; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=thumbv8.1m.main-none-none-eabi -mattr=+mve.fp -enable-arm-maskedldst %s -o - | FileCheck %s

define arm_aapcs_vfpcc <4 x i32> @gather_inc_mini_4i32(i32* noalias nocapture readonly %data, i32* noalias nocapture %dst, <4 x i32> %offs) {
; CHECK-LABEL: gather_inc_mini_4i32:
; CHECK:       @ %bb.0:
; CHECK-NEXT:    vmov.i32 q1, #0x4
; CHECK-NEXT:    vadd.i32 q1, q0, q1
; CHECK-NEXT:    vldrw.u32 q0, [r0, q1, uxtw #2]
; CHECK-NEXT:    bx lr
  %1 = add <4 x i32> %offs, <i32 4, i32 4, i32 4, i32 4>
  %2 = getelementptr inbounds i32, i32* %data, <4 x i32> %1
  %wide.masked.gather = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> %2, i32 4, <4 x i1> <i1 true, i1 true, i1 true, i1 true>, <4 x i32> undef)
  ret <4 x i32> %wide.masked.gather
}

define arm_aapcs_vfpcc <4 x i32> @gather_inc_minipred_4i32(i32* noalias nocapture readonly %data, i32* noalias nocapture %dst, <4 x i32> %offs) {
; CHECK-LABEL: gather_inc_minipred_4i32:
; CHECK:       @ %bb.0:
; CHECK-NEXT:    vmov.i32 q1, #0x4
; CHECK-NEXT:    movw r1, #3855
; CHECK-NEXT:    vadd.i32 q1, q0, q1
; CHECK-NEXT:    vmsr p0, r1
; CHECK-NEXT:    vpst
; CHECK-NEXT:    vldrwt.u32 q0, [r0, q1, uxtw #2]
; CHECK-NEXT:    bx lr
  %1 = add <4 x i32> %offs, <i32 4, i32 4, i32 4, i32 4>
  %2 = getelementptr inbounds i32, i32* %data, <4 x i32> %1
  %wide.masked.gather = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> %2, i32 4, <4 x i1> <i1 true, i1 false, i1 true, i1 false>, <4 x i32> undef)
  ret <4 x i32> %wide.masked.gather
}

define arm_aapcs_vfpcc <8 x i16> @gather_inc_mini_8i16(i16* noalias nocapture readonly %data, i16* noalias nocapture %dst, <8 x i32> %offs) {
; CHECK-LABEL: gather_inc_mini_8i16:
; CHECK:       @ %bb.0:
; CHECK-NEXT:    .save {r4, r5, r7, lr}
; CHECK-NEXT:    push {r4, r5, r7, lr}
; CHECK-NEXT:    vshl.i32 q0, q0, #1
; CHECK-NEXT:    vmov.i32 q2, #0x10
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:    vshl.i32 q1, q1, #1
; CHECK-NEXT:    vadd.i32 q0, q0, q2
; CHECK-NEXT:    vadd.i32 q1, q1, r0
; CHECK-NEXT:    vmov r4, r5, d0
; CHECK-NEXT:    vadd.i32 q1, q1, q2
; CHECK-NEXT:    vmov r0, r2, d1
; CHECK-NEXT:    vmov r1, lr, d2
; CHECK-NEXT:    vmov r3, r12, d3
; CHECK-NEXT:    ldrh r4, [r4]
; CHECK-NEXT:    ldrh r5, [r5]
; CHECK-NEXT:    vmov.16 q0[0], r4
; CHECK-NEXT:    ldrh r0, [r0]
; CHECK-NEXT:    vmov.16 q0[1], r5
; CHECK-NEXT:    ldrh r2, [r2]
; CHECK-NEXT:    vmov.16 q0[2], r0
; CHECK-NEXT:    ldrh r1, [r1]
; CHECK-NEXT:    vmov.16 q0[3], r2
; CHECK-NEXT:    ldrh.w lr, [lr]
; CHECK-NEXT:    vmov.16 q0[4], r1
; CHECK-NEXT:    ldrh r3, [r3]
; CHECK-NEXT:    vmov.16 q0[5], lr
; CHECK-NEXT:    ldrh.w r12, [r12]
; CHECK-NEXT:    vmov.16 q0[6], r3
; CHECK-NEXT:    vmov.16 q0[7], r12
; CHECK-NEXT:    pop {r4, r5, r7, pc}
  %1 = add <8 x i32> %offs, <i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8>
  %2 = getelementptr inbounds i16, i16* %data, <8 x i32> %1
  %wide.masked.gather = call <8 x i16> @llvm.masked.gather.v8i16.v8p0i16(<8 x i16*> %2, i32 4, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <8 x i16> undef)
  ret <8 x i16> %wide.masked.gather
}

define arm_aapcs_vfpcc <8 x i16> @gather_inc_minipred_8i16(i16* noalias nocapture readonly %data, i16* noalias nocapture %dst, <8 x i32> %offs) {
; CHECK-LABEL: gather_inc_minipred_8i16:
; CHECK:       @ %bb.0:
; CHECK-NEXT:    vshl.i32 q0, q0, #1
; CHECK-NEXT:    vmov.i32 q2, #0x10
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:    vshl.i32 q1, q1, #1
; CHECK-NEXT:    vadd.i32 q0, q0, q2
; CHECK-NEXT:    vadd.i32 q1, q1, r0
; CHECK-NEXT:    vmov r1, s0
; CHECK-NEXT:    vadd.i32 q1, q1, q2
; CHECK-NEXT:    vmov r3, s2
; CHECK-NEXT:    vmov r0, s4
; CHECK-NEXT:    vmov r2, s6
; CHECK-NEXT:    ldrh r1, [r1]
; CHECK-NEXT:    ldrh r3, [r3]
; CHECK-NEXT:    vmov.16 q0[0], r1
; CHECK-NEXT:    ldrh r0, [r0]
; CHECK-NEXT:    vmov.16 q0[2], r3
; CHECK-NEXT:    ldrh r2, [r2]
; CHECK-NEXT:    vmov.16 q0[4], r0
; CHECK-NEXT:    vmov.16 q0[6], r2
; CHECK-NEXT:    bx lr
  %1 = add <8 x i32> %offs, <i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8>
  %2 = getelementptr inbounds i16, i16* %data, <8 x i32> %1
  %wide.masked.gather = call <8 x i16> @llvm.masked.gather.v8i16.v8p0i16(<8 x i16*> %2, i32 4, <8 x i1> <i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false>, <8 x i16> undef)
  ret <8 x i16> %wide.masked.gather
}

define arm_aapcs_vfpcc <16 x i8> @gather_inc_mini_16i8(i8* noalias nocapture readonly %data, i8* noalias nocapture %dst, <16 x i32> %offs) {
; CHECK-LABEL: gather_inc_mini_16i8:
; CHECK:       @ %bb.0:
; CHECK-NEXT:    .save {r4, r5, r6, lr}
; CHECK-NEXT:    push {r4, r5, r6, lr}
; CHECK-NEXT:    .vsave {d8, d9}
; CHECK-NEXT:    vpush {d8, d9}
; CHECK-NEXT:    vmov.i32 q4, #0x10
; CHECK-NEXT:    vadd.i32 q3, q3, r0
; CHECK-NEXT:    vadd.i32 q2, q2, r0
; CHECK-NEXT:    vadd.i32 q3, q3, q4
; CHECK-NEXT:    vadd.i32 q2, q2, q4
; CHECK-NEXT:    vmov r1, r2, d7
; CHECK-NEXT:    vmov r3, r4, d6
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:    vmov r5, r6, d5
; CHECK-NEXT:    vadd.i32 q3, q0, q4
; CHECK-NEXT:    vadd.i32 q1, q1, r0
; CHECK-NEXT:    vadd.i32 q1, q1, q4
; CHECK-NEXT:    ldrb.w lr, [r1]
; CHECK-NEXT:    ldrb.w r12, [r2]
; CHECK-NEXT:    ldrb r1, [r4]
; CHECK-NEXT:    ldrb r4, [r5]
; CHECK-NEXT:    ldrb r2, [r6]
; CHECK-NEXT:    vmov r5, r6, d6
; CHECK-NEXT:    ldrb r3, [r3]
; CHECK-NEXT:    ldrb r5, [r5]
; CHECK-NEXT:    vmov.8 q0[0], r5
; CHECK-NEXT:    ldrb r5, [r6]
; CHECK-NEXT:    vmov.8 q0[1], r5
; CHECK-NEXT:    vmov r5, r6, d7
; CHECK-NEXT:    ldrb r0, [r5]
; CHECK-NEXT:    ldrb r6, [r6]
; CHECK-NEXT:    vmov.8 q0[2], r0
; CHECK-NEXT:    vmov r0, r5, d2
; CHECK-NEXT:    vmov.8 q0[3], r6
; CHECK-NEXT:    ldrb r0, [r0]
; CHECK-NEXT:    ldrb r5, [r5]
; CHECK-NEXT:    vmov.8 q0[4], r0
; CHECK-NEXT:    vmov.8 q0[5], r5
; CHECK-NEXT:    vmov r0, r5, d3
; CHECK-NEXT:    ldrb r0, [r0]
; CHECK-NEXT:    ldrb r5, [r5]
; CHECK-NEXT:    vmov.8 q0[6], r0
; CHECK-NEXT:    vmov.8 q0[7], r5
; CHECK-NEXT:    vmov r0, r5, d4
; CHECK-NEXT:    ldrb r0, [r0]
; CHECK-NEXT:    ldrb r5, [r5]
; CHECK-NEXT:    vmov.8 q0[8], r0
; CHECK-NEXT:    vmov.8 q0[9], r5
; CHECK-NEXT:    vmov.8 q0[10], r4
; CHECK-NEXT:    vmov.8 q0[11], r2
; CHECK-NEXT:    vmov.8 q0[12], r3
; CHECK-NEXT:    vmov.8 q0[13], r1
; CHECK-NEXT:    vmov.8 q0[14], lr
; CHECK-NEXT:    vmov.8 q0[15], r12
; CHECK-NEXT:    vpop {d8, d9}
; CHECK-NEXT:    pop {r4, r5, r6, pc}
  %1 = add <16 x i32> %offs, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
  %2 = getelementptr inbounds i8, i8* %data, <16 x i32> %1
  %wide.masked.gather = call <16 x i8> @llvm.masked.gather.v16i8.v16p0i8(<16 x i8*> %2, i32 2, <16 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <16 x i8> undef)
  ret <16 x i8> %wide.masked.gather
}

