[PATCH] D132344: [LV][SLP] Mark is_fpclass as vectorizable

[Matt Arsenault via Phabricator via llvm-commits]

arsenm requested changes to this revision.
arsenm added inline comments.
This revision now requires changes to proceed.
Herald added a subscriber: wdng.


================
Comment at: llvm/test/Transforms/SLPVectorizer/X86/intrinsic.ll:558
+  %arrayidx6 = getelementptr inbounds float, float* %a, i32 3
+  %i6 = load float, float* %arrayidx6, align 4
+  %arrayidx7 = getelementptr inbounds float, float* %b, i32 3
----------------
Should use opaque pointers


================
Comment at: llvm/test/Transforms/SLPVectorizer/X86/intrinsic.ll:573
+  ret void
+}
----------------
Should add tests where elements have different test masks


================
Comment at: llvm/test/Transforms/Scalarizer/intrinsics.ll:165-169
+; CHECK-LABEL: @scalarize_is_fpclass(
+; CHECK: %cond.i0 = call i1 @llvm.is.fpclass.f32(float %x.i0, i32 1)
+; CHECK: %cond.i1 = call i1 @llvm.is.fpclass.f32(float %x.i1, i32 1)
+; CHECK: %cond.upto0 = insertelement <2 x i1> poison, i1 %cond.i0, i32 0
+; CHECK: %cond = insertelement <2 x i1> %cond.upto0, i1 %cond.i1, i32 1
----------------
Scalarizer is a separate patch


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D132344/new/

https://reviews.llvm.org/D132344