sse: implement _mm_movelh_ps for Arm64

Inspired by https://github.com/DLTcollab/sse2neon/pull/579

sse: implement _mm_movelh_ps for Arm64
Inspired by https://github.com/DLTcollab/sse2neon/pull/579
98eb64b8 · Michael R. Crusoe · Michael R. Crusoe · 57197e8d · 98eb64b8
Commit 98eb64b8 authored Oct 12, 2023 by Michael R. Crusoe Committed by Michael R. Crusoe Oct 14, 2023
Hide whitespace changes
Inline Side-by-side

Showing with 3 additions and 1 deletion

x86/sse.h x86/sse.h +3 -1

No files found.
--- a/x86/sse.h
+++ b/x86/sse.h
@@ -3217,7 +3217,9 @@ simde_mm_movehl_ps (simde__m128 a, simde__m128 b) {
      a_ = simde__m128_to_private(a),
      b_ = simde__m128_to_private(b);

-    #if defined(SIMDE_ARM_NEON_A32V7_NATIVE)
+    #if defined(SIMDE_ARM_NEON_A64V8_NATIVE)
+      r_.neon_u64 = vzip2q_u64(b_.neon_u64, a_.neon_u64);
+    #elif defined(SIMDE_ARM_NEON_A32V7_NATIVE)
      float32x2_t a32 = vget_high_f32(a_.neon_f32);
      float32x2_t b32 = vget_high_f32(b_.neon_f32);
      r_.neon_f32 = vcombine_f32(b32, a32);