define arm_aapcs_vfpcc <16 x i8> @gather_inc_minipred_16i8(i8* noalias nocapture readonly %data, i8* noalias nocapture %dst, <16 x i32> %offs) {
; CHECK-LABEL: gather_inc_minipred_16i8:
; CHECK:       @ %bb.0:
; CHECK-NEXT:    .save {r4, r5, r7, lr}
; CHECK-NEXT:    push {r4, r5, r7, lr}
; CHECK-NEXT:    .vsave {d8, d9}
; CHECK-NEXT:    vpush {d8, d9}
; CHECK-NEXT:    vmov.i32 q4, #0x10
; CHECK-NEXT:    vadd.i32 q2, q2, r0
; CHECK-NEXT:    vadd.i32 q2, q2, q4
; CHECK-NEXT:    vadd.i32 q1, q1, r0
; CHECK-NEXT:    vmov r2, s8
; CHECK-NEXT:    vadd.i32 q1, q1, q4
; CHECK-NEXT:    vmov r1, s4
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:    vadd.i32 q0, q0, q4
; CHECK-NEXT:    vmov r3, s10
; CHECK-NEXT:    vmov r5, s2
; CHECK-NEXT:    ldrb.w lr, [r2]
; CHECK-NEXT:    vmov r2, s0
; CHECK-NEXT:    ldrb.w r12, [r1]
; CHECK-NEXT:    vmov r1, s6
; CHECK-NEXT:    vadd.i32 q1, q3, r0
; CHECK-NEXT:    ldrb r3, [r3]
; CHECK-NEXT:    vadd.i32 q1, q1, q4
; CHECK-NEXT:    ldrb r5, [r5]
; CHECK-NEXT:    vmov r0, s4
; CHECK-NEXT:    vmov r4, s6
; CHECK-NEXT:    ldrb r2, [r2]
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    vmov.8 q0[0], r2
; CHECK-NEXT:    vmov.8 q0[2], r5
; CHECK-NEXT:    vmov.8 q0[4], r12
; CHECK-NEXT:    ldrb r0, [r0]
; CHECK-NEXT:    vmov.8 q0[6], r1
; CHECK-NEXT:    ldrb r4, [r4]
; CHECK-NEXT:    vmov.8 q0[8], lr
; CHECK-NEXT:    vmov.8 q0[10], r3
; CHECK-NEXT:    vmov.8 q0[12], r0
; CHECK-NEXT:    vmov.8 q0[14], r4
; CHECK-NEXT:    vpop {d8, d9}
; CHECK-NEXT:    pop {r4, r5, r7, pc}
  %1 = add <16 x i32> %offs, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
  %2 = getelementptr inbounds i8, i8* %data, <16 x i32> %1
  %wide.masked.gather = call <16 x i8> @llvm.masked.gather.v16i8.v16p0i8(<16 x i8*> %2, i32 2, <16 x i1> <i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false>, <16 x i8> undef)
  ret <16 x i8> %wide.masked.gather
}

define arm_aapcs_vfpcc void @gather_pre_inc(i32* noalias nocapture readonly %data, i32* noalias nocapture %dst, i32 %n.vec) {
; CHECK-LABEL: gather_pre_inc:
; CHECK:       @ %bb.0: @ %vector.ph
; CHECK-NEXT:    adr r3, .LCPI6_0
; CHECK-NEXT:    vldrw.u32 q0, [r3]
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:  .LBB6_1: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q1, [q0, #96]!
; CHECK-NEXT:    subs r2, #4
; CHECK-NEXT:    vstrb.8 q1, [r1], #16
; CHECK-NEXT:    bne .LBB6_1
; CHECK-NEXT:  @ %bb.2: @ %end
; CHECK-NEXT:    bx lr
; CHECK-NEXT:    .p2align 4
; CHECK-NEXT:  @ %bb.3:
; CHECK-NEXT:  .LCPI6_0:
; CHECK-NEXT:    .long 4294967224 @ 0xffffffb8
; CHECK-NEXT:    .long 4294967248 @ 0xffffffd0
; CHECK-NEXT:    .long 4294967272 @ 0xffffffe8
; CHECK-NEXT:    .long 0 @ 0x0
vector.ph:                                        ; preds = %for.body.preheader
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %vec.ind = phi <4 x i32> [ <i32 0, i32 2, i32 4, i32 6>, %vector.ph ], [ %vec.ind.next, %vector.body ]
  %0 = mul <4 x i32> %vec.ind, <i32 3, i32 3, i32 3, i32 3>
  %1 = add <4 x i32> %0, <i32 6, i32 6, i32 6, i32 6>
  %2 = getelementptr inbounds i32, i32* %data, <4 x i32> %1
  %wide.masked.gather = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> %2, i32 4, <4 x i1> <i1 true, i1 true, i1 true, i1 true>, <4 x i32> undef)
  %3 = getelementptr inbounds i32, i32* %dst, i32 %index
  %4 = bitcast i32* %3 to <4 x i32>*
  store <4 x i32> %wide.masked.gather, <4 x i32>* %4, align 4
  %index.next = add i32 %index, 4
  %vec.ind.next = add <4 x i32> %vec.ind, <i32 8, i32 8, i32 8, i32 8>
  %5 = icmp eq i32 %index.next, %n.vec
  br i1 %5, label %end, label %vector.body

end:
  ret void;
}

define arm_aapcs_vfpcc void @gather_post_inc(i32* noalias nocapture readonly %data, i32* noalias nocapture %dst, i32 %n.vec43) {
; CHECK-LABEL: gather_post_inc:
; CHECK:       @ %bb.0: @ %vector.ph41
; CHECK-NEXT:    adr r3, .LCPI7_0
; CHECK-NEXT:    vldrw.u32 q0, [r3]
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:  .LBB7_1: @ %vector.body39
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q1, [q0, #96]!
; CHECK-NEXT:    subs r2, #4
; CHECK-NEXT:    vstrb.8 q1, [r1], #16
; CHECK-NEXT:    bne .LBB7_1
; CHECK-NEXT:  @ %bb.2: @ %end
; CHECK-NEXT:    bx lr
; CHECK-NEXT:    .p2align 4
; CHECK-NEXT:  @ %bb.3:
; CHECK-NEXT:  .LCPI7_0:
; CHECK-NEXT:    .long 4294967200 @ 0xffffffa0
; CHECK-NEXT:    .long 4294967224 @ 0xffffffb8
; CHECK-NEXT:    .long 4294967248 @ 0xffffffd0
; CHECK-NEXT:    .long 4294967272 @ 0xffffffe8
vector.ph41:                                      ; preds = %for.body6.preheader
  br label %vector.body39

vector.body39:                                    ; preds = %vector.body39, %vector.ph41
  %index44 = phi i32 [ 0, %vector.ph41 ], [ %index.next45, %vector.body39 ]
  %vec.ind50 = phi <4 x i32> [ <i32 0, i32 2, i32 4, i32 6>, %vector.ph41 ], [ %vec.ind.next51, %vector.body39 ]
  %0 = mul nuw nsw <4 x i32> %vec.ind50, <i32 3, i32 3, i32 3, i32 3>
  %1 = getelementptr inbounds i32, i32* %data, <4 x i32> %0
  %wide.masked.gather55 = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> %1, i32 4, <4 x i1> <i1 true, i1 true, i1 true, i1 true>, <4 x i32> undef)
  %2 = getelementptr inbounds i32, i32* %dst, i32 %index44
  %3 = bitcast i32* %2 to <4 x i32>*
  store <4 x i32> %wide.masked.gather55, <4 x i32>* %3, align 4
  %index.next45 = add i32 %index44, 4
  %vec.ind.next51 = add <4 x i32> %vec.ind50, <i32 8, i32 8, i32 8, i32 8>
  %4 = icmp eq i32 %index.next45, %n.vec43
  br i1 %4, label %end, label %vector.body39

end:
  ret void;
}

define arm_aapcs_vfpcc void @gather_inc_v4i32_simple(i32* noalias nocapture readonly %data, i32* noalias nocapture %dst, i32 %n) {
; CHECK-LABEL: gather_inc_v4i32_simple:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, lr}
; CHECK-NEXT:    push {r4, lr}
; CHECK-NEXT:    cmp r2, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    poplt {r4, pc}
; CHECK-NEXT:  .LBB8_1: @ %vector.ph.preheader
; CHECK-NEXT:    bic r12, r2, #3
; CHECK-NEXT:    movs r3, #1
; CHECK-NEXT:    sub.w lr, r12, #4
; CHECK-NEXT:    add.w r4, r3, lr, lsr #2
; CHECK-NEXT:    adr r3, .LCPI8_0
; CHECK-NEXT:    vldrw.u32 q0, [r3]
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:  .LBB8_2: @ %vector.ph
; CHECK-NEXT:    @ =>This Loop Header: Depth=1
; CHECK-NEXT:    @ Child Loop BB8_3 Depth 2
; CHECK-NEXT:    dls lr, r4
; CHECK-NEXT:    mov r0, r1
; CHECK-NEXT:    vmov q1, q0
; CHECK-NEXT:  .LBB8_3: @ %vector.body
; CHECK-NEXT:    @ Parent Loop BB8_2 Depth=1
; CHECK-NEXT:    @ => This Inner Loop Header: Depth=2
; CHECK-NEXT:    vldrw.u32 q2, [q1, #16]!
; CHECK-NEXT:    vstrb.8 q2, [r0], #16
; CHECK-NEXT:    le lr, .LBB8_3
; CHECK-NEXT:  @ %bb.4: @ %middle.block
; CHECK-NEXT:    @ in Loop: Header=BB8_2 Depth=1
; CHECK-NEXT:    cmp r12, r2
; CHECK-NEXT:    bne .LBB8_2
; CHECK-NEXT:  @ %bb.5: @ %for.cond.cleanup
; CHECK-NEXT:    pop {r4, pc}
; CHECK-NEXT:    .p2align 4
; CHECK-NEXT:  @ %bb.6:
; CHECK-NEXT:  .LCPI8_0:
; CHECK-NEXT:    .long 4294967280 @ 0xfffffff0
; CHECK-NEXT:    .long 4294967284 @ 0xfffffff4
; CHECK-NEXT:    .long 4294967288 @ 0xfffffff8
; CHECK-NEXT:    .long 4294967292 @ 0xfffffffc
entry:
  %cmp22 = icmp sgt i32 %n, 0
  br i1 %cmp22, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %for.body.preheader
  %n.vec = and i32 %n, -4
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %vec.ind = phi <4 x i32> [ <i32 0, i32 1, i32 2, i32 3>, %vector.ph ], [ %vec.ind.next, %vector.body ]
  %0 = getelementptr inbounds i32, i32* %data, <4 x i32> %vec.ind
  %wide.masked.gather = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> %0, i32 4, <4 x i1> <i1 true, i1 true, i1 true, i1 true>, <4 x i32> undef)
  %1 = getelementptr inbounds i32, i32* %dst, i32 %index
  %2 = bitcast i32* %1 to <4 x i32>*
  store <4 x i32> %wide.masked.gather, <4 x i32>* %2, align 4
  %index.next = add i32 %index, 4
  %vec.ind.next = add <4 x i32> %vec.ind, <i32 4, i32 4, i32 4, i32 4>
  %3 = icmp eq i32 %index.next, %n.vec
  br i1 %3, label %middle.block, label %vector.body

middle.block:                                     ; preds = %vector.body
  %cmp.n = icmp eq i32 %n.vec, %n
  br i1 %cmp.n, label %for.cond.cleanup, label %vector.ph

for.cond.cleanup:                                 ; preds = %for.body, %middle.block, %entry
  ret void
}

