[SR] Make sigrid_transforms fusion work on graph outputs #73091
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 is a re-work of D33669034; the change was backed out due to a data race causing crashes. The `output_types` vector was the culprit. It was previously lazily initialized on the first iteration. This was problematic because of static runtime's hidden assumption that ops are thread-safe. The re-work now only does the list unpack fusion if the output dtypes can be statically determined, e.g. if the sigrid transforms instance and `use_offsets` are both constant. Note that this is true for all the models we care about. Also, we were already partially making this assumption by dereferencing the `std::optional` sigrid transforms instance in most of the ops. Another advantage of this is that it makes the code simpler compared to D33669034. Once the output types are determined, they can be moved into the op lambda and shared as read-only data. Differential Revision: [D34290401](https://our.internmc.facebook.com/intern/diff/D34290401/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34290401/)! [ghstack-poisoned]
🔗 Helpful links
💊 CI failures summary and remediationsAs of commit 97f88da (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. |
CI Flow Status⚛️ CI FlowRuleset - Version:
|
facebook-github-bot
added
cla signed
oncall: jit
This is a re-work of D33669034; the change was backed out due to a data race causing crashes. The `output_types` vector was the culprit. It was previously lazily initialized on the first iteration. This was problematic because of static runtime's hidden assumption that ops are thread-safe. The re-work now only does the list unpack fusion if the output dtypes can be statically determined, e.g. if the sigrid transforms instance and `use_offsets` are both constant. Note that this is true for all the models we care about. Also, we were already partially making this assumption by dereferencing the `std::optional` sigrid transforms instance in most of the ops. Another advantage of this is that it makes the code simpler compared to D33669034. Once the output types are determined, they can be moved into the op lambda and shared as read-only data. Differential Revision: [D34290401](https://our.internmc.facebook.com/intern/diff/D34290401/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34290401/)! [ghstack-poisoned]
mikeiovine
pushed a commit
that referenced
this pull request
Feb 18, 2022
Pull Request resolved: #73091 This is a re-work of D33669034; the change was backed out due to a data race causing crashes. The `output_types` vector was the culprit. It was previously lazily initialized on the first iteration. This was problematic because of static runtime's hidden assumption that ops are thread-safe. The re-work now only does the list unpack fusion if the output dtypes can be statically determined, e.g. if the sigrid transforms instance and `use_offsets` are both constant. Note that this is true for all the models we care about. Also, we were already partially making this assumption by dereferencing the `std::optional` sigrid transforms instance in most of the ops. Another advantage of this is that it makes the code simpler compared to D33669034. Once the output types are determined, they can be moved into the op lambda and shared as read-only data. ghstack-source-id: 149529436 Differential Revision: [D34290401](https://our.internmc.facebook.com/intern/diff/D34290401/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34290401/)!
This is a re-work of D33669034; the change was backed out due to a data race causing crashes. The `output_types` vector was the culprit. It was previously lazily initialized on the first iteration. This was problematic because of static runtime's hidden assumption that ops are thread-safe. The re-work now only does the list unpack fusion if the output dtypes can be statically determined, e.g. if the sigrid transforms instance and `use_offsets` are both constant. Note that this is true for all the models we care about. Also, we were already partially making this assumption by dereferencing the `std::optional` sigrid transforms instance in most of the ops. Another advantage of this is that it makes the code simpler compared to D33669034. Once the output types are determined, they can be moved into the op lambda and shared as read-only data. Differential Revision: [D34290401](https://our.internmc.facebook.com/intern/diff/D34290401/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34290401/)! [ghstack-poisoned]
mikeiovine
pushed a commit
that referenced
this pull request
Feb 23, 2022
Pull Request resolved: #73091 This is a re-work of D33669034; the change was backed out due to a data race causing crashes. The `output_types` vector was the culprit. It was previously lazily initialized on the first iteration. This was problematic because of static runtime's hidden assumption that ops are thread-safe. The re-work now only does the list unpack fusion if the output dtypes can be statically determined, e.g. if the sigrid transforms instance and `use_offsets` are both constant. Note that this is true for all the models we care about. Also, we were already partially making this assumption by dereferencing the `std::optional` sigrid transforms instance in most of the ops. Another advantage of this is that it makes the code simpler compared to D33669034. Once the output types are determined, they can be moved into the op lambda and shared as read-only data. ghstack-source-id: 149756644 Differential Revision: [D34290401](https://our.internmc.facebook.com/intern/diff/D34290401/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34290401/)!
