; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=thumbv8.1m.main-none-none-eabi -mattr=+mve.fp,+fp64 -mve-max-interleave-factor=4 -verify-machineinstrs %s -o - | FileCheck %s

; i32

define void @vst4_v2i32(<2 x i32> *%src, <8 x i32> *%dst) {
; CHECK-LABEL: vst4_v2i32:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, lr}
; CHECK-NEXT:    push {r4, lr}
; CHECK-NEXT:    ldrd lr, r12, [r0]
; CHECK-NEXT:    ldrd r3, r2, [r0, #8]
; CHECK-NEXT:    ldrd r4, r0, [r0, #16]
; CHECK-NEXT:    vmov q1[2], q1[0], r4, r0
; CHECK-NEXT:    vmov.f64 d0, d2
; CHECK-NEXT:    vmov.f32 s1, s6
; CHECK-NEXT:    vmov.f32 s2, s4
; CHECK-NEXT:    vmov.f32 s3, s6
; CHECK-NEXT:    vmov q1[2], q1[0], lr, r3
; CHECK-NEXT:    vmov q1[3], q1[1], r12, r2
; CHECK-NEXT:    vmov.f64 d4, d2
; CHECK-NEXT:    vmov.f32 s9, s6
; CHECK-NEXT:    vmov.f32 s10, s0
; CHECK-NEXT:    vmov.f32 s11, s2
; CHECK-NEXT:    vstrw.32 q2, [r1]
; CHECK-NEXT:    vmov.f32 s8, s5
; CHECK-NEXT:    vmov.f32 s9, s7
; CHECK-NEXT:    vmov.f32 s10, s1
; CHECK-NEXT:    vmov.f32 s11, s3
; CHECK-NEXT:    vstrw.32 q2, [r1, #16]
; CHECK-NEXT:    pop {r4, pc}
entry:
  %s1 = getelementptr <2 x i32>, <2 x i32>* %src, i32 0
  %l1 = load <2 x i32>, <2 x i32>* %s1, align 4
  %s2 = getelementptr <2 x i32>, <2 x i32>* %src, i32 1
  %l2 = load <2 x i32>, <2 x i32>* %s2, align 4
  %s3 = getelementptr <2 x i32>, <2 x i32>* %src, i32 2
  %l3 = load <2 x i32>, <2 x i32>* %s3, align 4
  %s4 = getelementptr <2 x i32>, <2 x i32>* %src, i32 3
  %l4 = load <2 x i32>, <2 x i32>* %s3, align 4
  %t1 = shufflevector <2 x i32> %l1, <2 x i32> %l2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %t2 = shufflevector <2 x i32> %l3, <2 x i32> %l4, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %s = shufflevector <4 x i32> %t1, <4 x i32> %t2, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 1, i32 3, i32 5, i32 7>
  store <8 x i32> %s, <8 x i32> *%dst, align 4
  ret void
}

define void @vst4_v4i32(<4 x i32> *%src, <16 x i32> *%dst) {
; CHECK-LABEL: vst4_v4i32:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldrw.u32 q2, [r0, #32]
; CHECK-NEXT:    vldrw.u32 q1, [r0, #16]
; CHECK-NEXT:    vldrw.u32 q0, [r0]
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vst40.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst41.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst42.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst43.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <4 x i32>, <4 x i32>* %src, i32 0
  %l1 = load <4 x i32>, <4 x i32>* %s1, align 4
  %s2 = getelementptr <4 x i32>, <4 x i32>* %src, i32 1
  %l2 = load <4 x i32>, <4 x i32>* %s2, align 4
  %s3 = getelementptr <4 x i32>, <4 x i32>* %src, i32 2
  %l3 = load <4 x i32>, <4 x i32>* %s3, align 4
  %s4 = getelementptr <4 x i32>, <4 x i32>* %src, i32 3
  %l4 = load <4 x i32>, <4 x i32>* %s3, align 4
  %t1 = shufflevector <4 x i32> %l1, <4 x i32> %l2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %t2 = shufflevector <4 x i32> %l3, <4 x i32> %l4, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %s = shufflevector <8 x i32> %t1, <8 x i32> %t2, <16 x i32> <i32 0, i32 4, i32 8, i32 12, i32 1, i32 5, i32 9, i32 13, i32 2, i32 6, i32 10, i32 14, i32 3, i32 7, i32 11, i32 15>
  store <16 x i32> %s, <16 x i32> *%dst, align 4
  ret void
}

define void @vst4_v8i32(<8 x i32> *%src, <32 x i32> *%dst) {
; CHECK-LABEL: vst4_v8i32:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .vsave {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    vpush {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    vldrw.u32 q6, [r0, #64]
; CHECK-NEXT:    vldrw.u32 q2, [r0, #80]
; CHECK-NEXT:    vldrw.u32 q5, [r0, #32]
; CHECK-NEXT:    vldrw.u32 q1, [r0, #48]
; CHECK-NEXT:    vldrw.u32 q4, [r0]
; CHECK-NEXT:    vldrw.u32 q0, [r0, #16]
; CHECK-NEXT:    vmov q7, q6
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vst40.32 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst41.32 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst42.32 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst43.32 {q4, q5, q6, q7}, [r1]!
; CHECK-NEXT:    vst40.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst41.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst42.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst43.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <8 x i32>, <8 x i32>* %src, i32 0
  %l1 = load <8 x i32>, <8 x i32>* %s1, align 4
  %s2 = getelementptr <8 x i32>, <8 x i32>* %src, i32 1
  %l2 = load <8 x i32>, <8 x i32>* %s2, align 4
  %s3 = getelementptr <8 x i32>, <8 x i32>* %src, i32 2
  %l3 = load <8 x i32>, <8 x i32>* %s3, align 4
  %s4 = getelementptr <8 x i32>, <8 x i32>* %src, i32 3
  %l4 = load <8 x i32>, <8 x i32>* %s3, align 4
  %t1 = shufflevector <8 x i32> %l1, <8 x i32> %l2, <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>
  %t2 = shufflevector <8 x i32> %l3, <8 x i32> %l4, <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>
  %s = shufflevector <16 x i32> %t1, <16 x i32> %t2, <32 x i32> <i32 0, i32 8, i32 16, i32 24, i32 1, i32 9, i32 17, i32 25, i32 2, i32 10, i32 18, i32 26, i32 3, i32 11, i32 19, i32 27, i32 4, i32 12, i32 20, i32 28, i32 5, i32 13, i32 21, i32 29, i32 6, i32 14, i32 22, i32 30, i32 7, i32 15, i32 23, i32 31>
  store <32 x i32> %s, <32 x i32> *%dst, align 4
  ret void
}