define arm_aapcs_vfpcc void @gather_inc_v4i32_complex(i32* noalias nocapture readonly %data, i32* noalias nocapture %dst, i32 %n) {
; CHECK-LABEL: gather_inc_v4i32_complex:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, r5, r7, lr}
; CHECK-NEXT:    push {r4, r5, r7, lr}
; CHECK-NEXT:    .vsave {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    vpush {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    cmp r2, #1
; CHECK-NEXT:    blt .LBB9_5
; CHECK-NEXT:  @ %bb.1: @ %vector.ph.preheader
; CHECK-NEXT:    bic r12, r2, #3
; CHECK-NEXT:    movs r3, #1
; CHECK-NEXT:    sub.w lr, r12, #4
; CHECK-NEXT:    adr r4, .LCPI9_1
; CHECK-NEXT:    adr r5, .LCPI9_2
; CHECK-NEXT:    vldrw.u32 q1, [r4]
; CHECK-NEXT:    add.w r3, r3, lr, lsr #2
; CHECK-NEXT:    adr.w lr, .LCPI9_0
; CHECK-NEXT:    vldrw.u32 q0, [r5]
; CHECK-NEXT:    vldrw.u32 q2, [lr]
; CHECK-NEXT:    vadd.i32 q1, q1, r0
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:    vadd.i32 q2, q2, r0
; CHECK-NEXT:  .LBB9_2: @ %vector.ph
; CHECK-NEXT:    @ =>This Loop Header: Depth=1
; CHECK-NEXT:    @ Child Loop BB9_3 Depth 2
; CHECK-NEXT:    dls lr, r3
; CHECK-NEXT:    mov r0, r1
; CHECK-NEXT:    vmov q3, q1
; CHECK-NEXT:    vmov q4, q0
; CHECK-NEXT:    vmov q5, q2
; CHECK-NEXT:  .LBB9_3: @ %vector.body
; CHECK-NEXT:    @ Parent Loop BB9_2 Depth=1
; CHECK-NEXT:    @ => This Inner Loop Header: Depth=2
; CHECK-NEXT:    vldrw.u32 q6, [q5, #48]!
; CHECK-NEXT:    vldrw.u32 q7, [q3, #48]!
; CHECK-NEXT:    vadd.i32 q6, q7, q6
; CHECK-NEXT:    vldrw.u32 q7, [q4, #48]!
; CHECK-NEXT:    vadd.i32 q6, q6, q7
; CHECK-NEXT:    vstrb.8 q6, [r0], #16
; CHECK-NEXT:    le lr, .LBB9_3
; CHECK-NEXT:  @ %bb.4: @ %middle.block
; CHECK-NEXT:    @ in Loop: Header=BB9_2 Depth=1
; CHECK-NEXT:    cmp r12, r2
; CHECK-NEXT:    bne .LBB9_2
; CHECK-NEXT:  .LBB9_5: @ %for.cond.cleanup
; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    pop {r4, r5, r7, pc}
; CHECK-NEXT:    .p2align 4
; CHECK-NEXT:  @ %bb.6:
; CHECK-NEXT:  .LCPI9_0:
; CHECK-NEXT:    .long 4294967248 @ 0xffffffd0
; CHECK-NEXT:    .long 4294967260 @ 0xffffffdc
; CHECK-NEXT:    .long 4294967272 @ 0xffffffe8
; CHECK-NEXT:    .long 4294967284 @ 0xfffffff4
; CHECK-NEXT:  .LCPI9_1:
; CHECK-NEXT:    .long 4294967252 @ 0xffffffd4
; CHECK-NEXT:    .long 4294967264 @ 0xffffffe0
; CHECK-NEXT:    .long 4294967276 @ 0xffffffec
; CHECK-NEXT:    .long 4294967288 @ 0xfffffff8
; CHECK-NEXT:  .LCPI9_2:
; CHECK-NEXT:    .long 4294967256 @ 0xffffffd8
; CHECK-NEXT:    .long 4294967268 @ 0xffffffe4
; CHECK-NEXT:    .long 4294967280 @ 0xfffffff0
; CHECK-NEXT:    .long 4294967292 @ 0xfffffffc
entry:
  %cmp22 = icmp sgt i32 %n, 0
  br i1 %cmp22, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %for.body.preheader
  %n.vec = and i32 %n, -4
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %vec.ind = phi <4 x i32> [ <i32 0, i32 1, i32 2, i32 3>, %vector.ph ], [ %vec.ind.next, %vector.body ]
  %0 = mul nuw nsw <4 x i32> %vec.ind, <i32 3, i32 3, i32 3, i32 3>
  %1 = getelementptr inbounds i32, i32* %data, <4 x i32> %0
  %wide.masked.gather = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> %1, i32 4, <4 x i1> <i1 true, i1 true, i1 true, i1 true>, <4 x i32> undef)
  %2 = add nuw nsw <4 x i32> %0, <i32 1, i32 1, i32 1, i32 1>
  %3 = getelementptr inbounds i32, i32* %data, <4 x i32> %2
  %wide.masked.gather24 = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> %3, i32 4, <4 x i1> <i1 true, i1 true, i1 true, i1 true>, <4 x i32> undef)
  %4 = add nuw nsw <4 x i32> %0, <i32 2, i32 2, i32 2, i32 2>
  %5 = getelementptr inbounds i32, i32* %data, <4 x i32> %4
  %wide.masked.gather25 = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> %5, i32 4, <4 x i1> <i1 true, i1 true, i1 true, i1 true>, <4 x i32> undef)
  %6 = add nsw <4 x i32> %wide.masked.gather24, %wide.masked.gather
  %7 = add nsw <4 x i32> %6, %wide.masked.gather25
  %8 = getelementptr inbounds i32, i32* %dst, i32 %index
  %9 = bitcast i32* %8 to <4 x i32>*
  store <4 x i32> %7, <4 x i32>* %9, align 4
  %index.next = add i32 %index, 4
  %vec.ind.next = add <4 x i32> %vec.ind, <i32 4, i32 4, i32 4, i32 4>
  %10 = icmp eq i32 %index.next, %n.vec
  br i1 %10, label %middle.block, label %vector.body

middle.block:                                     ; preds = %vector.body
  %cmp.n = icmp eq i32 %n.vec, %n
  br i1 %cmp.n, label %for.cond.cleanup, label %vector.ph

for.cond.cleanup:                                 ; preds = %for.body, %middle.block, %entry
  ret void
}

define arm_aapcs_vfpcc void @gather_inc_v4i32_large(i32* noalias nocapture readonly %data, i32* noalias nocapture %dst, i32 %n) {
; CHECK-LABEL: gather_inc_v4i32_large:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, lr}
; CHECK-NEXT:    push {r4, lr}
; CHECK-NEXT:    cmp r2, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    poplt {r4, pc}
; CHECK-NEXT:  .LBB10_1: @ %vector.ph.preheader
; CHECK-NEXT:    bic r12, r2, #3
; CHECK-NEXT:    movs r3, #1
; CHECK-NEXT:    sub.w lr, r12, #4
; CHECK-NEXT:    add.w r4, r3, lr, lsr #2
; CHECK-NEXT:    adr r3, .LCPI10_0
; CHECK-NEXT:    vldrw.u32 q0, [r3]
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:  .LBB10_2: @ %vector.ph
; CHECK-NEXT:    @ =>This Loop Header: Depth=1
; CHECK-NEXT:    @ Child Loop BB10_3 Depth 2
; CHECK-NEXT:    dls lr, r4
; CHECK-NEXT:    mov r0, r1
; CHECK-NEXT:    vmov q1, q0
; CHECK-NEXT:  .LBB10_3: @ %vector.body
; CHECK-NEXT:    @ Parent Loop BB10_2 Depth=1
; CHECK-NEXT:    @ => This Inner Loop Header: Depth=2
; CHECK-NEXT:    vldrw.u32 q2, [q1, #508]!
; CHECK-NEXT:    vstrb.8 q2, [r0], #16
; CHECK-NEXT:    le lr, .LBB10_3
; CHECK-NEXT:  @ %bb.4: @ %middle.block
; CHECK-NEXT:    @ in Loop: Header=BB10_2 Depth=1
; CHECK-NEXT:    cmp r12, r2
; CHECK-NEXT:    bne .LBB10_2
; CHECK-NEXT:  @ %bb.5: @ %for.cond.cleanup
; CHECK-NEXT:    pop {r4, pc}
; CHECK-NEXT:    .p2align 4
; CHECK-NEXT:  @ %bb.6:
; CHECK-NEXT:  .LCPI10_0:
; CHECK-NEXT:    .long 4294966788 @ 0xfffffe04
; CHECK-NEXT:    .long 4294966792 @ 0xfffffe08
; CHECK-NEXT:    .long 4294966796 @ 0xfffffe0c
; CHECK-NEXT:    .long 4294966800 @ 0xfffffe10
entry:
  %cmp22 = icmp sgt i32 %n, 0
  br i1 %cmp22, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %for.body.preheader
  %n.vec = and i32 %n, -4
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %vec.ind = phi <4 x i32> [ <i32 0, i32 1, i32 2, i32 3>, %vector.ph ], [ %vec.ind.next, %vector.body ]
  %0 = getelementptr inbounds i32, i32* %data, <4 x i32> %vec.ind
  %wide.masked.gather = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> %0, i32 4, <4 x i1> <i1 true, i1 true, i1 true, i1 true>, <4 x i32> undef)
  %1 = getelementptr inbounds i32, i32* %dst, i32 %index
  %2 = bitcast i32* %1 to <4 x i32>*
  store <4 x i32> %wide.masked.gather, <4 x i32>* %2, align 4
  %index.next = add i32 %index, 4
  %vec.ind.next = add <4 x i32> %vec.ind, <i32 127, i32 127, i32 127, i32 127>
  %3 = icmp eq i32 %index.next, %n.vec
  br i1 %3, label %middle.block, label %vector.body

middle.block:                                     ; preds = %vector.body
  %cmp.n = icmp eq i32 %n.vec, %n
  br i1 %cmp.n, label %for.cond.cleanup, label %vector.ph

for.cond.cleanup:                                 ; preds = %for.body, %middle.block, %entry
  ret void
}

; TODO: uneven - I think it's not possible to create such an example, because vec.ind will always be increased by a vector with 4 elements (=> x*4 = even)

; TODO: What is sxth?
define arm_aapcs_vfpcc void @gather_inc_v8i16_simple(i16* noalias nocapture readonly %data, i16* noalias nocapture %dst, i32 %n) {
; CHECK-LABEL: gather_inc_v8i16_simple:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, r5, r6, r7, r8, r9, r10, r11, lr}
; CHECK-NEXT:    push.w {r4, r5, r6, r7, r8, r9, r10, r11, lr}
; CHECK-NEXT:    .pad #4
; CHECK-NEXT:    sub sp, #4
; CHECK-NEXT:    .vsave {d8, d9}
; CHECK-NEXT:    vpush {d8, d9}
; CHECK-NEXT:    .pad #24
; CHECK-NEXT:    sub sp, #24
; CHECK-NEXT:    cmp r2, #1
; CHECK-NEXT:    str r1, [sp] @ 4-byte Spill
; CHECK-NEXT:    mov r1, r2
; CHECK-NEXT:    str r2, [sp, #4] @ 4-byte Spill
; CHECK-NEXT:    blt .LBB11_5
; CHECK-NEXT:  @ %bb.1: @ %vector.ph.preheader
; CHECK-NEXT:    ldr r1, [sp, #4] @ 4-byte Reload
; CHECK-NEXT:    movs r5, #1
; CHECK-NEXT:    vmov.i16 q1, #0x8
; CHECK-NEXT:    bic r12, r1, #7
; CHECK-NEXT:    add r1, sp, #8
; CHECK-NEXT:    sub.w r3, r12, #8
; CHECK-NEXT:    add.w r8, r5, r3, lsr #3
; CHECK-NEXT:    adr r5, .LCPI11_0
; CHECK-NEXT:    vldrw.u32 q0, [r5]
; CHECK-NEXT:  .LBB11_2: @ %vector.ph
; CHECK-NEXT:    @ =>This Loop Header: Depth=1
; CHECK-NEXT:    @ Child Loop BB11_3 Depth 2
; CHECK-NEXT:    dls lr, r8
; CHECK-NEXT:    vmov q2, q0
; CHECK-NEXT:    ldr r5, [sp] @ 4-byte Reload
; CHECK-NEXT:  .LBB11_3: @ %vector.body
; CHECK-NEXT:    @ Parent Loop BB11_2 Depth=1
; CHECK-NEXT:    @ => This Inner Loop Header: Depth=2
; CHECK-NEXT:    vstrw.32 q2, [r1]
; CHECK-NEXT:    mov r10, r1
; CHECK-NEXT:    vldrh.s32 q4, [r1, #8]
; CHECK-NEXT:    vldrh.s32 q3, [r1]
; CHECK-NEXT:    vadd.i16 q2, q2, q1
; CHECK-NEXT:    vshl.i32 q4, q4, #1
; CHECK-NEXT:    vshl.i32 q3, q3, #1
; CHECK-NEXT:    vadd.i32 q4, q4, r0
; CHECK-NEXT:    vadd.i32 q3, q3, r0
; CHECK-NEXT:    vmov r1, r2, d9
; CHECK-NEXT:    vmov r6, r7, d7
; CHECK-NEXT:    vmov r3, r4, d8
; CHECK-NEXT:    ldrh.w r11, [r2]
; CHECK-NEXT:    vmov r2, r9, d6
; CHECK-NEXT:    ldrh r6, [r6]
; CHECK-NEXT:    ldrh r7, [r7]
; CHECK-NEXT:    ldrh r3, [r3]
; CHECK-NEXT:    ldrh r4, [r4]
; CHECK-NEXT:    ldrh r1, [r1]
; CHECK-NEXT:    ldrh r2, [r2]
; CHECK-NEXT:    ldrh.w r9, [r9]
; CHECK-NEXT:    vmov.16 q3[0], r2
; CHECK-NEXT:    vmov.16 q3[1], r9
; CHECK-NEXT:    vmov.16 q3[2], r6
; CHECK-NEXT:    vmov.16 q3[3], r7
; CHECK-NEXT:    vmov.16 q3[4], r3
; CHECK-NEXT:    vmov.16 q3[5], r4
; CHECK-NEXT:    vmov.16 q3[6], r1
; CHECK-NEXT:    mov r1, r10
; CHECK-NEXT:    vmov.16 q3[7], r11
; CHECK-NEXT:    vstrb.8 q3, [r5], #16
; CHECK-NEXT:    le lr, .LBB11_3
; CHECK-NEXT:  @ %bb.4: @ %middle.block
; CHECK-NEXT:    @ in Loop: Header=BB11_2 Depth=1
; CHECK-NEXT:    ldr r2, [sp, #4] @ 4-byte Reload
; CHECK-NEXT:    cmp r12, r2
; CHECK-NEXT:    bne .LBB11_2
; CHECK-NEXT:  .LBB11_5: @ %for.cond.cleanup
; CHECK-NEXT:    add sp, #24
; CHECK-NEXT:    vpop {d8, d9}
; CHECK-NEXT:    add sp, #4
; CHECK-NEXT:    pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc}
; CHECK-NEXT:    .p2align 4
; CHECK-NEXT:  @ %bb.6:
; CHECK-NEXT:  .LCPI11_0:
; CHECK-NEXT:    .short 0 @ 0x0
; CHECK-NEXT:    .short 1 @ 0x1
; CHECK-NEXT:    .short 2 @ 0x2
; CHECK-NEXT:    .short 3 @ 0x3
; CHECK-NEXT:    .short 4 @ 0x4
; CHECK-NEXT:    .short 5 @ 0x5
; CHECK-NEXT:    .short 6 @ 0x6
; CHECK-NEXT:    .short 7 @ 0x7


