; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=thumbv8.1m.main-none-none-eabi -mattr=+mve -verify-machineinstrs %s -o - | FileCheck %s --check-prefix=CHECK

define arm_aapcs_vfpcc i32 @add_v4i32_v4i32(<4 x i32> %x, <4 x i32> %y) {
; CHECK-LABEL: add_v4i32_v4i32:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlav.u32 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %m = mul <4 x i32> %x, %y
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %m)
  ret i32 %z
}

define arm_aapcs_vfpcc i64 @add_v4i32_v4i64_zext(<4 x i32> %x, <4 x i32> %y) {
; CHECK-LABEL: add_v4i32_v4i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlalv.u32 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i32> %x to <4 x i64>
  %yy = zext <4 x i32> %y to <4 x i64>
  %m = mul <4 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v4i32_v4i64_sext(<4 x i32> %x, <4 x i32> %y) {
; CHECK-LABEL: add_v4i32_v4i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlalv.s32 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i32> %x to <4 x i64>
  %yy = sext <4 x i32> %y to <4 x i64>
  %m = mul <4 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v2i32_v2i64_zext(<2 x i32> %x, <2 x i32> %y) {
; CHECK-LABEL: add_v2i32_v2i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmullb.u32 q2, q0, q1
; CHECK-NEXT:    vmov r0, r1, d5
; CHECK-NEXT:    vmov r2, r3, d4
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <2 x i32> %x to <2 x i64>
  %yy = zext <2 x i32> %y to <2 x i64>
  %m = mul <2 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v2i32_v2i64_sext(<2 x i32> %x, <2 x i32> %y) {
; CHECK-LABEL: add_v2i32_v2i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmullb.s32 q2, q0, q1
; CHECK-NEXT:    vmov r0, r1, d5
; CHECK-NEXT:    vmov r2, r3, d4
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <2 x i32> %x to <2 x i64>
  %yy = sext <2 x i32> %y to <2 x i64>
  %m = mul <2 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i32 @add_v8i16_v8i32_zext(<8 x i16> %x, <8 x i16> %y) {
; CHECK-LABEL: add_v8i16_v8i32_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlav.u16 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i16> %x to <8 x i32>
  %yy = zext <8 x i16> %y to <8 x i32>
  %m = mul <8 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %m)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v8i16_v8i32_sext(<8 x i16> %x, <8 x i16> %y) {
; CHECK-LABEL: add_v8i16_v8i32_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlav.s16 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i16> %x to <8 x i32>
  %yy = sext <8 x i16> %y to <8 x i32>
  %m = mul <8 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %m)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v4i16_v4i32_zext(<4 x i16> %x, <4 x i16> %y) {
; CHECK-LABEL: add_v4i16_v4i32_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u16 q1, q1
; CHECK-NEXT:    vmovlb.u16 q0, q0
; CHECK-NEXT:    vmlav.u32 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i16> %x to <4 x i32>
  %yy = zext <4 x i16> %y to <4 x i32>
  %m = mul <4 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %m)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v4i16_v4i32_sext(<4 x i16> %x, <4 x i16> %y) {
; CHECK-LABEL: add_v4i16_v4i32_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s16 q1, q1
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vmlav.u32 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i16> %x to <4 x i32>
  %yy = sext <4 x i16> %y to <4 x i32>
  %m = mul <4 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %m)
  ret i32 %z
}

define arm_aapcs_vfpcc zeroext i16 @add_v8i16_v8i16(<8 x i16> %x, <8 x i16> %y) {
; CHECK-LABEL: add_v8i16_v8i16:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlav.u16 r0, q0, q1
; CHECK-NEXT:    uxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %m = mul <8 x i16> %x, %y
  %z = call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> %m)
  ret i16 %z
}

