Skip to content

Conversation

@jreiffers
Copy link
Contributor

Currently, this pass drops the known_induction_variable and known_init_step fields in WhileLoopBackendConfig. This change keeps them if they were present before and updates them with the updated value, just like known_trip_count. This is only done for double buffered loops, not for fully unrolled ones.

@jreiffers
Copy link
Contributor Author

@golechwierowicz Might make sense for you to look at this one :-)

Thanks!

copybara-service bot pushed a commit that referenced this pull request Apr 22, 2025
Imported from GitHub PR #25275

Currently, this pass drops the `known_induction_variable` and `known_init_step` fields in `WhileLoopBackendConfig`. This change keeps them if they were present before and updates them with the updated value, just like `known_trip_count`. This is only done for double buffered loops, not for fully unrolled ones.
Copybara import of the project:

--
6bd4459 by Johannes Reifferscheid <jreiffers@nvidia.com>:

Keep loop metadata after loop unrolling.

--
369ad9b by Johannes Reifferscheid <jreiffers@nvidia.com>:

Fix init update and add tests.

Merging this change closes #25275

FUTURE_COPYBARA_INTEGRATE_REVIEW=#25275 from jreiffers:keep-metadata 369ad9b
PiperOrigin-RevId: 750112934
copybara-service bot pushed a commit to tensorflow/tensorflow that referenced this pull request Apr 22, 2025
Imported from GitHub PR openxla/xla#25275

Currently, this pass drops the `known_induction_variable` and `known_init_step` fields in `WhileLoopBackendConfig`. This change keeps them if they were present before and updates them with the updated value, just like `known_trip_count`. This is only done for double buffered loops, not for fully unrolled ones.
Copybara import of the project:

--
6bd44597c2e1dba443a367edf4a21024f6248cb6 by Johannes Reifferscheid <jreiffers@nvidia.com>:

Keep loop metadata after loop unrolling.

--
369ad9bedd031d1e1fd6abacb7a5002f290c72f2 by Johannes Reifferscheid <jreiffers@nvidia.com>:

Fix init update and add tests.

Merging this change closes #25275

FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#25275 from jreiffers:keep-metadata 369ad9bedd031d1e1fd6abacb7a5002f290c72f2
PiperOrigin-RevId: 750112934
copybara-service bot pushed a commit that referenced this pull request Apr 22, 2025
FUTURE_COPYBARA_INTEGRATE_REVIEW=#25275 from jreiffers:keep-metadata 369ad9b
PiperOrigin-RevId: 749972843
copybara-service bot pushed a commit to tensorflow/tensorflow that referenced this pull request Apr 22, 2025
FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#25275 from jreiffers:keep-metadata 369ad9bedd031d1e1fd6abacb7a5002f290c72f2
PiperOrigin-RevId: 749967019
copybara-service bot pushed a commit to tensorflow/tensorflow that referenced this pull request Apr 22, 2025
FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#25275 from jreiffers:keep-metadata 369ad9bedd031d1e1fd6abacb7a5002f290c72f2
PiperOrigin-RevId: 749972843
copybara-service bot pushed a commit to tensorflow/tensorflow that referenced this pull request Apr 22, 2025
Imported from GitHub PR openxla/xla#25275

Currently, this pass drops the `known_induction_variable` and `known_init_step` fields in `WhileLoopBackendConfig`. This change keeps them if they were present before and updates them with the updated value, just like `known_trip_count`. This is only done for double buffered loops, not for fully unrolled ones.
Copybara import of the project:

--
6bd44597c2e1dba443a367edf4a21024f6248cb6 by Johannes Reifferscheid <jreiffers@nvidia.com>:

Keep loop metadata after loop unrolling.

--
369ad9bedd031d1e1fd6abacb7a5002f290c72f2 by Johannes Reifferscheid <jreiffers@nvidia.com>:

Fix init update and add tests.

Merging this change closes #25275

PiperOrigin-RevId: 750122105
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants