[PATCH] D140599: AMDGPU: Promote array alloca if used by memmove/memcpy

Fri Dec 23 15:15:49 PST 2022

arsenm added a comment.

In D140599#4014542 <https://reviews.llvm.org/D140599#4014542>, @ruiling wrote:

> In D140599#4014540 <https://reviews.llvm.org/D140599#4014540>, @arsenm wrote:
>
>> Shouldn't these have been expanded into loads and stores already?
>
> We only split mem transfer intrinsics using a loop if its length is > 1024.

I think InstCombine and/or SROA do some too. Also that threshold is arbitrary, and we could start expanding the small ones in IR too

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D140599/new/

https://reviews.llvm.org/D140599