entry:
  %cmp22 = icmp sgt i32 %n, 0
  br i1 %cmp22, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %for.body.preheader
  %n.vec = and i32 %n, -8
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %vec.ind = phi <8 x i16> [ <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>, %vector.ph ], [ %vec.ind.next, %vector.body ]
  %0 = getelementptr inbounds i16, i16* %data, <8 x i16> %vec.ind
  %wide.masked.gather = call <8 x i16> @llvm.masked.gather.v8i16.v8p0i16(<8 x i16*> %0, i32 2, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <8 x i16> undef)
  %1 = getelementptr inbounds i16, i16* %dst, i32 %index
  %2 = bitcast i16* %1 to <8 x i16>*
  store <8 x i16> %wide.masked.gather, <8 x i16>* %2, align 2
  %index.next = add i32 %index, 8
  %vec.ind.next = add <8 x i16> %vec.ind, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
  %3 = icmp eq i32 %index.next, %n.vec
  br i1 %3, label %middle.block, label %vector.body

middle.block:                                     ; preds = %vector.body
  %cmp.n = icmp eq i32 %n.vec, %n
  br i1 %cmp.n, label %for.cond.cleanup, label %vector.ph

for.cond.cleanup:                                 ; preds = %for.body, %middle.block, %entry
  ret void
}

; TODO: This looks absolutely terrifying :(
define arm_aapcs_vfpcc void @gather_inc_v8i16_complex(i16* noalias nocapture readonly %data, i16* noalias nocapture %dst, i32 %n) {
; CHECK-LABEL: gather_inc_v8i16_complex:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, r5, r6, r7, r8, r9, r10, r11, lr}
; CHECK-NEXT:    push.w {r4, r5, r6, r7, r8, r9, r10, r11, lr}
; CHECK-NEXT:    .pad #4
; CHECK-NEXT:    sub sp, #4
; CHECK-NEXT:    .vsave {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    vpush {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    .pad #136
; CHECK-NEXT:    sub sp, #136
; CHECK-NEXT:    cmp r2, #1
; CHECK-NEXT:    str r1, [sp, #64] @ 4-byte Spill
; CHECK-NEXT:    mov r1, r2
; CHECK-NEXT:    str r2, [sp, #68] @ 4-byte Spill
; CHECK-NEXT:    blt.w .LBB12_5
; CHECK-NEXT:  @ %bb.1: @ %vector.ph.preheader
; CHECK-NEXT:    ldr r1, [sp, #68] @ 4-byte Reload
; CHECK-NEXT:    adr r3, .LCPI12_2
; CHECK-NEXT:    vldrw.u32 q0, [r3]
; CHECK-NEXT:    movs r2, #1
; CHECK-NEXT:    bic r1, r1, #7
; CHECK-NEXT:    str r1, [sp, #4] @ 4-byte Spill
; CHECK-NEXT:    subs r1, #8
; CHECK-NEXT:    vstrw.32 q0, [sp, #40] @ 16-byte Spill
; CHECK-NEXT:    vmov.i16 q2, #0x18
; CHECK-NEXT:    add.w r1, r2, r1, lsr #3
; CHECK-NEXT:    str r1, [sp, #60] @ 4-byte Spill
; CHECK-NEXT:    adr r1, .LCPI12_0
; CHECK-NEXT:    adr r2, .LCPI12_1
; CHECK-NEXT:    vldrw.u32 q0, [r1]
; CHECK-NEXT:    vstrw.32 q2, [sp, #72] @ 16-byte Spill
; CHECK-NEXT:    vstrw.32 q0, [sp, #24] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q0, [r2]
; CHECK-NEXT:    add r2, sp, #120
; CHECK-NEXT:    vstrw.32 q0, [sp, #8] @ 16-byte Spill
; CHECK-NEXT:  .LBB12_2: @ %vector.ph
; CHECK-NEXT:    @ =>This Loop Header: Depth=1
; CHECK-NEXT:    @ Child Loop BB12_3 Depth 2
; CHECK-NEXT:    ldr r1, [sp, #60] @ 4-byte Reload
; CHECK-NEXT:    add.w r10, sp, #104
; CHECK-NEXT:    dls lr, r1
; CHECK-NEXT:    ldr r7, [sp, #64] @ 4-byte Reload
; CHECK-NEXT:    vldrw.u32 q4, [sp, #24] @ 16-byte Reload
; CHECK-NEXT:    vldrw.u32 q5, [sp, #40] @ 16-byte Reload
; CHECK-NEXT:    vldrw.u32 q6, [sp, #8] @ 16-byte Reload
; CHECK-NEXT:  .LBB12_3: @ %vector.body
; CHECK-NEXT:    @ Parent Loop BB12_2 Depth=1
; CHECK-NEXT:    @ => This Inner Loop Header: Depth=2
; CHECK-NEXT:    vstrw.32 q5, [r2]
; CHECK-NEXT:    mov r8, r2
; CHECK-NEXT:    vldrh.s32 q0, [r2, #8]
; CHECK-NEXT:    vshl.i32 q0, q0, #1
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:    vmov r1, r3, d0
; CHECK-NEXT:    vmov r4, r5, d1
; CHECK-NEXT:    vldrh.s32 q0, [r2]
; CHECK-NEXT:    vshl.i32 q0, q0, #1
; CHECK-NEXT:    vadd.i32 q2, q0, r0
; CHECK-NEXT:    vmov r6, r2, d4
; CHECK-NEXT:    ldrh r1, [r1]
; CHECK-NEXT:    ldrh.w r12, [r4]
; CHECK-NEXT:    add r4, sp, #88
; CHECK-NEXT:    ldrh.w r11, [r5]
; CHECK-NEXT:    ldrh r3, [r3]
; CHECK-NEXT:    ldrh r5, [r6]
; CHECK-NEXT:    ldrh r2, [r2]
; CHECK-NEXT:    vstrw.32 q6, [r4]
; CHECK-NEXT:    vldrh.s32 q0, [r4]
; CHECK-NEXT:    vmov.16 q7[0], r5
; CHECK-NEXT:    vmov.16 q7[1], r2
; CHECK-NEXT:    vshl.i32 q0, q0, #1
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:    vmov r6, r9, d0
; CHECK-NEXT:    vmov r2, r5, d1
; CHECK-NEXT:    vldrh.s32 q0, [r4, #8]
; CHECK-NEXT:    vshl.i32 q0, q0, #1
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:    ldrh r6, [r6]
; CHECK-NEXT:    ldrh r2, [r2]
; CHECK-NEXT:    vmov.16 q1[0], r6
; CHECK-NEXT:    ldrh.w r6, [r9]
; CHECK-NEXT:    ldrh r5, [r5]
; CHECK-NEXT:    vmov.16 q1[1], r6
; CHECK-NEXT:    vmov.16 q1[2], r2
; CHECK-NEXT:    vmov r2, r6, d0
; CHECK-NEXT:    vmov.16 q1[3], r5
; CHECK-NEXT:    ldrh r2, [r2]
; CHECK-NEXT:    ldrh r6, [r6]
; CHECK-NEXT:    vmov.16 q1[4], r2
; CHECK-NEXT:    vmov r2, r5, d1
; CHECK-NEXT:    vmov.16 q1[5], r6
; CHECK-NEXT:    mov r6, r10
; CHECK-NEXT:    ldrh r2, [r2]
; CHECK-NEXT:    ldrh r5, [r5]
; CHECK-NEXT:    vstrw.32 q4, [r10]
; CHECK-NEXT:    vldrh.s32 q0, [r6]
; CHECK-NEXT:    vmov.16 q1[6], r2
; CHECK-NEXT:    vmov.16 q1[7], r5
; CHECK-NEXT:    vshl.i32 q0, q0, #1
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:    vmov r2, r5, d0
; CHECK-NEXT:    ldrh r2, [r2]
; CHECK-NEXT:    ldrh r5, [r5]
; CHECK-NEXT:    vmov.16 q3[0], r2
; CHECK-NEXT:    vmov.16 q3[1], r5
; CHECK-NEXT:    vmov r2, r5, d5
; CHECK-NEXT:    vldrw.u32 q2, [sp, #72] @ 16-byte Reload
; CHECK-NEXT:    vadd.i16 q6, q6, q2
; CHECK-NEXT:    vadd.i16 q5, q5, q2
; CHECK-NEXT:    vadd.i16 q4, q4, q2
; CHECK-NEXT:    ldrh.w r9, [r2]
; CHECK-NEXT:    vmov r2, r4, d1
; CHECK-NEXT:    vldrh.s32 q0, [r6, #8]
; CHECK-NEXT:    ldrh r5, [r5]
; CHECK-NEXT:    vmov.16 q7[2], r9
; CHECK-NEXT:    vshl.i32 q0, q0, #1
; CHECK-NEXT:    vmov.16 q7[3], r5
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:    vmov.16 q7[4], r1
; CHECK-NEXT:    vmov.16 q7[5], r3
; CHECK-NEXT:    vmov.16 q7[6], r12
; CHECK-NEXT:    vmov.16 q7[7], r11
; CHECK-NEXT:    ldrh r2, [r2]
; CHECK-NEXT:    ldrh r4, [r4]
; CHECK-NEXT:    vmov.16 q3[2], r2
; CHECK-NEXT:    vmov.16 q3[3], r4
; CHECK-NEXT:    vmov r2, r4, d0
; CHECK-NEXT:    ldrh r2, [r2]
; CHECK-NEXT:    ldrh r4, [r4]
; CHECK-NEXT:    vmov.16 q3[4], r2
; CHECK-NEXT:    vmov.16 q3[5], r4
; CHECK-NEXT:    vmov r2, r4, d1
; CHECK-NEXT:    ldrh r2, [r2]
; CHECK-NEXT:    ldrh r4, [r4]
; CHECK-NEXT:    vmov.16 q3[6], r2
; CHECK-NEXT:    mov r2, r8
; CHECK-NEXT:    vmov.16 q3[7], r4
; CHECK-NEXT:    vadd.i16 q0, q3, q1
; CHECK-NEXT:    vadd.i16 q0, q0, q7
; CHECK-NEXT:    vstrb.8 q0, [r7], #16
; CHECK-NEXT:    le lr, .LBB12_3
; CHECK-NEXT:  @ %bb.4: @ %middle.block
; CHECK-NEXT:    @ in Loop: Header=BB12_2 Depth=1
; CHECK-NEXT:    ldr r1, [sp, #4] @ 4-byte Reload
; CHECK-NEXT:    ldr r3, [sp, #68] @ 4-byte Reload
; CHECK-NEXT:    cmp r1, r3
; CHECK-NEXT:    bne.w .LBB12_2
; CHECK-NEXT:  .LBB12_5: @ %for.cond.cleanup
; CHECK-NEXT:    add sp, #136
; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    add sp, #4
; CHECK-NEXT:    pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc}
; CHECK-NEXT:    .p2align 4
; CHECK-NEXT:  @ %bb.6:
; CHECK-NEXT:  .LCPI12_0:
; CHECK-NEXT:    .short 1 @ 0x1
; CHECK-NEXT:    .short 4 @ 0x4
; CHECK-NEXT:    .short 7 @ 0x7
; CHECK-NEXT:    .short 10 @ 0xa
; CHECK-NEXT:    .short 13 @ 0xd
; CHECK-NEXT:    .short 16 @ 0x10
; CHECK-NEXT:    .short 19 @ 0x13
; CHECK-NEXT:    .short 22 @ 0x16
; CHECK-NEXT:  .LCPI12_1:
; CHECK-NEXT:    .short 0 @ 0x0
; CHECK-NEXT:    .short 3 @ 0x3
; CHECK-NEXT:    .short 6 @ 0x6
; CHECK-NEXT:    .short 9 @ 0x9
; CHECK-NEXT:    .short 12 @ 0xc
; CHECK-NEXT:    .short 15 @ 0xf
; CHECK-NEXT:    .short 18 @ 0x12
; CHECK-NEXT:    .short 21 @ 0x15
; CHECK-NEXT:  .LCPI12_2:
; CHECK-NEXT:    .short 2 @ 0x2
; CHECK-NEXT:    .short 5 @ 0x5
; CHECK-NEXT:    .short 8 @ 0x8
; CHECK-NEXT:    .short 11 @ 0xb
; CHECK-NEXT:    .short 14 @ 0xe
; CHECK-NEXT:    .short 17 @ 0x11
; CHECK-NEXT:    .short 20 @ 0x14
; CHECK-NEXT:    .short 23 @ 0x17


