[ET][Portable][Build Size] Introduce notion of compute type. Refactor add/clamp/where #6007
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
… add/clamp/where Introduced notion of compute type. We now perform the computation over `CTYPE_COMP` (the compute type) rather than `CTYPE_COMMON` (the common type) All of the occurrences of `CTYPE_COMMON` in elementwise_util.h need to be replaced with `CTYPE_COMP`, to properly reflect that we are dealing with the computation type, and not the common type. But we don't do that in this diff, to facilitate review. The previous `SupportedTensorDtypes::SAME_AS_COMMON` is transformed into `SupportedTensorDtypes::SAME_AS_COMP` and a newer `SupportedTensorDtypes::SAME_AS_COMMON` is written. `SupportedTensorDtypes::SAME_AS_COMMON` should perform the reverse mapping than get_compute_type(). In this case, this means that when `CTYPE_COMP` is anything but float, `SAME_AS_COMMON` is effectively the same as `SAME_AS_COMP`. But when `CTYPE_COMP` is float, `SAME_AS_COMMON` switches over `Float`, `Half` and `BFloat16` Build size reduction: - add: 21K -> 18K - clamp: 28K -> 23K - where: 16K -> 12K Differential Revision: [D63860791](https://our.internmc.facebook.com/intern/diff/D63860791/) [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/6007
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 326136c with merge base cb12061 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D63860791 |
This was referenced Oct 8, 2024
[ET][Portable][Build Size] Introduce comparison op pattern. Binary ops: eq, ge, gt, le, lt, ne
#6013
Closed
swolchok
approved these changes
Oct 8, 2024
…e. Refactor add/clamp/where" Introduced notion of compute type. We now perform the computation over `CTYPE_COMP` (the compute type) rather than `CTYPE_COMMON` (the common type) All of the occurrences of `CTYPE_COMMON` in elementwise_util.h need to be replaced with `CTYPE_COMP`, to properly reflect that we are dealing with the computation type, and not the common type. But we don't do that in this diff, to facilitate review. The previous `SupportedTensorDtypes::SAME_AS_COMMON` is transformed into `SupportedTensorDtypes::SAME_AS_COMP` and a newer `SupportedTensorDtypes::SAME_AS_COMMON` is written. `SupportedTensorDtypes::SAME_AS_COMMON` should perform the reverse mapping than get_compute_type(). In this case, this means that when `CTYPE_COMP` is anything but float, `SAME_AS_COMMON` is effectively the same as `SAME_AS_COMP`. But when `CTYPE_COMP` is float, `SAME_AS_COMMON` switches over `Float`, `Half` and `BFloat16` Build size reduction: - add: 21K -> 18K - clamp: 28K -> 23K - where: 16K -> 12K Differential Revision: [D63860791](https://our.internmc.facebook.com/intern/diff/D63860791/) [ghstack-poisoned]
|
This pull request was exported from Phabricator. Differential Revision: D63860791 |
…e. Refactor add/clamp/where" Introduced notion of compute type. We now perform the computation over `CTYPE_COMP` (the compute type) rather than `CTYPE_COMMON` (the common type) All of the occurrences of `CTYPE_COMMON` in elementwise_util.h need to be replaced with `CTYPE_COMP`, to properly reflect that we are dealing with the computation type, and not the common type. But we don't do that in this diff, to facilitate review. The previous `SupportedTensorDtypes::SAME_AS_COMMON` is transformed into `SupportedTensorDtypes::SAME_AS_COMP` and a newer `SupportedTensorDtypes::SAME_AS_COMMON` is written. `SupportedTensorDtypes::SAME_AS_COMMON` should perform the reverse mapping than get_compute_type(). In this case, this means that when `CTYPE_COMP` is anything but float, `SAME_AS_COMMON` is effectively the same as `SAME_AS_COMP`. But when `CTYPE_COMP` is float, `SAME_AS_COMMON` switches over `Float`, `Half` and `BFloat16` Build size reduction: - add: 21K -> 18K - clamp: 28K -> 23K - where: 16K -> 12K Differential Revision: [D63860791](https://our.internmc.facebook.com/intern/diff/D63860791/) [ghstack-poisoned]
