[PATCH] D134281: [CGP] Update MemIntrinsic alignment if possible

[Phoebe Wang via Phabricator via llvm-commits]

pengfei added inline comments.


================
Comment at: llvm/test/CodeGen/X86/memset-2.ll:8-18
+; CHECK-NEXT:    movaps %xmm0, 160
+; CHECK-NEXT:    movaps %xmm0, 144
+; CHECK-NEXT:    movaps %xmm0, 128
+; CHECK-NEXT:    movaps %xmm0, 112
+; CHECK-NEXT:    movaps %xmm0, 96
+; CHECK-NEXT:    movaps %xmm0, 80
+; CHECK-NEXT:    movaps %xmm0, 64
----------------
arichardson wrote:
> pengfei wrote:
> > Not sure if the use of `movaps` is a good idea considering the test in below file.
> I guess the real fix would be to not use null as the destination. I'm not sure what the test is actually checking, but I could change it to be an argument?
How about change `0` to `1`?
With a read on the history, I think the intention was to make sure no `movups` generated. See rG7998b1d6f and rGa048c83fe47


================
Comment at: llvm/test/CodeGen/X86/memset64-on-x86-32.ll:36
 ; SLOW_64-NEXT:    retq
   call void @llvm.memset.p0.i64(ptr align 4 null, i8 0, i64 80, i1 false)
   ret void
----------------
Ditto here. Change `0` to `4`?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D134281/new/

https://reviews.llvm.org/D134281