; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=thumbv8.1m.main-none-none-eabi -mattr=+mve,+fullfp16 -verify-machineinstrs %s -o - | FileCheck %s --check-prefix=CHECK-LE
; RUN: llc -mtriple=thumbebv8.1m.main-none-none-eabi -mattr=+mve,+fullfp16 -verify-machineinstrs %s -o - | FileCheck %s --check-prefix=CHECK-BE

define arm_aapcs_vfpcc void @masked_v4i32(ptr %dest, <4 x i32> %a) {
; CHECK-LE-LABEL: masked_v4i32:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vpt.s32 gt, q0, zr
; CHECK-LE-NEXT:    vstrwt.32 q0, [r0]
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v4i32:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vrev64.32 q1, q0
; CHECK-BE-NEXT:    vpt.s32 gt, q1, zr
; CHECK-BE-NEXT:    vstrwt.32 q1, [r0]
; CHECK-BE-NEXT:    bx lr
entry:
  %c = icmp sgt <4 x i32> %a, zeroinitializer
  call void @llvm.masked.store.v4i32.p0(<4 x i32> %a, ptr %dest, i32 4, <4 x i1> %c)
  ret void
}

define arm_aapcs_vfpcc void @masked_v4i32_align1(ptr %dest, <4 x i32> %a) {
; CHECK-LE-LABEL: masked_v4i32_align1:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    .pad #4
; CHECK-LE-NEXT:    sub sp, #4
; CHECK-LE-NEXT:    vcmp.s32 gt, q0, zr
; CHECK-LE-NEXT:    vmrs r2, p0
; CHECK-LE-NEXT:    and r1, r2, #1
; CHECK-LE-NEXT:    rsbs r3, r1, #0
; CHECK-LE-NEXT:    movs r1, #0
; CHECK-LE-NEXT:    bfi r1, r3, #0, #1
; CHECK-LE-NEXT:    ubfx r3, r2, #4, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r1, r3, #1, #1
; CHECK-LE-NEXT:    ubfx r3, r2, #8, #1
; CHECK-LE-NEXT:    ubfx r2, r2, #12, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r1, r3, #2, #1
; CHECK-LE-NEXT:    rsbs r2, r2, #0
; CHECK-LE-NEXT:    bfi r1, r2, #3, #1
; CHECK-LE-NEXT:    lsls r2, r1, #31
; CHECK-LE-NEXT:    itt ne
; CHECK-LE-NEXT:    vmovne r2, s0
; CHECK-LE-NEXT:    strne r2, [r0]
; CHECK-LE-NEXT:    lsls r2, r1, #30
; CHECK-LE-NEXT:    itt mi
; CHECK-LE-NEXT:    vmovmi r2, s1
; CHECK-LE-NEXT:    strmi r2, [r0, #4]
; CHECK-LE-NEXT:    lsls r2, r1, #29
; CHECK-LE-NEXT:    itt mi
; CHECK-LE-NEXT:    vmovmi r2, s2
; CHECK-LE-NEXT:    strmi r2, [r0, #8]
; CHECK-LE-NEXT:    lsls r1, r1, #28
; CHECK-LE-NEXT:    itt mi
; CHECK-LE-NEXT:    vmovmi r1, s3
; CHECK-LE-NEXT:    strmi r1, [r0, #12]
; CHECK-LE-NEXT:    add sp, #4
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v4i32_align1:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    .pad #4
; CHECK-BE-NEXT:    sub sp, #4
; CHECK-BE-NEXT:    vrev64.32 q1, q0
; CHECK-BE-NEXT:    vcmp.s32 gt, q1, zr
; CHECK-BE-NEXT:    vmrs r2, p0
; CHECK-BE-NEXT:    ubfx r1, r2, #12, #1
; CHECK-BE-NEXT:    rsbs r3, r1, #0
; CHECK-BE-NEXT:    movs r1, #0
; CHECK-BE-NEXT:    bfi r1, r3, #0, #1
; CHECK-BE-NEXT:    ubfx r3, r2, #8, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r1, r3, #1, #1
; CHECK-BE-NEXT:    ubfx r3, r2, #4, #1
; CHECK-BE-NEXT:    and r2, r2, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r1, r3, #2, #1
; CHECK-BE-NEXT:    rsbs r2, r2, #0
; CHECK-BE-NEXT:    bfi r1, r2, #3, #1
; CHECK-BE-NEXT:    lsls r2, r1, #28
; CHECK-BE-NEXT:    itt mi
; CHECK-BE-NEXT:    vmovmi r2, s4
; CHECK-BE-NEXT:    strmi r2, [r0]
; CHECK-BE-NEXT:    lsls r2, r1, #29
; CHECK-BE-NEXT:    itt mi
; CHECK-BE-NEXT:    vmovmi r2, s5
; CHECK-BE-NEXT:    strmi r2, [r0, #4]
; CHECK-BE-NEXT:    lsls r2, r1, #30
; CHECK-BE-NEXT:    itt mi
; CHECK-BE-NEXT:    vmovmi r2, s6
; CHECK-BE-NEXT:    strmi r2, [r0, #8]
; CHECK-BE-NEXT:    lsls r1, r1, #31
; CHECK-BE-NEXT:    itt ne
; CHECK-BE-NEXT:    vmovne r1, s7
; CHECK-BE-NEXT:    strne r1, [r0, #12]
; CHECK-BE-NEXT:    add sp, #4
; CHECK-BE-NEXT:    bx lr
entry:
  %c = icmp sgt <4 x i32> %a, zeroinitializer
  call void @llvm.masked.store.v4i32.p0(<4 x i32> %a, ptr %dest, i32 1, <4 x i1> %c)
  ret void
}

define ptr @masked_v4i32_pre(ptr %y, ptr %x, <4 x i32> %a) {
; CHECK-LE-LABEL: masked_v4i32_pre:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vldr d1, [sp]
; CHECK-LE-NEXT:    vmov d0, r2, r3
; CHECK-LE-NEXT:    vldrw.u32 q1, [r1]
; CHECK-LE-NEXT:    vpt.s32 gt, q0, zr
; CHECK-LE-NEXT:    vstrwt.32 q1, [r0, #4]!
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v4i32_pre:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vldr d1, [sp]
; CHECK-BE-NEXT:    vmov d0, r3, r2
; CHECK-BE-NEXT:    vldrw.u32 q1, [r1]
; CHECK-BE-NEXT:    vrev64.32 q2, q0
; CHECK-BE-NEXT:    vpt.s32 gt, q2, zr
; CHECK-BE-NEXT:    vstrwt.32 q1, [r0, #4]!
; CHECK-BE-NEXT:    bx lr
entry:
  %z = getelementptr inbounds i8, ptr %y, i32 4
  %0 = load <4 x i32>, ptr %x, align 4
  %c = icmp sgt <4 x i32> %a, zeroinitializer
  call void @llvm.masked.store.v4i32.p0(<4 x i32> %0, ptr %z, i32 4, <4 x i1> %c)
  ret ptr %z
}

define ptr @masked_v4i32_post(ptr %y, ptr %x, <4 x i32> %a) {
; CHECK-LE-LABEL: masked_v4i32_post:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vldr d1, [sp]
; CHECK-LE-NEXT:    vmov d0, r2, r3
; CHECK-LE-NEXT:    vldrw.u32 q1, [r1]
; CHECK-LE-NEXT:    vpt.s32 gt, q0, zr
; CHECK-LE-NEXT:    vstrwt.32 q1, [r0], #4
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v4i32_post:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vldr d1, [sp]
; CHECK-BE-NEXT:    vmov d0, r3, r2
; CHECK-BE-NEXT:    vldrw.u32 q1, [r1]
; CHECK-BE-NEXT:    vrev64.32 q2, q0
; CHECK-BE-NEXT:    vpt.s32 gt, q2, zr
; CHECK-BE-NEXT:    vstrwt.32 q1, [r0], #4
; CHECK-BE-NEXT:    bx lr
entry:
  %z = getelementptr inbounds i8, ptr %y, i32 4
  %0 = load <4 x i32>, ptr %x, align 4
  %c = icmp sgt <4 x i32> %a, zeroinitializer
  call void @llvm.masked.store.v4i32.p0(<4 x i32> %0, ptr %y, i32 4, <4 x i1> %c)
  ret ptr %z
}


define arm_aapcs_vfpcc void @masked_v8i16(ptr %dest, <8 x i16> %a) {
; CHECK-LE-LABEL: masked_v8i16:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vpt.s16 gt, q0, zr
; CHECK-LE-NEXT:    vstrht.16 q0, [r0]
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v8i16:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vrev64.16 q1, q0
; CHECK-BE-NEXT:    vpt.s16 gt, q1, zr
; CHECK-BE-NEXT:    vstrht.16 q1, [r0]
; CHECK-BE-NEXT:    bx lr
entry:
  %c = icmp sgt <8 x i16> %a, zeroinitializer
  call void @llvm.masked.store.v8i16.p0(<8 x i16> %a, ptr %dest, i32 2, <8 x i1> %c)
  ret void
}

