[Unity][Dlight] Fix decode-GeMV rule when spatial-inner without broadcasting #15330
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
Collaborator
|
Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from Reviewers by @-ing them in a comment.
Generated by tvm-bot |
Hzfengsy
approved these changes
Jul 16, 2023
…casting This PR fixes a bug of the previous decode-GeMV dlight scheduling. Previously, when the inner dimension of the largest tensor is spatial, in the end the fused epilogue block was not bound to any thread axis, which is wrong and will generate wrong GPU code with wrong numerical results. That is because after doing reverse-compute-at of the epilogue block, there are at lease one remaining spatial axis, and such axis is supposed to be bound to threadIdx. This PR fixes this issue, and add three test cases which can cover both the reduction-inner and spatial-inner cases with or without broadcasting.
MasterJH5574
force-pushed
the
unity-dev/2023-07-15-dlight-decode-gemv-spatial-inner
branch
from
3a31781 to
4c04c40
Compare
junrushao
pushed a commit
that referenced
this pull request
Jul 18, 2023
…casting (#15330) This PR fixes a bug of the previous decode-GeMV dlight scheduling. Previously, when the inner dimension of the largest tensor is spatial, in the end the fused epilogue block was not bound to any thread axis, which is wrong and will generate wrong GPU code with wrong numerical results. That is because after doing reverse-compute-at of the epilogue block, there are at lease one remaining spatial axis, and such axis is supposed to be bound to threadIdx. This PR fixes this issue, and add three test cases which can cover both the reduction-inner and spatial-inner cases with or without broadcasting.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 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.
This PR fixes a bug of the previous decode-GeMV dlight scheduling.
Previously, when the inner dimension of the largest tensor is spatial, in the end the fused epilogue block was not bound to any thread axis, which is wrong and will generate wrong GPU code with wrong numerical results. That is because after doing reverse-compute-at of the epilogue block, there are at lease one remaining spatial axis, and such axis is supposed to be bound to threadIdx.
This PR fixes this issue, and add three test cases which can cover both the reduction-inner and spatial-inner cases with or without broadcasting.