Library for SME vector operations #6863
robert3005
started this conversation in
Feature Requests
Replies: 1 comment 1 reply
-
|
Relevant rust RFC - rust-lang/rfcs#3838 |
Beta Was this translation helpful? Give feedback.
1 reply
-
|
FWIW that RFC is not enough by itself and you need the rfc mentioned as a follow up of this RFC, i.e. right now they have SVE and RVV but aarch64 M devices only have SME |
Beta Was this translation helpful? Give feedback.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
SME operations on M4+ chips seem promising performance wise. In benchmarks for #6135 SME version is %15-%25 faster than neon version. However, SME doesn't have any rust standard library support which means you end up writing naked assembly functions. Since the performance can be worthwhile not only for fastlanes transpose operations we should explore standalone SME operations crate that hides the naked assembly functions
Beta Was this translation helpful? Give feedback.
All reactions