define arm_aapcs_vfpcc void @masked_v8i16_align1(ptr %dest, <8 x i16> %a) {
; CHECK-LE-LABEL: masked_v8i16_align1:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    .pad #4
; CHECK-LE-NEXT:    sub sp, #4
; CHECK-LE-NEXT:    vcmp.s16 gt, q0, zr
; CHECK-LE-NEXT:    vmrs r1, p0
; CHECK-LE-NEXT:    and r2, r1, #1
; CHECK-LE-NEXT:    rsbs r3, r2, #0
; CHECK-LE-NEXT:    movs r2, #0
; CHECK-LE-NEXT:    bfi r2, r3, #0, #1
; CHECK-LE-NEXT:    ubfx r3, r1, #2, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r2, r3, #1, #1
; CHECK-LE-NEXT:    ubfx r3, r1, #4, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r2, r3, #2, #1
; CHECK-LE-NEXT:    ubfx r3, r1, #6, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r2, r3, #3, #1
; CHECK-LE-NEXT:    ubfx r3, r1, #8, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r2, r3, #4, #1
; CHECK-LE-NEXT:    ubfx r3, r1, #10, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r2, r3, #5, #1
; CHECK-LE-NEXT:    ubfx r3, r1, #12, #1
; CHECK-LE-NEXT:    ubfx r1, r1, #14, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r2, r3, #6, #1
; CHECK-LE-NEXT:    rsbs r1, r1, #0
; CHECK-LE-NEXT:    bfi r2, r1, #7, #1
; CHECK-LE-NEXT:    uxtb r1, r2
; CHECK-LE-NEXT:    lsls r2, r2, #31
; CHECK-LE-NEXT:    itt ne
; CHECK-LE-NEXT:    vmovne.u16 r2, q0[0]
; CHECK-LE-NEXT:    strhne r2, [r0]
; CHECK-LE-NEXT:    lsls r2, r1, #30
; CHECK-LE-NEXT:    itt mi
; CHECK-LE-NEXT:    vmovmi.u16 r2, q0[1]
; CHECK-LE-NEXT:    strhmi r2, [r0, #2]
; CHECK-LE-NEXT:    lsls r2, r1, #29
; CHECK-LE-NEXT:    itt mi
; CHECK-LE-NEXT:    vmovmi.u16 r2, q0[2]
; CHECK-LE-NEXT:    strhmi r2, [r0, #4]
; CHECK-LE-NEXT:    lsls r2, r1, #28
; CHECK-LE-NEXT:    itt mi
; CHECK-LE-NEXT:    vmovmi.u16 r2, q0[3]
; CHECK-LE-NEXT:    strhmi r2, [r0, #6]
; CHECK-LE-NEXT:    lsls r2, r1, #27
; CHECK-LE-NEXT:    itt mi
; CHECK-LE-NEXT:    vmovmi.u16 r2, q0[4]
; CHECK-LE-NEXT:    strhmi r2, [r0, #8]
; CHECK-LE-NEXT:    lsls r2, r1, #26
; CHECK-LE-NEXT:    itt mi
; CHECK-LE-NEXT:    vmovmi.u16 r2, q0[5]
; CHECK-LE-NEXT:    strhmi r2, [r0, #10]
; CHECK-LE-NEXT:    lsls r2, r1, #25
; CHECK-LE-NEXT:    itt mi
; CHECK-LE-NEXT:    vmovmi.u16 r2, q0[6]
; CHECK-LE-NEXT:    strhmi r2, [r0, #12]
; CHECK-LE-NEXT:    lsls r1, r1, #24
; CHECK-LE-NEXT:    itt mi
; CHECK-LE-NEXT:    vmovmi.u16 r1, q0[7]
; CHECK-LE-NEXT:    strhmi r1, [r0, #14]
; CHECK-LE-NEXT:    add sp, #4
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v8i16_align1:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    .pad #4
; CHECK-BE-NEXT:    sub sp, #4
; CHECK-BE-NEXT:    vrev64.16 q1, q0
; CHECK-BE-NEXT:    vcmp.s16 gt, q1, zr
; CHECK-BE-NEXT:    vmrs r1, p0
; CHECK-BE-NEXT:    ubfx r2, r1, #14, #1
; CHECK-BE-NEXT:    rsbs r3, r2, #0
; CHECK-BE-NEXT:    movs r2, #0
; CHECK-BE-NEXT:    bfi r2, r3, #0, #1
; CHECK-BE-NEXT:    ubfx r3, r1, #12, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r2, r3, #1, #1
; CHECK-BE-NEXT:    ubfx r3, r1, #10, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r2, r3, #2, #1
; CHECK-BE-NEXT:    ubfx r3, r1, #8, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r2, r3, #3, #1
; CHECK-BE-NEXT:    ubfx r3, r1, #6, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r2, r3, #4, #1
; CHECK-BE-NEXT:    ubfx r3, r1, #4, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r2, r3, #5, #1
; CHECK-BE-NEXT:    ubfx r3, r1, #2, #1
; CHECK-BE-NEXT:    and r1, r1, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r2, r3, #6, #1
; CHECK-BE-NEXT:    rsbs r1, r1, #0
; CHECK-BE-NEXT:    bfi r2, r1, #7, #1
; CHECK-BE-NEXT:    uxtb r1, r2
; CHECK-BE-NEXT:    lsls r2, r2, #24
; CHECK-BE-NEXT:    itt mi
; CHECK-BE-NEXT:    vmovmi.u16 r2, q1[0]
; CHECK-BE-NEXT:    strhmi r2, [r0]
; CHECK-BE-NEXT:    lsls r2, r1, #25
; CHECK-BE-NEXT:    itt mi
; CHECK-BE-NEXT:    vmovmi.u16 r2, q1[1]
; CHECK-BE-NEXT:    strhmi r2, [r0, #2]
; CHECK-BE-NEXT:    lsls r2, r1, #26
; CHECK-BE-NEXT:    itt mi
; CHECK-BE-NEXT:    vmovmi.u16 r2, q1[2]
; CHECK-BE-NEXT:    strhmi r2, [r0, #4]
; CHECK-BE-NEXT:    lsls r2, r1, #27
; CHECK-BE-NEXT:    itt mi
; CHECK-BE-NEXT:    vmovmi.u16 r2, q1[3]
; CHECK-BE-NEXT:    strhmi r2, [r0, #6]
; CHECK-BE-NEXT:    lsls r2, r1, #28
; CHECK-BE-NEXT:    itt mi
; CHECK-BE-NEXT:    vmovmi.u16 r2, q1[4]
; CHECK-BE-NEXT:    strhmi r2, [r0, #8]
; CHECK-BE-NEXT:    lsls r2, r1, #29
; CHECK-BE-NEXT:    itt mi
; CHECK-BE-NEXT:    vmovmi.u16 r2, q1[5]
; CHECK-BE-NEXT:    strhmi r2, [r0, #10]
; CHECK-BE-NEXT:    lsls r2, r1, #30
; CHECK-BE-NEXT:    itt mi
; CHECK-BE-NEXT:    vmovmi.u16 r2, q1[6]
; CHECK-BE-NEXT:    strhmi r2, [r0, #12]
; CHECK-BE-NEXT:    lsls r1, r1, #31
; CHECK-BE-NEXT:    itt ne
; CHECK-BE-NEXT:    vmovne.u16 r1, q1[7]
; CHECK-BE-NEXT:    strhne r1, [r0, #14]
; CHECK-BE-NEXT:    add sp, #4
; CHECK-BE-NEXT:    bx lr
entry:
  %c = icmp sgt <8 x i16> %a, zeroinitializer
  call void @llvm.masked.store.v8i16.p0(<8 x i16> %a, ptr %dest, i32 1, <8 x i1> %c)
  ret void
}

define ptr @masked_v8i16_pre(ptr %y, ptr %x, <8 x i16> %a) {
; CHECK-LE-LABEL: masked_v8i16_pre:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vldr d1, [sp]
; CHECK-LE-NEXT:    vmov d0, r2, r3
; CHECK-LE-NEXT:    vldrw.u32 q1, [r1]
; CHECK-LE-NEXT:    vpt.s16 gt, q0, zr
; CHECK-LE-NEXT:    vstrht.16 q1, [r0, #4]!
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v8i16_pre:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vldr d1, [sp]
; CHECK-BE-NEXT:    vmov d0, r3, r2
; CHECK-BE-NEXT:    vldrh.u16 q1, [r1]
; CHECK-BE-NEXT:    vrev64.16 q2, q0
; CHECK-BE-NEXT:    vpt.s16 gt, q2, zr
; CHECK-BE-NEXT:    vstrht.16 q1, [r0, #4]!
; CHECK-BE-NEXT:    bx lr
entry:
  %z = getelementptr inbounds i8, ptr %y, i32 4
  %0 = load <8 x i16>, ptr %x, align 4
  %c = icmp sgt <8 x i16> %a, zeroinitializer
  call void @llvm.masked.store.v8i16.p0(<8 x i16> %0, ptr %z, i32 2, <8 x i1> %c)
  ret ptr %z
}

define ptr @masked_v8i16_post(ptr %y, ptr %x, <8 x i16> %a) {
; CHECK-LE-LABEL: masked_v8i16_post:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vldr d1, [sp]
; CHECK-LE-NEXT:    vmov d0, r2, r3
; CHECK-LE-NEXT:    vldrw.u32 q1, [r1]
; CHECK-LE-NEXT:    vpt.s16 gt, q0, zr
; CHECK-LE-NEXT:    vstrht.16 q1, [r0], #4
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v8i16_post:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vldr d1, [sp]
; CHECK-BE-NEXT:    vmov d0, r3, r2
; CHECK-BE-NEXT:    vldrh.u16 q1, [r1]
; CHECK-BE-NEXT:    vrev64.16 q2, q0
; CHECK-BE-NEXT:    vpt.s16 gt, q2, zr
; CHECK-BE-NEXT:    vstrht.16 q1, [r0], #4
; CHECK-BE-NEXT:    bx lr
entry:
  %z = getelementptr inbounds i8, ptr %y, i32 4
  %0 = load <8 x i16>, ptr %x, align 4
  %c = icmp sgt <8 x i16> %a, zeroinitializer
  call void @llvm.masked.store.v8i16.p0(<8 x i16> %0, ptr %y, i32 2, <8 x i1> %c)
  ret ptr %z
}


define arm_aapcs_vfpcc void @masked_v16i8(ptr %dest, <16 x i8> %a) {
; CHECK-LE-LABEL: masked_v16i8:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vpt.s8 gt, q0, zr
; CHECK-LE-NEXT:    vstrbt.8 q0, [r0]
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v16i8:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vrev64.8 q1, q0
; CHECK-BE-NEXT:    vpt.s8 gt, q1, zr
; CHECK-BE-NEXT:    vstrbt.8 q1, [r0]
; CHECK-BE-NEXT:    bx lr
entry:
  %c = icmp sgt <16 x i8> %a, zeroinitializer
  call void @llvm.masked.store.v16i8.p0(<16 x i8> %a, ptr %dest, i32 1, <16 x i1> %c)
  ret void
}

define ptr @masked_v16i8_pre(ptr %y, ptr %x, <16 x i8> %a) {
; CHECK-LE-LABEL: masked_v16i8_pre:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vldr d1, [sp]
; CHECK-LE-NEXT:    vmov d0, r2, r3
; CHECK-LE-NEXT:    vldrw.u32 q1, [r1]
; CHECK-LE-NEXT:    vpt.s8 gt, q0, zr
; CHECK-LE-NEXT:    vstrbt.8 q1, [r0, #4]!
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v16i8_pre:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vldr d1, [sp]
; CHECK-BE-NEXT:    vmov d0, r3, r2
; CHECK-BE-NEXT:    vldrb.u8 q1, [r1]
; CHECK-BE-NEXT:    vrev64.8 q2, q0
; CHECK-BE-NEXT:    vpt.s8 gt, q2, zr
; CHECK-BE-NEXT:    vstrbt.8 q1, [r0, #4]!
; CHECK-BE-NEXT:    bx lr
entry:
  %z = getelementptr inbounds i8, ptr %y, i32 4
  %0 = load <16 x i8>, ptr %x, align 4
  %c = icmp sgt <16 x i8> %a, zeroinitializer
  call void @llvm.masked.store.v16i8.p0(<16 x i8> %0, ptr %z, i32 1, <16 x i1> %c)
  ret ptr %z
}