define arm_aapcs_vfpcc i64 @add_v8i16_v8i64_zext(<8 x i16> %x, <8 x i16> %y) {
; CHECK-LABEL: add_v8i16_v8i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlalv.u16 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i16> %x to <8 x i64>
  %yy = zext <8 x i16> %y to <8 x i64>
  %m = mul <8 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v8i16_v8i64_sext(<8 x i16> %x, <8 x i16> %y) {
; CHECK-LABEL: add_v8i16_v8i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlalv.s16 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i16> %x to <8 x i64>
  %yy = sext <8 x i16> %y to <8 x i64>
  %m = mul <8 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v8i8i16_v8i64_zext(<8 x i16> %x, <8 x i8> %y) {
; CHECK-LABEL: add_v8i8i16_v8i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u8 q1, q1
; CHECK-NEXT:    vmlalv.u16 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i16> %x to <8 x i64>
  %yy = zext <8 x i8> %y to <8 x i64>
  %m = mul <8 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v8i8i16_v8i64_sext(<8 x i16> %x, <8 x i8> %y) {
; CHECK-LABEL: add_v8i8i16_v8i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q1, q1
; CHECK-NEXT:    vmlalv.s16 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i16> %x to <8 x i64>
  %yy = sext <8 x i8> %y to <8 x i64>
  %m = mul <8 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v4i16_v4i64_zext(<4 x i16> %x, <4 x i16> %y) {
; CHECK-LABEL: add_v4i16_v4i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u16 q1, q1
; CHECK-NEXT:    vmovlb.u16 q0, q0
; CHECK-NEXT:    vmlalv.u32 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i16> %x to <4 x i64>
  %yy = zext <4 x i16> %y to <4 x i64>
  %m = mul <4 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v4i16_v4i64_sext(<4 x i16> %x, <4 x i16> %y) {
; CHECK-LABEL: add_v4i16_v4i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s16 q1, q1
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vmlalv.s32 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i16> %x to <4 x i64>
  %yy = sext <4 x i16> %y to <4 x i64>
  %m = mul <4 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v8i16_v8i32_v8i64_zext(<8 x i16> %x, <8 x i16> %y) {
; CHECK-LABEL: add_v8i16_v8i32_v8i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlalv.u16 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i16> %x to <8 x i32>
  %yy = zext <8 x i16> %y to <8 x i32>
  %m = mul <8 x i32> %xx, %yy
  %ma = zext <8 x i32> %m to <8 x i64>
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %ma)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v8i16_v8i32_v8i64_sext(<8 x i16> %x, <8 x i16> %y) {
; CHECK-LABEL: add_v8i16_v8i32_v8i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlalv.s16 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i16> %x to <8 x i32>
  %yy = sext <8 x i16> %y to <8 x i32>
  %m = mul <8 x i32> %xx, %yy
  %ma = sext <8 x i32> %m to <8 x i64>
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %ma)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v8i16_v8i32_v8i64_sextzext(<8 x i16> %x, <8 x i16> %y) {
; CHECK-LABEL: add_v8i16_v8i32_v8i64_sextzext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlalv.s16 r0, r1, q0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i16> %x to <8 x i32>
  %m = mul <8 x i32> %xx, %xx
  %ma = zext <8 x i32> %m to <8 x i64>
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %ma)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v2i16_v2i64_zext(<2 x i16> %x, <2 x i16> %y) {
; CHECK-LABEL: add_v2i16_v2i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.i64 q2, #0xffff
; CHECK-NEXT:    vand q1, q1, q2
; CHECK-NEXT:    vand q0, q0, q2
; CHECK-NEXT:    vmov r0, s6
; CHECK-NEXT:    vmov r1, s2
; CHECK-NEXT:    vmov r2, s4
; CHECK-NEXT:    vmov r3, s0
; CHECK-NEXT:    umull r0, r1, r1, r0
; CHECK-NEXT:    umlal r0, r1, r3, r2
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <2 x i16> %x to <2 x i64>
  %yy = zext <2 x i16> %y to <2 x i64>
  %m = mul <2 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v2i16_v2i64_sext(<2 x i16> %x, <2 x i16> %y) {
; CHECK-LABEL: add_v2i16_v2i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov r0, s6
; CHECK-NEXT:    vmov r1, s2
; CHECK-NEXT:    vmov r2, s4
; CHECK-NEXT:    vmov r3, s0
; CHECK-NEXT:    sxth r0, r0
; CHECK-NEXT:    sxth r1, r1
; CHECK-NEXT:    smull r0, r1, r1, r0
; CHECK-NEXT:    sxth r2, r2
; CHECK-NEXT:    sxth r3, r3
; CHECK-NEXT:    smlal r0, r1, r3, r2
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <2 x i16> %x to <2 x i64>
  %yy = sext <2 x i16> %y to <2 x i64>
  %m = mul <2 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i32 @add_v16i8_v16i32_zext(<16 x i8> %x, <16 x i8> %y) {
; CHECK-LABEL: add_v16i8_v16i32_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlav.u8 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <16 x i8> %x to <16 x i32>
  %yy = zext <16 x i8> %y to <16 x i32>
  %m = mul <16 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %m)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v16i8_v16i32_sext(<16 x i8> %x, <16 x i8> %y) {
; CHECK-LABEL: add_v16i8_v16i32_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlav.s8 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <16 x i8> %x to <16 x i32>
  %yy = sext <16 x i8> %y to <16 x i32>
  %m = mul <16 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %m)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v8i8_v8i32_zext(<8 x i8> %x, <8 x i8> %y) {
; CHECK-LABEL: add_v8i8_v8i32_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u8 q1, q1
; CHECK-NEXT:    vmovlb.u8 q0, q0
; CHECK-NEXT:    vmlav.u16 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i8> %x to <8 x i32>
  %yy = zext <8 x i8> %y to <8 x i32>
  %m = mul <8 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %m)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v8i8_v8i32_sext(<8 x i8> %x, <8 x i8> %y) {
; CHECK-LABEL: add_v8i8_v8i32_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q1, q1
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vmlav.s16 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i8> %x to <8 x i32>
  %yy = sext <8 x i8> %y to <8 x i32>
  %m = mul <8 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %m)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v8i8i16_v8i32_zext(<8 x i8> %x, <8 x i16> %y) {
; CHECK-LABEL: add_v8i8i16_v8i32_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u8 q0, q0
; CHECK-NEXT:    vmlav.u16 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i8> %x to <8 x i32>
  %yy = zext <8 x i16> %y to <8 x i32>
  %m = mul <8 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %m)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v8i8i16_v8i32_sext(<8 x i8> %x, <8 x i16> %y) {
; CHECK-LABEL: add_v8i8i16_v8i32_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vmlav.s16 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i8> %x to <8 x i32>
  %yy = sext <8 x i16> %y to <8 x i32>
  %m = mul <8 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %m)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v16i8_v16i16_v16i32_zext(<16 x i8> %x, <16 x i8> %y) {
; CHECK-LABEL: add_v16i8_v16i16_v16i32_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlav.u8 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <16 x i8> %x to <16 x i16>
  %yy = zext <16 x i8> %y to <16 x i16>
  %m = mul <16 x i16> %xx, %yy
  %ma = zext <16 x i16> %m to <16 x i32>
  %z = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %ma)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v16i8_v16i16_v16i32_sext(<16 x i8> %x, <16 x i8> %y) {
; CHECK-LABEL: add_v16i8_v16i16_v16i32_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlav.s8 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <16 x i8> %x to <16 x i16>
  %yy = sext <16 x i8> %y to <16 x i16>
  %m = mul <16 x i16> %xx, %yy
  %ma = sext <16 x i16> %m to <16 x i32>
  %z = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %ma)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v16i8_v16i16_v16i32_sextzext(<16 x i8> %x, <16 x i8> %y) {
; CHECK-LABEL: add_v16i8_v16i16_v16i32_sextzext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlav.s8 r0, q0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <16 x i8> %x to <16 x i16>
  %m = mul <16 x i16> %xx, %xx
  %ma = zext <16 x i16> %m to <16 x i32>
  %z = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %ma)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v4i8_v4i32_zext(<4 x i8> %x, <4 x i8> %y) {
; CHECK-LABEL: add_v4i8_v4i32_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.i32 q2, #0xff
; CHECK-NEXT:    vand q1, q1, q2
; CHECK-NEXT:    vand q0, q0, q2
; CHECK-NEXT:    vmlav.u32 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i8> %x to <4 x i32>
  %yy = zext <4 x i8> %y to <4 x i32>
  %m = mul <4 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %m)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v4i8_v4i32_sext(<4 x i8> %x, <4 x i8> %y) {
; CHECK-LABEL: add_v4i8_v4i32_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q1, q1
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vmovlb.s16 q1, q1
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vmlav.u32 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i8> %x to <4 x i32>
  %yy = sext <4 x i8> %y to <4 x i32>
  %m = mul <4 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %m)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v4i8_v4i32_szext(<4 x i8> %x, <4 x i8> %y) {
; CHECK-LABEL: add_v4i8_v4i32_szext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vmov.i32 q2, #0xff
; CHECK-NEXT:    vand q1, q1, q2
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vmlav.u32 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i8> %x to <4 x i32>
  %yy = zext <4 x i8> %y to <4 x i32>
  %m = mul <4 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %m)
  ret i32 %z
}