entry:
  %cmp22 = icmp sgt i32 %n, 0
  br i1 %cmp22, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %for.body.preheader
  %n.vec = and i32 %n, -8
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %vec.ind = phi <8 x i16> [ <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>, %vector.ph ], [ %vec.ind.next, %vector.body ]
  %0 = mul nuw nsw <8 x i16> %vec.ind, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
  %1 = getelementptr inbounds i16, i16* %data, <8 x i16> %0
  %wide.masked.gather = call <8 x i16> @llvm.masked.gather.v8i16.v8p0i16(<8 x i16*> %1, i32 2, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <8 x i16> undef)
  %2 = add nuw nsw <8 x i16> %0, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
  %3 = getelementptr inbounds i16, i16* %data, <8 x i16> %2
  %wide.masked.gather24 = call <8 x i16> @llvm.masked.gather.v8i16.v8p0i16(<8 x i16*> %3, i32 2, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <8 x i16> undef)
  %4 = add nuw nsw <8 x i16> %0, <i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2>
  %5 = getelementptr inbounds i16, i16* %data, <8 x i16> %4
  %wide.masked.gather25 = call <8 x i16> @llvm.masked.gather.v8i16.v8p0i16(<8 x i16*> %5, i32 2, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <8 x i16> undef)
  %6 = add nsw <8 x i16> %wide.masked.gather24, %wide.masked.gather
  %7 = add nsw <8 x i16> %6, %wide.masked.gather25
  %8 = getelementptr inbounds i16, i16* %dst, i32 %index
  %9 = bitcast i16* %8 to <8 x i16>*
  store <8 x i16> %7, <8 x i16>* %9, align 2
  %index.next = add i32 %index, 8
  %vec.ind.next = add <8 x i16> %vec.ind, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
  %10 = icmp eq i32 %index.next, %n.vec
  br i1 %10, label %middle.block, label %vector.body

middle.block:                                     ; preds = %vector.body
  %cmp.n = icmp eq i32 %n.vec, %n
  br i1 %cmp.n, label %for.cond.cleanup, label %vector.ph

for.cond.cleanup:                                 ; preds = %for.body, %middle.block, %entry
  ret void
}