define ptr @masked_v16i8_post(ptr %y, ptr %x, <16 x i8> %a) {
; CHECK-LE-LABEL: masked_v16i8_post:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vldr d1, [sp]
; CHECK-LE-NEXT:    vmov d0, r2, r3
; CHECK-LE-NEXT:    vldrw.u32 q1, [r1]
; CHECK-LE-NEXT:    vpt.s8 gt, q0, zr
; CHECK-LE-NEXT:    vstrbt.8 q1, [r0], #4
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v16i8_post:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vldr d1, [sp]
; CHECK-BE-NEXT:    vmov d0, r3, r2
; CHECK-BE-NEXT:    vldrb.u8 q1, [r1]
; CHECK-BE-NEXT:    vrev64.8 q2, q0
; CHECK-BE-NEXT:    vpt.s8 gt, q2, zr
; CHECK-BE-NEXT:    vstrbt.8 q1, [r0], #4
; CHECK-BE-NEXT:    bx lr
entry:
  %z = getelementptr inbounds i8, ptr %y, i32 4
  %0 = load <16 x i8>, ptr %x, align 4
  %c = icmp sgt <16 x i8> %a, zeroinitializer
  call void @llvm.masked.store.v16i8.p0(<16 x i8> %0, ptr %y, i32 1, <16 x i1> %c)
  ret ptr %z
}


define arm_aapcs_vfpcc void @masked_v4f32(ptr %dest, <4 x float> %a, <4 x i32> %b) {
; CHECK-LE-LABEL: masked_v4f32:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vpt.i32 ne, q1, zr
; CHECK-LE-NEXT:    vstrwt.32 q0, [r0]
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v4f32:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vrev64.32 q2, q1
; CHECK-BE-NEXT:    vrev64.32 q1, q0
; CHECK-BE-NEXT:    vpt.i32 ne, q2, zr
; CHECK-BE-NEXT:    vstrwt.32 q1, [r0]
; CHECK-BE-NEXT:    bx lr
entry:
  %c = icmp ugt <4 x i32> %b, zeroinitializer
  call void @llvm.masked.store.v4f32.p0(<4 x float> %a, ptr %dest, i32 4, <4 x i1> %c)
  ret void
}

define arm_aapcs_vfpcc void @masked_v4f32_align1(ptr %dest, <4 x float> %a, <4 x i32> %b) {
; CHECK-LE-LABEL: masked_v4f32_align1:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    .pad #4
; CHECK-LE-NEXT:    sub sp, #4
; CHECK-LE-NEXT:    vcmp.i32 ne, q1, zr
; CHECK-LE-NEXT:    movs r1, #0
; CHECK-LE-NEXT:    vmrs r2, p0
; CHECK-LE-NEXT:    and r3, r2, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r1, r3, #0, #1
; CHECK-LE-NEXT:    ubfx r3, r2, #4, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r1, r3, #1, #1
; CHECK-LE-NEXT:    ubfx r3, r2, #8, #1
; CHECK-LE-NEXT:    ubfx r2, r2, #12, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r1, r3, #2, #1
; CHECK-LE-NEXT:    rsbs r2, r2, #0
; CHECK-LE-NEXT:    bfi r1, r2, #3, #1
; CHECK-LE-NEXT:    lsls r2, r1, #31
; CHECK-LE-NEXT:    itt ne
; CHECK-LE-NEXT:    vmovne r2, s0
; CHECK-LE-NEXT:    strne r2, [r0]
; CHECK-LE-NEXT:    lsls r2, r1, #30
; CHECK-LE-NEXT:    itt mi
; CHECK-LE-NEXT:    vmovmi r2, s1
; CHECK-LE-NEXT:    strmi r2, [r0, #4]
; CHECK-LE-NEXT:    lsls r2, r1, #29
; CHECK-LE-NEXT:    itt mi
; CHECK-LE-NEXT:    vmovmi r2, s2
; CHECK-LE-NEXT:    strmi r2, [r0, #8]
; CHECK-LE-NEXT:    lsls r1, r1, #28
; CHECK-LE-NEXT:    itt mi
; CHECK-LE-NEXT:    vmovmi r1, s3
; CHECK-LE-NEXT:    strmi r1, [r0, #12]
; CHECK-LE-NEXT:    add sp, #4
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v4f32_align1:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    .pad #4
; CHECK-BE-NEXT:    sub sp, #4
; CHECK-BE-NEXT:    vrev64.32 q2, q1
; CHECK-BE-NEXT:    movs r1, #0
; CHECK-BE-NEXT:    vcmp.i32 ne, q2, zr
; CHECK-BE-NEXT:    vrev64.32 q1, q0
; CHECK-BE-NEXT:    vmrs r2, p0
; CHECK-BE-NEXT:    ubfx r3, r2, #12, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r1, r3, #0, #1
; CHECK-BE-NEXT:    ubfx r3, r2, #8, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r1, r3, #1, #1
; CHECK-BE-NEXT:    ubfx r3, r2, #4, #1
; CHECK-BE-NEXT:    and r2, r2, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r1, r3, #2, #1
; CHECK-BE-NEXT:    rsbs r2, r2, #0
; CHECK-BE-NEXT:    bfi r1, r2, #3, #1
; CHECK-BE-NEXT:    lsls r2, r1, #28
; CHECK-BE-NEXT:    itt mi
; CHECK-BE-NEXT:    vmovmi r2, s4
; CHECK-BE-NEXT:    strmi r2, [r0]
; CHECK-BE-NEXT:    lsls r2, r1, #29
; CHECK-BE-NEXT:    itt mi
; CHECK-BE-NEXT:    vmovmi r2, s5
; CHECK-BE-NEXT:    strmi r2, [r0, #4]
; CHECK-BE-NEXT:    lsls r2, r1, #30
; CHECK-BE-NEXT:    itt mi
; CHECK-BE-NEXT:    vmovmi r2, s6
; CHECK-BE-NEXT:    strmi r2, [r0, #8]
; CHECK-BE-NEXT:    lsls r1, r1, #31
; CHECK-BE-NEXT:    itt ne
; CHECK-BE-NEXT:    vmovne r1, s7
; CHECK-BE-NEXT:    strne r1, [r0, #12]
; CHECK-BE-NEXT:    add sp, #4
; CHECK-BE-NEXT:    bx lr
entry:
  %c = icmp ugt <4 x i32> %b, zeroinitializer
  call void @llvm.masked.store.v4f32.p0(<4 x float> %a, ptr %dest, i32 1, <4 x i1> %c)
  ret void
}

define ptr @masked_v4f32_pre(ptr %y, ptr %x, <4 x i32> %a) {
; CHECK-LE-LABEL: masked_v4f32_pre:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vldr d1, [sp]
; CHECK-LE-NEXT:    vmov d0, r2, r3
; CHECK-LE-NEXT:    vldrw.u32 q1, [r1]
; CHECK-LE-NEXT:    vpt.s32 gt, q0, zr
; CHECK-LE-NEXT:    vstrwt.32 q1, [r0, #4]!
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v4f32_pre:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vldr d1, [sp]
; CHECK-BE-NEXT:    vmov d0, r3, r2
; CHECK-BE-NEXT:    vldrw.u32 q1, [r1]
; CHECK-BE-NEXT:    vrev64.32 q2, q0
; CHECK-BE-NEXT:    vpt.s32 gt, q2, zr
; CHECK-BE-NEXT:    vstrwt.32 q1, [r0, #4]!
; CHECK-BE-NEXT:    bx lr
entry:
  %z = getelementptr inbounds i8, ptr %y, i32 4
  %0 = load <4 x float>, ptr %x, align 4
  %c = icmp sgt <4 x i32> %a, zeroinitializer
  call void @llvm.masked.store.v4f32.p0(<4 x float> %0, ptr %z, i32 4, <4 x i1> %c)
  ret ptr %z
}

define ptr @masked_v4f32_post(ptr %y, ptr %x, <4 x i32> %a) {
; CHECK-LE-LABEL: masked_v4f32_post:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vldr d1, [sp]
; CHECK-LE-NEXT:    vmov d0, r2, r3
; CHECK-LE-NEXT:    vldrw.u32 q1, [r1]
; CHECK-LE-NEXT:    vpt.s32 gt, q0, zr
; CHECK-LE-NEXT:    vstrwt.32 q1, [r0], #4
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v4f32_post:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vldr d1, [sp]
; CHECK-BE-NEXT:    vmov d0, r3, r2
; CHECK-BE-NEXT:    vldrw.u32 q1, [r1]
; CHECK-BE-NEXT:    vrev64.32 q2, q0
; CHECK-BE-NEXT:    vpt.s32 gt, q2, zr
; CHECK-BE-NEXT:    vstrwt.32 q1, [r0], #4
; CHECK-BE-NEXT:    bx lr
entry:
  %z = getelementptr inbounds i8, ptr %y, i32 4
  %0 = load <4 x float>, ptr %x, align 4
  %c = icmp sgt <4 x i32> %a, zeroinitializer
  call void @llvm.masked.store.v4f32.p0(<4 x float> %0, ptr %y, i32 4, <4 x i1> %c)
  ret ptr %z
}


define arm_aapcs_vfpcc void @masked_v8f16(ptr %dest, <8 x half> %a, <8 x i16> %b) {
; CHECK-LE-LABEL: masked_v8f16:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vpt.i16 ne, q1, zr
; CHECK-LE-NEXT:    vstrht.16 q0, [r0]
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v8f16:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vrev64.16 q2, q1
; CHECK-BE-NEXT:    vrev64.16 q1, q0
; CHECK-BE-NEXT:    vpt.i16 ne, q2, zr
; CHECK-BE-NEXT:    vstrht.16 q1, [r0]
; CHECK-BE-NEXT:    bx lr
entry:
  %c = icmp ugt <8 x i16> %b, zeroinitializer
  call void @llvm.masked.store.v8f16.p0(<8 x half> %a, ptr %dest, i32 2, <8 x i1> %c)
  ret void
}