define arm_aapcs_vfpcc zeroext i16 @add_v16i8_v16i16_zext(<16 x i8> %x, <16 x i8> %y) {
; CHECK-LABEL: add_v16i8_v16i16_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlav.u8 r0, q0, q1
; CHECK-NEXT:    uxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <16 x i8> %x to <16 x i16>
  %yy = zext <16 x i8> %y to <16 x i16>
  %m = mul <16 x i16> %xx, %yy
  %z = call i16 @llvm.vector.reduce.add.v16i16(<16 x i16> %m)
  ret i16 %z
}

define arm_aapcs_vfpcc signext i16 @add_v16i8_v16i16_sext(<16 x i8> %x, <16 x i8> %y) {
; CHECK-LABEL: add_v16i8_v16i16_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlav.s8 r0, q0, q1
; CHECK-NEXT:    sxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <16 x i8> %x to <16 x i16>
  %yy = sext <16 x i8> %y to <16 x i16>
  %m = mul <16 x i16> %xx, %yy
  %z = call i16 @llvm.vector.reduce.add.v16i16(<16 x i16> %m)
  ret i16 %z
}

define arm_aapcs_vfpcc signext i16 @add_v16i8_v16i16_szext(<16 x i8> %x, <16 x i8> %y) {
; CHECK-LABEL: add_v16i8_v16i16_szext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .pad #32
; CHECK-NEXT:    sub sp, #32
; CHECK-NEXT:    add r0, sp, #16
; CHECK-NEXT:    mov r1, sp
; CHECK-NEXT:    vstrw.32 q1, [r0]
; CHECK-NEXT:    vstrw.32 q0, [r1]
; CHECK-NEXT:    vldrb.u16 q0, [r0, #8]
; CHECK-NEXT:    vldrb.s16 q1, [r1, #8]
; CHECK-NEXT:    vmlav.u16 r2, q1, q0
; CHECK-NEXT:    vldrb.u16 q0, [r0]
; CHECK-NEXT:    vldrb.s16 q1, [r1]
; CHECK-NEXT:    vmlava.u16 r2, q1, q0
; CHECK-NEXT:    sxth r0, r2
; CHECK-NEXT:    add sp, #32
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <16 x i8> %x to <16 x i16>
  %yy = zext <16 x i8> %y to <16 x i16>
  %m = mul <16 x i16> %xx, %yy
  %z = call i16 @llvm.vector.reduce.add.v16i16(<16 x i16> %m)
  ret i16 %z
}

define arm_aapcs_vfpcc zeroext i16 @add_v8i8_v8i16_zext(<8 x i8> %x, <8 x i8> %y) {
; CHECK-LABEL: add_v8i8_v8i16_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u8 q1, q1
; CHECK-NEXT:    vmovlb.u8 q0, q0
; CHECK-NEXT:    vmlav.u16 r0, q0, q1
; CHECK-NEXT:    uxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i8> %x to <8 x i16>
  %yy = zext <8 x i8> %y to <8 x i16>
  %m = mul <8 x i16> %xx, %yy
  %z = call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> %m)
  ret i16 %z
}

define arm_aapcs_vfpcc signext i16 @add_v8i8_v8i16_sext(<8 x i8> %x, <8 x i8> %y) {
; CHECK-LABEL: add_v8i8_v8i16_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q1, q1
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vmlav.u16 r0, q0, q1
; CHECK-NEXT:    sxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i8> %x to <8 x i16>
  %yy = sext <8 x i8> %y to <8 x i16>
  %m = mul <8 x i16> %xx, %yy
  %z = call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> %m)
  ret i16 %z
}

define arm_aapcs_vfpcc zeroext i8 @add_v16i8_v16i8(<16 x i8> %x, <16 x i8> %y) {
; CHECK-LABEL: add_v16i8_v16i8:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlav.u8 r0, q0, q1
; CHECK-NEXT:    uxtb r0, r0
; CHECK-NEXT:    bx lr
entry:
  %m = mul <16 x i8> %x, %y
  %z = call i8 @llvm.vector.reduce.add.v16i8(<16 x i8> %m)
  ret i8 %z
}