define void @vst4_v16i32(<16 x i32> *%src, <64 x i32> *%dst) {
; CHECK-LABEL: vst4_v16i32:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, r5}
; CHECK-NEXT:    push {r4, r5}
; 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 #216
; CHECK-NEXT:    sub sp, #216
; CHECK-NEXT:    vldrw.u32 q2, [r0, #144]
; CHECK-NEXT:    add r2, sp, #64
; CHECK-NEXT:    vldrw.u32 q1, [r0, #80]
; CHECK-NEXT:    vldrw.u32 q6, [r0, #176]
; CHECK-NEXT:    vldrw.u32 q0, [r0, #16]
; CHECK-NEXT:    vstmia r2, {d0, d1, d2, d3, d4, d5, d6, d7} @ 64-byte Spill
; CHECK-NEXT:    add r2, sp, #128
; CHECK-NEXT:    vldrw.u32 q2, [r0, #128]
; CHECK-NEXT:    vstmia r2, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Spill
; CHECK-NEXT:    add r2, sp, #64
; CHECK-NEXT:    vldrw.u32 q1, [r0, #64]
; CHECK-NEXT:    vldmia r2, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Reload
; CHECK-NEXT:    add r2, sp, #64
; CHECK-NEXT:    vmov q7, q6
; CHECK-NEXT:    vldrw.u32 q0, [r0]
; CHECK-NEXT:    vstmia r2, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Spill
; CHECK-NEXT:    add r2, sp, #128
; CHECK-NEXT:    vldrw.u32 q6, [r0, #160]
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vstmia sp, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Spill
; CHECK-NEXT:    vldmia r2, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Reload
; CHECK-NEXT:    add r2, sp, #128
; CHECK-NEXT:    vldrw.u32 q5, [r0, #112]
; CHECK-NEXT:    vstmia r2, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Spill
; CHECK-NEXT:    add r2, sp, #128
; CHECK-NEXT:    vldmia sp, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Reload
; CHECK-NEXT:    vldrw.u32 q5, [r0, #96]
; CHECK-NEXT:    vstmia sp, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Spill
; CHECK-NEXT:    vldmia r2, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Reload
; CHECK-NEXT:    add r2, sp, #128
; CHECK-NEXT:    vldrw.u32 q4, [r0, #48]
; CHECK-NEXT:    vstmia r2, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Spill
; CHECK-NEXT:    add r2, sp, #64
; CHECK-NEXT:    vldmia sp, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Reload
; CHECK-NEXT:    vldrw.u32 q4, [r0, #32]
; CHECK-NEXT:    mov r0, r1
; CHECK-NEXT:    vst40.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst41.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst42.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst43.32 {q0, q1, q2, q3}, [r0]!
; CHECK-NEXT:    vldmia r2, {d0, d1, d2, d3, d4, d5, d6, d7} @ 64-byte Reload
; CHECK-NEXT:    add r2, sp, #128
; CHECK-NEXT:    vmov q7, q6
; CHECK-NEXT:    vst40.32 {q0, q1, q2, q3}, [r0]
; CHECK-NEXT:    vst41.32 {q0, q1, q2, q3}, [r0]
; CHECK-NEXT:    vst42.32 {q0, q1, q2, q3}, [r0]
; CHECK-NEXT:    vst43.32 {q0, q1, q2, q3}, [r0]
; CHECK-NEXT:    vldmia r2, {d0, d1, d2, d3, d4, d5, d6, d7} @ 64-byte Reload
; CHECK-NEXT:    add.w r0, r1, #192
; CHECK-NEXT:    adds r1, #128
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vst40.32 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst41.32 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst42.32 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst43.32 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst40.32 {q0, q1, q2, q3}, [r0]
; CHECK-NEXT:    vst41.32 {q0, q1, q2, q3}, [r0]
; CHECK-NEXT:    vst42.32 {q0, q1, q2, q3}, [r0]
; CHECK-NEXT:    vst43.32 {q0, q1, q2, q3}, [r0]
; CHECK-NEXT:    add sp, #216
; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    pop {r4, r5}
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <16 x i32>, <16 x i32>* %src, i32 0
  %l1 = load <16 x i32>, <16 x i32>* %s1, align 4
  %s2 = getelementptr <16 x i32>, <16 x i32>* %src, i32 1
  %l2 = load <16 x i32>, <16 x i32>* %s2, align 4
  %s3 = getelementptr <16 x i32>, <16 x i32>* %src, i32 2
  %l3 = load <16 x i32>, <16 x i32>* %s3, align 4
  %s4 = getelementptr <16 x i32>, <16 x i32>* %src, i32 3
  %l4 = load <16 x i32>, <16 x i32>* %s3, align 4
  %t1 = shufflevector <16 x i32> %l1, <16 x i32> %l2, <32 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, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
  %t2 = shufflevector <16 x i32> %l3, <16 x i32> %l4, <32 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, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
  %s = shufflevector <32 x i32> %t1, <32 x i32> %t2, <64 x i32> <i32 0, i32 16, i32 32, i32 48, i32 1, i32 17, i32 33, i32 49, i32 2, i32 18, i32 34, i32 50, i32 3, i32 19, i32 35, i32 51, i32 4, i32 20, i32 36, i32 52, i32 5, i32 21, i32 37, i32 53, i32 6, i32 22, i32 38, i32 54, i32 7, i32 23, i32 39, i32 55, i32 8, i32 24, i32 40, i32 56, i32 9, i32 25, i32 41, i32 57, i32 10, i32 26, i32 42, i32 58, i32 11, i32 27, i32 43, i32 59, i32 12, i32 28, i32 44, i32 60, i32 13, i32 29, i32 45, i32 61, i32 14, i32 30, i32 46, i32 62, i32 15, i32 31, i32 47, i32 63>
  store <64 x i32> %s, <64 x i32> *%dst, align 4
  ret void
}

