test/CodeGen/AArch64/double_reduct.ll

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc --mtriple=aarch64-eabi < %s | FileCheck %s

define float @add_f32(<8 x float> %a, <4 x float> %b) {
; CHECK-LABEL: add_f32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    fadd v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    fadd v0.4s, v0.4s, v2.4s
; CHECK-NEXT:    faddp v0.4s, v0.4s, v0.4s
; CHECK-NEXT:    faddp s0, v0.2s
; CHECK-NEXT:    ret
  %r1 = call fast float @llvm.vector.reduce.fadd.f32.v8f32(float -0.0, <8 x float> %a)
  %r2 = call fast float @llvm.vector.reduce.fadd.f32.v4f32(float -0.0, <4 x float> %b)
  %r = fadd fast float %r1, %r2
  ret float %r
}

define float @fmul_f32(<8 x float> %a, <4 x float> %b) {
; CHECK-LABEL: fmul_f32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    fmul v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    fmul v0.4s, v0.4s, v2.4s
; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
; CHECK-NEXT:    fmul v0.2s, v0.2s, v1.2s
; CHECK-NEXT:    fmul s0, s0, v0.s[1]
; CHECK-NEXT:    ret
  %r1 = call fast float @llvm.vector.reduce.fmul.f32.v8f32(float 1.0, <8 x float> %a)
  %r2 = call fast float @llvm.vector.reduce.fmul.f32.v4f32(float 1.0, <4 x float> %b)
  %r = fmul fast float %r1, %r2
  ret float %r
}

define float @fmin_f32(<8 x float> %a, <4 x float> %b) {
; CHECK-LABEL: fmin_f32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    fminnm v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    fminnm v0.4s, v0.4s, v2.4s
; CHECK-NEXT:    fminnmv s0, v0.4s
; CHECK-NEXT:    ret
  %r1 = call float @llvm.vector.reduce.fmin.v8f32(<8 x float> %a)
  %r2 = call float @llvm.vector.reduce.fmin.v4f32(<4 x float> %b)
  %r = call float @llvm.minnum.f32(float %r1, float %r2)
  ret float %r
}

define float @fmax_f32(<8 x float> %a, <4 x float> %b) {
; CHECK-LABEL: fmax_f32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    fmaxnm v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    fmaxnm v0.4s, v0.4s, v2.4s
; CHECK-NEXT:    fmaxnmv s0, v0.4s
; CHECK-NEXT:    ret
  %r1 = call float @llvm.vector.reduce.fmax.v8f32(<8 x float> %a)
  %r2 = call float @llvm.vector.reduce.fmax.v4f32(<4 x float> %b)
  %r = call float @llvm.maxnum.f32(float %r1, float %r2)
  ret float %r
}

define float @fminimum_f32(<8 x float> %a, <4 x float> %b) {
; CHECK-LABEL: fminimum_f32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    fmin v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    fmin v0.4s, v0.4s, v2.4s
; CHECK-NEXT:    fminv s0, v0.4s
; CHECK-NEXT:    ret
  %r1 = call float @llvm.vector.reduce.fminimum.v8f32(<8 x float> %a)
  %r2 = call float @llvm.vector.reduce.fminimum.v4f32(<4 x float> %b)
  %r = call float @llvm.minimum.f32(float %r1, float %r2)
  ret float %r
}

define float @fmaximum_f32(<8 x float> %a, <4 x float> %b) {
; CHECK-LABEL: fmaximum_f32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    fmax v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    fmax v0.4s, v0.4s, v2.4s
; CHECK-NEXT:    fmaxv s0, v0.4s
; CHECK-NEXT:    ret
  %r1 = call float @llvm.vector.reduce.fmaximum.v8f32(<8 x float> %a)
  %r2 = call float @llvm.vector.reduce.fmaximum.v4f32(<4 x float> %b)
  %r = call float @llvm.maximum.f32(float %r1, float %r2)
  ret float %r
}