define arm_aapcs_vfpcc i64 @add_v16i8_v16i64_zext(<16 x i8> %x, <16 x i8> %y) {
; CHECK-LABEL: add_v16i8_v16i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .pad #32
; CHECK-NEXT:    sub sp, #32
; CHECK-NEXT:    add r2, sp, #16
; CHECK-NEXT:    mov r3, sp
; CHECK-NEXT:    vstrw.32 q1, [r2]
; CHECK-NEXT:    vstrw.32 q0, [r3]
; CHECK-NEXT:    vldrb.u16 q0, [r2]
; CHECK-NEXT:    vldrb.u16 q1, [r3]
; CHECK-NEXT:    vmlalv.u16 r0, r1, q1, q0
; CHECK-NEXT:    vldrb.u16 q0, [r2, #8]
; CHECK-NEXT:    vldrb.u16 q1, [r3, #8]
; CHECK-NEXT:    vmlalva.u16 r0, r1, q1, q0
; CHECK-NEXT:    add sp, #32
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <16 x i8> %x to <16 x i64>
  %yy = zext <16 x i8> %y to <16 x i64>
  %m = mul <16 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v16i64(<16 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v16i8_v16i64_sext(<16 x i8> %x, <16 x i8> %y) {
; CHECK-LABEL: add_v16i8_v16i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .pad #32
; CHECK-NEXT:    sub sp, #32
; CHECK-NEXT:    add r2, sp, #16
; CHECK-NEXT:    mov r3, sp
; CHECK-NEXT:    vstrw.32 q1, [r2]
; CHECK-NEXT:    vstrw.32 q0, [r3]
; CHECK-NEXT:    vldrb.s16 q0, [r2]
; CHECK-NEXT:    vldrb.s16 q1, [r3]
; CHECK-NEXT:    vmlalv.s16 r0, r1, q1, q0
; CHECK-NEXT:    vldrb.s16 q0, [r2, #8]
; CHECK-NEXT:    vldrb.s16 q1, [r3, #8]
; CHECK-NEXT:    vmlalva.s16 r0, r1, q1, q0
; CHECK-NEXT:    add sp, #32
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <16 x i8> %x to <16 x i64>
  %yy = sext <16 x i8> %y to <16 x i64>
  %m = mul <16 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v16i64(<16 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v16i8_v16i64_zext_load(ptr %xp, ptr %yp) {
; CHECK-LABEL: add_v16i8_v16i64_zext_load:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldrb.u16 q0, [r1]
; CHECK-NEXT:    vldrb.u16 q1, [r0]
; CHECK-NEXT:    vmlalv.u16 r2, r3, q1, q0
; CHECK-NEXT:    vldrb.u16 q0, [r1, #8]
; CHECK-NEXT:    vldrb.u16 q1, [r0, #8]
; CHECK-NEXT:    vmlalva.u16 r2, r3, q1, q0
; CHECK-NEXT:    mov r0, r2
; CHECK-NEXT:    mov r1, r3
; CHECK-NEXT:    bx lr
entry:
  %x = load <16 x i8>, ptr %xp
  %y = load <16 x i8>, ptr %yp
  %xx = zext <16 x i8> %x to <16 x i64>
  %yy = zext <16 x i8> %y to <16 x i64>
  %m = mul <16 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v16i64(<16 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v16i8_v16i64_sext_load(ptr %xp, ptr %yp) {
; CHECK-LABEL: add_v16i8_v16i64_sext_load:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldrb.s16 q0, [r1]
; CHECK-NEXT:    vldrb.s16 q1, [r0]
; CHECK-NEXT:    vmlalv.s16 r2, r3, q1, q0
; CHECK-NEXT:    vldrb.s16 q0, [r1, #8]
; CHECK-NEXT:    vldrb.s16 q1, [r0, #8]
; CHECK-NEXT:    vmlalva.s16 r2, r3, q1, q0
; CHECK-NEXT:    mov r0, r2
; CHECK-NEXT:    mov r1, r3
; CHECK-NEXT:    bx lr
entry:
  %x = load <16 x i8>, ptr %xp
  %y = load <16 x i8>, ptr %yp
  %xx = sext <16 x i8> %x to <16 x i64>
  %yy = sext <16 x i8> %y to <16 x i64>
  %m = mul <16 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v16i64(<16 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v8i8_v8i64_zext(<8 x i8> %x, <8 x i8> %y) {
; CHECK-LABEL: add_v8i8_v8i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u8 q1, q1
; CHECK-NEXT:    vmovlb.u8 q0, q0
; CHECK-NEXT:    vmlalv.u16 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i8> %x to <8 x i64>
  %yy = zext <8 x i8> %y to <8 x i64>
  %m = mul <8 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v8i8_v8i64_sext(<8 x i8> %x, <8 x i8> %y) {
; CHECK-LABEL: add_v8i8_v8i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q1, q1
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vmlalv.s16 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i8> %x to <8 x i64>
  %yy = sext <8 x i8> %y to <8 x i64>
  %m = mul <8 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v4i8_v4i64_zext(<4 x i8> %x, <4 x i8> %y) {
; CHECK-LABEL: add_v4i8_v4i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.i32 q2, #0xff
; CHECK-NEXT:    vand q1, q1, q2
; CHECK-NEXT:    vand q0, q0, q2
; CHECK-NEXT:    vmlalv.u32 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i8> %x to <4 x i64>
  %yy = zext <4 x i8> %y to <4 x i64>
  %m = mul <4 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v4i8_v4i64_sext(<4 x i8> %x, <4 x i8> %y) {
; CHECK-LABEL: add_v4i8_v4i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q1, q1
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vmovlb.s16 q1, q1
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vmlalv.s32 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i8> %x to <4 x i64>
  %yy = sext <4 x i8> %y to <4 x i64>
  %m = mul <4 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v4i8i16_v4i64_zext(<4 x i8> %x, <4 x i16> %y) {
; CHECK-LABEL: add_v4i8i16_v4i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.i32 q2, #0xff
; CHECK-NEXT:    vmovlb.u16 q1, q1
; CHECK-NEXT:    vand q0, q0, q2
; CHECK-NEXT:    vmlalv.u32 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i8> %x to <4 x i64>
  %yy = zext <4 x i16> %y to <4 x i64>
  %m = mul <4 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v4i8i16_v4i64_sext(<4 x i8> %x, <4 x i16> %y) {
; CHECK-LABEL: add_v4i8i16_v4i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vmovlb.s16 q1, q1
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vmlalv.s32 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i8> %x to <4 x i64>
  %yy = sext <4 x i16> %y to <4 x i64>
  %m = mul <4 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v4i8i16_v4i32_v4i64_zext(<4 x i8> %x, <4 x i16> %y) {
; CHECK-LABEL: add_v4i8i16_v4i32_v4i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.i32 q2, #0xff
; CHECK-NEXT:    vmovlb.u16 q1, q1
; CHECK-NEXT:    vand q0, q0, q2
; CHECK-NEXT:    vmlalv.u32 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i8> %x to <4 x i32>
  %yy = zext <4 x i16> %y to <4 x i32>
  %mm = mul <4 x i32> %xx, %yy
  %m = zext <4 x i32> %mm to <4 x i64>
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v4i8i16_v4i32_v4i64_sext(<4 x i8> %x, <4 x i16> %y) {
; CHECK-LABEL: add_v4i8i16_v4i32_v4i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vmovlb.s16 q1, q1
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vmlalv.s32 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i8> %x to <4 x i32>
  %yy = sext <4 x i16> %y to <4 x i32>
  %mm = mul <4 x i32> %xx, %yy
  %m = sext <4 x i32> %mm to <4 x i64>
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v2i8_v2i64_zext(<2 x i8> %x, <2 x i8> %y) {
; CHECK-LABEL: add_v2i8_v2i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.i64 q2, #0xff
; CHECK-NEXT:    vand q1, q1, q2
; CHECK-NEXT:    vand q0, q0, q2
; CHECK-NEXT:    vmov r0, s6
; CHECK-NEXT:    vmov r1, s2
; CHECK-NEXT:    vmov r2, s4
; CHECK-NEXT:    vmov r3, s0
; CHECK-NEXT:    umull r0, r1, r1, r0
; CHECK-NEXT:    umull r2, r3, r3, r2
; CHECK-NEXT:    add r0, r2
; CHECK-NEXT:    orrs r1, r3
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <2 x i8> %x to <2 x i64>
  %yy = zext <2 x i8> %y to <2 x i64>
  %m = mul <2 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v2i8_v2i64_sext(<2 x i8> %x, <2 x i8> %y) {
; CHECK-LABEL: add_v2i8_v2i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov r0, s6
; CHECK-NEXT:    vmov r1, s2
; CHECK-NEXT:    vmov r2, s4
; CHECK-NEXT:    vmov r3, s0
; CHECK-NEXT:    sxtb r0, r0
; CHECK-NEXT:    sxtb r1, r1
; CHECK-NEXT:    smull r0, r1, r1, r0
; CHECK-NEXT:    sxtb r2, r2
; CHECK-NEXT:    sxtb r3, r3
; CHECK-NEXT:    smlal r0, r1, r3, r2
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <2 x i8> %x to <2 x i64>
  %yy = sext <2 x i8> %y to <2 x i64>
  %m = mul <2 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v2i64_v2i64(<2 x i64> %x, <2 x i64> %y) {
; CHECK-LABEL: add_v2i64_v2i64:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, r5, r7, lr}
; CHECK-NEXT:    push {r4, r5, r7, lr}
; CHECK-NEXT:    vmov r0, lr, d3
; CHECK-NEXT:    vmov r2, r3, d1
; CHECK-NEXT:    umull r12, r1, r2, r0
; CHECK-NEXT:    mla r1, r2, lr, r1
; CHECK-NEXT:    mla lr, r3, r0, r1
; CHECK-NEXT:    vmov r0, r2, d2
; CHECK-NEXT:    vmov r3, r1, d0
; CHECK-NEXT:    umull r4, r5, r3, r0
; CHECK-NEXT:    mla r2, r3, r2, r5
; CHECK-NEXT:    mla r1, r1, r0, r2
; CHECK-NEXT:    adds.w r0, r4, r12
; CHECK-NEXT:    adc.w r1, r1, lr
; CHECK-NEXT:    pop {r4, r5, r7, pc}
entry:
  %m = mul <2 x i64> %x, %y
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %m)
  ret i64 %z
}

