[PATCH] D99951: [NFC] Add scalable vectorisation tests for int/FP <> int/FP conversions

Kerry McLaughlin via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Fri Apr 23 03:40:00 PDT 2021


kmclaughlin added inline comments.


================
Comment at: llvm/test/Transforms/LoopVectorize/AArch64/sve-type-conv.ll:156-157
+  %0 = load i64, i64* %arrayidx, align 8
+  %conv = uitofp i64 %0 to float
+  %conv1 = fptrunc float %conv to half
+  %arrayidx2 = getelementptr inbounds half, half* %dst, i64 %i.08
----------------
I'm not sure why this test and the one below are using both a `[u|s]itofp` and an `fptrunc`, is it possible to just use `uitofp i64 %0 to half` here?


================
Comment at: llvm/test/Transforms/LoopVectorize/AArch64/sve-type-conv.ll:217
+
+define void @u8_to_s16(i16* noalias nocapture %dst, i8* noalias nocapture readonly %src, i64 %N) #0 {
+; CHECK-LABEL: @u8_to_s16(
----------------
nit: Should this test be called `@u8_to_u16`?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D99951/new/

https://reviews.llvm.org/D99951



More information about the llvm-commits mailing list