define arm_aapcs_vfpcc void @masked_v8f16_align1(ptr %dest, <8 x half> %a, <8 x i16> %b) {
; CHECK-LE-LABEL: masked_v8f16_align1:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    .pad #36
; CHECK-LE-NEXT:    sub sp, #36
; CHECK-LE-NEXT:    vcmp.i16 ne, q1, zr
; CHECK-LE-NEXT:    movs r2, #0
; CHECK-LE-NEXT:    vmrs r1, p0
; CHECK-LE-NEXT:    and r3, r1, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r2, r3, #0, #1
; CHECK-LE-NEXT:    ubfx r3, r1, #2, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r2, r3, #1, #1
; CHECK-LE-NEXT:    ubfx r3, r1, #4, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r2, r3, #2, #1
; CHECK-LE-NEXT:    ubfx r3, r1, #6, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r2, r3, #3, #1
; CHECK-LE-NEXT:    ubfx r3, r1, #8, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r2, r3, #4, #1
; CHECK-LE-NEXT:    ubfx r3, r1, #10, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r2, r3, #5, #1
; CHECK-LE-NEXT:    ubfx r3, r1, #12, #1
; CHECK-LE-NEXT:    ubfx r1, r1, #14, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r2, r3, #6, #1
; CHECK-LE-NEXT:    rsbs r1, r1, #0
; CHECK-LE-NEXT:    bfi r2, r1, #7, #1
; CHECK-LE-NEXT:    uxtb r1, r2
; CHECK-LE-NEXT:    lsls r2, r2, #31
; CHECK-LE-NEXT:    bne .LBB16_9
; CHECK-LE-NEXT:  @ %bb.1: @ %else
; CHECK-LE-NEXT:    lsls r2, r1, #30
; CHECK-LE-NEXT:    bmi .LBB16_10
; CHECK-LE-NEXT:  .LBB16_2: @ %else2
; CHECK-LE-NEXT:    lsls r2, r1, #29
; CHECK-LE-NEXT:    bmi .LBB16_11
; CHECK-LE-NEXT:  .LBB16_3: @ %else4
; CHECK-LE-NEXT:    lsls r2, r1, #28
; CHECK-LE-NEXT:    bmi .LBB16_12
; CHECK-LE-NEXT:  .LBB16_4: @ %else6
; CHECK-LE-NEXT:    lsls r2, r1, #27
; CHECK-LE-NEXT:    bmi .LBB16_13
; CHECK-LE-NEXT:  .LBB16_5: @ %else8
; CHECK-LE-NEXT:    lsls r2, r1, #26
; CHECK-LE-NEXT:    bmi .LBB16_14
; CHECK-LE-NEXT:  .LBB16_6: @ %else10
; CHECK-LE-NEXT:    lsls r2, r1, #25
; CHECK-LE-NEXT:    bmi .LBB16_15
; CHECK-LE-NEXT:  .LBB16_7: @ %else12
; CHECK-LE-NEXT:    lsls r1, r1, #24
; CHECK-LE-NEXT:    bmi .LBB16_16
; CHECK-LE-NEXT:  .LBB16_8: @ %else14
; CHECK-LE-NEXT:    add sp, #36
; CHECK-LE-NEXT:    bx lr
; CHECK-LE-NEXT:  .LBB16_9: @ %cond.store
; CHECK-LE-NEXT:    vstr.16 s0, [sp, #28]
; CHECK-LE-NEXT:    ldrh.w r2, [sp, #28]
; CHECK-LE-NEXT:    strh r2, [r0]
; CHECK-LE-NEXT:    lsls r2, r1, #30
; CHECK-LE-NEXT:    bpl .LBB16_2
; CHECK-LE-NEXT:  .LBB16_10: @ %cond.store1
; CHECK-LE-NEXT:    vmovx.f16 s0, s0
; CHECK-LE-NEXT:    vstr.16 s0, [sp, #24]
; CHECK-LE-NEXT:    ldrh.w r2, [sp, #24]
; CHECK-LE-NEXT:    strh r2, [r0, #2]
; CHECK-LE-NEXT:    lsls r2, r1, #29
; CHECK-LE-NEXT:    bpl .LBB16_3
; CHECK-LE-NEXT:  .LBB16_11: @ %cond.store3
; CHECK-LE-NEXT:    vstr.16 s1, [sp, #20]
; CHECK-LE-NEXT:    ldrh.w r2, [sp, #20]
; CHECK-LE-NEXT:    strh r2, [r0, #4]
; CHECK-LE-NEXT:    lsls r2, r1, #28
; CHECK-LE-NEXT:    bpl .LBB16_4
; CHECK-LE-NEXT:  .LBB16_12: @ %cond.store5
; CHECK-LE-NEXT:    vmovx.f16 s0, s1
; CHECK-LE-NEXT:    vstr.16 s0, [sp, #16]
; CHECK-LE-NEXT:    ldrh.w r2, [sp, #16]
; CHECK-LE-NEXT:    strh r2, [r0, #6]
; CHECK-LE-NEXT:    lsls r2, r1, #27
; CHECK-LE-NEXT:    bpl .LBB16_5
; CHECK-LE-NEXT:  .LBB16_13: @ %cond.store7
; CHECK-LE-NEXT:    vstr.16 s2, [sp, #12]
; CHECK-LE-NEXT:    ldrh.w r2, [sp, #12]
; CHECK-LE-NEXT:    strh r2, [r0, #8]
; CHECK-LE-NEXT:    lsls r2, r1, #26
; CHECK-LE-NEXT:    bpl .LBB16_6
; CHECK-LE-NEXT:  .LBB16_14: @ %cond.store9
; CHECK-LE-NEXT:    vmovx.f16 s0, s2
; CHECK-LE-NEXT:    vstr.16 s0, [sp, #8]
; CHECK-LE-NEXT:    ldrh.w r2, [sp, #8]
; CHECK-LE-NEXT:    strh r2, [r0, #10]
; CHECK-LE-NEXT:    lsls r2, r1, #25
; CHECK-LE-NEXT:    bpl .LBB16_7
; CHECK-LE-NEXT:  .LBB16_15: @ %cond.store11
; CHECK-LE-NEXT:    vstr.16 s3, [sp, #4]
; CHECK-LE-NEXT:    ldrh.w r2, [sp, #4]
; CHECK-LE-NEXT:    strh r2, [r0, #12]
; CHECK-LE-NEXT:    lsls r1, r1, #24
; CHECK-LE-NEXT:    bpl .LBB16_8
; CHECK-LE-NEXT:  .LBB16_16: @ %cond.store13
; CHECK-LE-NEXT:    vmovx.f16 s0, s3
; CHECK-LE-NEXT:    vstr.16 s0, [sp]
; CHECK-LE-NEXT:    ldrh.w r1, [sp]
; CHECK-LE-NEXT:    strh r1, [r0, #14]
; CHECK-LE-NEXT:    add sp, #36
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v8f16_align1:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    .pad #36
; CHECK-BE-NEXT:    sub sp, #36
; CHECK-BE-NEXT:    vrev64.16 q2, q1
; CHECK-BE-NEXT:    vrev64.16 q1, q0
; CHECK-BE-NEXT:    vcmp.i16 ne, q2, zr
; CHECK-BE-NEXT:    vmrs r1, p0
; CHECK-BE-NEXT:    ubfx r2, r1, #14, #1
; CHECK-BE-NEXT:    rsbs r3, r2, #0
; CHECK-BE-NEXT:    movs r2, #0
; CHECK-BE-NEXT:    bfi r2, r3, #0, #1
; CHECK-BE-NEXT:    ubfx r3, r1, #12, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r2, r3, #1, #1
; CHECK-BE-NEXT:    ubfx r3, r1, #10, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r2, r3, #2, #1
; CHECK-BE-NEXT:    ubfx r3, r1, #8, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r2, r3, #3, #1
; CHECK-BE-NEXT:    ubfx r3, r1, #6, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r2, r3, #4, #1
; CHECK-BE-NEXT:    ubfx r3, r1, #4, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r2, r3, #5, #1
; CHECK-BE-NEXT:    ubfx r3, r1, #2, #1
; CHECK-BE-NEXT:    and r1, r1, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r2, r3, #6, #1
; CHECK-BE-NEXT:    rsbs r1, r1, #0
; CHECK-BE-NEXT:    bfi r2, r1, #7, #1
; CHECK-BE-NEXT:    uxtb r1, r2
; CHECK-BE-NEXT:    lsls r2, r2, #24
; CHECK-BE-NEXT:    bmi .LBB16_9
; CHECK-BE-NEXT:  @ %bb.1: @ %else
; CHECK-BE-NEXT:    lsls r2, r1, #25
; CHECK-BE-NEXT:    bmi .LBB16_10
; CHECK-BE-NEXT:  .LBB16_2: @ %else2
; CHECK-BE-NEXT:    lsls r2, r1, #26
; CHECK-BE-NEXT:    bmi .LBB16_11
; CHECK-BE-NEXT:  .LBB16_3: @ %else4
; CHECK-BE-NEXT:    lsls r2, r1, #27
; CHECK-BE-NEXT:    bmi .LBB16_12
; CHECK-BE-NEXT:  .LBB16_4: @ %else6
; CHECK-BE-NEXT:    lsls r2, r1, #28
; CHECK-BE-NEXT:    bmi .LBB16_13
; CHECK-BE-NEXT:  .LBB16_5: @ %else8
; CHECK-BE-NEXT:    lsls r2, r1, #29
; CHECK-BE-NEXT:    bmi .LBB16_14
; CHECK-BE-NEXT:  .LBB16_6: @ %else10
; CHECK-BE-NEXT:    lsls r2, r1, #30
; CHECK-BE-NEXT:    bmi .LBB16_15
; CHECK-BE-NEXT:  .LBB16_7: @ %else12
; CHECK-BE-NEXT:    lsls r1, r1, #31
; CHECK-BE-NEXT:    bne .LBB16_16
; CHECK-BE-NEXT:  .LBB16_8: @ %else14
; CHECK-BE-NEXT:    add sp, #36
; CHECK-BE-NEXT:    bx lr
; CHECK-BE-NEXT:  .LBB16_9: @ %cond.store
; CHECK-BE-NEXT:    vstr.16 s4, [sp, #28]
; CHECK-BE-NEXT:    ldrh.w r2, [sp, #28]
; CHECK-BE-NEXT:    strh r2, [r0]
; CHECK-BE-NEXT:    lsls r2, r1, #25
; CHECK-BE-NEXT:    bpl .LBB16_2
; CHECK-BE-NEXT:  .LBB16_10: @ %cond.store1
; CHECK-BE-NEXT:    vmovx.f16 s0, s4
; CHECK-BE-NEXT:    vstr.16 s0, [sp, #24]
; CHECK-BE-NEXT:    ldrh.w r2, [sp, #24]
; CHECK-BE-NEXT:    strh r2, [r0, #2]
; CHECK-BE-NEXT:    lsls r2, r1, #26
; CHECK-BE-NEXT:    bpl .LBB16_3
; CHECK-BE-NEXT:  .LBB16_11: @ %cond.store3
; CHECK-BE-NEXT:    vstr.16 s5, [sp, #20]
; CHECK-BE-NEXT:    ldrh.w r2, [sp, #20]
; CHECK-BE-NEXT:    strh r2, [r0, #4]
; CHECK-BE-NEXT:    lsls r2, r1, #27
; CHECK-BE-NEXT:    bpl .LBB16_4
; CHECK-BE-NEXT:  .LBB16_12: @ %cond.store5
; CHECK-BE-NEXT:    vmovx.f16 s0, s5
; CHECK-BE-NEXT:    vstr.16 s0, [sp, #16]
; CHECK-BE-NEXT:    ldrh.w r2, [sp, #16]
; CHECK-BE-NEXT:    strh r2, [r0, #6]
; CHECK-BE-NEXT:    lsls r2, r1, #28
; CHECK-BE-NEXT:    bpl .LBB16_5
; CHECK-BE-NEXT:  .LBB16_13: @ %cond.store7
; CHECK-BE-NEXT:    vstr.16 s6, [sp, #12]
; CHECK-BE-NEXT:    ldrh.w r2, [sp, #12]
; CHECK-BE-NEXT:    strh r2, [r0, #8]
; CHECK-BE-NEXT:    lsls r2, r1, #29
; CHECK-BE-NEXT:    bpl .LBB16_6
; CHECK-BE-NEXT:  .LBB16_14: @ %cond.store9
; CHECK-BE-NEXT:    vmovx.f16 s0, s6
; CHECK-BE-NEXT:    vstr.16 s0, [sp, #8]
; CHECK-BE-NEXT:    ldrh.w r2, [sp, #8]
; CHECK-BE-NEXT:    strh r2, [r0, #10]
; CHECK-BE-NEXT:    lsls r2, r1, #30
; CHECK-BE-NEXT:    bpl .LBB16_7
; CHECK-BE-NEXT:  .LBB16_15: @ %cond.store11
; CHECK-BE-NEXT:    vstr.16 s7, [sp, #4]
; CHECK-BE-NEXT:    ldrh.w r2, [sp, #4]
; CHECK-BE-NEXT:    strh r2, [r0, #12]
; CHECK-BE-NEXT:    lsls r1, r1, #31
; CHECK-BE-NEXT:    beq .LBB16_8
; CHECK-BE-NEXT:  .LBB16_16: @ %cond.store13
; CHECK-BE-NEXT:    vmovx.f16 s0, s7
; CHECK-BE-NEXT:    vstr.16 s0, [sp]
; CHECK-BE-NEXT:    ldrh.w r1, [sp]
; CHECK-BE-NEXT:    strh r1, [r0, #14]
; CHECK-BE-NEXT:    add sp, #36
; CHECK-BE-NEXT:    bx lr
entry:
  %c = icmp ugt <8 x i16> %b, zeroinitializer
  call void @llvm.masked.store.v8f16.p0(<8 x half> %a, ptr %dest, i32 1, <8 x i1> %c)
  ret void
}