define arm_aapcs_vfpcc void @gather_inc_v16i8_complex(i8* noalias nocapture readonly %data, i8* noalias nocapture %dst, i32 %n) {
; CHECK-LABEL: gather_inc_v16i8_complex:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, r5, r6, r7, r8, r9, r10, r11, lr}
; CHECK-NEXT:    push.w {r4, r5, r6, r7, r8, r9, r10, r11, lr}
; CHECK-NEXT:    .pad #4
; CHECK-NEXT:    sub sp, #4
; CHECK-NEXT:    .vsave {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    vpush {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    .pad #312
; CHECK-NEXT:    sub sp, #312
; CHECK-NEXT:    cmp r2, #1
; CHECK-NEXT:    str r1, [sp, #116] @ 4-byte Spill
; CHECK-NEXT:    blt.w .LBB13_5
; CHECK-NEXT:  @ %bb.1: @ %vector.ph.preheader
; CHECK-NEXT:    adr r1, .LCPI13_0
; CHECK-NEXT:    adr r6, .LCPI13_8
; CHECK-NEXT:    vldrw.u32 q0, [r1]
; CHECK-NEXT:    adr r1, .LCPI13_1
; CHECK-NEXT:    adr r7, .LCPI13_7
; CHECK-NEXT:    adr r3, .LCPI13_6
; CHECK-NEXT:    vstrw.32 q0, [sp, #96] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q0, [r1]
; CHECK-NEXT:    adr r1, .LCPI13_5
; CHECK-NEXT:    bic r10, r2, #7
; CHECK-NEXT:    vstrw.32 q0, [sp, #80] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q0, [r6]
; CHECK-NEXT:    adr r6, .LCPI13_9
; CHECK-NEXT:    vmov.i32 q2, #0x30
; CHECK-NEXT:    vstrw.32 q0, [sp, #64] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q0, [r7]
; CHECK-NEXT:    vstrw.32 q0, [sp, #48] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q0, [r6]
; CHECK-NEXT:    vstrw.32 q0, [sp, #32] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q0, [r1]
; CHECK-NEXT:    vstrw.32 q0, [sp, #16] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q0, [r3]
; CHECK-NEXT:    vstrw.32 q0, [sp] @ 16-byte Spill
; CHECK-NEXT:  .LBB13_2: @ %vector.ph
; CHECK-NEXT:    @ =>This Loop Header: Depth=1
; CHECK-NEXT:    @ Child Loop BB13_3 Depth 2
; CHECK-NEXT:    adr r1, .LCPI13_3
; CHECK-NEXT:    vldrw.u32 q6, [sp, #16] @ 16-byte Reload
; CHECK-NEXT:    vldrw.u32 q0, [r1]
; CHECK-NEXT:    adr r1, .LCPI13_4
; CHECK-NEXT:    vldrw.u32 q5, [r1]
; CHECK-NEXT:    adr r1, .LCPI13_2
; CHECK-NEXT:    vldrw.u32 q3, [r1]
; CHECK-NEXT:    adr r1, .LCPI13_10
; CHECK-NEXT:    vstrw.32 q6, [sp, #280] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q6, [sp, #32] @ 16-byte Reload
; CHECK-NEXT:    vstrw.32 q3, [sp, #296] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q3, [r1]
; CHECK-NEXT:    adr r1, .LCPI13_11
; CHECK-NEXT:    ldr.w r8, [sp, #116] @ 4-byte Reload
; CHECK-NEXT:    vstrw.32 q3, [sp, #248] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q3, [sp, #80] @ 16-byte Reload
; CHECK-NEXT:    vstrw.32 q6, [sp, #264] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q6, [sp, #48] @ 16-byte Reload
; CHECK-NEXT:    vstrw.32 q3, [sp, #216] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q3, [sp, #64] @ 16-byte Reload
; CHECK-NEXT:    vldrw.u32 q7, [r1]
; CHECK-NEXT:    vldrw.u32 q1, [sp] @ 16-byte Reload
; CHECK-NEXT:    vstrw.32 q3, [sp, #200] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q3, [sp, #96] @ 16-byte Reload
; CHECK-NEXT:    mov r11, r10
; CHECK-NEXT:    vstrw.32 q6, [sp, #232] @ 16-byte Spill
; CHECK-NEXT:    vstrw.32 q3, [sp, #184] @ 16-byte Spill
; CHECK-NEXT:  .LBB13_3: @ %vector.body
; CHECK-NEXT:    @ Parent Loop BB13_2 Depth=1
; CHECK-NEXT:    @ => This Inner Loop Header: Depth=2
; CHECK-NEXT:    vadd.i32 q4, q1, r0
; CHECK-NEXT:    vstrw.32 q7, [sp, #136] @ 16-byte Spill
; CHECK-NEXT:    vmov r1, lr, d8
; CHECK-NEXT:    vadd.i32 q7, q7, r0
; CHECK-NEXT:    vmov r5, r4, d15
; CHECK-NEXT:    vadd.i32 q6, q0, r0
; CHECK-NEXT:    vmov r6, r7, d13
; CHECK-NEXT:    vstrw.32 q1, [sp, #152] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q1, [sp, #296] @ 16-byte Reload
; CHECK-NEXT:    vstrw.32 q0, [sp, #168] @ 16-byte Spill
; CHECK-NEXT:    vmov q0, q2
; CHECK-NEXT:    vmov q3, q5
; CHECK-NEXT:    vadd.i32 q1, q1, r0
; CHECK-NEXT:    vldrw.u32 q0, [sp, #248] @ 16-byte Reload
; CHECK-NEXT:    vldrw.u32 q3, [sp, #216] @ 16-byte Reload
; CHECK-NEXT:    vstrw.32 q5, [sp, #120] @ 16-byte Spill
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:    subs.w r11, r11, #16
; CHECK-NEXT:    ldrb.w r9, [r1]
; CHECK-NEXT:    vmov r1, r3, d14
; CHECK-NEXT:    ldrb r5, [r5]
; CHECK-NEXT:    ldrb r7, [r7]
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    vmov.8 q7[0], r1
; CHECK-NEXT:    ldrb r1, [r3]
; CHECK-NEXT:    vmov.8 q7[1], r1
; CHECK-NEXT:    vmov r1, r3, d12
; CHECK-NEXT:    vmov.8 q7[2], r5
; CHECK-NEXT:    ldrb r5, [r6]
; CHECK-NEXT:    ldrb r6, [r4]
; CHECK-NEXT:    vmov.8 q7[3], r6
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    ldrb r3, [r3]
; CHECK-NEXT:    vmov.8 q6[0], r1
; CHECK-NEXT:    vmov r6, r1, d2
; CHECK-NEXT:    vmov.8 q6[1], r3
; CHECK-NEXT:    vmov.8 q6[2], r5
; CHECK-NEXT:    vmov.8 q6[3], r7
; CHECK-NEXT:    ldrb.w r7, [lr]
; CHECK-NEXT:    vmov.8 q6[4], r9
; CHECK-NEXT:    vmov.8 q6[5], r7
; CHECK-NEXT:    ldrb r4, [r1]
; CHECK-NEXT:    vmov r1, r5, d3
; CHECK-NEXT:    vldrw.u32 q1, [sp, #232] @ 16-byte Reload
; CHECK-NEXT:    ldrb.w r12, [r1]
; CHECK-NEXT:    vmov r1, r3, d9
; CHECK-NEXT:    ldrb r5, [r5]
; CHECK-NEXT:    vldrw.u32 q4, [sp, #184] @ 16-byte Reload
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    ldrb r3, [r3]
; CHECK-NEXT:    vmov.8 q6[6], r1
; CHECK-NEXT:    vmov r1, r7, d0
; CHECK-NEXT:    vmov.8 q6[7], r3
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    ldrb r7, [r7]
; CHECK-NEXT:    vmov.8 q7[4], r1
; CHECK-NEXT:    vmov r1, r3, d1
; CHECK-NEXT:    vldrw.u32 q0, [sp, #264] @ 16-byte Reload
; CHECK-NEXT:    vmov.8 q7[5], r7
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    ldrb r3, [r3]
; CHECK-NEXT:    vmov.8 q7[6], r1
; CHECK-NEXT:    ldrb r1, [r6]
; CHECK-NEXT:    vmov r7, r6, d0
; CHECK-NEXT:    vmov.8 q7[7], r3
; CHECK-NEXT:    vmov r3, lr, d1
; CHECK-NEXT:    vldrw.u32 q0, [sp, #280] @ 16-byte Reload
; CHECK-NEXT:    vmov.8 q7[8], r1
; CHECK-NEXT:    vadd.i32 q0, q0, r0
; CHECK-NEXT:    vmov.8 q7[9], r4
; CHECK-NEXT:    vmov r4, r1, d0
; CHECK-NEXT:    vmov.8 q7[10], r12
; CHECK-NEXT:    vmov.8 q7[11], r5
; CHECK-NEXT:    ldrb r7, [r7]
; CHECK-NEXT:    ldrb r6, [r6]
; CHECK-NEXT:    ldrb r3, [r3]
; CHECK-NEXT:    ldrb r4, [r4]
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    vmov.8 q6[8], r4
; CHECK-NEXT:    vmov r5, r4, d1
; CHECK-NEXT:    vmov.8 q6[9], r1
; CHECK-NEXT:    vadd.i32 q0, q5, r0
; CHECK-NEXT:    vldrw.u32 q5, [sp, #200] @ 16-byte Reload
; CHECK-NEXT:    ldrb r5, [r5]
; CHECK-NEXT:    ldrb r4, [r4]
; CHECK-NEXT:    vmov.8 q6[10], r5
; CHECK-NEXT:    vmov.8 q6[11], r4
; CHECK-NEXT:    vmov.8 q6[12], r7
; CHECK-NEXT:    vmov.8 q6[13], r6
; CHECK-NEXT:    vmov.8 q6[14], r3
; CHECK-NEXT:    vmov r1, r3, d0
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    vmov.8 q7[12], r1
; CHECK-NEXT:    ldrb r1, [r3]
; CHECK-NEXT:    vmov.8 q7[13], r1
; CHECK-NEXT:    vmov r1, r3, d1
; CHECK-NEXT:    vadd.i32 q0, q1, r0
; CHECK-NEXT:    vadd.i32 q1, q1, q2
; CHECK-NEXT:    vstrw.32 q1, [sp, #232] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q1, [sp, #248] @ 16-byte Reload
; CHECK-NEXT:    vadd.i32 q1, q1, q2
; CHECK-NEXT:    vstrw.32 q1, [sp, #248] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q1, [sp, #152] @ 16-byte Reload
; CHECK-NEXT:    vadd.i32 q1, q1, q2
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    vmov.8 q7[14], r1
; CHECK-NEXT:    ldrb r1, [r3]
; CHECK-NEXT:    vmov.8 q7[15], r1
; CHECK-NEXT:    ldrb.w r1, [lr]
; CHECK-NEXT:    vmov.8 q6[15], r1
; CHECK-NEXT:    vmov r1, r3, d0
; CHECK-NEXT:    vadd.i8 q6, q6, q7
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    ldrb r3, [r3]
; CHECK-NEXT:    vmov.8 q7[0], r1
; CHECK-NEXT:    vmov.8 q7[1], r3
; CHECK-NEXT:    vmov r1, r3, d1
; CHECK-NEXT:    vadd.i32 q0, q3, r0
; CHECK-NEXT:    vadd.i32 q3, q3, q2
; CHECK-NEXT:    vstrw.32 q3, [sp, #216] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q3, [sp, #296] @ 16-byte Reload
; CHECK-NEXT:    vadd.i32 q3, q3, q2
; CHECK-NEXT:    vstrw.32 q3, [sp, #296] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q3, [sp, #280] @ 16-byte Reload
; CHECK-NEXT:    vadd.i32 q3, q3, q2
; CHECK-NEXT:    vstrw.32 q3, [sp, #280] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q3, [sp, #264] @ 16-byte Reload
; CHECK-NEXT:    vadd.i32 q3, q3, q2
; CHECK-NEXT:    vstrw.32 q3, [sp, #264] @ 16-byte Spill
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    vmov.8 q7[2], r1
; CHECK-NEXT:    ldrb r1, [r3]
; CHECK-NEXT:    vmov.8 q7[3], r1
; CHECK-NEXT:    vmov r1, r3, d0
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    vmov.8 q7[4], r1
; CHECK-NEXT:    ldrb r1, [r3]
; CHECK-NEXT:    vmov.8 q7[5], r1
; CHECK-NEXT:    vmov r1, r3, d1
; CHECK-NEXT:    vadd.i32 q0, q5, r0
; CHECK-NEXT:    vadd.i32 q5, q5, q2
; CHECK-NEXT:    vstrw.32 q5, [sp, #200] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q5, [sp, #120] @ 16-byte Reload
; CHECK-NEXT:    vadd.i32 q5, q5, q2
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    vmov.8 q7[6], r1
; CHECK-NEXT:    ldrb r1, [r3]
; CHECK-NEXT:    vmov.8 q7[7], r1
; CHECK-NEXT:    vmov r1, r3, d0
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    vmov.8 q7[8], r1
; CHECK-NEXT:    ldrb r1, [r3]
; CHECK-NEXT:    vmov.8 q7[9], r1
; CHECK-NEXT:    vmov r1, r3, d1
; CHECK-NEXT:    vadd.i32 q0, q4, r0
; CHECK-NEXT:    vadd.i32 q4, q4, q2
; CHECK-NEXT:    vstrw.32 q4, [sp, #184] @ 16-byte Spill
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    vmov.8 q7[10], r1
; CHECK-NEXT:    ldrb r1, [r3]
; CHECK-NEXT:    vmov.8 q7[11], r1
; CHECK-NEXT:    vmov r1, r3, d0
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    vmov.8 q7[12], r1
; CHECK-NEXT:    ldrb r1, [r3]
; CHECK-NEXT:    vmov.8 q7[13], r1
; CHECK-NEXT:    vmov r1, r3, d1
; CHECK-NEXT:    ldrb r1, [r1]
; CHECK-NEXT:    vmov.8 q7[14], r1
; CHECK-NEXT:    ldrb r1, [r3]
; CHECK-NEXT:    vmov.8 q7[15], r1
; CHECK-NEXT:    vadd.i8 q0, q6, q7
; CHECK-NEXT:    vldrw.u32 q7, [sp, #136] @ 16-byte Reload
; CHECK-NEXT:    vstrb.8 q0, [r8], #16
; CHECK-NEXT:    vldrw.u32 q0, [sp, #168] @ 16-byte Reload
; CHECK-NEXT:    vadd.i32 q7, q7, q2
; CHECK-NEXT:    vadd.i32 q0, q0, q2
; CHECK-NEXT:    bne.w .LBB13_3
; CHECK-NEXT:  @ %bb.4: @ %middle.block
; CHECK-NEXT:    @ in Loop: Header=BB13_2 Depth=1
; CHECK-NEXT:    cmp r10, r2
; CHECK-NEXT:    bne.w .LBB13_2
; CHECK-NEXT:  .LBB13_5: @ %for.cond.cleanup
; CHECK-NEXT:    add sp, #312
; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    add sp, #4
; CHECK-NEXT:    pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc}
; CHECK-NEXT:    .p2align 4
; CHECK-NEXT:  @ %bb.6:
; CHECK-NEXT:  .LCPI13_0:
; CHECK-NEXT:    .long 38 @ 0x26
; CHECK-NEXT:    .long 41 @ 0x29
; CHECK-NEXT:    .long 44 @ 0x2c
; CHECK-NEXT:    .long 47 @ 0x2f
; CHECK-NEXT:  .LCPI13_1:
; CHECK-NEXT:    .long 14 @ 0xe
; CHECK-NEXT:    .long 17 @ 0x11
; CHECK-NEXT:    .long 20 @ 0x14
; CHECK-NEXT:    .long 23 @ 0x17
; CHECK-NEXT:  .LCPI13_2:
; CHECK-NEXT:    .long 24 @ 0x18
; CHECK-NEXT:    .long 27 @ 0x1b
; CHECK-NEXT:    .long 30 @ 0x1e
; CHECK-NEXT:    .long 33 @ 0x21
; CHECK-NEXT:  .LCPI13_3:
; CHECK-NEXT:    .long 1 @ 0x1
; CHECK-NEXT:    .long 4 @ 0x4
; CHECK-NEXT:    .long 7 @ 0x7
; CHECK-NEXT:    .long 10 @ 0xa
; CHECK-NEXT:  .LCPI13_4:
; CHECK-NEXT:    .long 36 @ 0x24
; CHECK-NEXT:    .long 39 @ 0x27
; CHECK-NEXT:    .long 42 @ 0x2a
; CHECK-NEXT:    .long 45 @ 0x2d
; CHECK-NEXT:  .LCPI13_5:
; CHECK-NEXT:    .long 25 @ 0x19
; CHECK-NEXT:    .long 28 @ 0x1c
; CHECK-NEXT:    .long 31 @ 0x1f
; CHECK-NEXT:    .long 34 @ 0x22
; CHECK-NEXT:  .LCPI13_6:
; CHECK-NEXT:    .long 13 @ 0xd
; CHECK-NEXT:    .long 16 @ 0x10
; CHECK-NEXT:    .long 19 @ 0x13
; CHECK-NEXT:    .long 22 @ 0x16
; CHECK-NEXT:  .LCPI13_7:
; CHECK-NEXT:    .long 2 @ 0x2
; CHECK-NEXT:    .long 5 @ 0x5
; CHECK-NEXT:    .long 8 @ 0x8
; CHECK-NEXT:    .long 11 @ 0xb
; CHECK-NEXT:  .LCPI13_8:
; CHECK-NEXT:    .long 26 @ 0x1a
; CHECK-NEXT:    .long 29 @ 0x1d
; CHECK-NEXT:    .long 32 @ 0x20
; CHECK-NEXT:    .long 35 @ 0x23
; CHECK-NEXT:  .LCPI13_9:
; CHECK-NEXT:    .long 37 @ 0x25
; CHECK-NEXT:    .long 40 @ 0x28
; CHECK-NEXT:    .long 43 @ 0x2b
; CHECK-NEXT:    .long 46 @ 0x2e
; CHECK-NEXT:  .LCPI13_10:
; CHECK-NEXT:    .long 12 @ 0xc
; CHECK-NEXT:    .long 15 @ 0xf
; CHECK-NEXT:    .long 18 @ 0x12
; CHECK-NEXT:    .long 21 @ 0x15
; CHECK-NEXT:  .LCPI13_11:
; CHECK-NEXT:    .long 0 @ 0x0
; CHECK-NEXT:    .long 3 @ 0x3
; CHECK-NEXT:    .long 6 @ 0x6
; CHECK-NEXT:    .long 9 @ 0x9