define arm_aapcs_vfpcc i32 @add_v4i32_v4i32_acc(<4 x i32> %x, <4 x i32> %y, i32 %a) {
; CHECK-LABEL: add_v4i32_v4i32_acc:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlava.u32 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %m = mul <4 x i32> %x, %y
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %m)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i64 @add_v4i32_v4i64_acc_zext(<4 x i32> %x, <4 x i32> %y, i64 %a) {
; CHECK-LABEL: add_v4i32_v4i64_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlalva.u32 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i32> %x to <4 x i64>
  %yy = zext <4 x i32> %y to <4 x i64>
  %m = mul <4 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %m)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v4i32_v4i64_acc_sext(<4 x i32> %x, <4 x i32> %y, i64 %a) {
; CHECK-LABEL: add_v4i32_v4i64_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlalva.s32 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i32> %x to <4 x i64>
  %yy = sext <4 x i32> %y to <4 x i64>
  %m = mul <4 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %m)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v2i32_v2i64_acc_zext(<2 x i32> %x, <2 x i32> %y, i64 %a) {
; CHECK-LABEL: add_v2i32_v2i64_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    vmullb.u32 q2, q0, q1
; CHECK-NEXT:    vmov lr, r12, d5
; CHECK-NEXT:    vmov r3, r2, d4
; CHECK-NEXT:    adds.w r3, r3, lr
; CHECK-NEXT:    adc.w r2, r2, r12
; CHECK-NEXT:    adds r0, r0, r3
; CHECK-NEXT:    adcs r1, r2
; CHECK-NEXT:    pop {r7, pc}
entry:
  %xx = zext <2 x i32> %x to <2 x i64>
  %yy = zext <2 x i32> %y to <2 x i64>
  %m = mul <2 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %m)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v2i32_v2i64_acc_sext(<2 x i32> %x, <2 x i32> %y, i64 %a) {
; CHECK-LABEL: add_v2i32_v2i64_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    vmullb.s32 q2, q0, q1
; CHECK-NEXT:    vmov lr, r12, d5
; CHECK-NEXT:    vmov r3, r2, d4
; CHECK-NEXT:    adds.w r3, r3, lr
; CHECK-NEXT:    adc.w r2, r2, r12
; CHECK-NEXT:    adds r0, r0, r3
; CHECK-NEXT:    adcs r1, r2
; CHECK-NEXT:    pop {r7, pc}
entry:
  %xx = sext <2 x i32> %x to <2 x i64>
  %yy = sext <2 x i32> %y to <2 x i64>
  %m = mul <2 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %m)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i32 @add_v8i16_v8i32_acc_zext(<8 x i16> %x, <8 x i16> %y, i32 %a) {
; CHECK-LABEL: add_v8i16_v8i32_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlava.u16 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i16> %x to <8 x i32>
  %yy = zext <8 x i16> %y to <8 x i32>
  %m = mul <8 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %m)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v8i16_v8i32_acc_sext(<8 x i16> %x, <8 x i16> %y, i32 %a) {
; CHECK-LABEL: add_v8i16_v8i32_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlava.s16 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i16> %x to <8 x i32>
  %yy = sext <8 x i16> %y to <8 x i32>
  %m = mul <8 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %m)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v4i16_v4i32_acc_zext(<4 x i16> %x, <4 x i16> %y, i32 %a) {
; CHECK-LABEL: add_v4i16_v4i32_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u16 q1, q1
; CHECK-NEXT:    vmovlb.u16 q0, q0
; CHECK-NEXT:    vmlava.u32 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i16> %x to <4 x i32>
  %yy = zext <4 x i16> %y to <4 x i32>
  %m = mul <4 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %m)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v4i16_v4i32_acc_sext(<4 x i16> %x, <4 x i16> %y, i32 %a) {
; CHECK-LABEL: add_v4i16_v4i32_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s16 q1, q1
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vmlava.u32 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i16> %x to <4 x i32>
  %yy = sext <4 x i16> %y to <4 x i32>
  %m = mul <4 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %m)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc zeroext i16 @add_v8i16_v8i16_acc(<8 x i16> %x, <8 x i16> %y, i16 %a) {
; CHECK-LABEL: add_v8i16_v8i16_acc:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlava.u16 r0, q0, q1
; CHECK-NEXT:    uxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %m = mul <8 x i16> %x, %y
  %z = call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> %m)
  %r = add i16 %z, %a
  ret i16 %r
}