define ptr @masked_v8f16_pre(ptr %y, ptr %x, <8 x i16> %a) {
; CHECK-LE-LABEL: masked_v8f16_pre:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vldr d1, [sp]
; CHECK-LE-NEXT:    vmov d0, r2, r3
; CHECK-LE-NEXT:    vldrw.u32 q1, [r1]
; CHECK-LE-NEXT:    vpt.s16 gt, q0, zr
; CHECK-LE-NEXT:    vstrht.16 q1, [r0, #4]!
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v8f16_pre:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vldr d1, [sp]
; CHECK-BE-NEXT:    vmov d0, r3, r2
; CHECK-BE-NEXT:    vldrh.u16 q1, [r1]
; CHECK-BE-NEXT:    vrev64.16 q2, q0
; CHECK-BE-NEXT:    vpt.s16 gt, q2, zr
; CHECK-BE-NEXT:    vstrht.16 q1, [r0, #4]!
; CHECK-BE-NEXT:    bx lr
entry:
  %z = getelementptr inbounds i8, ptr %y, i32 4
  %0 = load <8 x half>, ptr %x, align 4
  %c = icmp sgt <8 x i16> %a, zeroinitializer
  call void @llvm.masked.store.v8f16.p0(<8 x half> %0, ptr %z, i32 2, <8 x i1> %c)
  ret ptr %z
}

define ptr @masked_v8f16_post(ptr %y, ptr %x, <8 x i16> %a) {
; CHECK-LE-LABEL: masked_v8f16_post:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vldr d1, [sp]
; CHECK-LE-NEXT:    vmov d0, r2, r3
; CHECK-LE-NEXT:    vldrw.u32 q1, [r1]
; CHECK-LE-NEXT:    vpt.s16 gt, q0, zr
; CHECK-LE-NEXT:    vstrht.16 q1, [r0], #4
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v8f16_post:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vldr d1, [sp]
; CHECK-BE-NEXT:    vmov d0, r3, r2
; CHECK-BE-NEXT:    vldrh.u16 q1, [r1]
; CHECK-BE-NEXT:    vrev64.16 q2, q0
; CHECK-BE-NEXT:    vpt.s16 gt, q2, zr
; CHECK-BE-NEXT:    vstrht.16 q1, [r0], #4
; CHECK-BE-NEXT:    bx lr
entry:
  %z = getelementptr inbounds i8, ptr %y, i32 4
  %0 = load <8 x half>, ptr %x, align 4
  %c = icmp sgt <8 x i16> %a, zeroinitializer
  call void @llvm.masked.store.v8f16.p0(<8 x half> %0, ptr %y, i32 2, <8 x i1> %c)
  ret ptr %z
}


define arm_aapcs_vfpcc void @masked_v2i64(ptr %dest, <2 x i64> %a) {
; CHECK-LE-LABEL: masked_v2i64:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    .save {r7, lr}
; CHECK-LE-NEXT:    push {r7, lr}
; CHECK-LE-NEXT:    .pad #4
; CHECK-LE-NEXT:    sub sp, #4
; CHECK-LE-NEXT:    vmov r2, r3, d0
; CHECK-LE-NEXT:    movs r1, #0
; CHECK-LE-NEXT:    vmov r12, lr, d1
; CHECK-LE-NEXT:    rsbs r2, r2, #0
; CHECK-LE-NEXT:    sbcs.w r2, r1, r3
; CHECK-LE-NEXT:    csetm r2, lt
; CHECK-LE-NEXT:    rsbs.w r3, r12, #0
; CHECK-LE-NEXT:    sbcs.w r3, r1, lr
; CHECK-LE-NEXT:    bfi r1, r2, #0, #1
; CHECK-LE-NEXT:    csetm r2, lt
; CHECK-LE-NEXT:    bfi r1, r2, #1, #1
; CHECK-LE-NEXT:    lsls r2, r1, #31
; CHECK-LE-NEXT:    it ne
; CHECK-LE-NEXT:    vstrne d0, [r0]
; CHECK-LE-NEXT:    lsls r1, r1, #30
; CHECK-LE-NEXT:    it mi
; CHECK-LE-NEXT:    vstrmi d1, [r0, #8]
; CHECK-LE-NEXT:    add sp, #4
; CHECK-LE-NEXT:    pop {r7, pc}
;
; CHECK-BE-LABEL: masked_v2i64:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    .save {r7, lr}
; CHECK-BE-NEXT:    push {r7, lr}
; CHECK-BE-NEXT:    .pad #4
; CHECK-BE-NEXT:    sub sp, #4
; CHECK-BE-NEXT:    vrev64.32 q1, q0
; CHECK-BE-NEXT:    movs r1, #0
; CHECK-BE-NEXT:    vmov r2, r3, d3
; CHECK-BE-NEXT:    vmov r12, lr, d2
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    sbcs.w r2, r1, r2
; CHECK-BE-NEXT:    csetm r2, lt
; CHECK-BE-NEXT:    rsbs.w r3, lr, #0
; CHECK-BE-NEXT:    sbcs.w r3, r1, r12
; CHECK-BE-NEXT:    bfi r1, r2, #0, #1
; CHECK-BE-NEXT:    csetm r2, lt
; CHECK-BE-NEXT:    bfi r1, r2, #1, #1
; CHECK-BE-NEXT:    lsls r2, r1, #30
; CHECK-BE-NEXT:    it mi
; CHECK-BE-NEXT:    vstrmi d0, [r0]
; CHECK-BE-NEXT:    lsls r1, r1, #31
; CHECK-BE-NEXT:    it ne
; CHECK-BE-NEXT:    vstrne d1, [r0, #8]
; CHECK-BE-NEXT:    add sp, #4
; CHECK-BE-NEXT:    pop {r7, pc}
entry:
  %c = icmp sgt <2 x i64> %a, zeroinitializer
  call void @llvm.masked.store.v2i64.p0(<2 x i64> %a, ptr %dest, i32 8, <2 x i1> %c)
  ret void
}

define arm_aapcs_vfpcc void @masked_v2f64(ptr %dest, <2 x double> %a, <2 x i64> %b) {
; CHECK-LE-LABEL: masked_v2f64:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    .save {r7, lr}
; CHECK-LE-NEXT:    push {r7, lr}
; CHECK-LE-NEXT:    .pad #4
; CHECK-LE-NEXT:    sub sp, #4
; CHECK-LE-NEXT:    vmov r2, r3, d2
; CHECK-LE-NEXT:    movs r1, #0
; CHECK-LE-NEXT:    vmov r12, lr, d3
; CHECK-LE-NEXT:    rsbs r2, r2, #0
; CHECK-LE-NEXT:    sbcs.w r2, r1, r3
; CHECK-LE-NEXT:    csetm r2, lt
; CHECK-LE-NEXT:    rsbs.w r3, r12, #0
; CHECK-LE-NEXT:    sbcs.w r3, r1, lr
; CHECK-LE-NEXT:    bfi r1, r2, #0, #1
; CHECK-LE-NEXT:    csetm r2, lt
; CHECK-LE-NEXT:    bfi r1, r2, #1, #1
; CHECK-LE-NEXT:    lsls r2, r1, #31
; CHECK-LE-NEXT:    it ne
; CHECK-LE-NEXT:    vstrne d0, [r0]
; CHECK-LE-NEXT:    lsls r1, r1, #30
; CHECK-LE-NEXT:    it mi
; CHECK-LE-NEXT:    vstrmi d1, [r0, #8]
; CHECK-LE-NEXT:    add sp, #4
; CHECK-LE-NEXT:    pop {r7, pc}
;
; CHECK-BE-LABEL: masked_v2f64:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    .save {r7, lr}
; CHECK-BE-NEXT:    push {r7, lr}
; CHECK-BE-NEXT:    .pad #4
; CHECK-BE-NEXT:    sub sp, #4
; CHECK-BE-NEXT:    vrev64.32 q2, q1
; CHECK-BE-NEXT:    movs r1, #0
; CHECK-BE-NEXT:    vmov r2, r3, d5
; CHECK-BE-NEXT:    vmov r12, lr, d4
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    sbcs.w r2, r1, r2
; CHECK-BE-NEXT:    csetm r2, lt
; CHECK-BE-NEXT:    rsbs.w r3, lr, #0
; CHECK-BE-NEXT:    sbcs.w r3, r1, r12
; CHECK-BE-NEXT:    bfi r1, r2, #0, #1
; CHECK-BE-NEXT:    csetm r2, lt
; CHECK-BE-NEXT:    bfi r1, r2, #1, #1
; CHECK-BE-NEXT:    lsls r2, r1, #30
; CHECK-BE-NEXT:    it mi
; CHECK-BE-NEXT:    vstrmi d0, [r0]
; CHECK-BE-NEXT:    lsls r1, r1, #31
; CHECK-BE-NEXT:    it ne
; CHECK-BE-NEXT:    vstrne d1, [r0, #8]
; CHECK-BE-NEXT:    add sp, #4
; CHECK-BE-NEXT:    pop {r7, pc}
entry:
  %c = icmp sgt <2 x i64> %b, zeroinitializer
  call void @llvm.masked.store.v2f64.p0(<2 x double> %a, ptr %dest, i32 8, <2 x i1> %c)
  ret void
}

