Remove fp16 target feature from some intrinsics #1978
Merged
+218
−142
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 seems to affect codegen on a lot of Arm intrinsics so I've avoided making the change there.
Collaborator
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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Outdated suggestions cannot be applied.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.
The instructions for some f16 intrinsics are available in the base neon feature (DUP/INS variants, store/loads and the fp conversion instructions FCVTL/FCVTN). This relaxes the requirement for the fp16 target feature on aarch64 for these intrinsics, which doesn't affect codegen/ABI.
This change makes it possible to create the Neon types on non-fp16 hardware, but since there's very little that can be done with these types outside of the intrinsics this shouldn't require the compiler to handle fallback implementations operations like it does with the scalar f16 type.