define arm_aapcs_vfpcc i64 @add_v8i16_v8i64_acc_zext(<8 x i16> %x, <8 x i16> %y, i64 %a) {
; CHECK-LABEL: add_v8i16_v8i64_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlalva.u16 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i16> %x to <8 x i64>
  %yy = zext <8 x i16> %y to <8 x i64>
  %m = mul <8 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %m)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v8i16_v8i64_acc_sext(<8 x i16> %x, <8 x i16> %y, i64 %a) {
; CHECK-LABEL: add_v8i16_v8i64_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlalva.s16 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i16> %x to <8 x i64>
  %yy = sext <8 x i16> %y to <8 x i64>
  %m = mul <8 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %m)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v8i16_v8i32_v8i64_acc_zext(<8 x i16> %x, <8 x i16> %y, i64 %a) {
; CHECK-LABEL: add_v8i16_v8i32_v8i64_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlalva.u16 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i16> %x to <8 x i32>
  %yy = zext <8 x i16> %y to <8 x i32>
  %m = mul <8 x i32> %xx, %yy
  %ma = zext <8 x i32> %m to <8 x i64>
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %ma)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v8i16_v8i32_v8i64_acc_sext(<8 x i16> %x, <8 x i16> %y, i64 %a) {
; CHECK-LABEL: add_v8i16_v8i32_v8i64_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlalva.s16 r0, r1, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i16> %x to <8 x i32>
  %yy = sext <8 x i16> %y to <8 x i32>
  %m = mul <8 x i32> %xx, %yy
  %ma = sext <8 x i32> %m to <8 x i64>
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %ma)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v8i16_v8i32_v8i64_acc_sextzext(<8 x i16> %x, <8 x i16> %y, i64 %a) {
; CHECK-LABEL: add_v8i16_v8i32_v8i64_acc_sextzext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlalva.s16 r0, r1, q0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i16> %x to <8 x i32>
  %m = mul <8 x i32> %xx, %xx
  %ma = zext <8 x i32> %m to <8 x i64>
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %ma)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v2i16_v2i64_acc_zext(<2 x i16> %x, <2 x i16> %y, i64 %a) {
; CHECK-LABEL: add_v2i16_v2i64_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    vmov.i64 q2, #0xffff
; CHECK-NEXT:    vand q1, q1, q2
; CHECK-NEXT:    vand q0, q0, q2
; CHECK-NEXT:    vmov r2, s6
; CHECK-NEXT:    vmov r3, s2
; CHECK-NEXT:    vmov r12, s4
; CHECK-NEXT:    umull r2, lr, r3, r2
; CHECK-NEXT:    vmov r3, s0
; CHECK-NEXT:    umlal r2, lr, r3, r12
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, lr
; CHECK-NEXT:    pop {r7, pc}
entry:
  %xx = zext <2 x i16> %x to <2 x i64>
  %yy = zext <2 x i16> %y to <2 x i64>
  %m = mul <2 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %m)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v2i16_v2i64_acc_sext(<2 x i16> %x, <2 x i16> %y, i64 %a) {
; CHECK-LABEL: add_v2i16_v2i64_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    vmov r2, s6
; CHECK-NEXT:    vmov r3, s2
; CHECK-NEXT:    sxth r2, r2
; CHECK-NEXT:    sxth r3, r3
; CHECK-NEXT:    smull r2, r12, r3, r2
; CHECK-NEXT:    vmov r3, s4
; CHECK-NEXT:    sxth.w lr, r3
; CHECK-NEXT:    vmov r3, s0
; CHECK-NEXT:    sxth r3, r3
; CHECK-NEXT:    smlal r2, r12, r3, lr
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r12
; CHECK-NEXT:    pop {r7, pc}
entry:
  %xx = sext <2 x i16> %x to <2 x i64>
  %yy = sext <2 x i16> %y to <2 x i64>
  %m = mul <2 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %m)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i32 @add_v16i8_v16i32_acc_zext(<16 x i8> %x, <16 x i8> %y, i32 %a) {
; CHECK-LABEL: add_v16i8_v16i32_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlava.u8 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <16 x i8> %x to <16 x i32>
  %yy = zext <16 x i8> %y to <16 x i32>
  %m = mul <16 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %m)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v16i8_v16i32_acc_sext(<16 x i8> %x, <16 x i8> %y, i32 %a) {
; CHECK-LABEL: add_v16i8_v16i32_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlava.s8 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <16 x i8> %x to <16 x i32>
  %yy = sext <16 x i8> %y to <16 x i32>
  %m = mul <16 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %m)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v16i8_v16i16_v16i32_acc_zext(<16 x i8> %x, <16 x i8> %y, i32 %a) {
; CHECK-LABEL: add_v16i8_v16i16_v16i32_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlava.u8 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <16 x i8> %x to <16 x i16>
  %yy = zext <16 x i8> %y to <16 x i16>
  %m = mul <16 x i16> %xx, %yy
  %ma = zext <16 x i16> %m to <16 x i32>
  %z = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %ma)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v16i8_v16i16_v16i32_acc_sext(<16 x i8> %x, <16 x i8> %y, i32 %a) {
; CHECK-LABEL: add_v16i8_v16i16_v16i32_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlava.s8 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <16 x i8> %x to <16 x i16>
  %yy = sext <16 x i8> %y to <16 x i16>
  %m = mul <16 x i16> %xx, %yy
  %ma = sext <16 x i16> %m to <16 x i32>
  %z = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %ma)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v16i8_v16i16_v16i32_acc_sextzext(<16 x i8> %x, i32 %a) {
; CHECK-LABEL: add_v16i8_v16i16_v16i32_acc_sextzext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlava.s8 r0, q0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <16 x i8> %x to <16 x i16>
  %m = mul <16 x i16> %xx, %xx
  %ma = zext <16 x i16> %m to <16 x i32>
  %z = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %ma)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v4i8_v4i32_acc_zext(<4 x i8> %x, <4 x i8> %y, i32 %a) {
; CHECK-LABEL: add_v4i8_v4i32_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.i32 q2, #0xff
; CHECK-NEXT:    vand q1, q1, q2
; CHECK-NEXT:    vand q0, q0, q2
; CHECK-NEXT:    vmlava.u32 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i8> %x to <4 x i32>
  %yy = zext <4 x i8> %y to <4 x i32>
  %m = mul <4 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %m)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v4i8_v4i32_acc_sext(<4 x i8> %x, <4 x i8> %y, i32 %a) {
; CHECK-LABEL: add_v4i8_v4i32_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q1, q1
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vmovlb.s16 q1, q1
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vmlava.u32 r0, q0, q1
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i8> %x to <4 x i32>
  %yy = sext <4 x i8> %y to <4 x i32>
  %m = mul <4 x i32> %xx, %yy
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %m)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc zeroext i16 @add_v16i8_v16i16_acc_zext(<16 x i8> %x, <16 x i8> %y, i16 %a) {
; CHECK-LABEL: add_v16i8_v16i16_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlava.u8 r0, q0, q1
; CHECK-NEXT:    uxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <16 x i8> %x to <16 x i16>
  %yy = zext <16 x i8> %y to <16 x i16>
  %m = mul <16 x i16> %xx, %yy
  %z = call i16 @llvm.vector.reduce.add.v16i16(<16 x i16> %m)
  %r = add i16 %z, %a
  ret i16 %r
}