; These next two tests have incorrect minnum/minimum combinations
define float @fminimumnum_f32(<8 x float> %a, <4 x float> %b) {
; CHECK-LABEL: fminimumnum_f32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    fmin v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    fminv s1, v2.4s
; CHECK-NEXT:    fminv s0, v0.4s
; CHECK-NEXT:    fminnm s0, s0, s1
; CHECK-NEXT:    ret
  %r1 = call float @llvm.vector.reduce.fminimum.v8f32(<8 x float> %a)
  %r2 = call float @llvm.vector.reduce.fminimum.v4f32(<4 x float> %b)
  %r = call float @llvm.minnum.f32(float %r1, float %r2)
  ret float %r
}

define float @fmaxnumimum_f32(<8 x float> %a, <4 x float> %b) {
; CHECK-LABEL: fmaxnumimum_f32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    fmaxnm v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    fmaxnmv s1, v2.4s
; CHECK-NEXT:    fmaxnmv s0, v0.4s
; CHECK-NEXT:    fmax s0, s0, s1
; CHECK-NEXT:    ret
  %r1 = call float @llvm.vector.reduce.fmax.v8f32(<8 x float> %a)
  %r2 = call float @llvm.vector.reduce.fmax.v4f32(<4 x float> %b)
  %r = call float @llvm.maximum.f32(float %r1, float %r2)
  ret float %r
}


define i32 @add_i32(<8 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: add_i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    add v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    add v0.4s, v0.4s, v2.4s
; CHECK-NEXT:    addv s0, v0.4s
; CHECK-NEXT:    fmov w0, s0
; CHECK-NEXT:    ret
  %r1 = call i32 @llvm.vector.reduce.add.i32.v8i32(<8 x i32> %a)
  %r2 = call i32 @llvm.vector.reduce.add.i32.v4i32(<4 x i32> %b)
  %r = add i32 %r1, %r2
  ret i32 %r
}

define i16 @add_ext_i16(<16 x i8> %a, <16 x i8> %b) {
; CHECK-LABEL: add_ext_i16:
; CHECK:       // %bb.0:
; CHECK-NEXT:    uaddlp v1.8h, v1.16b
; CHECK-NEXT:    uadalp v1.8h, v0.16b
; CHECK-NEXT:    addv h0, v1.8h
; CHECK-NEXT:    fmov w0, s0
; CHECK-NEXT:    ret
  %ae = zext <16 x i8> %a to <16 x i16>
  %be = zext <16 x i8> %b to <16 x i16>
  %r1 = call i16 @llvm.vector.reduce.add.i16.v16i16(<16 x i16> %ae)
  %r2 = call i16 @llvm.vector.reduce.add.i16.v16i16(<16 x i16> %be)
  %r = add i16 %r1, %r2
  ret i16 %r
}

define i16 @add_ext_v32i16(<32 x i8> %a, <16 x i8> %b) {
; CHECK-LABEL: add_ext_v32i16:
; CHECK:       // %bb.0:
; CHECK-NEXT:    uaddl2 v3.8h, v0.16b, v1.16b
; CHECK-NEXT:    uaddl v0.8h, v0.8b, v1.8b
; CHECK-NEXT:    add v0.8h, v0.8h, v3.8h
; CHECK-NEXT:    uadalp v0.8h, v2.16b
; CHECK-NEXT:    addv h0, v0.8h
; CHECK-NEXT:    fmov w0, s0
; CHECK-NEXT:    ret
  %ae = zext <32 x i8> %a to <32 x i16>
  %be = zext <16 x i8> %b to <16 x i16>
  %r1 = call i16 @llvm.vector.reduce.add.i16.v32i16(<32 x i16> %ae)
  %r2 = call i16 @llvm.vector.reduce.add.i16.v16i16(<16 x i16> %be)
  %r = add i16 %r1, %r2
  ret i16 %r
}

define i32 @mul_i32(<8 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: mul_i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    mul v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    mul v0.4s, v0.4s, v2.4s
; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
; CHECK-NEXT:    mul v0.2s, v0.2s, v1.2s
; CHECK-NEXT:    mov w8, v0.s[1]
; CHECK-NEXT:    fmov w9, s0
; CHECK-NEXT:    mul w0, w9, w8
; CHECK-NEXT:    ret
  %r1 = call i32 @llvm.vector.reduce.mul.i32.v8i32(<8 x i32> %a)
  %r2 = call i32 @llvm.vector.reduce.mul.i32.v4i32(<4 x i32> %b)
  %r = mul i32 %r1, %r2
  ret i32 %r
}