entry:
  %cmp22 = icmp sgt i32 %n, 0
  br i1 %cmp22, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %for.body.preheader
  %n.vec = and i32 %n, -8
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %vec.ind = phi <16 x i32> [ <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>, %vector.ph ], [ %vec.ind.next, %vector.body ]
  %0 = mul nuw nsw <16 x i32> %vec.ind, <i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3>
  %1 = getelementptr inbounds i8, i8* %data, <16 x i32> %0
  %wide.masked.gather = call <16 x i8> @llvm.masked.gather.v16i8.v16p0i8(<16 x i8*> %1, i32 2, <16 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <16 x i8> undef)
  %2 = add nuw nsw <16 x i32> %0, <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
  %3 = getelementptr inbounds i8, i8* %data, <16 x i32> %2
  %wide.masked.gather24 = call <16 x i8> @llvm.masked.gather.v16i8.v16p0i8(<16 x i8*> %3, i32 2, <16 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <16 x i8> undef)
  %4 = add nuw nsw <16 x i32> %0, <i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2>
  %5 = getelementptr inbounds i8, i8* %data, <16 x i32> %4
  %wide.masked.gather25 = call <16 x i8> @llvm.masked.gather.v16i8.v16p0i8(<16 x i8*> %5, i32 2, <16 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <16 x i8> undef)
  %6 = add nsw <16 x i8> %wide.masked.gather24, %wide.masked.gather
  %7 = add nsw <16 x i8> %6, %wide.masked.gather25
  %8 = getelementptr inbounds i8, i8* %dst, i32 %index
  %9 = bitcast i8* %8 to <16 x i8>*
  store <16 x i8> %7, <16 x i8>* %9, align 2
  %index.next = add i32 %index, 16
  %vec.ind.next = add <16 x i32> %vec.ind, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
  %10 = icmp eq i32 %index.next, %n.vec
  br i1 %10, label %middle.block, label %vector.body

middle.block:                                     ; preds = %vector.body
  %cmp.n = icmp eq i32 %n.vec, %n
  br i1 %cmp.n, label %for.cond.cleanup, label %vector.ph

for.cond.cleanup:                                 ; preds = %for.body, %middle.block, %entry
  ret void
}

define arm_aapcs_vfpcc void @gather_inc_v16i8_simple(i8* noalias nocapture readonly %data, i8* noalias nocapture %dst, i32 %n) {
; CHECK-LABEL: gather_inc_v16i8_simple:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, r5, r6, r7, r8, r9, r10, r11, lr}
; CHECK-NEXT:    push.w {r4, r5, r6, r7, r8, r9, r10, r11, lr}
; CHECK-NEXT:    .pad #4
; CHECK-NEXT:    sub sp, #4
; CHECK-NEXT:    .vsave {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    vpush {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    .pad #72
; CHECK-NEXT:    sub sp, #72
; CHECK-NEXT:    cmp r2, #1
; CHECK-NEXT:    str r1, [sp, #68] @ 4-byte Spill
; CHECK-NEXT:    blt.w .LBB14_5
; CHECK-NEXT:  @ %bb.1: @ %vector.ph.preheader
; CHECK-NEXT:    adr r5, .LCPI14_3
; CHECK-NEXT:    adr r7, .LCPI14_1
; CHECK-NEXT:    vldrw.u32 q0, [r5]
; CHECK-NEXT:    adr r6, .LCPI14_2
; CHECK-NEXT:    adr r3, .LCPI14_0
; CHECK-NEXT:    bic r1, r2, #7
; CHECK-NEXT:    vstrw.32 q0, [sp, #48] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q0, [r7]
; CHECK-NEXT:    vmov.i32 q4, #0x10
; CHECK-NEXT:    vstrw.32 q0, [sp, #32] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q0, [r6]
; CHECK-NEXT:    vstrw.32 q0, [sp, #16] @ 16-byte Spill
; CHECK-NEXT:    vldrw.u32 q0, [r3]
; CHECK-NEXT:    vstrw.32 q0, [sp] @ 16-byte Spill
; CHECK-NEXT:  .LBB14_2: @ %vector.ph
; CHECK-NEXT:    @ =>This Loop Header: Depth=1
; CHECK-NEXT:    @ Child Loop BB14_3 Depth 2
; CHECK-NEXT:    ldr.w lr, [sp, #68] @ 4-byte Reload
; CHECK-NEXT:    vldrw.u32 q5, [sp] @ 16-byte Reload
; CHECK-NEXT:    vldrw.u32 q6, [sp, #16] @ 16-byte Reload
; CHECK-NEXT:    vldrw.u32 q7, [sp, #32] @ 16-byte Reload
; CHECK-NEXT:    vldrw.u32 q0, [sp, #48] @ 16-byte Reload
; CHECK-NEXT:    mov r8, r1
; CHECK-NEXT:  .LBB14_3: @ %vector.body
; CHECK-NEXT:    @ Parent Loop BB14_2 Depth=1
; CHECK-NEXT:    @ => This Inner Loop Header: Depth=2
; CHECK-NEXT:    vadd.i32 q2, q6, r0
; CHECK-NEXT:    vadd.i32 q1, q5, r0
; CHECK-NEXT:    vmov r6, r7, d5
; CHECK-NEXT:    vadd.i32 q3, q0, r0
; CHECK-NEXT:    vmov r4, r5, d3
; CHECK-NEXT:    subs.w r8, r8, #16
; CHECK-NEXT:    vmov r3, r9, d4
; CHECK-NEXT:    vadd.i32 q2, q7, r0
; CHECK-NEXT:    vadd.i32 q5, q5, q4
; CHECK-NEXT:    vadd.i32 q6, q6, q4
; CHECK-NEXT:    vadd.i32 q7, q7, q4
; CHECK-NEXT:    vadd.i32 q0, q0, q4
; CHECK-NEXT:    ldrb.w r11, [r6]
; CHECK-NEXT:    ldrb.w r10, [r7]
; CHECK-NEXT:    vmov r6, r7, d2
; CHECK-NEXT:    ldrb r4, [r4]
; CHECK-NEXT:    ldrb r5, [r5]
; CHECK-NEXT:    ldrb r3, [r3]
; CHECK-NEXT:    ldrb.w r9, [r9]
; CHECK-NEXT:    ldrb r6, [r6]
; CHECK-NEXT:    ldrb r7, [r7]
; CHECK-NEXT:    vmov.8 q1[0], r6
; CHECK-NEXT:    vmov.8 q1[1], r7
; CHECK-NEXT:    vmov r6, r7, d5
; CHECK-NEXT:    vmov.8 q1[2], r4
; CHECK-NEXT:    vmov.8 q1[3], r5
; CHECK-NEXT:    vmov.8 q1[4], r3
; CHECK-NEXT:    vmov.8 q1[5], r9
; CHECK-NEXT:    vmov.8 q1[6], r11
; CHECK-NEXT:    vmov.8 q1[7], r10
; CHECK-NEXT:    ldrb.w r12, [r7]
; CHECK-NEXT:    vmov r5, r7, d7
; CHECK-NEXT:    ldrb r6, [r6]
; CHECK-NEXT:    ldrb.w r9, [r7]
; CHECK-NEXT:    vmov r7, r3, d6
; CHECK-NEXT:    ldrb r5, [r5]
; CHECK-NEXT:    ldrb.w r11, [r3]
; CHECK-NEXT:    vmov r3, r4, d4
; CHECK-NEXT:    ldrb r7, [r7]
; CHECK-NEXT:    ldrb r3, [r3]
; CHECK-NEXT:    ldrb r4, [r4]
; CHECK-NEXT:    vmov.8 q1[8], r3
; CHECK-NEXT:    vmov.8 q1[9], r4
; CHECK-NEXT:    vmov.8 q1[10], r6
; CHECK-NEXT:    vmov.8 q1[11], r12
; CHECK-NEXT:    vmov.8 q1[12], r7
; CHECK-NEXT:    vmov.8 q1[13], r11
; CHECK-NEXT:    vmov.8 q1[14], r5
; CHECK-NEXT:    vmov.8 q1[15], r9
; CHECK-NEXT:    vstrb.8 q1, [lr], #16
; CHECK-NEXT:    bne .LBB14_3
; CHECK-NEXT:  @ %bb.4: @ %middle.block
; CHECK-NEXT:    @ in Loop: Header=BB14_2 Depth=1
; CHECK-NEXT:    cmp r1, r2
; CHECK-NEXT:    bne .LBB14_2
; CHECK-NEXT:  .LBB14_5: @ %for.cond.cleanup
; CHECK-NEXT:    add sp, #72
; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    add sp, #4
; CHECK-NEXT:    pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc}
; CHECK-NEXT:    .p2align 4
; CHECK-NEXT:  @ %bb.6:
; CHECK-NEXT:  .LCPI14_0:
; CHECK-NEXT:    .long 0 @ 0x0
; CHECK-NEXT:    .long 1 @ 0x1
; CHECK-NEXT:    .long 2 @ 0x2
; CHECK-NEXT:    .long 3 @ 0x3
; CHECK-NEXT:  .LCPI14_1:
; CHECK-NEXT:    .long 8 @ 0x8
; CHECK-NEXT:    .long 9 @ 0x9
; CHECK-NEXT:    .long 10 @ 0xa
; CHECK-NEXT:    .long 11 @ 0xb
; CHECK-NEXT:  .LCPI14_2:
; CHECK-NEXT:    .long 4 @ 0x4
; CHECK-NEXT:    .long 5 @ 0x5
; CHECK-NEXT:    .long 6 @ 0x6
; CHECK-NEXT:    .long 7 @ 0x7
; CHECK-NEXT:  .LCPI14_3:
; CHECK-NEXT:    .long 12 @ 0xc
; CHECK-NEXT:    .long 13 @ 0xd
; CHECK-NEXT:    .long 14 @ 0xe
; CHECK-NEXT:    .long 15 @ 0xf


entry:
  %cmp22 = icmp sgt i32 %n, 0
  br i1 %cmp22, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %for.body.preheader
  %n.vec = and i32 %n, -8
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %vec.ind = phi <16 x i32> [ <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>, %vector.ph ], [ %vec.ind.next, %vector.body ]
  %0 = getelementptr inbounds i8, i8* %data, <16 x i32> %vec.ind
  %wide.masked.gather = call <16 x i8> @llvm.masked.gather.v16i8.v16p0i8(<16 x i8*> %0, i32 2, <16 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <16 x i8> undef)
  %1 = getelementptr inbounds i8, i8* %dst, i32 %index
  %2 = bitcast i8* %1 to <16 x i8>*
  store <16 x i8> %wide.masked.gather, <16 x i8>* %2, align 2
  %index.next = add i32 %index, 16
  %vec.ind.next = add <16 x i32> %vec.ind, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
  %3 = icmp eq i32 %index.next, %n.vec
  br i1 %3, label %middle.block, label %vector.body