define void @vst4_v4i32_align1(<4 x i32> *%src, <16 x i32> *%dst) {
; CHECK-LABEL: vst4_v4i32_align1:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldrw.u32 q2, [r0, #32]
; CHECK-NEXT:    vldrw.u32 q1, [r0, #16]
; CHECK-NEXT:    vldrw.u32 q0, [r0]
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vst40.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst41.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst42.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst43.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <4 x i32>, <4 x i32>* %src, i32 0
  %l1 = load <4 x i32>, <4 x i32>* %s1, align 4
  %s2 = getelementptr <4 x i32>, <4 x i32>* %src, i32 1
  %l2 = load <4 x i32>, <4 x i32>* %s2, align 4
  %s3 = getelementptr <4 x i32>, <4 x i32>* %src, i32 2
  %l3 = load <4 x i32>, <4 x i32>* %s3, align 4
  %s4 = getelementptr <4 x i32>, <4 x i32>* %src, i32 3
  %l4 = load <4 x i32>, <4 x i32>* %s3, align 4
  %t1 = shufflevector <4 x i32> %l1, <4 x i32> %l2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %t2 = shufflevector <4 x i32> %l3, <4 x i32> %l4, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %s = shufflevector <8 x i32> %t1, <8 x i32> %t2, <16 x i32> <i32 0, i32 4, i32 8, i32 12, i32 1, i32 5, i32 9, i32 13, i32 2, i32 6, i32 10, i32 14, i32 3, i32 7, i32 11, i32 15>
  store <16 x i32> %s, <16 x i32> *%dst, align 1
  ret void
}

; i16

define void @vst4_v2i16(<2 x i16> *%src, <8 x i16> *%dst) {
; CHECK-LABEL: vst4_v2i16:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, lr}
; CHECK-NEXT:    push {r4, lr}
; CHECK-NEXT:    ldrh r2, [r0, #2]
; CHECK-NEXT:    ldrh.w r12, [r0, #4]
; CHECK-NEXT:    ldrh r3, [r0, #8]
; CHECK-NEXT:    ldrh.w lr, [r0, #6]
; CHECK-NEXT:    ldrh r4, [r0, #10]
; CHECK-NEXT:    ldrh r0, [r0]
; CHECK-NEXT:    vmov q0[2], q0[0], r0, r2
; CHECK-NEXT:    vmov r0, s0
; CHECK-NEXT:    vmov.16 q0[0], r0
; CHECK-NEXT:    vmov.16 q0[1], r12
; CHECK-NEXT:    vmov.16 q0[2], r3
; CHECK-NEXT:    vmov.16 q0[3], r3
; CHECK-NEXT:    vmov.16 q0[4], r2
; CHECK-NEXT:    vmov.16 q0[5], lr
; CHECK-NEXT:    vmov.16 q0[6], r4
; CHECK-NEXT:    vmov.16 q0[7], r4
; CHECK-NEXT:    vstrh.16 q0, [r1]
; CHECK-NEXT:    pop {r4, pc}
entry:
  %s1 = getelementptr <2 x i16>, <2 x i16>* %src, i32 0
  %l1 = load <2 x i16>, <2 x i16>* %s1, align 4
  %s2 = getelementptr <2 x i16>, <2 x i16>* %src, i32 1
  %l2 = load <2 x i16>, <2 x i16>* %s2, align 4
  %s3 = getelementptr <2 x i16>, <2 x i16>* %src, i32 2
  %l3 = load <2 x i16>, <2 x i16>* %s3, align 4
  %s4 = getelementptr <2 x i16>, <2 x i16>* %src, i32 3
  %l4 = load <2 x i16>, <2 x i16>* %s3, align 4
  %t1 = shufflevector <2 x i16> %l1, <2 x i16> %l2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %t2 = shufflevector <2 x i16> %l3, <2 x i16> %l4, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %s = shufflevector <4 x i16> %t1, <4 x i16> %t2, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 1, i32 3, i32 5, i32 7>
  store <8 x i16> %s, <8 x i16> *%dst, align 2
  ret void
}

define void @vst4_v4i16(<4 x i16> *%src, <16 x i16> *%dst) {
; CHECK-LABEL: vst4_v4i16:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .vsave {d8, d9}
; CHECK-NEXT:    vpush {d8, d9}
; CHECK-NEXT:    vldrh.u32 q1, [r0]
; CHECK-NEXT:    vldrh.u32 q2, [r0, #8]
; CHECK-NEXT:    vldrh.u32 q3, [r0, #16]
; CHECK-NEXT:    vmov r2, s6
; CHECK-NEXT:    vmov.16 q0[0], r2
; CHECK-NEXT:    vmov r2, s10
; CHECK-NEXT:    vmov.16 q0[1], r2
; CHECK-NEXT:    vmov r0, s14
; CHECK-NEXT:    vmov.16 q0[2], r0
; CHECK-NEXT:    vmov.16 q0[3], r0
; CHECK-NEXT:    vmov r0, s7
; CHECK-NEXT:    vmov.16 q0[4], r0
; CHECK-NEXT:    vmov r0, s11
; CHECK-NEXT:    vmov.16 q0[5], r0
; CHECK-NEXT:    vmov r0, s15
; CHECK-NEXT:    vmov.16 q0[6], r0
; CHECK-NEXT:    vmov.16 q0[7], r0
; CHECK-NEXT:    vmov r0, s4
; CHECK-NEXT:    vmov.16 q4[0], r0
; CHECK-NEXT:    vmov r0, s8
; CHECK-NEXT:    vmov.16 q4[1], r0
; CHECK-NEXT:    vmov r0, s12
; CHECK-NEXT:    vmov.16 q4[2], r0
; CHECK-NEXT:    vstrh.16 q0, [r1, #16]
; CHECK-NEXT:    vmov.16 q4[3], r0
; CHECK-NEXT:    vmov r0, s5
; CHECK-NEXT:    vmov.16 q4[4], r0
; CHECK-NEXT:    vmov r0, s9
; CHECK-NEXT:    vmov.16 q4[5], r0
; CHECK-NEXT:    vmov r0, s13
; CHECK-NEXT:    vmov.16 q4[6], r0
; CHECK-NEXT:    vmov.16 q4[7], r0
; CHECK-NEXT:    vstrh.16 q4, [r1]
; CHECK-NEXT:    vpop {d8, d9}
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <4 x i16>, <4 x i16>* %src, i32 0
  %l1 = load <4 x i16>, <4 x i16>* %s1, align 4
  %s2 = getelementptr <4 x i16>, <4 x i16>* %src, i32 1
  %l2 = load <4 x i16>, <4 x i16>* %s2, align 4
  %s3 = getelementptr <4 x i16>, <4 x i16>* %src, i32 2
  %l3 = load <4 x i16>, <4 x i16>* %s3, align 4
  %s4 = getelementptr <4 x i16>, <4 x i16>* %src, i32 3
  %l4 = load <4 x i16>, <4 x i16>* %s3, align 4
  %t1 = shufflevector <4 x i16> %l1, <4 x i16> %l2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %t2 = shufflevector <4 x i16> %l3, <4 x i16> %l4, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %s = shufflevector <8 x i16> %t1, <8 x i16> %t2, <16 x i32> <i32 0, i32 4, i32 8, i32 12, i32 1, i32 5, i32 9, i32 13, i32 2, i32 6, i32 10, i32 14, i32 3, i32 7, i32 11, i32 15>
  store <16 x i16> %s, <16 x i16> *%dst, align 2
  ret void
}