define arm_aapcs_vfpcc void @masked_v4i16(ptr %dest, <4 x i32> %a) {
; CHECK-LE-LABEL: masked_v4i16:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vpt.s32 gt, q0, zr
; CHECK-LE-NEXT:    vstrht.32 q0, [r0]
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v4i16:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vrev64.32 q1, q0
; CHECK-BE-NEXT:    vpt.s32 gt, q1, zr
; CHECK-BE-NEXT:    vstrht.32 q1, [r0]
; CHECK-BE-NEXT:    bx lr
entry:
  %c = icmp sgt <4 x i32> %a, zeroinitializer
  %trunc = trunc <4 x i32> %a to <4 x i16>
  call void @llvm.masked.store.v4i16.p0(<4 x i16> %trunc, ptr %dest, i32 2, <4 x i1> %c)
  ret void
}

define arm_aapcs_vfpcc void @masked_v4i8(ptr %dest, <4 x i32> %a) {
; CHECK-LE-LABEL: masked_v4i8:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vpt.s32 gt, q0, zr
; CHECK-LE-NEXT:    vstrbt.32 q0, [r0]
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v4i8:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vrev64.32 q1, q0
; CHECK-BE-NEXT:    vpt.s32 gt, q1, zr
; CHECK-BE-NEXT:    vstrbt.32 q1, [r0]
; CHECK-BE-NEXT:    bx lr
entry:
  %c = icmp sgt <4 x i32> %a, zeroinitializer
  %trunc = trunc <4 x i32> %a to <4 x i8>
  call void @llvm.masked.store.v4i8.p0(<4 x i8> %trunc, ptr %dest, i32 1, <4 x i1> %c)
  ret void
}

define arm_aapcs_vfpcc void @masked_v8i8(ptr %dest, <8 x i16> %a) {
; CHECK-LE-LABEL: masked_v8i8:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    vpt.s16 gt, q0, zr
; CHECK-LE-NEXT:    vstrbt.16 q0, [r0]
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v8i8:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    vrev64.16 q1, q0
; CHECK-BE-NEXT:    vpt.s16 gt, q1, zr
; CHECK-BE-NEXT:    vstrbt.16 q1, [r0]
; CHECK-BE-NEXT:    bx lr
entry:
  %c = icmp sgt <8 x i16> %a, zeroinitializer
  %trunc = trunc <8 x i16> %a to <8 x i8>
  call void @llvm.masked.store.v8i8.p0(<8 x i8> %trunc, ptr %dest, i32 1, <8 x i1> %c)
  ret void
}

define arm_aapcs_vfpcc void @masked_v4i16_align1(ptr %dest, <4 x i32> %a) {
; CHECK-LE-LABEL: masked_v4i16_align1:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    .pad #4
; CHECK-LE-NEXT:    sub sp, #4
; CHECK-LE-NEXT:    vcmp.s32 gt, q0, zr
; CHECK-LE-NEXT:    vmrs r2, p0
; CHECK-LE-NEXT:    and r1, r2, #1
; CHECK-LE-NEXT:    rsbs r3, r1, #0
; CHECK-LE-NEXT:    movs r1, #0
; CHECK-LE-NEXT:    bfi r1, r3, #0, #1
; CHECK-LE-NEXT:    ubfx r3, r2, #4, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r1, r3, #1, #1
; CHECK-LE-NEXT:    ubfx r3, r2, #8, #1
; CHECK-LE-NEXT:    ubfx r2, r2, #12, #1
; CHECK-LE-NEXT:    rsbs r3, r3, #0
; CHECK-LE-NEXT:    bfi r1, r3, #2, #1
; CHECK-LE-NEXT:    rsbs r2, r2, #0
; CHECK-LE-NEXT:    bfi r1, r2, #3, #1
; CHECK-LE-NEXT:    lsls r2, r1, #31
; CHECK-LE-NEXT:    itt ne
; CHECK-LE-NEXT:    vmovne r2, s0
; CHECK-LE-NEXT:    strhne r2, [r0]
; CHECK-LE-NEXT:    lsls r2, r1, #30
; CHECK-LE-NEXT:    itt mi
; CHECK-LE-NEXT:    vmovmi r2, s1
; CHECK-LE-NEXT:    strhmi r2, [r0, #2]
; CHECK-LE-NEXT:    lsls r2, r1, #29
; CHECK-LE-NEXT:    itt mi
; CHECK-LE-NEXT:    vmovmi r2, s2
; CHECK-LE-NEXT:    strhmi r2, [r0, #4]
; CHECK-LE-NEXT:    lsls r1, r1, #28
; CHECK-LE-NEXT:    itt mi
; CHECK-LE-NEXT:    vmovmi r1, s3
; CHECK-LE-NEXT:    strhmi r1, [r0, #6]
; CHECK-LE-NEXT:    add sp, #4
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v4i16_align1:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    .pad #4
; CHECK-BE-NEXT:    sub sp, #4
; CHECK-BE-NEXT:    vrev64.32 q1, q0
; CHECK-BE-NEXT:    vcmp.s32 gt, q1, zr
; CHECK-BE-NEXT:    vmrs r2, p0
; CHECK-BE-NEXT:    ubfx r1, r2, #12, #1
; CHECK-BE-NEXT:    rsbs r3, r1, #0
; CHECK-BE-NEXT:    movs r1, #0
; CHECK-BE-NEXT:    bfi r1, r3, #0, #1
; CHECK-BE-NEXT:    ubfx r3, r2, #8, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r1, r3, #1, #1
; CHECK-BE-NEXT:    ubfx r3, r2, #4, #1
; CHECK-BE-NEXT:    and r2, r2, #1
; CHECK-BE-NEXT:    rsbs r3, r3, #0
; CHECK-BE-NEXT:    bfi r1, r3, #2, #1
; CHECK-BE-NEXT:    rsbs r2, r2, #0
; CHECK-BE-NEXT:    bfi r1, r2, #3, #1
; CHECK-BE-NEXT:    lsls r2, r1, #28
; CHECK-BE-NEXT:    itt mi
; CHECK-BE-NEXT:    vmovmi r2, s4
; CHECK-BE-NEXT:    strhmi r2, [r0]
; CHECK-BE-NEXT:    lsls r2, r1, #29
; CHECK-BE-NEXT:    itt mi
; CHECK-BE-NEXT:    vmovmi r2, s5
; CHECK-BE-NEXT:    strhmi r2, [r0, #2]
; CHECK-BE-NEXT:    lsls r2, r1, #30
; CHECK-BE-NEXT:    itt mi
; CHECK-BE-NEXT:    vmovmi r2, s6
; CHECK-BE-NEXT:    strhmi r2, [r0, #4]
; CHECK-BE-NEXT:    lsls r1, r1, #31
; CHECK-BE-NEXT:    itt ne
; CHECK-BE-NEXT:    vmovne r1, s7
; CHECK-BE-NEXT:    strhne r1, [r0, #6]
; CHECK-BE-NEXT:    add sp, #4
; CHECK-BE-NEXT:    bx lr
entry:
  %c = icmp sgt <4 x i32> %a, zeroinitializer
  %trunc = trunc <4 x i32> %a to <4 x i16>
  call void @llvm.masked.store.v4i16.p0(<4 x i16> %trunc, ptr %dest, i32 1, <4 x i1> %c)
  ret void
}

