Cleans up call inliner in the XLA shared code path #23964
Closed
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
yashk2810
approved these changes
Mar 20, 2025
ezhulenev
approved these changes
Mar 20, 2025
copybara-service bot
pushed a commit
that referenced
this pull request
Mar 20, 2025
Imported from GitHub PR #23964 1. Remove all the GPU-specific logic inside Call inliner. 2. Rewrite the IsInlineableCallOp to make the code more readable and less error-prone. The previous impl has some implicit priorities between all the checks and leads to bug. 3. Removes the test for stream annotation. Discussed with @yashk2810 , actually now we have two controls over the `compute_on` boxes from JAX. The flag in XLA `xla_gpu_experimental_stream_annotation` seems to be confusing and we should remove that. Because the control is explicitly placed in JAX, but users will get confused if there is another flag to control this. Copybara import of the project: -- 1ee6a74 by Yunlong Liu <[email protected]>: cleanup Merging this change closes #23964 FUTURE_COPYBARA_INTEGRATE_REVIEW=#23964 from yliu120:cleanup_call_inliner 1ee6a74 PiperOrigin-RevId: 738911497
copybara-service bot
pushed a commit
that referenced
this pull request
Mar 20, 2025
Imported from GitHub PR #23964 1. Remove all the GPU-specific logic inside Call inliner. 2. Rewrite the IsInlineableCallOp to make the code more readable and less error-prone. The previous impl has some implicit priorities between all the checks and leads to bug. 3. Removes the test for stream annotation. Discussed with @yashk2810 , actually now we have two controls over the `compute_on` boxes from JAX. The flag in XLA `xla_gpu_experimental_stream_annotation` seems to be confusing and we should remove that. Because the control is explicitly placed in JAX, but users will get confused if there is another flag to control this. Copybara import of the project: -- 1ee6a74 by Yunlong Liu <[email protected]>: cleanup Merging this change closes #23964 FUTURE_COPYBARA_INTEGRATE_REVIEW=#23964 from yliu120:cleanup_call_inliner 1ee6a74 PiperOrigin-RevId: 738911497
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Mar 20, 2025
Imported from GitHub PR openxla/xla#23964 1. Remove all the GPU-specific logic inside Call inliner. 2. Rewrite the IsInlineableCallOp to make the code more readable and less error-prone. The previous impl has some implicit priorities between all the checks and leads to bug. 3. Removes the test for stream annotation. Discussed with @yashk2810 , actually now we have two controls over the `compute_on` boxes from JAX. The flag in XLA `xla_gpu_experimental_stream_annotation` seems to be confusing and we should remove that. Because the control is explicitly placed in JAX, but users will get confused if there is another flag to control this. Copybara import of the project: -- 1ee6a74ec622d2a152e52f3f3e793e6a2f7af256 by Yunlong Liu <[email protected]>: cleanup Merging this change closes #23964 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#23964 from yliu120:cleanup_call_inliner 1ee6a74ec622d2a152e52f3f3e793e6a2f7af256 PiperOrigin-RevId: 738911497
ezhulenev
approved these changes
Mar 20, 2025
copybara-service bot
pushed a commit
that referenced
this pull request
Mar 20, 2025
Imported from GitHub PR #23964 1. Remove all the GPU-specific logic inside Call inliner. 2. Rewrite the IsInlineableCallOp to make the code more readable and less error-prone. The previous impl has some implicit priorities between all the checks and leads to bug. 3. Removes the test for stream annotation. Discussed with @yashk2810 , actually now we have two controls over the `compute_on` boxes from JAX. The flag in XLA `xla_gpu_experimental_stream_annotation` seems to be confusing and we should remove that. Because the control is explicitly placed in JAX, but users will get confused if there is another flag to control this. Copybara import of the project: -- 1ee6a74 by Yunlong Liu <[email protected]>: cleanup Merging this change closes #23964 FUTURE_COPYBARA_INTEGRATE_REVIEW=#23964 from yliu120:cleanup_call_inliner 1ee6a74 PiperOrigin-RevId: 738911497
