Don't construct Embed{}s for 1x1 filters in convolutions #1669
+66
−6
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
Fixes ROCm/rocMLIR-internal#1642 The [dN] + [dN]o (that is, index in the filter (from the K dimension) + index in the output (from the N dimension)) embedding in our conv2gemm is needed. However, in the case of 1x1 filters, that should simplify to just [dN]o (possibly scaled by the stride) that Embed{} operation was throwing off collapseContiguousMerges(), leading to inefficient codegen compared to the mathematically equivalent GEMM operation as described in the ticket. This commit resolves the issue and adds tests.
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## develop #1669 +/- ##
===========================================
- Coverage 83.69% 78.01% -5.69%
===========================================
Files 97 98 +1
Lines 26445 26501 +56
Branches 3796 3809 +13
===========================================
- Hits 22134 20674 -1460
- Misses 4276 4294 +18
- Partials 35 1533 +1498
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
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Outdated suggestions cannot be applied.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.
Fixes https://github.com/ROCm/rocMLIR-internal/issues/1642
The [dN] + [dN]o (that is, index in the filter (from the K dimension)
However, in the case of 1x1 filters, that should simplify to just [dN]o (possibly scaled by the stride) that Embed{} operation was throwing off collapseContiguousMerges(), leading to inefficient codegen compared to the mathematically equivalent GEMM operation as described in the ticket.
This commit resolves the issue and adds tests.