define arm_aapcs_vfpcc signext i16 @add_v16i8_v16i16_acc_sext(<16 x i8> %x, <16 x i8> %y, i16 %a) {
; CHECK-LABEL: add_v16i8_v16i16_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlava.s8 r0, q0, q1
; CHECK-NEXT:    sxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <16 x i8> %x to <16 x i16>
  %yy = sext <16 x i8> %y to <16 x i16>
  %m = mul <16 x i16> %xx, %yy
  %z = call i16 @llvm.vector.reduce.add.v16i16(<16 x i16> %m)
  %r = add i16 %z, %a
  ret i16 %r
}

define arm_aapcs_vfpcc zeroext i16 @add_v8i8_v8i16_acc_zext(<8 x i8> %x, <8 x i8> %y, i16 %a) {
; CHECK-LABEL: add_v8i8_v8i16_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u8 q1, q1
; CHECK-NEXT:    vmovlb.u8 q0, q0
; CHECK-NEXT:    vmlava.u16 r0, q0, q1
; CHECK-NEXT:    uxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i8> %x to <8 x i16>
  %yy = zext <8 x i8> %y to <8 x i16>
  %m = mul <8 x i16> %xx, %yy
  %z = call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> %m)
  %r = add i16 %z, %a
  ret i16 %r
}

define arm_aapcs_vfpcc signext i16 @add_v8i8_v8i16_acc_sext(<8 x i8> %x, <8 x i8> %y, i16 %a) {
; CHECK-LABEL: add_v8i8_v8i16_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q1, q1
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vmlava.u16 r0, q0, q1
; CHECK-NEXT:    sxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i8> %x to <8 x i16>
  %yy = sext <8 x i8> %y to <8 x i16>
  %m = mul <8 x i16> %xx, %yy
  %z = call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> %m)
  %r = add i16 %z, %a
  ret i16 %r
}

define arm_aapcs_vfpcc zeroext i8 @add_v16i8_v16i8_acc(<16 x i8> %x, <16 x i8> %y, i8 %a) {
; CHECK-LABEL: add_v16i8_v16i8_acc:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmlava.u8 r0, q0, q1
; CHECK-NEXT:    uxtb r0, r0
; CHECK-NEXT:    bx lr
entry:
  %m = mul <16 x i8> %x, %y
  %z = call i8 @llvm.vector.reduce.add.v16i8(<16 x i8> %m)
  %r = add i8 %z, %a
  ret i8 %r
}

