[PATCH] D84034: [SVE] Add lowering for fixed length vector fdiv, fma, fmul and fsub operations.

[Paul Walker via Phabricator via llvm-commits]

paulwalker-arm added a comment.

In D84034#2159377 <https://reviews.llvm.org/D84034#2159377>, @cameron.mcinally wrote:

> Do we need to worry about fixed types less than 64b? E.g. <1 x float>.


At some point I think yes.  Currently SVE fixed length code generation tries to remain compatible with the way NEON does things for the types it already supported.  That means <1 x float> will likely be widen to <2 x float> before selection.  It's an area that really needs an ABI defining before we can break away from NEON.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D84034/new/

https://reviews.llvm.org/D84034