define arm_aapcs_vfpcc void @masked_v4f16_align4(ptr %dest, <4 x float> %a) {
; CHECK-LE-LABEL: masked_v4f16_align4:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    .pad #4
; CHECK-LE-NEXT:    sub sp, #4
; CHECK-LE-NEXT:    vcmp.f32 s0, #0
; CHECK-LE-NEXT:    movs r1, #0
; CHECK-LE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-LE-NEXT:    vcmp.f32 s1, #0
; CHECK-LE-NEXT:    vcvtb.f16.f32 s4, s0
; CHECK-LE-NEXT:    vcvtb.f16.f32 s5, s2
; CHECK-LE-NEXT:    vcvtt.f16.f32 s4, s1
; CHECK-LE-NEXT:    vcvtt.f16.f32 s5, s3
; CHECK-LE-NEXT:    csetm r2, gt
; CHECK-LE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-LE-NEXT:    vcmp.f32 s2, #0
; CHECK-LE-NEXT:    bfi r1, r2, #0, #1
; CHECK-LE-NEXT:    csetm r2, gt
; CHECK-LE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-LE-NEXT:    vcmp.f32 s3, #0
; CHECK-LE-NEXT:    bfi r1, r2, #1, #1
; CHECK-LE-NEXT:    csetm r2, gt
; CHECK-LE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-LE-NEXT:    bfi r1, r2, #2, #1
; CHECK-LE-NEXT:    csetm r2, gt
; CHECK-LE-NEXT:    bfi r1, r2, #3, #1
; CHECK-LE-NEXT:    lsls r2, r1, #31
; CHECK-LE-NEXT:    bne .LBB25_5
; CHECK-LE-NEXT:  @ %bb.1: @ %else
; CHECK-LE-NEXT:    lsls r2, r1, #30
; CHECK-LE-NEXT:    bmi .LBB25_6
; CHECK-LE-NEXT:  .LBB25_2: @ %else2
; CHECK-LE-NEXT:    lsls r2, r1, #29
; CHECK-LE-NEXT:    bmi .LBB25_7
; CHECK-LE-NEXT:  .LBB25_3: @ %else4
; CHECK-LE-NEXT:    lsls r1, r1, #28
; CHECK-LE-NEXT:    bmi .LBB25_8
; CHECK-LE-NEXT:  .LBB25_4: @ %else6
; CHECK-LE-NEXT:    add sp, #4
; CHECK-LE-NEXT:    bx lr
; CHECK-LE-NEXT:  .LBB25_5: @ %cond.store
; CHECK-LE-NEXT:    vstr.16 s4, [r0]
; CHECK-LE-NEXT:    lsls r2, r1, #30
; CHECK-LE-NEXT:    bpl .LBB25_2
; CHECK-LE-NEXT:  .LBB25_6: @ %cond.store1
; CHECK-LE-NEXT:    vmovx.f16 s0, s4
; CHECK-LE-NEXT:    vstr.16 s0, [r0, #2]
; CHECK-LE-NEXT:    lsls r2, r1, #29
; CHECK-LE-NEXT:    bpl .LBB25_3
; CHECK-LE-NEXT:  .LBB25_7: @ %cond.store3
; CHECK-LE-NEXT:    vstr.16 s5, [r0, #4]
; CHECK-LE-NEXT:    lsls r1, r1, #28
; CHECK-LE-NEXT:    bpl .LBB25_4
; CHECK-LE-NEXT:  .LBB25_8: @ %cond.store5
; CHECK-LE-NEXT:    vmovx.f16 s0, s5
; CHECK-LE-NEXT:    vstr.16 s0, [r0, #6]
; CHECK-LE-NEXT:    add sp, #4
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v4f16_align4:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    .pad #4
; CHECK-BE-NEXT:    sub sp, #4
; CHECK-BE-NEXT:    vrev64.32 q1, q0
; CHECK-BE-NEXT:    movs r1, #0
; CHECK-BE-NEXT:    vcmp.f32 s7, #0
; CHECK-BE-NEXT:    vcvtb.f16.f32 s0, s4
; CHECK-BE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-BE-NEXT:    vcmp.f32 s6, #0
; CHECK-BE-NEXT:    vcvtb.f16.f32 s1, s6
; CHECK-BE-NEXT:    vcvtt.f16.f32 s0, s5
; CHECK-BE-NEXT:    vcvtt.f16.f32 s1, s7
; CHECK-BE-NEXT:    csetm r2, gt
; CHECK-BE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-BE-NEXT:    vcmp.f32 s5, #0
; CHECK-BE-NEXT:    bfi r1, r2, #0, #1
; CHECK-BE-NEXT:    csetm r2, gt
; CHECK-BE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-BE-NEXT:    vcmp.f32 s4, #0
; CHECK-BE-NEXT:    bfi r1, r2, #1, #1
; CHECK-BE-NEXT:    csetm r2, gt
; CHECK-BE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-BE-NEXT:    bfi r1, r2, #2, #1
; CHECK-BE-NEXT:    csetm r2, gt
; CHECK-BE-NEXT:    bfi r1, r2, #3, #1
; CHECK-BE-NEXT:    lsls r2, r1, #28
; CHECK-BE-NEXT:    bmi .LBB25_5
; CHECK-BE-NEXT:  @ %bb.1: @ %else
; CHECK-BE-NEXT:    lsls r2, r1, #29
; CHECK-BE-NEXT:    bmi .LBB25_6
; CHECK-BE-NEXT:  .LBB25_2: @ %else2
; CHECK-BE-NEXT:    lsls r2, r1, #30
; CHECK-BE-NEXT:    bmi .LBB25_7
; CHECK-BE-NEXT:  .LBB25_3: @ %else4
; CHECK-BE-NEXT:    lsls r1, r1, #31
; CHECK-BE-NEXT:    bne .LBB25_8
; CHECK-BE-NEXT:  .LBB25_4: @ %else6
; CHECK-BE-NEXT:    add sp, #4
; CHECK-BE-NEXT:    bx lr
; CHECK-BE-NEXT:  .LBB25_5: @ %cond.store
; CHECK-BE-NEXT:    vstr.16 s0, [r0]
; CHECK-BE-NEXT:    lsls r2, r1, #29
; CHECK-BE-NEXT:    bpl .LBB25_2
; CHECK-BE-NEXT:  .LBB25_6: @ %cond.store1
; CHECK-BE-NEXT:    vmovx.f16 s0, s0
; CHECK-BE-NEXT:    vstr.16 s0, [r0, #2]
; CHECK-BE-NEXT:    lsls r2, r1, #30
; CHECK-BE-NEXT:    bpl .LBB25_3
; CHECK-BE-NEXT:  .LBB25_7: @ %cond.store3
; CHECK-BE-NEXT:    vstr.16 s1, [r0, #4]
; CHECK-BE-NEXT:    lsls r1, r1, #31
; CHECK-BE-NEXT:    beq .LBB25_4
; CHECK-BE-NEXT:  .LBB25_8: @ %cond.store5
; CHECK-BE-NEXT:    vmovx.f16 s0, s1
; CHECK-BE-NEXT:    vstr.16 s0, [r0, #6]
; CHECK-BE-NEXT:    add sp, #4
; CHECK-BE-NEXT:    bx lr
entry:
  %c = fcmp ogt <4 x float> %a, zeroinitializer
  %trunc = fptrunc <4 x float> %a to <4 x half>
  call void @llvm.masked.store.v4f16.p0(<4 x half> %trunc, ptr %dest, i32 4, <4 x i1> %c)
  ret void
}

define arm_aapcs_vfpcc void @masked_v4f16_align2(ptr %dest, <4 x float> %a) {
; CHECK-LE-LABEL: masked_v4f16_align2:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    .pad #4
; CHECK-LE-NEXT:    sub sp, #4
; CHECK-LE-NEXT:    vcmp.f32 s0, #0
; CHECK-LE-NEXT:    movs r1, #0
; CHECK-LE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-LE-NEXT:    vcmp.f32 s1, #0
; CHECK-LE-NEXT:    vcvtb.f16.f32 s4, s0
; CHECK-LE-NEXT:    vcvtb.f16.f32 s5, s2
; CHECK-LE-NEXT:    vcvtt.f16.f32 s4, s1
; CHECK-LE-NEXT:    vcvtt.f16.f32 s5, s3
; CHECK-LE-NEXT:    csetm r2, gt
; CHECK-LE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-LE-NEXT:    vcmp.f32 s2, #0
; CHECK-LE-NEXT:    bfi r1, r2, #0, #1
; CHECK-LE-NEXT:    csetm r2, gt
; CHECK-LE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-LE-NEXT:    vcmp.f32 s3, #0
; CHECK-LE-NEXT:    bfi r1, r2, #1, #1
; CHECK-LE-NEXT:    csetm r2, gt
; CHECK-LE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-LE-NEXT:    bfi r1, r2, #2, #1
; CHECK-LE-NEXT:    csetm r2, gt
; CHECK-LE-NEXT:    bfi r1, r2, #3, #1
; CHECK-LE-NEXT:    lsls r2, r1, #31
; CHECK-LE-NEXT:    bne .LBB26_5
; CHECK-LE-NEXT:  @ %bb.1: @ %else
; CHECK-LE-NEXT:    lsls r2, r1, #30
; CHECK-LE-NEXT:    bmi .LBB26_6
; CHECK-LE-NEXT:  .LBB26_2: @ %else2
; CHECK-LE-NEXT:    lsls r2, r1, #29
; CHECK-LE-NEXT:    bmi .LBB26_7
; CHECK-LE-NEXT:  .LBB26_3: @ %else4
; CHECK-LE-NEXT:    lsls r1, r1, #28
; CHECK-LE-NEXT:    bmi .LBB26_8
; CHECK-LE-NEXT:  .LBB26_4: @ %else6
; CHECK-LE-NEXT:    add sp, #4
; CHECK-LE-NEXT:    bx lr
; CHECK-LE-NEXT:  .LBB26_5: @ %cond.store
; CHECK-LE-NEXT:    vstr.16 s4, [r0]
; CHECK-LE-NEXT:    lsls r2, r1, #30
; CHECK-LE-NEXT:    bpl .LBB26_2
; CHECK-LE-NEXT:  .LBB26_6: @ %cond.store1
; CHECK-LE-NEXT:    vmovx.f16 s0, s4
; CHECK-LE-NEXT:    vstr.16 s0, [r0, #2]
; CHECK-LE-NEXT:    lsls r2, r1, #29
; CHECK-LE-NEXT:    bpl .LBB26_3
; CHECK-LE-NEXT:  .LBB26_7: @ %cond.store3
; CHECK-LE-NEXT:    vstr.16 s5, [r0, #4]
; CHECK-LE-NEXT:    lsls r1, r1, #28
; CHECK-LE-NEXT:    bpl .LBB26_4
; CHECK-LE-NEXT:  .LBB26_8: @ %cond.store5
; CHECK-LE-NEXT:    vmovx.f16 s0, s5
; CHECK-LE-NEXT:    vstr.16 s0, [r0, #6]
; CHECK-LE-NEXT:    add sp, #4
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v4f16_align2:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    .pad #4
; CHECK-BE-NEXT:    sub sp, #4
; CHECK-BE-NEXT:    vrev64.32 q1, q0
; CHECK-BE-NEXT:    movs r1, #0
; CHECK-BE-NEXT:    vcmp.f32 s7, #0
; CHECK-BE-NEXT:    vcvtb.f16.f32 s0, s4
; CHECK-BE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-BE-NEXT:    vcmp.f32 s6, #0
; CHECK-BE-NEXT:    vcvtb.f16.f32 s1, s6
; CHECK-BE-NEXT:    vcvtt.f16.f32 s0, s5
; CHECK-BE-NEXT:    vcvtt.f16.f32 s1, s7
; CHECK-BE-NEXT:    csetm r2, gt
; CHECK-BE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-BE-NEXT:    vcmp.f32 s5, #0
; CHECK-BE-NEXT:    bfi r1, r2, #0, #1
; CHECK-BE-NEXT:    csetm r2, gt
; CHECK-BE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-BE-NEXT:    vcmp.f32 s4, #0
; CHECK-BE-NEXT:    bfi r1, r2, #1, #1
; CHECK-BE-NEXT:    csetm r2, gt
; CHECK-BE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-BE-NEXT:    bfi r1, r2, #2, #1
; CHECK-BE-NEXT:    csetm r2, gt
; CHECK-BE-NEXT:    bfi r1, r2, #3, #1
; CHECK-BE-NEXT:    lsls r2, r1, #28
; CHECK-BE-NEXT:    bmi .LBB26_5
; CHECK-BE-NEXT:  @ %bb.1: @ %else
; CHECK-BE-NEXT:    lsls r2, r1, #29
; CHECK-BE-NEXT:    bmi .LBB26_6
; CHECK-BE-NEXT:  .LBB26_2: @ %else2
; CHECK-BE-NEXT:    lsls r2, r1, #30
; CHECK-BE-NEXT:    bmi .LBB26_7
; CHECK-BE-NEXT:  .LBB26_3: @ %else4
; CHECK-BE-NEXT:    lsls r1, r1, #31
; CHECK-BE-NEXT:    bne .LBB26_8
; CHECK-BE-NEXT:  .LBB26_4: @ %else6
; CHECK-BE-NEXT:    add sp, #4
; CHECK-BE-NEXT:    bx lr
; CHECK-BE-NEXT:  .LBB26_5: @ %cond.store
; CHECK-BE-NEXT:    vstr.16 s0, [r0]
; CHECK-BE-NEXT:    lsls r2, r1, #29
; CHECK-BE-NEXT:    bpl .LBB26_2
; CHECK-BE-NEXT:  .LBB26_6: @ %cond.store1
; CHECK-BE-NEXT:    vmovx.f16 s0, s0
; CHECK-BE-NEXT:    vstr.16 s0, [r0, #2]
; CHECK-BE-NEXT:    lsls r2, r1, #30
; CHECK-BE-NEXT:    bpl .LBB26_3
; CHECK-BE-NEXT:  .LBB26_7: @ %cond.store3
; CHECK-BE-NEXT:    vstr.16 s1, [r0, #4]
; CHECK-BE-NEXT:    lsls r1, r1, #31
; CHECK-BE-NEXT:    beq .LBB26_4
; CHECK-BE-NEXT:  .LBB26_8: @ %cond.store5
; CHECK-BE-NEXT:    vmovx.f16 s0, s1
; CHECK-BE-NEXT:    vstr.16 s0, [r0, #6]
; CHECK-BE-NEXT:    add sp, #4
; CHECK-BE-NEXT:    bx lr
entry:
  %c = fcmp ogt <4 x float> %a, zeroinitializer
  %trunc = fptrunc <4 x float> %a to <4 x half>
  call void @llvm.masked.store.v4f16.p0(<4 x half> %trunc, ptr %dest, i32 2, <4 x i1> %c)
  ret void
}