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Mar 20, 2025
Imported from GitHub PR openxla/xla#23964 1. Remove all the GPU-specific logic inside Call inliner. 2. Rewrite the IsInlineableCallOp to make the code more readable and less error-prone. The previous impl has some implicit priorities between all the checks and leads to bug. 3. Removes the test for stream annotation. Discussed with @yashk2810 , actually now we have two controls over the `compute_on` boxes from JAX. The flag in XLA `xla_gpu_experimental_stream_annotation` seems to be confusing and we should remove that. Because the control is explicitly placed in JAX, but users will get confused if there is another flag to control this. Copybara import of the project: -- 1ee6a74ec622d2a152e52f3f3e793e6a2f7af256 by Yunlong Liu <[email protected]>: cleanup Merging this change closes #23964 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#23964 from yliu120:cleanup_call_inliner 1ee6a74ec622d2a152e52f3f3e793e6a2f7af256 PiperOrigin-RevId: 738911497
copybara-service bot
pushed a commit
that referenced
this pull request
Mar 20, 2025
Imported from GitHub PR #23964 1. Remove all the GPU-specific logic inside Call inliner. 2. Rewrite the IsInlineableCallOp to make the code more readable and less error-prone. The previous impl has some implicit priorities between all the checks and leads to bug. 3. Removes the test for stream annotation. Discussed with @yashk2810 , actually now we have two controls over the `compute_on` boxes from JAX. The flag in XLA `xla_gpu_experimental_stream_annotation` seems to be confusing and we should remove that. Because the control is explicitly placed in JAX, but users will get confused if there is another flag to control this. Copybara import of the project: -- 1ee6a74 by Yunlong Liu <[email protected]>: cleanup Merging this change closes #23964 FUTURE_COPYBARA_INTEGRATE_REVIEW=#23964 from yliu120:cleanup_call_inliner 1ee6a74 PiperOrigin-RevId: 738911497
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Mar 20, 2025
Imported from GitHub PR openxla/xla#23964 1. Remove all the GPU-specific logic inside Call inliner. 2. Rewrite the IsInlineableCallOp to make the code more readable and less error-prone. The previous impl has some implicit priorities between all the checks and leads to bug. 3. Removes the test for stream annotation. Discussed with @yashk2810 , actually now we have two controls over the `compute_on` boxes from JAX. The flag in XLA `xla_gpu_experimental_stream_annotation` seems to be confusing and we should remove that. Because the control is explicitly placed in JAX, but users will get confused if there is another flag to control this. Copybara import of the project: -- 1ee6a74ec622d2a152e52f3f3e793e6a2f7af256 by Yunlong Liu <[email protected]>: cleanup Merging this change closes #23964 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#23964 from yliu120:cleanup_call_inliner 1ee6a74ec622d2a152e52f3f3e793e6a2f7af256 PiperOrigin-RevId: 738911497
copybara-service bot
pushed a commit
that referenced
this pull request
Mar 20, 2025
Imported from GitHub PR #23964 1. Remove all the GPU-specific logic inside Call inliner. 2. Rewrite the IsInlineableCallOp to make the code more readable and less error-prone. The previous impl has some implicit priorities between all the checks and leads to bug. 3. Removes the test for stream annotation. Discussed with @yashk2810 , actually now we have two controls over the `compute_on` boxes from JAX. The flag in XLA `xla_gpu_experimental_stream_annotation` seems to be confusing and we should remove that. Because the control is explicitly placed in JAX, but users will get confused if there is another flag to control this. Copybara import of the project: -- 1ee6a74 by Yunlong Liu <[email protected]>: cleanup Merging this change closes #23964 FUTURE_COPYBARA_INTEGRATE_REVIEW=#23964 from yliu120:cleanup_call_inliner 1ee6a74 PiperOrigin-RevId: 738911497
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Mar 20, 2025
