Add C10_EMBEDDED to gate ostream usage in Half/BFloat16 #140566
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
We want to use Half/BFloat16 in ExecuTorch to support shared kernel code. They will need to be used in ExecuTorch core, so they can't have streams. This diff introduces a macro to gate the stream code off. Differential Revision: [D65888035](https://our.internmc.facebook.com/intern/diff/D65888035/) [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/140566
Note: Links to docs will display an error until the docs builds have been completed. ❗ 2 Active SEVsThere are 2 currently active SEVs. If your PR is affected, please view them below:
✅ No FailuresAs of commit 43c24d0 with merge base 891ba2e ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
This was referenced Nov 13, 2024
swolchok
added a commit
that referenced
this pull request
Nov 13, 2024
We want to use Half/BFloat16 in ExecuTorch to support shared kernel code. They will need to be used in ExecuTorch core, so they can't have streams. This diff introduces a macro to gate the stream code off. Differential Revision: [D65888035](https://our.internmc.facebook.com/intern/diff/D65888035/) ghstack-source-id: 253352122 Pull Request resolved: #140566
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D65888035 |
Contributor
Author
|
existing executorch half/bfloat16 contain |
ezyang
approved these changes
Nov 14, 2024
We want to use Half/BFloat16 in ExecuTorch to support shared kernel code. They will need to be used in ExecuTorch core, so they can't have streams. This diff introduces a macro to gate the stream code off. Differential Revision: [D65888035](https://our.internmc.facebook.com/intern/diff/D65888035/) [ghstack-poisoned]
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D65888035 |
We want to use Half/BFloat16 in ExecuTorch to support shared kernel code. They will need to be used in ExecuTorch core, so they can't have streams. This diff introduces a macro to gate the stream code off. Differential Revision: [D65888035](https://our.internmc.facebook.com/intern/diff/D65888035/) [ghstack-poisoned]
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D65888035 |
malfet
approved these changes
Nov 18, 2024
pytorchmergebot
pushed a commit
that referenced
this pull request
Nov 18, 2024
Accessing the inactive member of a union is undefined behavior. Fortunately, we have c10::bit_cast. Differential Revision: [D65888680](https://our.internmc.facebook.com/intern/diff/D65888680/) Pull Request resolved: #140567 Approved by: https://github.com/Skylion007, https://github.com/malfet ghstack dependencies: #140564, #140565, #140566
pytorchmergebot
pushed a commit
that referenced
this pull request
Nov 18, 2024
This was added in pytorch/executorch#1789 . I'm working on sharing Half.h with ExecuTorch, and this is a missing feature. Differential Revision: [D65949409](https://our.internmc.facebook.com/intern/diff/D65949409/) Pull Request resolved: #140720 Approved by: https://github.com/malfet ghstack dependencies: #140564, #140565, #140566, #140567
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D65888035 |
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D65888035 |
pytorchmergebot
pushed a commit
that referenced
this pull request
Nov 19, 2024
It passed PyTorch CI, but internally we saw failures from this. Differential Revision: [D66137897](https://our.internmc.facebook.com/intern/diff/D66137897/) Pull Request resolved: #140994 Approved by: https://github.com/malfet ghstack dependencies: #140564, #140565, #140566, #140567, #140720
pytorchmergebot
pushed a commit
that referenced
this pull request
Nov 19, 2024
Make what we're doing as obvious as possible to the compiler. Differential Revision: [D66108811](https://our.internmc.facebook.com/intern/diff/D66108811/) Pull Request resolved: #141035 Approved by: https://github.com/Skylion007, https://github.com/ezyang, https://github.com/malfet ghstack dependencies: #140564, #140565, #140566, #140567, #140720, #140994
pobin6
pushed a commit
to pobin6/pytorch
that referenced
this pull request
Dec 5, 2024
We want to use Half/BFloat16 in ExecuTorch to support shared kernel code. They will need to be used in ExecuTorch core, so they can't have streams. This diff introduces a macro to gate the stream code off. Differential Revision: [D65888035](https://our.internmc.facebook.com/intern/diff/D65888035/) Pull Request resolved: pytorch#140566 Approved by: https://github.com/ezyang, https://github.com/malfet ghstack dependencies: pytorch#140564, pytorch#140565
pobin6
pushed a commit
to pobin6/pytorch
that referenced
this pull request
Dec 5, 2024
…140567) Accessing the inactive member of a union is undefined behavior. Fortunately, we have c10::bit_cast. Differential Revision: [D65888680](https://our.internmc.facebook.com/intern/diff/D65888680/) Pull Request resolved: pytorch#140567 Approved by: https://github.com/Skylion007, https://github.com/malfet ghstack dependencies: pytorch#140564, pytorch#140565, pytorch#140566
pobin6
pushed a commit
to pobin6/pytorch
that referenced
this pull request
Dec 5, 2024
This was added in pytorch/executorch#1789 . I'm working on sharing Half.h with ExecuTorch, and this is a missing feature. Differential Revision: [D65949409](https://our.internmc.facebook.com/intern/diff/D65949409/) Pull Request resolved: pytorch#140720 Approved by: https://github.com/malfet ghstack dependencies: pytorch#140564, pytorch#140565, pytorch#140566, pytorch#140567
pobin6
pushed a commit
to pobin6/pytorch
that referenced
this pull request
Dec 5, 2024
It passed PyTorch CI, but internally we saw failures from this. Differential Revision: [D66137897](https://our.internmc.facebook.com/intern/diff/D66137897/) Pull Request resolved: pytorch#140994 Approved by: https://github.com/malfet ghstack dependencies: pytorch#140564, pytorch#140565, pytorch#140566, pytorch#140567, pytorch#140720
pobin6
pushed a commit
to pobin6/pytorch
that referenced
this pull request
Dec 5, 2024
Make what we're doing as obvious as possible to the compiler. Differential Revision: [D66108811](https://our.internmc.facebook.com/intern/diff/D66108811/) Pull Request resolved: pytorch#141035 Approved by: https://github.com/Skylion007, https://github.com/ezyang, https://github.com/malfet ghstack dependencies: pytorch#140564, pytorch#140565, pytorch#140566, pytorch#140567, pytorch#140720, pytorch#140994
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):
We want to use Half/BFloat16 in ExecuTorch to support shared kernel code. They will need to be used in ExecuTorch core, so they can't have streams. This diff introduces a macro to gate the stream code off.
Differential Revision: D65888035