define i32 @and_i32(<8 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: and_i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    and v0.16b, v0.16b, v1.16b
; CHECK-NEXT:    and v0.16b, v0.16b, v2.16b
; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
; CHECK-NEXT:    and v0.8b, v0.8b, v1.8b
; CHECK-NEXT:    fmov x8, d0
; CHECK-NEXT:    lsr x9, x8, #32
; CHECK-NEXT:    and w0, w8, w9
; CHECK-NEXT:    ret
  %r1 = call i32 @llvm.vector.reduce.and.i32.v8i32(<8 x i32> %a)
  %r2 = call i32 @llvm.vector.reduce.and.i32.v4i32(<4 x i32> %b)
  %r = and i32 %r1, %r2
  ret i32 %r
}

define i32 @or_i32(<8 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: or_i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    orr v0.16b, v0.16b, v1.16b
; CHECK-NEXT:    orr v0.16b, v0.16b, v2.16b
; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
; CHECK-NEXT:    orr v0.8b, v0.8b, v1.8b
; CHECK-NEXT:    fmov x8, d0
; CHECK-NEXT:    lsr x9, x8, #32
; CHECK-NEXT:    orr w0, w8, w9
; CHECK-NEXT:    ret
  %r1 = call i32 @llvm.vector.reduce.or.i32.v8i32(<8 x i32> %a)
  %r2 = call i32 @llvm.vector.reduce.or.i32.v4i32(<4 x i32> %b)
  %r = or i32 %r1, %r2
  ret i32 %r
}

define i32 @xor_i32(<8 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: xor_i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    eor v0.16b, v0.16b, v1.16b
; CHECK-NEXT:    eor v0.16b, v0.16b, v2.16b
; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
; CHECK-NEXT:    eor v0.8b, v0.8b, v1.8b
; CHECK-NEXT:    fmov x8, d0
; CHECK-NEXT:    lsr x9, x8, #32
; CHECK-NEXT:    eor w0, w8, w9
; CHECK-NEXT:    ret
  %r1 = call i32 @llvm.vector.reduce.xor.i32.v8i32(<8 x i32> %a)
  %r2 = call i32 @llvm.vector.reduce.xor.i32.v4i32(<4 x i32> %b)
  %r = xor i32 %r1, %r2
  ret i32 %r
}

define i32 @umin_i32(<8 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: umin_i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    umin v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    umin v0.4s, v0.4s, v2.4s
; CHECK-NEXT:    uminv s0, v0.4s
; CHECK-NEXT:    fmov w0, s0
; CHECK-NEXT:    ret
  %r1 = call i32 @llvm.vector.reduce.umin.i32.v8i32(<8 x i32> %a)
  %r2 = call i32 @llvm.vector.reduce.umin.i32.v4i32(<4 x i32> %b)
  %r = call i32 @llvm.umin.i32(i32 %r1, i32 %r2)
  ret i32 %r
}

define i32 @umax_i32(<8 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: umax_i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    umax v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    umax v0.4s, v0.4s, v2.4s
; CHECK-NEXT:    umaxv s0, v0.4s
; CHECK-NEXT:    fmov w0, s0
; CHECK-NEXT:    ret
  %r1 = call i32 @llvm.vector.reduce.umax.i32.v8i32(<8 x i32> %a)
  %r2 = call i32 @llvm.vector.reduce.umax.i32.v4i32(<4 x i32> %b)
  %r = call i32 @llvm.umax.i32(i32 %r1, i32 %r2)
  ret i32 %r
}

define i32 @smin_i32(<8 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: smin_i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    smin v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    smin v0.4s, v0.4s, v2.4s
; CHECK-NEXT:    sminv s0, v0.4s
; CHECK-NEXT:    fmov w0, s0
; CHECK-NEXT:    ret
  %r1 = call i32 @llvm.vector.reduce.smin.i32.v8i32(<8 x i32> %a)
  %r2 = call i32 @llvm.vector.reduce.smin.i32.v4i32(<4 x i32> %b)
  %r = call i32 @llvm.smin.i32(i32 %r1, i32 %r2)
  ret i32 %r
}