Imported from GitHub PR openxla/xla#23964 1. Remove all the GPU-specific logic inside Call inliner. 2. Rewrite the IsInlineableCallOp to make the code more readable and less error-prone. The previous impl has some implicit priorities between all the checks and leads to bug. 3. Removes the test for stream annotation. Discussed with @yashk2810 , actually now we have two controls over the `compute_on` boxes from JAX. The flag in XLA `xla_gpu_experimental_stream_annotation` seems to be confusing and we should remove that. Because the control is explicitly placed in JAX, but users will get confused if there is another flag to control this. Copybara import of the project: -- 1ee6a74ec622d2a152e52f3f3e793e6a2f7af256 by Yunlong Liu <[email protected]>: cleanup Merging this change closes #23964 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#23964 from yliu120:cleanup_call_inliner 1ee6a74ec622d2a152e52f3f3e793e6a2f7af256 PiperOrigin-RevId: 738911497
copybara-service bot
pushed a commit
that referenced
this pull request
Mar 21, 2025
Imported from GitHub PR #23964 1. Remove all the GPU-specific logic inside Call inliner. 2. Rewrite the IsInlineableCallOp to make the code more readable and less error-prone. The previous impl has some implicit priorities between all the checks and leads to bug. 3. Removes the test for stream annotation. Discussed with @yashk2810 , actually now we have two controls over the `compute_on` boxes from JAX. The flag in XLA `xla_gpu_experimental_stream_annotation` seems to be confusing and we should remove that. Because the control is explicitly placed in JAX, but users will get confused if there is another flag to control this. Copybara import of the project: -- 1ee6a74 by Yunlong Liu <[email protected]>: cleanup Merging this change closes #23964 FUTURE_COPYBARA_INTEGRATE_REVIEW=#23964 from yliu120:cleanup_call_inliner 1ee6a74 PiperOrigin-RevId: 738911497
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Mar 21, 2025
Imported from GitHub PR openxla/xla#23964 1. Remove all the GPU-specific logic inside Call inliner. 2. Rewrite the IsInlineableCallOp to make the code more readable and less error-prone. The previous impl has some implicit priorities between all the checks and leads to bug. 3. Removes the test for stream annotation. Discussed with @yashk2810 , actually now we have two controls over the `compute_on` boxes from JAX. The flag in XLA `xla_gpu_experimental_stream_annotation` seems to be confusing and we should remove that. Because the control is explicitly placed in JAX, but users will get confused if there is another flag to control this. Copybara import of the project: -- 1ee6a74ec622d2a152e52f3f3e793e6a2f7af256 by Yunlong Liu <[email protected]>: cleanup Merging this change closes #23964 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#23964 from yliu120:cleanup_call_inliner 1ee6a74ec622d2a152e52f3f3e793e6a2f7af256 PiperOrigin-RevId: 738911497
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Mar 21, 2025
Imported from GitHub PR openxla/xla#23964 1. Remove all the GPU-specific logic inside Call inliner. 2. Rewrite the IsInlineableCallOp to make the code more readable and less error-prone. The previous impl has some implicit priorities between all the checks and leads to bug. 3. Removes the test for stream annotation. Discussed with @yashk2810 , actually now we have two controls over the `compute_on` boxes from JAX. The flag in XLA `xla_gpu_experimental_stream_annotation` seems to be confusing and we should remove that. Because the control is explicitly placed in JAX, but users will get confused if there is another flag to control this. Copybara import of the project: -- 1ee6a74ec622d2a152e52f3f3e793e6a2f7af256 by Yunlong Liu <[email protected]>: cleanup Merging this change closes #23964 PiperOrigin-RevId: 739113714
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.
Discussed with @yashk2810 , actually now we have two controls over the
compute_onboxes from JAX. The flag in XLAxla_gpu_experimental_stream_annotationseems to be confusing and we should remove that.Because the control is explicitly placed in JAX, but users will get confused if there is another flag to control this.