middle.block:                                     ; preds = %vector.body
  %cmp.n = icmp eq i32 %n.vec, %n
  br i1 %cmp.n, label %for.cond.cleanup, label %vector.ph

for.cond.cleanup:                                 ; preds = %for.body, %middle.block, %entry
  ret void
}

define void @shl(i32* nocapture %x, i32* noalias nocapture readonly %y, i32 %n) {
; CHECK-LABEL: shl:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    cmp r2, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    poplt {r7, pc}
; CHECK-NEXT:  .LBB15_1: @ %vector.ph
; CHECK-NEXT:    adr r3, .LCPI15_0
; CHECK-NEXT:    vldrw.u32 q0, [r3]
; CHECK-NEXT:    vadd.i32 q0, q0, r1
; CHECK-NEXT:    dlstp.32 lr, r2
; CHECK-NEXT:  .LBB15_2: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q1, [q0, #64]!
; CHECK-NEXT:    vstrw.32 q1, [r0], #16
; CHECK-NEXT:    letp lr, .LBB15_2
; CHECK-NEXT:  @ %bb.3: @ %for.cond.cleanup
; CHECK-NEXT:    pop {r7, pc}
; CHECK-NEXT:    .p2align 4
; CHECK-NEXT:  @ %bb.4:
; CHECK-NEXT:  .LCPI15_0:
; CHECK-NEXT:    .long 4294967232 @ 0xffffffc0
; CHECK-NEXT:    .long 4294967248 @ 0xffffffd0
; CHECK-NEXT:    .long 4294967264 @ 0xffffffe0
; CHECK-NEXT:    .long 4294967280 @ 0xfffffff0
entry:
  %cmp6 = icmp sgt i32 %n, 0
  br i1 %cmp6, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %entry
  %n.rnd.up = add i32 %n, 3
  %n.vec = and i32 %n.rnd.up, -4
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %vec.ind = phi <4 x i32> [ <i32 0, i32 1, i32 2, i32 3>, %vector.ph ], [ %vec.ind.next, %vector.body ]
  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
  %0 = shl nsw <4 x i32> %vec.ind, <i32 2, i32 2, i32 2, i32 2>
  %1 = getelementptr inbounds i32, i32* %y, <4 x i32> %0
  %wide.masked.gather = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> %1, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
  %2 = getelementptr inbounds i32, i32* %x, i32 %index
  %3 = bitcast i32* %2 to <4 x i32>*
  call void @llvm.masked.store.v4i32.p0v4i32(<4 x i32> %wide.masked.gather, <4 x i32>* %3, i32 4, <4 x i1> %active.lane.mask)
  %index.next = add i32 %index, 4
  %vec.ind.next = add <4 x i32> %vec.ind, <i32 4, i32 4, i32 4, i32 4>
  %4 = icmp eq i32 %index.next, %n.vec
  br i1 %4, label %for.cond.cleanup, label %vector.body

for.cond.cleanup:                                 ; preds = %vector.body, %entry
  ret void
}

define void @shlor(i32* nocapture %x, i32* noalias nocapture readonly %y, i32 %n) {
; CHECK-LABEL: shlor:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, r5, r6, lr}
; CHECK-NEXT:    push {r4, r5, r6, lr}
; CHECK-NEXT:    .vsave {d8, d9, d10, d11, d12, d13}
; CHECK-NEXT:    vpush {d8, d9, d10, d11, d12, d13}
; CHECK-NEXT:    cmp r2, #1
; CHECK-NEXT:    blt .LBB16_3
; CHECK-NEXT:  @ %bb.1: @ %vector.ph
; CHECK-NEXT:    adr.w lr, .LCPI16_0
; CHECK-NEXT:    adr r4, .LCPI16_1
; CHECK-NEXT:    adr r5, .LCPI16_2
; CHECK-NEXT:    adr r6, .LCPI16_3
; CHECK-NEXT:    vldrw.u32 q0, [r6]
; CHECK-NEXT:    vldrw.u32 q1, [r5]
; CHECK-NEXT:    vldrw.u32 q2, [r4]
; CHECK-NEXT:    vldrw.u32 q3, [lr]
; CHECK-NEXT:    vadd.i32 q0, q0, r1
; CHECK-NEXT:    vadd.i32 q1, q1, r1
; CHECK-NEXT:    vadd.i32 q2, q2, r1
; CHECK-NEXT:    vadd.i32 q3, q3, r1
; CHECK-NEXT:    dlstp.32 lr, r2
; CHECK-NEXT:  .LBB16_2: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q4, [q3, #128]!
; CHECK-NEXT:    vldrw.u32 q5, [q2, #128]!
; CHECK-NEXT:    vldrw.u32 q6, [q0, #128]!
; CHECK-NEXT:    vadd.i32 q4, q5, q4
; CHECK-NEXT:    vldrw.u32 q5, [q1, #128]!
; CHECK-NEXT:    vadd.i32 q4, q4, q5
; CHECK-NEXT:    vadd.i32 q4, q4, q6
; CHECK-NEXT:    vstrw.32 q4, [r0], #16
; CHECK-NEXT:    letp lr, .LBB16_2
; CHECK-NEXT:  .LBB16_3: @ %for.cond.cleanup
; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13}
; CHECK-NEXT:    pop {r4, r5, r6, pc}
; CHECK-NEXT:    .p2align 4
; CHECK-NEXT:  @ %bb.4:
; CHECK-NEXT:  .LCPI16_0:
; CHECK-NEXT:    .long 4294967168 @ 0xffffff80
; CHECK-NEXT:    .long 4294967200 @ 0xffffffa0
; CHECK-NEXT:    .long 4294967232 @ 0xffffffc0
; CHECK-NEXT:    .long 4294967264 @ 0xffffffe0
; CHECK-NEXT:  .LCPI16_1:
; CHECK-NEXT:    .long 4294967176 @ 0xffffff88
; CHECK-NEXT:    .long 4294967208 @ 0xffffffa8
; CHECK-NEXT:    .long 4294967240 @ 0xffffffc8
; CHECK-NEXT:    .long 4294967272 @ 0xffffffe8
; CHECK-NEXT:  .LCPI16_2:
; CHECK-NEXT:    .long 4294967184 @ 0xffffff90
; CHECK-NEXT:    .long 4294967216 @ 0xffffffb0
; CHECK-NEXT:    .long 4294967248 @ 0xffffffd0
; CHECK-NEXT:    .long 4294967280 @ 0xfffffff0
; CHECK-NEXT:  .LCPI16_3:
; CHECK-NEXT:    .long 4294967192 @ 0xffffff98
; CHECK-NEXT:    .long 4294967224 @ 0xffffffb8
; CHECK-NEXT:    .long 4294967256 @ 0xffffffd8
; CHECK-NEXT:    .long 4294967288 @ 0xfffffff8
entry:
  %cmp23 = icmp sgt i32 %n, 0
  br i1 %cmp23, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %entry
  %n.rnd.up = add i32 %n, 3
  %n.vec = and i32 %n.rnd.up, -4
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %vec.ind = phi <4 x i32> [ <i32 0, i32 1, i32 2, i32 3>, %vector.ph ], [ %vec.ind.next, %vector.body ]
  %active.lane.mask = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32 %index, i32 %n)
  %0 = shl nsw <4 x i32> %vec.ind, <i32 3, i32 3, i32 3, i32 3>
  %1 = getelementptr inbounds i32, i32* %y, <4 x i32> %0
  %wide.masked.gather = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> %1, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
  %2 = or <4 x i32> %0, <i32 2, i32 2, i32 2, i32 2>
  %3 = getelementptr inbounds i32, i32* %y, <4 x i32> %2
  %wide.masked.gather25 = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> %3, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
  %4 = add nsw <4 x i32> %wide.masked.gather25, %wide.masked.gather
  %5 = or <4 x i32> %0, <i32 4, i32 4, i32 4, i32 4>
  %6 = getelementptr inbounds i32, i32* %y, <4 x i32> %5
  %wide.masked.gather26 = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> %6, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
  %7 = add nsw <4 x i32> %4, %wide.masked.gather26
  %8 = or <4 x i32> %0, <i32 6, i32 6, i32 6, i32 6>
  %9 = getelementptr inbounds i32, i32* %y, <4 x i32> %8
  %wide.masked.gather27 = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> %9, i32 4, <4 x i1> %active.lane.mask, <4 x i32> undef)
  %10 = add nsw <4 x i32> %7, %wide.masked.gather27
  %11 = getelementptr inbounds i32, i32* %x, i32 %index
  %12 = bitcast i32* %11 to <4 x i32>*
  call void @llvm.masked.store.v4i32.p0v4i32(<4 x i32> %10, <4 x i32>* %12, i32 4, <4 x i1> %active.lane.mask)
  %index.next = add i32 %index, 4
  %vec.ind.next = add <4 x i32> %vec.ind, <i32 4, i32 4, i32 4, i32 4>
  %13 = icmp eq i32 %index.next, %n.vec
  br i1 %13, label %for.cond.cleanup, label %vector.body

for.cond.cleanup:                                 ; preds = %vector.body, %entry
  ret void
}


declare <2 x i32> @llvm.masked.gather.v2i32.v2p0i32(<2 x i32*>, i32, <2 x i1>, <2 x i32>)
declare <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*>, i32, <4 x i1>, <4 x i32>)
declare <8 x i32> @llvm.masked.gather.v8i32.v8p0i32(<8 x i32*>, i32, <8 x i1>, <8 x i32>)
declare <16 x i32> @llvm.masked.gather.v16i32.v16p0i32(<16 x i32*>, i32, <16 x i1>, <16 x i32>)
declare <2 x float> @llvm.masked.gather.v2f32.v2p0f32(<2 x float*>, i32, <2 x i1>, <2 x float>)
declare <4 x float> @llvm.masked.gather.v4f32.v4p0f32(<4 x float*>, i32, <4 x i1>, <4 x float>)
declare <8 x float> @llvm.masked.gather.v8f32.v8p0f32(<8 x float*>, i32, <8 x i1>, <8 x float>)
declare <2 x i16> @llvm.masked.gather.v2i16.v2p0i16(<2 x i16*>, i32, <2 x i1>, <2 x i16>)
declare <4 x i16> @llvm.masked.gather.v4i16.v4p0i16(<4 x i16*>, i32, <4 x i1>, <4 x i16>)
declare <8 x i16> @llvm.masked.gather.v8i16.v8p0i16(<8 x i16*>, i32, <8 x i1>, <8 x i16>)
declare <16 x i16> @llvm.masked.gather.v16i16.v16p0i16(<16 x i16*>, i32, <16 x i1>, <16 x i16>)
declare <4 x half> @llvm.masked.gather.v4f16.v4p0f16(<4 x half*>, i32, <4 x i1>, <4 x half>)
declare <8 x half> @llvm.masked.gather.v8f16.v8p0f16(<8 x half*>, i32, <8 x i1>, <8 x half>)
declare <16 x half> @llvm.masked.gather.v16f16.v16p0f16(<16 x half*>, i32, <16 x i1>, <16 x half>)
declare <4 x i8> @llvm.masked.gather.v4i8.v4p0i8(<4 x i8*>, i32, <4 x i1>, <4 x i8>)
declare <8 x i8> @llvm.masked.gather.v8i8.v8p0i8(<8 x i8*>, i32, <8 x i1>, <8 x i8>)
declare <16 x i8> @llvm.masked.gather.v16i8.v16p0i8(<16 x i8*>, i32, <16 x i1>, <16 x i8>)
declare <32 x i8> @llvm.masked.gather.v32i8.v32p0i8(<32 x i8*>, i32, <32 x i1>, <32 x i8>)
declare void @llvm.masked.store.v4i32.p0v4i32(<4 x i32>, <4 x i32>*, i32, <4 x i1>)
declare <4 x i1> @llvm.get.active.lane.mask.v4i1.i32(i32, i32)