[llvm] [X86] Require legal result types for f32 estimates (PR #217810)

[Simon Pilgrim via llvm-commits]

================
@@ -0,0 +1,69 @@
+; RUN: llc -mtriple=x86_64-unknown-linux-gnu -mcpu=skx -mattr=+soft-float \
+; RUN:   -verify-machineinstrs -o /dev/null < %s
+; RUN: llc -mtriple=x86_64-unknown-linux-gnu -mcpu=x86-64 \
+; RUN:   -mattr=+avx,+soft-float -verify-machineinstrs -o /dev/null < %s
+
+declare float @llvm.sqrt.f32(float)
+declare <4 x float> @llvm.sqrt.v4f32(<4 x float>)
+declare <8 x float> @llvm.sqrt.v8f32(<8 x float>)
+declare <16 x float> @llvm.sqrt.v16f32(<16 x float>)
----------------
RKSimon wrote:

Should be unnecessary

https://github.com/llvm/llvm-project/pull/217810