define void @vst4_v8i16(<8 x i16> *%src, <32 x i16> *%dst) {
; CHECK-LABEL: vst4_v8i16:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldrw.u32 q2, [r0, #32]
; CHECK-NEXT:    vldrw.u32 q1, [r0, #16]
; CHECK-NEXT:    vldrw.u32 q0, [r0]
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vst40.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst41.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst42.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst43.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <8 x i16>, <8 x i16>* %src, i32 0
  %l1 = load <8 x i16>, <8 x i16>* %s1, align 4
  %s2 = getelementptr <8 x i16>, <8 x i16>* %src, i32 1
  %l2 = load <8 x i16>, <8 x i16>* %s2, align 4
  %s3 = getelementptr <8 x i16>, <8 x i16>* %src, i32 2
  %l3 = load <8 x i16>, <8 x i16>* %s3, align 4
  %s4 = getelementptr <8 x i16>, <8 x i16>* %src, i32 3
  %l4 = load <8 x i16>, <8 x i16>* %s3, align 4
  %t1 = shufflevector <8 x i16> %l1, <8 x i16> %l2, <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>
  %t2 = shufflevector <8 x i16> %l3, <8 x i16> %l4, <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>
  %s = shufflevector <16 x i16> %t1, <16 x i16> %t2, <32 x i32> <i32 0, i32 8, i32 16, i32 24, i32 1, i32 9, i32 17, i32 25, i32 2, i32 10, i32 18, i32 26, i32 3, i32 11, i32 19, i32 27, i32 4, i32 12, i32 20, i32 28, i32 5, i32 13, i32 21, i32 29, i32 6, i32 14, i32 22, i32 30, i32 7, i32 15, i32 23, i32 31>
  store <32 x i16> %s, <32 x i16> *%dst, align 2
  ret void
}

define void @vst4_v16i16(<16 x i16> *%src, <64 x i16> *%dst) {
; CHECK-LABEL: vst4_v16i16:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .vsave {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    vpush {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    vldrw.u32 q6, [r0, #64]
; CHECK-NEXT:    vldrw.u32 q2, [r0, #80]
; CHECK-NEXT:    vldrw.u32 q5, [r0, #32]
; CHECK-NEXT:    vldrw.u32 q1, [r0, #48]
; CHECK-NEXT:    vldrw.u32 q4, [r0]
; CHECK-NEXT:    vldrw.u32 q0, [r0, #16]
; CHECK-NEXT:    vmov q7, q6
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vst40.16 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst41.16 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst42.16 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst43.16 {q4, q5, q6, q7}, [r1]!
; CHECK-NEXT:    vst40.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst41.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst42.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst43.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <16 x i16>, <16 x i16>* %src, i32 0
  %l1 = load <16 x i16>, <16 x i16>* %s1, align 4
  %s2 = getelementptr <16 x i16>, <16 x i16>* %src, i32 1
  %l2 = load <16 x i16>, <16 x i16>* %s2, align 4
  %s3 = getelementptr <16 x i16>, <16 x i16>* %src, i32 2
  %l3 = load <16 x i16>, <16 x i16>* %s3, align 4
  %s4 = getelementptr <16 x i16>, <16 x i16>* %src, i32 3
  %l4 = load <16 x i16>, <16 x i16>* %s3, align 4
  %t1 = shufflevector <16 x i16> %l1, <16 x i16> %l2, <32 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, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
  %t2 = shufflevector <16 x i16> %l3, <16 x i16> %l4, <32 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, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
  %s = shufflevector <32 x i16> %t1, <32 x i16> %t2, <64 x i32> <i32 0, i32 16, i32 32, i32 48, i32 1, i32 17, i32 33, i32 49, i32 2, i32 18, i32 34, i32 50, i32 3, i32 19, i32 35, i32 51, i32 4, i32 20, i32 36, i32 52, i32 5, i32 21, i32 37, i32 53, i32 6, i32 22, i32 38, i32 54, i32 7, i32 23, i32 39, i32 55, i32 8, i32 24, i32 40, i32 56, i32 9, i32 25, i32 41, i32 57, i32 10, i32 26, i32 42, i32 58, i32 11, i32 27, i32 43, i32 59, i32 12, i32 28, i32 44, i32 60, i32 13, i32 29, i32 45, i32 61, i32 14, i32 30, i32 46, i32 62, i32 15, i32 31, i32 47, i32 63>
  store <64 x i16> %s, <64 x i16> *%dst, align 2
  ret void
}

define void @vst4_v8i16_align1(<8 x i16> *%src, <32 x i16> *%dst) {
; CHECK-LABEL: vst4_v8i16_align1:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldrw.u32 q2, [r0, #32]
; CHECK-NEXT:    vldrw.u32 q1, [r0, #16]
; CHECK-NEXT:    vldrw.u32 q0, [r0]
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vst40.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst41.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst42.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst43.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <8 x i16>, <8 x i16>* %src, i32 0
  %l1 = load <8 x i16>, <8 x i16>* %s1, align 4
  %s2 = getelementptr <8 x i16>, <8 x i16>* %src, i32 1
  %l2 = load <8 x i16>, <8 x i16>* %s2, align 4
  %s3 = getelementptr <8 x i16>, <8 x i16>* %src, i32 2
  %l3 = load <8 x i16>, <8 x i16>* %s3, align 4
  %s4 = getelementptr <8 x i16>, <8 x i16>* %src, i32 3
  %l4 = load <8 x i16>, <8 x i16>* %s3, align 4
  %t1 = shufflevector <8 x i16> %l1, <8 x i16> %l2, <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>
  %t2 = shufflevector <8 x i16> %l3, <8 x i16> %l4, <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>
  %s = shufflevector <16 x i16> %t1, <16 x i16> %t2, <32 x i32> <i32 0, i32 8, i32 16, i32 24, i32 1, i32 9, i32 17, i32 25, i32 2, i32 10, i32 18, i32 26, i32 3, i32 11, i32 19, i32 27, i32 4, i32 12, i32 20, i32 28, i32 5, i32 13, i32 21, i32 29, i32 6, i32 14, i32 22, i32 30, i32 7, i32 15, i32 23, i32 31>
  store <32 x i16> %s, <32 x i16> *%dst, align 1
  ret void
}