define arm_aapcs_vfpcc void @masked_v4f16_align1(ptr %dest, <4 x float> %a) {
; CHECK-LE-LABEL: masked_v4f16_align1:
; CHECK-LE:       @ %bb.0: @ %entry
; CHECK-LE-NEXT:    .pad #20
; CHECK-LE-NEXT:    sub sp, #20
; CHECK-LE-NEXT:    vcmp.f32 s0, #0
; CHECK-LE-NEXT:    movs r1, #0
; CHECK-LE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-LE-NEXT:    vcmp.f32 s1, #0
; CHECK-LE-NEXT:    vcvtb.f16.f32 s4, s0
; CHECK-LE-NEXT:    vcvtb.f16.f32 s5, s2
; CHECK-LE-NEXT:    vcvtt.f16.f32 s4, s1
; CHECK-LE-NEXT:    vcvtt.f16.f32 s5, s3
; CHECK-LE-NEXT:    csetm r2, gt
; CHECK-LE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-LE-NEXT:    vcmp.f32 s2, #0
; CHECK-LE-NEXT:    bfi r1, r2, #0, #1
; CHECK-LE-NEXT:    csetm r2, gt
; CHECK-LE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-LE-NEXT:    vcmp.f32 s3, #0
; CHECK-LE-NEXT:    bfi r1, r2, #1, #1
; CHECK-LE-NEXT:    csetm r2, gt
; CHECK-LE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-LE-NEXT:    bfi r1, r2, #2, #1
; CHECK-LE-NEXT:    csetm r2, gt
; CHECK-LE-NEXT:    bfi r1, r2, #3, #1
; CHECK-LE-NEXT:    lsls r2, r1, #31
; CHECK-LE-NEXT:    bne .LBB27_5
; CHECK-LE-NEXT:  @ %bb.1: @ %else
; CHECK-LE-NEXT:    lsls r2, r1, #30
; CHECK-LE-NEXT:    bmi .LBB27_6
; CHECK-LE-NEXT:  .LBB27_2: @ %else2
; CHECK-LE-NEXT:    lsls r2, r1, #29
; CHECK-LE-NEXT:    bmi .LBB27_7
; CHECK-LE-NEXT:  .LBB27_3: @ %else4
; CHECK-LE-NEXT:    lsls r1, r1, #28
; CHECK-LE-NEXT:    bmi .LBB27_8
; CHECK-LE-NEXT:  .LBB27_4: @ %else6
; CHECK-LE-NEXT:    add sp, #20
; CHECK-LE-NEXT:    bx lr
; CHECK-LE-NEXT:  .LBB27_5: @ %cond.store
; CHECK-LE-NEXT:    vstr.16 s4, [sp, #12]
; CHECK-LE-NEXT:    ldrh.w r2, [sp, #12]
; CHECK-LE-NEXT:    strh r2, [r0]
; CHECK-LE-NEXT:    lsls r2, r1, #30
; CHECK-LE-NEXT:    bpl .LBB27_2
; CHECK-LE-NEXT:  .LBB27_6: @ %cond.store1
; CHECK-LE-NEXT:    vmovx.f16 s0, s4
; CHECK-LE-NEXT:    vstr.16 s0, [sp, #8]
; CHECK-LE-NEXT:    ldrh.w r2, [sp, #8]
; CHECK-LE-NEXT:    strh r2, [r0, #2]
; CHECK-LE-NEXT:    lsls r2, r1, #29
; CHECK-LE-NEXT:    bpl .LBB27_3
; CHECK-LE-NEXT:  .LBB27_7: @ %cond.store3
; CHECK-LE-NEXT:    vstr.16 s5, [sp, #4]
; CHECK-LE-NEXT:    ldrh.w r2, [sp, #4]
; CHECK-LE-NEXT:    strh r2, [r0, #4]
; CHECK-LE-NEXT:    lsls r1, r1, #28
; CHECK-LE-NEXT:    bpl .LBB27_4
; CHECK-LE-NEXT:  .LBB27_8: @ %cond.store5
; CHECK-LE-NEXT:    vmovx.f16 s0, s5
; CHECK-LE-NEXT:    vstr.16 s0, [sp]
; CHECK-LE-NEXT:    ldrh.w r1, [sp]
; CHECK-LE-NEXT:    strh r1, [r0, #6]
; CHECK-LE-NEXT:    add sp, #20
; CHECK-LE-NEXT:    bx lr
;
; CHECK-BE-LABEL: masked_v4f16_align1:
; CHECK-BE:       @ %bb.0: @ %entry
; CHECK-BE-NEXT:    .pad #20
; CHECK-BE-NEXT:    sub sp, #20
; CHECK-BE-NEXT:    vrev64.32 q1, q0
; CHECK-BE-NEXT:    movs r1, #0
; CHECK-BE-NEXT:    vcmp.f32 s7, #0
; CHECK-BE-NEXT:    vcvtb.f16.f32 s0, s4
; CHECK-BE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-BE-NEXT:    vcmp.f32 s6, #0
; CHECK-BE-NEXT:    vcvtb.f16.f32 s1, s6
; CHECK-BE-NEXT:    vcvtt.f16.f32 s0, s5
; CHECK-BE-NEXT:    vcvtt.f16.f32 s1, s7
; CHECK-BE-NEXT:    csetm r2, gt
; CHECK-BE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-BE-NEXT:    vcmp.f32 s5, #0
; CHECK-BE-NEXT:    bfi r1, r2, #0, #1
; CHECK-BE-NEXT:    csetm r2, gt
; CHECK-BE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-BE-NEXT:    vcmp.f32 s4, #0
; CHECK-BE-NEXT:    bfi r1, r2, #1, #1
; CHECK-BE-NEXT:    csetm r2, gt
; CHECK-BE-NEXT:    vmrs APSR_nzcv, fpscr
; CHECK-BE-NEXT:    bfi r1, r2, #2, #1
; CHECK-BE-NEXT:    csetm r2, gt
; CHECK-BE-NEXT:    bfi r1, r2, #3, #1
; CHECK-BE-NEXT:    lsls r2, r1, #28
; CHECK-BE-NEXT:    bmi .LBB27_5
; CHECK-BE-NEXT:  @ %bb.1: @ %else
; CHECK-BE-NEXT:    lsls r2, r1, #29
; CHECK-BE-NEXT:    bmi .LBB27_6
; CHECK-BE-NEXT:  .LBB27_2: @ %else2
; CHECK-BE-NEXT:    lsls r2, r1, #30
; CHECK-BE-NEXT:    bmi .LBB27_7
; CHECK-BE-NEXT:  .LBB27_3: @ %else4
; CHECK-BE-NEXT:    lsls r1, r1, #31
; CHECK-BE-NEXT:    bne .LBB27_8
; CHECK-BE-NEXT:  .LBB27_4: @ %else6
; CHECK-BE-NEXT:    add sp, #20
; CHECK-BE-NEXT:    bx lr
; CHECK-BE-NEXT:  .LBB27_5: @ %cond.store
; CHECK-BE-NEXT:    vstr.16 s0, [sp, #12]
; CHECK-BE-NEXT:    ldrh.w r2, [sp, #12]
; CHECK-BE-NEXT:    strh r2, [r0]
; CHECK-BE-NEXT:    lsls r2, r1, #29
; CHECK-BE-NEXT:    bpl .LBB27_2
; CHECK-BE-NEXT:  .LBB27_6: @ %cond.store1
; CHECK-BE-NEXT:    vmovx.f16 s0, s0
; CHECK-BE-NEXT:    vstr.16 s0, [sp, #8]
; CHECK-BE-NEXT:    ldrh.w r2, [sp, #8]
; CHECK-BE-NEXT:    strh r2, [r0, #2]
; CHECK-BE-NEXT:    lsls r2, r1, #30
; CHECK-BE-NEXT:    bpl .LBB27_3
; CHECK-BE-NEXT:  .LBB27_7: @ %cond.store3
; CHECK-BE-NEXT:    vstr.16 s1, [sp, #4]
; CHECK-BE-NEXT:    ldrh.w r2, [sp, #4]
; CHECK-BE-NEXT:    strh r2, [r0, #4]
; CHECK-BE-NEXT:    lsls r1, r1, #31
; CHECK-BE-NEXT:    beq .LBB27_4
; CHECK-BE-NEXT:  .LBB27_8: @ %cond.store5
; CHECK-BE-NEXT:    vmovx.f16 s0, s1
; CHECK-BE-NEXT:    vstr.16 s0, [sp]
; CHECK-BE-NEXT:    ldrh.w r1, [sp]
; CHECK-BE-NEXT:    strh r1, [r0, #6]
; CHECK-BE-NEXT:    add sp, #20
; CHECK-BE-NEXT:    bx lr
entry:
  %c = fcmp ogt <4 x float> %a, zeroinitializer
  %trunc = fptrunc <4 x float> %a to <4 x half>
  call void @llvm.masked.store.v4f16.p0(<4 x half> %trunc, ptr %dest, i32 1, <4 x i1> %c)
  ret void
}

declare void @llvm.masked.store.v4i8.p0(<4 x i8>, ptr, i32, <4 x i1>)
declare void @llvm.masked.store.v8i8.p0(<8 x i8>, ptr, i32, <8 x i1>)
declare void @llvm.masked.store.v4i16.p0(<4 x i16>, ptr, i32, <4 x i1>)
declare void @llvm.masked.store.v4i32.p0(<4 x i32>, ptr, i32, <4 x i1>)
declare void @llvm.masked.store.v8i16.p0(<8 x i16>, ptr, i32, <8 x i1>)
declare void @llvm.masked.store.v16i8.p0(<16 x i8>, ptr, i32, <16 x i1>)
declare void @llvm.masked.store.v4f32.p0(<4 x float>, ptr, i32, <4 x i1>)
declare void @llvm.masked.store.v4f16.p0(<4 x half>, ptr, i32, <4 x i1>)
declare void @llvm.masked.store.v8f16.p0(<8 x half>, ptr, i32, <8 x i1>)
declare void @llvm.masked.store.v2i64.p0(<2 x i64>, ptr, i32, <2 x i1>)
declare void @llvm.masked.store.v2f64.p0(<2 x double>, ptr, i32, <2 x i1>)