define i32 @smax_i32(<8 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: smax_i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    smax v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    smax v0.4s, v0.4s, v2.4s
; CHECK-NEXT:    smaxv s0, v0.4s
; CHECK-NEXT:    fmov w0, s0
; CHECK-NEXT:    ret
  %r1 = call i32 @llvm.vector.reduce.smax.i32.v8i32(<8 x i32> %a)
  %r2 = call i32 @llvm.vector.reduce.smax.i32.v4i32(<4 x i32> %b)
  %r = call i32 @llvm.smax.i32(i32 %r1, i32 %r2)
  ret i32 %r
}

declare float @llvm.vector.reduce.fadd.f32.v8f32(float, <8 x float>)
declare float @llvm.vector.reduce.fadd.f32.v4f32(float, <4 x float>)
declare float @llvm.vector.reduce.fmul.f32.v8f32(float, <8 x float>)
declare float @llvm.vector.reduce.fmul.f32.v4f32(float, <4 x float>)
declare float @llvm.vector.reduce.fmin.v8f32(<8 x float>)
declare float @llvm.vector.reduce.fmin.v4f32(<4 x float>)
declare float @llvm.vector.reduce.fmax.v8f32(<8 x float>)
declare float @llvm.vector.reduce.fmax.v4f32(<4 x float>)
declare float @llvm.vector.reduce.fminimum.v8f32(<8 x float>)
declare float @llvm.vector.reduce.fminimum.v4f32(<4 x float>)
declare float @llvm.vector.reduce.fmaximum.v8f32(<8 x float>)
declare float @llvm.vector.reduce.fmaximum.v4f32(<4 x float>)
declare i32 @llvm.vector.reduce.add.i32.v8i32(<8 x i32>)
declare i32 @llvm.vector.reduce.add.i32.v4i32(<4 x i32>)
declare i16 @llvm.vector.reduce.add.i16.v32i16(<32 x i16>)
declare i16 @llvm.vector.reduce.add.i16.v16i16(<16 x i16>)
declare i32 @llvm.vector.reduce.mul.i32.v8i32(<8 x i32>)
declare i32 @llvm.vector.reduce.mul.i32.v4i32(<4 x i32>)
declare i32 @llvm.vector.reduce.and.i32.v8i32(<8 x i32>)
declare i32 @llvm.vector.reduce.and.i32.v4i32(<4 x i32>)
declare i32 @llvm.vector.reduce.or.i32.v8i32(<8 x i32>)
declare i32 @llvm.vector.reduce.or.i32.v4i32(<4 x i32>)
declare i32 @llvm.vector.reduce.xor.i32.v8i32(<8 x i32>)
declare i32 @llvm.vector.reduce.xor.i32.v4i32(<4 x i32>)
declare i32 @llvm.vector.reduce.umin.i32.v8i32(<8 x i32>)
declare i32 @llvm.vector.reduce.umin.i32.v4i32(<4 x i32>)
declare i32 @llvm.vector.reduce.umax.i32.v8i32(<8 x i32>)
declare i32 @llvm.vector.reduce.umax.i32.v4i32(<4 x i32>)
declare i32 @llvm.vector.reduce.smin.i32.v8i32(<8 x i32>)
declare i32 @llvm.vector.reduce.smin.i32.v4i32(<4 x i32>)
declare i32 @llvm.vector.reduce.smax.i32.v8i32(<8 x i32>)
declare i32 @llvm.vector.reduce.smax.i32.v4i32(<4 x i32>)
declare float @llvm.minnum.f32(float, float)
declare float @llvm.maxnum.f32(float, float)
declare float @llvm.minimum.f32(float, float)
declare float @llvm.maximum.f32(float, float)
declare i32 @llvm.umin.i32(i32, i32)
declare i32 @llvm.umax.i32(i32, i32)
declare i32 @llvm.smin.i32(i32, i32)
declare i32 @llvm.smax.i32(i32, i32)