; i8

define void @vst4_v2i8(<2 x i8> *%src, <8 x i8> *%dst) {
; CHECK-LABEL: vst4_v2i8:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, lr}
; CHECK-NEXT:    push {r4, lr}
; CHECK-NEXT:    ldrb r2, [r0]
; CHECK-NEXT:    ldrb r3, [r0, #1]
; CHECK-NEXT:    ldrb.w r12, [r0, #2]
; CHECK-NEXT:    vmov q0[2], q0[0], r2, r3
; CHECK-NEXT:    ldrb.w lr, [r0, #3]
; CHECK-NEXT:    vmov r2, s0
; CHECK-NEXT:    ldrb r4, [r0, #5]
; CHECK-NEXT:    vmov.16 q0[0], r2
; CHECK-NEXT:    ldrb r0, [r0, #4]
; CHECK-NEXT:    vmov.16 q0[1], r12
; CHECK-NEXT:    vmov.16 q0[2], r0
; CHECK-NEXT:    vmov.16 q0[3], r0
; CHECK-NEXT:    vmov.16 q0[4], r3
; CHECK-NEXT:    vmov.16 q0[5], lr
; CHECK-NEXT:    vmov.16 q0[6], r4
; CHECK-NEXT:    vmov.16 q0[7], r4
; CHECK-NEXT:    vstrb.16 q0, [r1]
; CHECK-NEXT:    pop {r4, pc}
entry:
  %s1 = getelementptr <2 x i8>, <2 x i8>* %src, i32 0
  %l1 = load <2 x i8>, <2 x i8>* %s1, align 4
  %s2 = getelementptr <2 x i8>, <2 x i8>* %src, i32 1
  %l2 = load <2 x i8>, <2 x i8>* %s2, align 4
  %s3 = getelementptr <2 x i8>, <2 x i8>* %src, i32 2
  %l3 = load <2 x i8>, <2 x i8>* %s3, align 4
  %s4 = getelementptr <2 x i8>, <2 x i8>* %src, i32 3
  %l4 = load <2 x i8>, <2 x i8>* %s3, align 4
  %t1 = shufflevector <2 x i8> %l1, <2 x i8> %l2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %t2 = shufflevector <2 x i8> %l3, <2 x i8> %l4, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %s = shufflevector <4 x i8> %t1, <4 x i8> %t2, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 1, i32 3, i32 5, i32 7>
  store <8 x i8> %s, <8 x i8> *%dst, align 1
  ret void
}

define void @vst4_v4i8(<4 x i8> *%src, <16 x i8> *%dst) {
; CHECK-LABEL: vst4_v4i8:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldrb.u32 q1, [r0]
; CHECK-NEXT:    vldrb.u32 q2, [r0, #4]
; CHECK-NEXT:    vldrb.u32 q3, [r0, #8]
; CHECK-NEXT:    vmov r2, s4
; CHECK-NEXT:    vmov.8 q0[0], r2
; CHECK-NEXT:    vmov r2, s8
; CHECK-NEXT:    vmov.8 q0[1], r2
; CHECK-NEXT:    vmov r0, s12
; CHECK-NEXT:    vmov.8 q0[2], r0
; CHECK-NEXT:    vmov.8 q0[3], r0
; CHECK-NEXT:    vmov r0, s5
; CHECK-NEXT:    vmov.8 q0[4], r0
; CHECK-NEXT:    vmov r0, s9
; CHECK-NEXT:    vmov.8 q0[5], r0
; CHECK-NEXT:    vmov r0, s13
; CHECK-NEXT:    vmov.8 q0[6], r0
; CHECK-NEXT:    vmov.8 q0[7], r0
; CHECK-NEXT:    vmov r0, s6
; CHECK-NEXT:    vmov.8 q0[8], r0
; CHECK-NEXT:    vmov r0, s10
; CHECK-NEXT:    vmov.8 q0[9], r0
; CHECK-NEXT:    vmov r0, s14
; CHECK-NEXT:    vmov.8 q0[10], r0
; CHECK-NEXT:    vmov.8 q0[11], r0
; CHECK-NEXT:    vmov r0, s7
; CHECK-NEXT:    vmov.8 q0[12], r0
; CHECK-NEXT:    vmov r0, s11
; CHECK-NEXT:    vmov.8 q0[13], r0
; CHECK-NEXT:    vmov r0, s15
; CHECK-NEXT:    vmov.8 q0[14], r0
; CHECK-NEXT:    vmov.8 q0[15], r0
; CHECK-NEXT:    vstrb.8 q0, [r1]
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <4 x i8>, <4 x i8>* %src, i32 0
  %l1 = load <4 x i8>, <4 x i8>* %s1, align 4
  %s2 = getelementptr <4 x i8>, <4 x i8>* %src, i32 1
  %l2 = load <4 x i8>, <4 x i8>* %s2, align 4
  %s3 = getelementptr <4 x i8>, <4 x i8>* %src, i32 2
  %l3 = load <4 x i8>, <4 x i8>* %s3, align 4
  %s4 = getelementptr <4 x i8>, <4 x i8>* %src, i32 3
  %l4 = load <4 x i8>, <4 x i8>* %s3, align 4
  %t1 = shufflevector <4 x i8> %l1, <4 x i8> %l2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %t2 = shufflevector <4 x i8> %l3, <4 x i8> %l4, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %s = shufflevector <8 x i8> %t1, <8 x i8> %t2, <16 x i32> <i32 0, i32 4, i32 8, i32 12, i32 1, i32 5, i32 9, i32 13, i32 2, i32 6, i32 10, i32 14, i32 3, i32 7, i32 11, i32 15>
  store <16 x i8> %s, <16 x i8> *%dst, align 1
  ret void
}

