[SR] Eliminate extra permute ops before aten::sum
#74481
Closed
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
This diff fixes an interesting performance issue related to `permute_copy`. We see this pattern frequently: ``` y = torch.permute(x, (0, 2, 1)) z = torch.sum(y, dim=-1) ``` With copy variants off, we get a strided output from `permute`, and we hit this (faster) kernel in `sum`: https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cpu/SumKernel.cpp#L589 But with copy variants on, we get a contiguous output from `permute_copy`, which causes us to hit the slower reduction: https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cpu/SumKernel.cpp#L597 But the permute is actually unnecessary, we can just statically turn the graph into this to ensure that the fast kernel is hit with copy variants on: ``` z = torch.sum(x, dim=1) ``` Differential Revision: [D34992319](https://our.internmc.facebook.com/intern/diff/D34992319/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34992319/)! [ghstack-poisoned]
Contributor
🔗 Helpful links
💊 CI failures summary and remediationsAs of commit 9b55df5 (more details on the Dr. CI page): 💚 💚 Looks good so far! There are no failures yet. 💚 💚 This comment was automatically generated by Dr. CI (expand for details).Please report bugs/suggestions to the (internal) Dr. CI Users group. |
facebook-github-bot
added
cla signed
oncall: jit
This diff fixes an interesting performance issue related to `permute_copy`. We see this pattern frequently: ``` y = torch.permute(x, (0, 2, 1)) z = torch.sum(y, dim=-1) ``` With copy variants off, we get a strided output from `permute`, and we hit this (faster) kernel in `sum`: https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cpu/SumKernel.cpp#L589 But with copy variants on, we get a contiguous output from `permute_copy`, which causes us to hit the slower reduction: https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cpu/SumKernel.cpp#L597 But the permute is actually unnecessary, we can just statically turn the graph into this to ensure that the fast kernel is hit with copy variants on: ``` z = torch.sum(x, dim=1) ``` Differential Revision: [D34992319](https://our.internmc.facebook.com/intern/diff/D34992319/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34992319/)! [ghstack-poisoned]
mikeiovine
pushed a commit
that referenced
this pull request
Mar 21, 2022
Pull Request resolved: #74481 This diff fixes an interesting performance issue related to `permute_copy`. We see this pattern frequently: ``` y = torch.permute(x, (0, 2, 1)) z = torch.sum(y, dim=-1) ``` With copy variants off, we get a strided output from `permute`, and we hit this (faster) kernel in `sum`: https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cpu/SumKernel.cpp#L589 But with copy variants on, we get a contiguous output from `permute_copy`, which causes us to hit the slower reduction: https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cpu/SumKernel.cpp#L597 But the permute is actually unnecessary, we can just statically turn the graph into this to ensure that the fast kernel is hit with copy variants on: ``` z = torch.sum(x, dim=1) ``` ghstack-source-id: 151822137 Differential Revision: [D34992319](https://our.internmc.facebook.com/intern/diff/D34992319/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34992319/)!
This diff fixes an interesting performance issue related to `permute_copy`. We see this pattern frequently: ``` y = torch.permute(x, (0, 2, 1)) z = torch.sum(y, dim=-1) ``` With copy variants off, we get a strided output from `permute`, and we hit this (faster) kernel in `sum`: https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cpu/SumKernel.cpp#L589 But with copy variants on, we get a contiguous output from `permute_copy`, which causes us to hit the slower reduction: https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cpu/SumKernel.cpp#L597 But the permute is actually unnecessary, we can just statically turn the graph into this to ensure that the fast kernel is hit with copy variants on: ``` z = torch.sum(x, dim=1) ``` Differential Revision: [D34992319](https://our.internmc.facebook.com/intern/diff/D34992319/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34992319/)! [ghstack-poisoned]
mikeiovine
pushed a commit
that referenced
this pull request
Mar 23, 2022
Pull Request resolved: #74481 This diff fixes an interesting performance issue related to `permute_copy`. We see this pattern frequently: ``` y = torch.permute(x, (0, 2, 1)) z = torch.sum(y, dim=-1) ``` With copy variants off, we get a strided output from `permute`, and we hit this (faster) kernel in `sum`: https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cpu/SumKernel.cpp#L589 But with copy variants on, we get a contiguous output from `permute_copy`, which causes us to hit the slower reduction: https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cpu/SumKernel.cpp#L597 But the permute is actually unnecessary, we can just statically turn the graph into this to ensure that the fast kernel is hit with copy variants on: ``` z = torch.sum(x, dim=1) ``` ghstack-source-id: 152003888 Differential Revision: [D34992319](https://our.internmc.facebook.com/intern/diff/D34992319/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34992319/)!
facebook-github-bot
pushed a commit
that referenced
this pull request
Mar 23, 2022
Summary: Pull Request resolved: #74481 This diff fixes an interesting performance issue related to `permute_copy`. We see this pattern frequently: ``` y = torch.permute(x, (0, 2, 1)) z = torch.sum(y, dim=-1) ``` With copy variants off, we get a strided output from `permute`, and we hit this (faster) kernel in `sum`: https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cpu/SumKernel.cpp#L589 But with copy variants on, we get a contiguous output from `permute_copy`, which causes us to hit the slower reduction: https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cpu/SumKernel.cpp#L597 But the permute is actually unnecessary, we can just statically turn the graph into this to ensure that the fast kernel is hit with copy variants on: ``` z = torch.sum(x, dim=1) ``` ghstack-source-id: 152003888 Reviewed By: navahgar Differential Revision: D34992319 fbshipit-source-id: 0baf493708ee2180c899814a954d220d88ba1d4f
Contributor
|
Hey @mikeiovine. |
shahofblah
pushed a commit
that referenced
this pull request
Mar 25, 2022
Summary: Pull Request resolved: #74481 This diff fixes an interesting performance issue related to `permute_copy`. We see this pattern frequently: ``` y = torch.permute(x, (0, 2, 1)) z = torch.sum(y, dim=-1) ``` With copy variants off, we get a strided output from `permute`, and we hit this (faster) kernel in `sum`: https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cpu/SumKernel.cpp#L589 But with copy variants on, we get a contiguous output from `permute_copy`, which causes us to hit the slower reduction: https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cpu/SumKernel.cpp#L597 But the permute is actually unnecessary, we can just statically turn the graph into this to ensure that the fast kernel is hit with copy variants on: ``` z = torch.sum(x, dim=1) ``` ghstack-source-id: 152003888 Reviewed By: navahgar Differential Revision: D34992319 fbshipit-source-id: 0baf493708ee2180c899814a954d220d88ba1d4f (cherry picked from commit 797b6be)
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.
Stack from ghstack (oldest at bottom):
aten::sum#74481This diff fixes an interesting performance issue related to
permute_copy.We see this pattern frequently:
With copy variants off, we get a strided output from
permute, and we hit this (faster) kernel insum: https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cpu/SumKernel.cpp#L589But with copy variants on, we get a contiguous output from
permute_copy, which causes us to hit the slower reduction:https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/cpu/SumKernel.cpp#L597
But the permute is actually unnecessary, we can just statically turn the graph into this to ensure that the fast kernel is hit with copy variants on:
Differential Revision: D34992319
NOTE FOR REVIEWERS: This PR has internal Facebook specific changes or comments, please review them on Phabricator!