Add rewrite for matmul when only one of the inputs has batched dimensions #558
Conversation
This file contains 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
ricardoV94
added
enhancement
ricardoV94
force-pushed
the
batched_matmul_to_reshaped_dot_rewrite
branch
from
0f6e911
to
d28b35f
Compare
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## main #558 +/- ##
=======================================
Coverage 80.91% 80.92%
=======================================
Files 162 162
Lines 46436 46458 +22
Branches 11357 11364 +7
=======================================
+ Hits 37576 37597 +21
Misses 6638 6638
- Partials 2222 2223 +1
|
…ions This rewrites replaces a batched matmul by a core matmul by raveling the batched dimensions of the batched input, doing a 2D matmul and then unravelling the batched dimensions of the output. This sidesteps the Blockwise Dot operation and allows specialization into BLAS routines. The idea was taken from these two discussions: numpy/numpy#7569 numpy/numpy#8957
ricardoV94
force-pushed
the
batched_matmul_to_reshaped_dot_rewrite
branch
from
d28b35f
to
b19c6ab
Compare
twiecki
approved these changes
Dec 20, 2023
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Outdated suggestions cannot be applied.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.
Description
This rewrites replaces a batched matmul by a core matmul by raveling the batched dimensions of the batched input, doing a 2D matmul and then unravelling the batched dimensions of the output.
This sidesteps the Blockwise Dot operation and allows specialization into BLAS routines.
The idea was taken from these two discussions:
numpy/numpy#7569
numpy/numpy#8957
Related Issue
Checklist
Type of change