define void @vst4_v8i8(<8 x i8> *%src, <32 x i8> *%dst) {
; CHECK-LABEL: vst4_v8i8:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .vsave {d8, d9}
; CHECK-NEXT:    vpush {d8, d9}
; CHECK-NEXT:    vldrb.u16 q1, [r0]
; CHECK-NEXT:    vldrb.u16 q2, [r0, #8]
; CHECK-NEXT:    vldrb.u16 q3, [r0, #16]
; CHECK-NEXT:    vmov.u16 r2, q1[4]
; CHECK-NEXT:    vmov.8 q0[0], r2
; CHECK-NEXT:    vmov.u16 r2, q2[4]
; CHECK-NEXT:    vmov.8 q0[1], r2
; CHECK-NEXT:    vmov.u16 r0, q3[4]
; CHECK-NEXT:    vmov.8 q0[2], r0
; CHECK-NEXT:    vmov.8 q0[3], r0
; CHECK-NEXT:    vmov.u16 r0, q1[5]
; CHECK-NEXT:    vmov.8 q0[4], r0
; CHECK-NEXT:    vmov.u16 r0, q2[5]
; CHECK-NEXT:    vmov.8 q0[5], r0
; CHECK-NEXT:    vmov.u16 r0, q3[5]
; CHECK-NEXT:    vmov.8 q0[6], r0
; CHECK-NEXT:    vmov.8 q0[7], r0
; CHECK-NEXT:    vmov.u16 r0, q1[6]
; CHECK-NEXT:    vmov.8 q0[8], r0
; CHECK-NEXT:    vmov.u16 r0, q2[6]
; CHECK-NEXT:    vmov.8 q0[9], r0
; CHECK-NEXT:    vmov.u16 r0, q3[6]
; CHECK-NEXT:    vmov.8 q0[10], r0
; CHECK-NEXT:    vmov.8 q0[11], r0
; CHECK-NEXT:    vmov.u16 r0, q1[7]
; CHECK-NEXT:    vmov.8 q0[12], r0
; CHECK-NEXT:    vmov.u16 r0, q2[7]
; CHECK-NEXT:    vmov.8 q0[13], r0
; CHECK-NEXT:    vmov.u16 r0, q3[7]
; CHECK-NEXT:    vmov.8 q0[14], r0
; CHECK-NEXT:    vmov.8 q0[15], r0
; CHECK-NEXT:    vmov.u16 r0, q1[0]
; CHECK-NEXT:    vmov.8 q4[0], r0
; CHECK-NEXT:    vmov.u16 r0, q2[0]
; CHECK-NEXT:    vmov.8 q4[1], r0
; CHECK-NEXT:    vmov.u16 r0, q3[0]
; CHECK-NEXT:    vmov.8 q4[2], r0
; CHECK-NEXT:    vstrb.8 q0, [r1, #16]
; CHECK-NEXT:    vmov.8 q4[3], r0
; CHECK-NEXT:    vmov.u16 r0, q1[1]
; CHECK-NEXT:    vmov.8 q4[4], r0
; CHECK-NEXT:    vmov.u16 r0, q2[1]
; CHECK-NEXT:    vmov.8 q4[5], r0
; CHECK-NEXT:    vmov.u16 r0, q3[1]
; CHECK-NEXT:    vmov.8 q4[6], r0
; CHECK-NEXT:    vmov.8 q4[7], r0
; CHECK-NEXT:    vmov.u16 r0, q1[2]
; CHECK-NEXT:    vmov.8 q4[8], r0
; CHECK-NEXT:    vmov.u16 r0, q2[2]
; CHECK-NEXT:    vmov.8 q4[9], r0
; CHECK-NEXT:    vmov.u16 r0, q3[2]
; CHECK-NEXT:    vmov.8 q4[10], r0
; CHECK-NEXT:    vmov.8 q4[11], r0
; CHECK-NEXT:    vmov.u16 r0, q1[3]
; CHECK-NEXT:    vmov.8 q4[12], r0
; CHECK-NEXT:    vmov.u16 r0, q2[3]
; CHECK-NEXT:    vmov.8 q4[13], r0
; CHECK-NEXT:    vmov.u16 r0, q3[3]
; CHECK-NEXT:    vmov.8 q4[14], r0
; CHECK-NEXT:    vmov.8 q4[15], r0
; CHECK-NEXT:    vstrb.8 q4, [r1]
; CHECK-NEXT:    vpop {d8, d9}
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <8 x i8>, <8 x i8>* %src, i32 0
  %l1 = load <8 x i8>, <8 x i8>* %s1, align 4
  %s2 = getelementptr <8 x i8>, <8 x i8>* %src, i32 1
  %l2 = load <8 x i8>, <8 x i8>* %s2, align 4
  %s3 = getelementptr <8 x i8>, <8 x i8>* %src, i32 2
  %l3 = load <8 x i8>, <8 x i8>* %s3, align 4
  %s4 = getelementptr <8 x i8>, <8 x i8>* %src, i32 3
  %l4 = load <8 x i8>, <8 x i8>* %s3, align 4
  %t1 = shufflevector <8 x i8> %l1, <8 x i8> %l2, <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>
  %t2 = shufflevector <8 x i8> %l3, <8 x i8> %l4, <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>
  %s = shufflevector <16 x i8> %t1, <16 x i8> %t2, <32 x i32> <i32 0, i32 8, i32 16, i32 24, i32 1, i32 9, i32 17, i32 25, i32 2, i32 10, i32 18, i32 26, i32 3, i32 11, i32 19, i32 27, i32 4, i32 12, i32 20, i32 28, i32 5, i32 13, i32 21, i32 29, i32 6, i32 14, i32 22, i32 30, i32 7, i32 15, i32 23, i32 31>
  store <32 x i8> %s, <32 x i8> *%dst, align 1
  ret void
}

define void @vst4_v16i8(<16 x i8> *%src, <64 x i8> *%dst) {
; CHECK-LABEL: vst4_v16i8:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldrw.u32 q2, [r0, #32]
; CHECK-NEXT:    vldrw.u32 q1, [r0, #16]
; CHECK-NEXT:    vldrw.u32 q0, [r0]
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vst40.8 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst41.8 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst42.8 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst43.8 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <16 x i8>, <16 x i8>* %src, i32 0
  %l1 = load <16 x i8>, <16 x i8>* %s1, align 4
  %s2 = getelementptr <16 x i8>, <16 x i8>* %src, i32 1
  %l2 = load <16 x i8>, <16 x i8>* %s2, align 4
  %s3 = getelementptr <16 x i8>, <16 x i8>* %src, i32 2
  %l3 = load <16 x i8>, <16 x i8>* %s3, align 4
  %s4 = getelementptr <16 x i8>, <16 x i8>* %src, i32 3
  %l4 = load <16 x i8>, <16 x i8>* %s3, align 4
  %t1 = shufflevector <16 x i8> %l1, <16 x i8> %l2, <32 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, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
  %t2 = shufflevector <16 x i8> %l3, <16 x i8> %l4, <32 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, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
  %s = shufflevector <32 x i8> %t1, <32 x i8> %t2, <64 x i32> <i32 0, i32 16, i32 32, i32 48, i32 1, i32 17, i32 33, i32 49, i32 2, i32 18, i32 34, i32 50, i32 3, i32 19, i32 35, i32 51, i32 4, i32 20, i32 36, i32 52, i32 5, i32 21, i32 37, i32 53, i32 6, i32 22, i32 38, i32 54, i32 7, i32 23, i32 39, i32 55, i32 8, i32 24, i32 40, i32 56, i32 9, i32 25, i32 41, i32 57, i32 10, i32 26, i32 42, i32 58, i32 11, i32 27, i32 43, i32 59, i32 12, i32 28, i32 44, i32 60, i32 13, i32 29, i32 45, i32 61, i32 14, i32 30, i32 46, i32 62, i32 15, i32 31, i32 47, i32 63>
  store <64 x i8> %s, <64 x i8> *%dst, align 1
  ret void
}