define arm_aapcs_vfpcc i64 @add_v16i8_v16i64_acc_zext(<16 x i8> %x, <16 x i8> %y, i64 %a) {
; CHECK-LABEL: add_v16i8_v16i64_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .pad #32
; CHECK-NEXT:    sub sp, #32
; CHECK-NEXT:    add r2, sp, #16
; CHECK-NEXT:    mov r3, sp
; CHECK-NEXT:    vstrw.32 q1, [r2]
; CHECK-NEXT:    vstrw.32 q0, [r3]
; CHECK-NEXT:    vldrb.u16 q0, [r2]
; CHECK-NEXT:    vldrb.u16 q1, [r3]
; CHECK-NEXT:    vmlalva.u16 r0, r1, q1, q0
; CHECK-NEXT:    vldrb.u16 q0, [r2, #8]
; CHECK-NEXT:    vldrb.u16 q1, [r3, #8]
; CHECK-NEXT:    vmlalva.u16 r0, r1, q1, q0
; CHECK-NEXT:    add sp, #32
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <16 x i8> %x to <16 x i64>
  %yy = zext <16 x i8> %y to <16 x i64>
  %m = mul <16 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v16i64(<16 x i64> %m)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v16i8_v16i64_acc_sext(<16 x i8> %x, <16 x i8> %y, i64 %a) {
; CHECK-LABEL: add_v16i8_v16i64_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .pad #32
; CHECK-NEXT:    sub sp, #32
; CHECK-NEXT:    add r2, sp, #16
; CHECK-NEXT:    mov r3, sp
; CHECK-NEXT:    vstrw.32 q1, [r2]
; CHECK-NEXT:    vstrw.32 q0, [r3]
; CHECK-NEXT:    vldrb.s16 q0, [r2]
; CHECK-NEXT:    vldrb.s16 q1, [r3]
; CHECK-NEXT:    vmlalva.s16 r0, r1, q1, q0
; CHECK-NEXT:    vldrb.s16 q0, [r2, #8]
; CHECK-NEXT:    vldrb.s16 q1, [r3, #8]
; CHECK-NEXT:    vmlalva.s16 r0, r1, q1, q0
; CHECK-NEXT:    add sp, #32
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <16 x i8> %x to <16 x i64>
  %yy = sext <16 x i8> %y to <16 x i64>
  %m = mul <16 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v16i64(<16 x i64> %m)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v16i8_v16i64_acc_zext_load(ptr %xp, ptr %yp, i64 %a) {
; CHECK-LABEL: add_v16i8_v16i64_acc_zext_load:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldrb.u16 q0, [r1]
; CHECK-NEXT:    vldrb.u16 q1, [r0]
; CHECK-NEXT:    vmlalva.u16 r2, r3, q1, q0
; CHECK-NEXT:    vldrb.u16 q0, [r1, #8]
; CHECK-NEXT:    vldrb.u16 q1, [r0, #8]
; CHECK-NEXT:    vmlalva.u16 r2, r3, q1, q0
; CHECK-NEXT:    mov r0, r2
; CHECK-NEXT:    mov r1, r3
; CHECK-NEXT:    bx lr
entry:
  %x = load <16 x i8>, ptr %xp
  %y = load <16 x i8>, ptr %yp
  %xx = zext <16 x i8> %x to <16 x i64>
  %yy = zext <16 x i8> %y to <16 x i64>
  %m = mul <16 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v16i64(<16 x i64> %m)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v16i8_v16i64_acc_sext_load(ptr %xp, ptr %yp, i64 %a) {
; CHECK-LABEL: add_v16i8_v16i64_acc_sext_load:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldrb.s16 q0, [r1]
; CHECK-NEXT:    vldrb.s16 q1, [r0]
; CHECK-NEXT:    vmlalva.s16 r2, r3, q1, q0
; CHECK-NEXT:    vldrb.s16 q0, [r1, #8]
; CHECK-NEXT:    vldrb.s16 q1, [r0, #8]
; CHECK-NEXT:    vmlalva.s16 r2, r3, q1, q0
; CHECK-NEXT:    mov r0, r2
; CHECK-NEXT:    mov r1, r3
; CHECK-NEXT:    bx lr
entry:
  %x = load <16 x i8>, ptr %xp
  %y = load <16 x i8>, ptr %yp
  %xx = sext <16 x i8> %x to <16 x i64>
  %yy = sext <16 x i8> %y to <16 x i64>
  %m = mul <16 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v16i64(<16 x i64> %m)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v2i8_v2i64_acc_zext(<2 x i8> %x, <2 x i8> %y, i64 %a) {
; CHECK-LABEL: add_v2i8_v2i64_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    vmov.i64 q2, #0xff
; CHECK-NEXT:    vand q1, q1, q2
; CHECK-NEXT:    vand q0, q0, q2
; CHECK-NEXT:    vmov r2, s6
; CHECK-NEXT:    vmov r3, s2
; CHECK-NEXT:    umull r12, lr, r3, r2
; CHECK-NEXT:    vmov r2, s4
; CHECK-NEXT:    vmov r3, s0
; CHECK-NEXT:    umull r2, r3, r3, r2
; CHECK-NEXT:    add r2, r12
; CHECK-NEXT:    orr.w r3, r3, lr
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    pop {r7, pc}
entry:
  %xx = zext <2 x i8> %x to <2 x i64>
  %yy = zext <2 x i8> %y to <2 x i64>
  %m = mul <2 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %m)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v2i8_v2i64_acc_sext(<2 x i8> %x, <2 x i8> %y, i64 %a) {
; CHECK-LABEL: add_v2i8_v2i64_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    vmov r2, s6
; CHECK-NEXT:    vmov r3, s2
; CHECK-NEXT:    sxtb r2, r2
; CHECK-NEXT:    sxtb r3, r3
; CHECK-NEXT:    smull r2, r12, r3, r2
; CHECK-NEXT:    vmov r3, s4
; CHECK-NEXT:    sxtb.w lr, r3
; CHECK-NEXT:    vmov r3, s0
; CHECK-NEXT:    sxtb r3, r3
; CHECK-NEXT:    smlal r2, r12, r3, lr
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r12
; CHECK-NEXT:    pop {r7, pc}
entry:
  %xx = sext <2 x i8> %x to <2 x i64>
  %yy = sext <2 x i8> %y to <2 x i64>
  %m = mul <2 x i64> %xx, %yy
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %m)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v2i64_v2i64_acc(<2 x i64> %x, <2 x i64> %y, i64 %a) {
; CHECK-LABEL: add_v2i64_v2i64_acc:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, r5, r6, r7, r8, lr}
; CHECK-NEXT:    push.w {r4, r5, r6, r7, r8, lr}
; CHECK-NEXT:    vmov r2, r12, d3
; CHECK-NEXT:    vmov r3, lr, d1
; CHECK-NEXT:    vmov r4, r6, d0
; CHECK-NEXT:    umull r8, r5, r3, r2
; CHECK-NEXT:    mla r3, r3, r12, r5
; CHECK-NEXT:    mla r12, lr, r2, r3
; CHECK-NEXT:    vmov r3, r5, d2
; CHECK-NEXT:    umull r7, r2, r4, r3
; CHECK-NEXT:    mla r2, r4, r5, r2
; CHECK-NEXT:    mla r2, r6, r3, r2
; CHECK-NEXT:    adds.w r3, r7, r8
; CHECK-NEXT:    adc.w r2, r2, r12
; CHECK-NEXT:    adds r0, r0, r3
; CHECK-NEXT:    adcs r1, r2
; CHECK-NEXT:    pop.w {r4, r5, r6, r7, r8, pc}
entry:
  %m = mul <2 x i64> %x, %y
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %m)
  %r = add i64 %z, %a
  ret i64 %r
}

declare i16 @llvm.vector.reduce.add.v16i16(<16 x i16>)
declare i16 @llvm.vector.reduce.add.v8i16(<8 x i16>)
declare i32 @llvm.vector.reduce.add.v16i32(<16 x i32>)
declare i32 @llvm.vector.reduce.add.v4i32(<4 x i32>)
declare i32 @llvm.vector.reduce.add.v8i32(<8 x i32>)
declare i64 @llvm.vector.reduce.add.v16i64(<16 x i64>)
declare i64 @llvm.vector.reduce.add.v2i64(<2 x i64>)
declare i64 @llvm.vector.reduce.add.v4i64(<4 x i64>)
declare i64 @llvm.vector.reduce.add.v8i64(<8 x i64>)
declare i8 @llvm.vector.reduce.add.v16i8(<16 x i8>)