Fix A-scale preshuffle LDS reads for multi-wave M configurations#1056
Merged
harsh-nod merged 4 commits intoiree-org:mainfrom Mar 6, 2026
Merged
Fix A-scale preshuffle LDS reads for multi-wave M configurations#1056harsh-nod merged 4 commits intoiree-org:mainfrom
harsh-nod merged 4 commits intoiree-org:mainfrom
Conversation
Signed-off-by: Sanket Pandit <sanket.pandit@amd.com>
Signed-off-by: Sanket Pandit <sanket.pandit@amd.com>
xintin
pushed a commit
that referenced
this pull request
Mar 6, 2026
- Fix A-scale LDS addressing in `preshuffle_scale_to_shared` that caused wrong results (50% mismatch) when multiple waves cover different M-blocks (e.g., `wave_shape=(2,2)` with `block=(128,32,256)`) - Handle `_Rational` values inside `Max`/`Min` and comparison operators in the emitter, fixing codegen failures with dynamic shapes - Add `(128, 32, 256)` and `(256, 224, 256)` macrotiles with `wave_shape=(2,2)` to preshuffle tests, split tile lists by wave shape (`MACROTILES_PRESHUFFLE_1x4`, `MACROTILES_PRESHUFFLE_2x2`), and parametrize both static and dynamic tests over `(block_shape, wave_shape)` pairs --------- Signed-off-by: Sanket Pandit <sanket.pandit@amd.com> Signed-off-by: xintin <gaurav.verma@amd.com>
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
preshuffle_scale_to_sharedthat caused wrong results (50% mismatch) when multiple waves cover different M-blocks (e.g.,wave_shape=(2,2)withblock=(128,32,256))_Rationalvalues insideMax/Minand comparison operators in the emitter, fixing codegen failures with dynamic shapes(128, 32, 256)and(256, 224, 256)macrotiles withwave_shape=(2,2)to preshuffle tests, split tile lists by wave shape (MACROTILES_PRESHUFFLE_1x4,MACROTILES_PRESHUFFLE_2x2), and parametrize both static and dynamic tests over(block_shape, wave_shape)pairs