEncodingIntoPaddingPassand its associatedaddEncodingToPaddingPassesfunction.MaterializeDeviceEncodingPasswith a newtest-gpu-encoding-resolveroption that supports bothgpu_data_tilingandgpu_paddingresolver kinds for testing purpose.Fixes #20160