|
This pull request was exported from Phabricator. Differential Revision: D63860791 |
…e. Refactor add/clamp/where" Introduced notion of compute type. We now perform the computation over `CTYPE_COMP` (the compute type) rather than `CTYPE_COMMON` (the common type) All of the occurrences of `CTYPE_COMMON` in elementwise_util.h need to be replaced with `CTYPE_COMP`, to properly reflect that we are dealing with the computation type, and not the common type. But we don't do that in this diff, to facilitate review. The previous `SupportedTensorDtypes::SAME_AS_COMMON` is transformed into `SupportedTensorDtypes::SAME_AS_COMP` and a newer `SupportedTensorDtypes::SAME_AS_COMMON` is written. `SupportedTensorDtypes::SAME_AS_COMMON` should perform the reverse mapping than get_compute_type(). In this case, this means that when `CTYPE_COMP` is anything but float, `SAME_AS_COMMON` is effectively the same as `SAME_AS_COMP`. But when `CTYPE_COMP` is float, `SAME_AS_COMMON` switches over `Float`, `Half` and `BFloat16` Build size reduction: - add: 21K -> 18K - clamp: 28K -> 23K - where: 16K -> 12K Differential Revision: [D63860791](https://our.internmc.facebook.com/intern/diff/D63860791/) [ghstack-poisoned]
|
This pull request was exported from Phabricator. Differential Revision: D63860791 |
…e. Refactor add/clamp/where" Introduced notion of compute type. We now perform the computation over `CTYPE_COMP` (the compute type) rather than `CTYPE_COMMON` (the common type) All of the occurrences of `CTYPE_COMMON` in elementwise_util.h need to be replaced with `CTYPE_COMP`, to properly reflect that we are dealing with the computation type, and not the common type. But we don't do that in this diff, to facilitate review. The previous `SupportedTensorDtypes::SAME_AS_COMMON` is transformed into `SupportedTensorDtypes::SAME_AS_COMP` and a newer `SupportedTensorDtypes::SAME_AS_COMMON` is written. `SupportedTensorDtypes::SAME_AS_COMMON` should perform the reverse mapping than get_compute_type(). In this case, this means that when `CTYPE_COMP` is anything but float, `SAME_AS_COMMON` is effectively the same as `SAME_AS_COMP`. But when `CTYPE_COMP` is float, `SAME_AS_COMMON` switches over `Float`, `Half` and `BFloat16` Build size reduction: - add: 21K -> 18K - clamp: 28K -> 23K - where: 16K -> 12K Differential Revision: [D63860791](https://our.internmc.facebook.com/intern/diff/D63860791/) [ghstack-poisoned]
|
This pull request was exported from Phabricator. Differential Revision: D63860791 |
|
This pull request has been merged in 607d4a3. |
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):
Introduced notion of compute type. We now perform the computation over
CTYPE_COMP(the compute type) rather thanCTYPE_COMMON(the common type)All of the occurrences of
CTYPE_COMMONin elementwise_util.h need to be replaced withCTYPE_COMP, to properly reflect that we are dealing with the computation type, and not the common type. But we don't do that in this diff, to facilitate review.The previous
SupportedTensorDtypes::SAME_AS_COMMONis transformed intoSupportedTensorDtypes::SAME_AS_COMPand a newerSupportedTensorDtypes::SAME_AS_COMMONis written.SupportedTensorDtypes::SAME_AS_COMMONshould perform the reverse mapping than get_compute_type(). In this case, this means that whenCTYPE_COMPis anything but float,SAME_AS_COMMONis effectively the same asSAME_AS_COMP. But whenCTYPE_COMPis float,SAME_AS_COMMONswitches overFloat,HalfandBFloat16Build size reduction:
Differential Revision: D63860791