llvm/llvm/test/CodeGen/AArch64/arm64-raddhn-combine.ll

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple aarch64-none-linux-gnu | FileCheck %s

define <16 x i8> @test_combine_v8i16_to_v16i8(<8 x i16> %x, <8 x i16> %y) {
; CHECK-LABEL: test_combine_v8i16_to_v16i8:
; CHECK:       // %bb.0: // %entry
; CHECK-NEXT:    movi v2.2d, #0000000000000000
; CHECK-NEXT:    raddhn v0.8b, v0.8h, v2.8h
; CHECK-NEXT:    raddhn2 v0.16b, v1.8h, v2.8h
; CHECK-NEXT:    ret
entry:
  %res = call <8 x i8> @llvm.aarch64.neon.rshrn.v8i8(<8 x i16> %x, i32 8)
  %res2 = call <8 x i8> @llvm.aarch64.neon.rshrn.v8i8(<8 x i16> %y, i32 8)
  %shuffle = shufflevector <8 x i8> %res, <8 x i8> %res2, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
  ret <16 x i8> %shuffle
}

define <8 x i16> @test_combine_v4i32_to_v8i16(<4 x i32> %x, <4 x i32> %y) {
; CHECK-LABEL: test_combine_v4i32_to_v8i16:
; CHECK:       // %bb.0: // %entry
; CHECK-NEXT:    movi v2.2d, #0000000000000000
; CHECK-NEXT:    raddhn v0.4h, v0.4s, v2.4s
; CHECK-NEXT:    raddhn2 v0.8h, v1.4s, v2.4s
; CHECK-NEXT:    ret
entry:
  %res = call <4 x i16> @llvm.aarch64.neon.rshrn.v4i16(<4 x i32> %x, i32 16)
  %res2 = call <4 x i16> @llvm.aarch64.neon.rshrn.v4i16(<4 x i32> %y, i32 16)
  %shuffle = shufflevector <4 x i16> %res, <4 x i16> %res2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  ret <8 x i16> %shuffle
}

define <4 x i32> @test_combine_v2i64_to_v4i32(<2 x i64> %x, <2 x i64> %y) {
; CHECK-LABEL: test_combine_v2i64_to_v4i32:
; CHECK:       // %bb.0: // %entry
; CHECK-NEXT:    movi v2.2d, #0000000000000000
; CHECK-NEXT:    raddhn v0.2s, v0.2d, v2.2d
; CHECK-NEXT:    raddhn2 v0.4s, v1.2d, v2.2d
; CHECK-NEXT:    ret
entry:
  %res = call <2 x i32> @llvm.aarch64.neon.rshrn.v2i32(<2 x i64> %x, i32 32)
  %res2 = call <2 x i32> @llvm.aarch64.neon.rshrn.v2i32(<2 x i64> %y, i32 32)
  %shuffle = shufflevector <2 x i32> %res, <2 x i32> %res2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  ret <4 x i32> %shuffle
}

declare <8 x i8>  @llvm.aarch64.neon.rshrn.v8i8(<8 x i16>, i32)
declare <4 x i16> @llvm.aarch64.neon.rshrn.v4i16(<4 x i32>, i32)
declare <2 x i32> @llvm.aarch64.neon.rshrn.v2i32(<2 x i64>, i32)