[PATCH] D32916: [DAGCombine] (addcarry 0, 0, X) -> (ext/trunc X)

Amaury SECHET via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Wed May 17 23:07:28 PDT 2017


deadalnix added inline comments.


================
Comment at: test/CodeGen/X86/mul-i1024.ll:5
 
 define void @test_1024(i1024* %a, i1024* %b, i1024* %out) nounwind {
 ; X32-LABEL: test_1024:
----------------
spatel wrote:
> spatel wrote:
> > What does testing with i1024 tell us that testing with i512 (or i256 for that matter) would not?
> > This is a massive file and a massive diff, and I have no idea how to make sense of it. If others can follow what's happening in this codegen, I suppose we should leave it. Otherwise, let's just get rid of this file?
> I didn't see an answer to this question. Do we need to test i512 / i1024? How are they different than i256?
Really, I have no idea. These tests weren't added for this diff specifically, so that's a bit outside of the scope of this change.


https://reviews.llvm.org/D32916





More information about the llvm-commits mailing list