This is a re-work of D33669034; the change was backed out due to a data race causing crashes. The `output_types` vector was the culprit. It was previously lazily initialized on the first iteration. This was problematic because of static runtime's hidden assumption that ops are thread-safe. The re-work now only does the list unpack fusion if the output dtypes can be statically determined, e.g. if the sigrid transforms instance and `use_offsets` are both constant. Note that this is true for all the models we care about. Also, we were already partially making this assumption by dereferencing the `std::optional` sigrid transforms instance in most of the ops. Another advantage of this is that it makes the code simpler compared to D33669034. Once the output types are determined, they can be moved into the op lambda and shared as read-only data. Differential Revision: [D34290401](https://our.internmc.facebook.com/intern/diff/D34290401/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34290401/)! [ghstack-poisoned]
mikeiovine
pushed a commit
that referenced
this pull request
Feb 24, 2022
Pull Request resolved: #73091 This is a re-work of D33669034; the change was backed out due to a data race causing crashes. The `output_types` vector was the culprit. It was previously lazily initialized on the first iteration. This was problematic because of static runtime's hidden assumption that ops are thread-safe. The re-work now only does the list unpack fusion if the output dtypes can be statically determined, e.g. if the sigrid transforms instance and `use_offsets` are both constant. Note that this is true for all the models we care about. Also, we were already partially making this assumption by dereferencing the `std::optional` sigrid transforms instance in most of the ops. Another advantage of this is that it makes the code simpler compared to D33669034. Once the output types are determined, they can be moved into the op lambda and shared as read-only data. ghstack-source-id: 149896109 Differential Revision: [D34290401](https://our.internmc.facebook.com/intern/diff/D34290401/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34290401/)!
This is a re-work of D33669034; the change was backed out due to a data race causing crashes. The `output_types` vector was the culprit. It was previously lazily initialized on the first iteration. This was problematic because of static runtime's hidden assumption that ops are thread-safe. The re-work now only does the list unpack fusion if the output dtypes can be statically determined, e.g. if the sigrid transforms instance and `use_offsets` are both constant. Note that this is true for all the models we care about. Also, we were already partially making this assumption by dereferencing the `std::optional` sigrid transforms instance in most of the ops. Another advantage of this is that it makes the code simpler compared to D33669034. Once the output types are determined, they can be moved into the op lambda and shared as read-only data. Differential Revision: [D34290401](https://our.internmc.facebook.com/intern/diff/D34290401/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34290401/)! [ghstack-poisoned]
This is a re-work of D33669034; the change was backed out due to a data race causing crashes. The `output_types` vector was the culprit. It was previously lazily initialized on the first iteration. This was problematic because of static runtime's hidden assumption that ops are thread-safe. The re-work now only does the list unpack fusion if the output dtypes can be statically determined, e.g. if the sigrid transforms instance and `use_offsets` are both constant. Note that this is true for all the models we care about. Also, we were already partially making this assumption by dereferencing the `std::optional` sigrid transforms instance in most of the ops. Another advantage of this is that it makes the code simpler compared to D33669034. Once the output types are determined, they can be moved into the op lambda and shared as read-only data. Differential Revision: [D34290401](https://our.internmc.facebook.com/intern/diff/D34290401/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34290401/)! [ghstack-poisoned]
mikeiovine
pushed a commit
that referenced
this pull request
Mar 7, 2022
Pull Request resolved: #73091 This is a re-work of D33669034; the change was backed out due to a data race causing crashes. The `output_types` vector was the culprit. It was previously lazily initialized on the first iteration. This was problematic because of static runtime's hidden assumption that ops are thread-safe. The re-work now only does the list unpack fusion if the output dtypes can be statically determined, e.g. if the sigrid transforms instance and `use_offsets` are both constant. Note that this is true for all the models we care about. Also, we were already partially making this assumption by dereferencing the `std::optional` sigrid transforms instance in most of the ops. Another advantage of this is that it makes the code simpler compared to D33669034. Once the output types are determined, they can be moved into the op lambda and shared as read-only data. ghstack-source-id: 150704445 Differential Revision: [D34290401](https://our.internmc.facebook.com/intern/diff/D34290401/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D34290401/)!