; i64

define void @vst4_v2i64(<2 x i64> *%src, <8 x i64> *%dst) {
; CHECK-LABEL: vst4_v2i64:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldrw.u32 q3, [r0]
; CHECK-NEXT:    vldrw.u32 q0, [r0, #16]
; CHECK-NEXT:    vldrw.u32 q1, [r0, #32]
; CHECK-NEXT:    vmov.f64 d4, d6
; CHECK-NEXT:    vmov.f32 s9, s13
; CHECK-NEXT:    vmov.f32 s10, s0
; CHECK-NEXT:    vmov.f32 s11, s1
; CHECK-NEXT:    vmov.f32 s0, s14
; CHECK-NEXT:    vstrw.32 q2, [r1]
; CHECK-NEXT:    vmov.f32 s1, s15
; CHECK-NEXT:    vmov.f64 d6, d2
; CHECK-NEXT:    vstrw.32 q0, [r1, #32]
; CHECK-NEXT:    vmov.f32 s13, s5
; CHECK-NEXT:    vmov.f32 s14, s4
; CHECK-NEXT:    vmov.f32 s15, s5
; CHECK-NEXT:    vmov.f32 s4, s6
; CHECK-NEXT:    vstrw.32 q3, [r1, #16]
; CHECK-NEXT:    vmov.f32 s5, s7
; CHECK-NEXT:    vstrw.32 q1, [r1, #48]
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <2 x i64>, <2 x i64>* %src, i32 0
  %l1 = load <2 x i64>, <2 x i64>* %s1, align 4
  %s2 = getelementptr <2 x i64>, <2 x i64>* %src, i32 1
  %l2 = load <2 x i64>, <2 x i64>* %s2, align 4
  %s3 = getelementptr <2 x i64>, <2 x i64>* %src, i32 2
  %l3 = load <2 x i64>, <2 x i64>* %s3, align 4
  %s4 = getelementptr <2 x i64>, <2 x i64>* %src, i32 3
  %l4 = load <2 x i64>, <2 x i64>* %s3, align 4
  %t1 = shufflevector <2 x i64> %l1, <2 x i64> %l2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %t2 = shufflevector <2 x i64> %l3, <2 x i64> %l4, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %s = shufflevector <4 x i64> %t1, <4 x i64> %t2, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 1, i32 3, i32 5, i32 7>
  store <8 x i64> %s, <8 x i64> *%dst, align 8
  ret void
}

define void @vst4_v4i64(<4 x i64> *%src, <16 x i64> *%dst) {
; CHECK-LABEL: vst4_v4i64:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .vsave {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    vpush {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    vldrw.u32 q5, [r0]
; CHECK-NEXT:    vldrw.u32 q0, [r0, #32]
; CHECK-NEXT:    vldrw.u32 q6, [r0, #16]
; CHECK-NEXT:    vldrw.u32 q2, [r0, #48]
; CHECK-NEXT:    vmov.f64 d6, d10
; CHECK-NEXT:    vldrw.u32 q4, [r0, #80]
; CHECK-NEXT:    vldrw.u32 q1, [r0, #64]
; CHECK-NEXT:    vmov.f32 s13, s21
; CHECK-NEXT:    vmov.f32 s14, s0
; CHECK-NEXT:    vmov.f32 s15, s1
; CHECK-NEXT:    vmov.f32 s0, s22
; CHECK-NEXT:    vstrw.32 q3, [r1]
; CHECK-NEXT:    vmov.f32 s1, s23
; CHECK-NEXT:    vmov.f64 d10, d12
; CHECK-NEXT:    vstrw.32 q0, [r1, #32]
; CHECK-NEXT:    vmov.f32 s21, s25
; CHECK-NEXT:    vmov.f32 s22, s8
; CHECK-NEXT:    vmov.f32 s23, s9
; CHECK-NEXT:    vmov.f32 s8, s26
; CHECK-NEXT:    vstrw.32 q5, [r1, #64]
; CHECK-NEXT:    vmov.f32 s9, s27
; CHECK-NEXT:    vmov.f64 d12, d2
; CHECK-NEXT:    vstrw.32 q2, [r1, #96]
; CHECK-NEXT:    vmov.f64 d14, d8
; CHECK-NEXT:    vmov.f32 s25, s5
; CHECK-NEXT:    vmov.f32 s29, s17
; CHECK-NEXT:    vmov.f32 s26, s4
; CHECK-NEXT:    vmov.f32 s30, s16
; CHECK-NEXT:    vmov.f32 s27, s5
; CHECK-NEXT:    vmov.f32 s4, s6
; CHECK-NEXT:    vstrw.32 q6, [r1, #16]
; CHECK-NEXT:    vmov.f32 s31, s17
; CHECK-NEXT:    vmov.f32 s16, s18
; CHECK-NEXT:    vstrw.32 q7, [r1, #80]
; CHECK-NEXT:    vmov.f32 s5, s7
; CHECK-NEXT:    vmov.f32 s17, s19
; CHECK-NEXT:    vstrw.32 q1, [r1, #48]
; CHECK-NEXT:    vstrw.32 q4, [r1, #112]
; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <4 x i64>, <4 x i64>* %src, i32 0
  %l1 = load <4 x i64>, <4 x i64>* %s1, align 4
  %s2 = getelementptr <4 x i64>, <4 x i64>* %src, i32 1
  %l2 = load <4 x i64>, <4 x i64>* %s2, align 4
  %s3 = getelementptr <4 x i64>, <4 x i64>* %src, i32 2
  %l3 = load <4 x i64>, <4 x i64>* %s3, align 4
  %s4 = getelementptr <4 x i64>, <4 x i64>* %src, i32 3
  %l4 = load <4 x i64>, <4 x i64>* %s3, align 4
  %t1 = shufflevector <4 x i64> %l1, <4 x i64> %l2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %t2 = shufflevector <4 x i64> %l3, <4 x i64> %l4, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %s = shufflevector <8 x i64> %t1, <8 x i64> %t2, <16 x i32> <i32 0, i32 4, i32 8, i32 12, i32 1, i32 5, i32 9, i32 13, i32 2, i32 6, i32 10, i32 14, i32 3, i32 7, i32 11, i32 15>
  store <16 x i64> %s, <16 x i64> *%dst, align 8
  ret void
}

