[llvm] [RISCV] Adjust RISCVVectorMaskDAGMutation to look for copy to V0 (PR #129296)
Pengcheng Wang via llvm-commits
llvm-commits at lists.llvm.org
Tue Mar 4 20:17:20 PST 2025
https://github.com/wangpc-pp approved this pull request.
LGTM.
> As an aside, I suspect we're still going to want some kind of scheduling false dependency even on the register class representation. Expecting the scheduler to perfectly handle register pressure on size=1 register classes seems like a big ask. I'd suggest starting with the scheduling dependency in place (over the new representation), and then remove it in a separate change.
I don't fully understand this, can you please explain your thought? I also am interested in this.
https://github.com/llvm/llvm-project/pull/129296
More information about the llvm-commits
mailing list