[llvm] [AArch64] Avoid single-element vector fp converts in streaming[-compatible] functions (PR #112213)

[Benjamin Maxwell via llvm-commits]

================
@@ -0,0 +1,121 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -force-streaming-compatible  < %s | FileCheck %s
+; RUN: llc < %s | FileCheck %s --check-prefix=NON-STREAMING
+
+target triple = "aarch64-unknown-linux-gnu"
+
+define double @t1(double %x) {
+; CHECK-LABEL: t1:
+; CHECK:       // %bb.0: // %entry
+; CHECK-NEXT:    fcvtzs x8, d0
+; CHECK-NEXT:    scvtf d0, x8
----------------
MacDue wrote:

I'll look into that in a follow up patch :+1: I need to check how to do something ~equivalent for SVE since for Neon this are a load of .td patterns (matching both the to/from int parts). 

https://github.com/llvm/llvm-project/pull/112213