[llvm] [AArch64][SVE] Use SVE for scalar FP converts in streaming[-compatible] functions (1/n) (PR #118505)

[Benjamin Maxwell via llvm-commits]

================
@@ -2951,7 +2965,10 @@ define half @ucvtf_i16_f16(ptr %0) {
 ; CHECK-LABEL: ucvtf_i16_f16:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    ldrh w8, [x0]
-; CHECK-NEXT:    ucvtf h0, w8
+; CHECK-NEXT:    ptrue p0.s
+; CHECK-NEXT:    fmov s0, w8
----------------
MacDue wrote:

So this is just lowered the same as a i32 to f32 conversion. 

https://github.com/llvm/llvm-project/pull/118505