facebook-github-bot
pushed a commit
that referenced
this pull request
Mar 8, 2022
Summary: Pull Request resolved: #73091 This is a re-work of D33669034 (c76c491); the change was backed out due to a data race causing crashes. The `output_types` vector was the culprit. It was previously lazily initialized on the first iteration. This was problematic because of static runtime's hidden assumption that ops are thread-safe. The re-work now only does the list unpack fusion if the output dtypes can be statically determined, e.g. if the sigrid transforms instance and `use_offsets` are both constant. Note that this is true for all the models we care about. Also, we were already partially making this assumption by dereferencing the `std::optional` sigrid transforms instance in most of the ops. Another advantage of this is that it makes the code simpler compared to D33669034 (c76c491). Once the output types are determined, they can be moved into the op lambda and shared as read-only data. ghstack-source-id: 150704445 Reviewed By: d1jang Differential Revision: D34290401 fbshipit-source-id: 9573e6f08ee9e8282de961bf5f5cc8d32b81e601
|
Hey @mikeiovine. |
cyyever
pushed a commit
to cyyever/pytorch_private
that referenced
this pull request
Mar 9, 2022
Summary: Pull Request resolved: pytorch/pytorch#73091 This is a re-work of D33669034 (pytorch/pytorch@c76c491); the change was backed out due to a data race causing crashes. The `output_types` vector was the culprit. It was previously lazily initialized on the first iteration. This was problematic because of static runtime's hidden assumption that ops are thread-safe. The re-work now only does the list unpack fusion if the output dtypes can be statically determined, e.g. if the sigrid transforms instance and `use_offsets` are both constant. Note that this is true for all the models we care about. Also, we were already partially making this assumption by dereferencing the `std::optional` sigrid transforms instance in most of the ops. Another advantage of this is that it makes the code simpler compared to D33669034 (pytorch/pytorch@c76c491). Once the output types are determined, they can be moved into the op lambda and shared as read-only data. ghstack-source-id: 150704445 Reviewed By: d1jang Differential Revision: D34290401 fbshipit-source-id: 9573e6f08ee9e8282de961bf5f5cc8d32b81e601 (cherry picked from commit 715b0077bd18cb144b9653f5f51057b9440252ad)
cyyever
pushed a commit
to cyyever/pytorch_private
that referenced
this pull request
Mar 9, 2022
Summary: Pull Request resolved: pytorch/pytorch#73091 This is a re-work of D33669034 (pytorch/pytorch@c76c491); the change was backed out due to a data race causing crashes. The `output_types` vector was the culprit. It was previously lazily initialized on the first iteration. This was problematic because of static runtime's hidden assumption that ops are thread-safe. The re-work now only does the list unpack fusion if the output dtypes can be statically determined, e.g. if the sigrid transforms instance and `use_offsets` are both constant. Note that this is true for all the models we care about. Also, we were already partially making this assumption by dereferencing the `std::optional` sigrid transforms instance in most of the ops. Another advantage of this is that it makes the code simpler compared to D33669034 (pytorch/pytorch@c76c491). Once the output types are determined, they can be moved into the op lambda and shared as read-only data. ghstack-source-id: 150704445 Reviewed By: d1jang Differential Revision: D34290401 fbshipit-source-id: 9573e6f08ee9e8282de961bf5f5cc8d32b81e601 (cherry picked from commit 715b0077bd18cb144b9653f5f51057b9440252ad)
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):
This is a re-work of D33669034; the change was backed out due to a data race causing crashes.
The
output_typesvector was the culprit. It was previously lazily initialized on the first iteration. This was problematic because of static runtime's hidden assumption that ops are thread-safe.The re-work now only does the list unpack fusion if the output dtypes can be statically determined, e.g. if the sigrid transforms instance and
use_offsetsare both constant. Note that this is true for all the models we care about. Also, we were already partially making this assumption by dereferencing thestd::optionalsigrid transforms instance in most of the ops. Another advantage of this is that it makes the code simpler compared to D33669034.Once the output types are determined, they can be moved into the op lambda and shared as read-only data.
Differential Revision: D34290401
NOTE FOR REVIEWERS: This PR has internal Facebook specific changes or comments, please review them on Phabricator!