[PATCH] D25551: AMDGPU: Implement SGPR spilling with scalar stores

[Matt Arsenault via llvm-commits]

arsenm added a comment.

In https://reviews.llvm.org/D25551#591815, @nhaehnle wrote:

> I guess the comment on https://reviews.llvm.org/D26005 about leaving 63 elements wasted belongs here? It would be nice to have a TODO about this here, since it seems to be just an arbitrary limitation of how the spill slots are currently allocated.


https://reviews.llvm.org/D26104 is the patch to use the wide stores which fixes this


https://reviews.llvm.org/D25551