Add Arm SVE 8-wide (256b) implementation #480
Merged
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
|
This should check SVE width at runtime and error. As with other ISA checks that is done in the CLI front-end, and not in the library. |
|
SVE runtime checks now added. |
solidpixel
force-pushed
the
sve_support
branch
4 times, most recently
from
f1f6c04
to
b9216ca
Compare
bengaineyarm
reviewed
Aug 1, 2024
|
Confirmed passes all tests on V1 (as we can't test SVE using GH actions). |
bengaineyarm
approved these changes
Aug 5, 2024
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Outdated suggestions cannot be applied.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.
This PR adds support for Arm SVE via a fixed-width 256-bit implementation, as well as extending the fixed-width 128-bit implementation (which is mostly NEON) with a few targeted SVE operations such as native gathers.
Due to the style of 128-bit accumulator we use for floating-point invariance, it's not really possible to write a true vector-length-agnostic SVE implementation, so this implementation is a compile-time choice that will only work on 256b SVE implementations.
On a Neoverse V1 this code is ~30% faster than the equivalent NEON build.