; f32

define void @vst4_v2f32(<2 x float> *%src, <8 x float> *%dst) {
; CHECK-LABEL: vst4_v2f32:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldr s0, [r0]
; CHECK-NEXT:    vldr s4, [r0, #4]
; CHECK-NEXT:    vldr s1, [r0, #8]
; CHECK-NEXT:    vldr s5, [r0, #12]
; CHECK-NEXT:    vldr s2, [r0, #16]
; CHECK-NEXT:    vldr s6, [r0, #20]
; CHECK-NEXT:    vmov.f32 s3, s2
; CHECK-NEXT:    vmov.f32 s7, s6
; CHECK-NEXT:    vstrw.32 q0, [r1]
; CHECK-NEXT:    vstrw.32 q1, [r1, #16]
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <2 x float>, <2 x float>* %src, i32 0
  %l1 = load <2 x float>, <2 x float>* %s1, align 4
  %s2 = getelementptr <2 x float>, <2 x float>* %src, i32 1
  %l2 = load <2 x float>, <2 x float>* %s2, align 4
  %s3 = getelementptr <2 x float>, <2 x float>* %src, i32 2
  %l3 = load <2 x float>, <2 x float>* %s3, align 4
  %s4 = getelementptr <2 x float>, <2 x float>* %src, i32 3
  %l4 = load <2 x float>, <2 x float>* %s3, align 4
  %t1 = shufflevector <2 x float> %l1, <2 x float> %l2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %t2 = shufflevector <2 x float> %l3, <2 x float> %l4, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %s = shufflevector <4 x float> %t1, <4 x float> %t2, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 1, i32 3, i32 5, i32 7>
  store <8 x float> %s, <8 x float> *%dst, align 4
  ret void
}

define void @vst4_v4f32(<4 x float> *%src, <16 x float> *%dst) {
; CHECK-LABEL: vst4_v4f32:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldrw.u32 q2, [r0, #32]
; CHECK-NEXT:    vldrw.u32 q1, [r0, #16]
; CHECK-NEXT:    vldrw.u32 q0, [r0]
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vst40.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst41.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst42.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst43.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <4 x float>, <4 x float>* %src, i32 0
  %l1 = load <4 x float>, <4 x float>* %s1, align 4
  %s2 = getelementptr <4 x float>, <4 x float>* %src, i32 1
  %l2 = load <4 x float>, <4 x float>* %s2, align 4
  %s3 = getelementptr <4 x float>, <4 x float>* %src, i32 2
  %l3 = load <4 x float>, <4 x float>* %s3, align 4
  %s4 = getelementptr <4 x float>, <4 x float>* %src, i32 3
  %l4 = load <4 x float>, <4 x float>* %s3, align 4
  %t1 = shufflevector <4 x float> %l1, <4 x float> %l2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %t2 = shufflevector <4 x float> %l3, <4 x float> %l4, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %s = shufflevector <8 x float> %t1, <8 x float> %t2, <16 x i32> <i32 0, i32 4, i32 8, i32 12, i32 1, i32 5, i32 9, i32 13, i32 2, i32 6, i32 10, i32 14, i32 3, i32 7, i32 11, i32 15>
  store <16 x float> %s, <16 x float> *%dst, align 4
  ret void
}

define void @vst4_v8f32(<8 x float> *%src, <32 x float> *%dst) {
; CHECK-LABEL: vst4_v8f32:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .vsave {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    vpush {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    vldrw.u32 q6, [r0, #64]
; CHECK-NEXT:    vldrw.u32 q2, [r0, #80]
; CHECK-NEXT:    vldrw.u32 q5, [r0, #32]
; CHECK-NEXT:    vldrw.u32 q1, [r0, #48]
; CHECK-NEXT:    vldrw.u32 q4, [r0]
; CHECK-NEXT:    vldrw.u32 q0, [r0, #16]
; CHECK-NEXT:    vmov q7, q6
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vst40.32 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst41.32 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst42.32 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst43.32 {q4, q5, q6, q7}, [r1]!
; CHECK-NEXT:    vst40.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst41.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst42.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst43.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <8 x float>, <8 x float>* %src, i32 0
  %l1 = load <8 x float>, <8 x float>* %s1, align 4
  %s2 = getelementptr <8 x float>, <8 x float>* %src, i32 1
  %l2 = load <8 x float>, <8 x float>* %s2, align 4
  %s3 = getelementptr <8 x float>, <8 x float>* %src, i32 2
  %l3 = load <8 x float>, <8 x float>* %s3, align 4
  %s4 = getelementptr <8 x float>, <8 x float>* %src, i32 3
  %l4 = load <8 x float>, <8 x float>* %s3, align 4
  %t1 = shufflevector <8 x float> %l1, <8 x float> %l2, <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>
  %t2 = shufflevector <8 x float> %l3, <8 x float> %l4, <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>
  %s = shufflevector <16 x float> %t1, <16 x float> %t2, <32 x i32> <i32 0, i32 8, i32 16, i32 24, i32 1, i32 9, i32 17, i32 25, i32 2, i32 10, i32 18, i32 26, i32 3, i32 11, i32 19, i32 27, i32 4, i32 12, i32 20, i32 28, i32 5, i32 13, i32 21, i32 29, i32 6, i32 14, i32 22, i32 30, i32 7, i32 15, i32 23, i32 31>
  store <32 x float> %s, <32 x float> *%dst, align 4
  ret void
}

define void @vst4_v16f32(<16 x float> *%src, <64 x float> *%dst) {
; CHECK-LABEL: vst4_v16f32:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, r5}
; CHECK-NEXT:    push {r4, r5}
; 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 #216
; CHECK-NEXT:    sub sp, #216
; CHECK-NEXT:    vldrw.u32 q2, [r0, #144]
; CHECK-NEXT:    add r2, sp, #64
; CHECK-NEXT:    vldrw.u32 q1, [r0, #80]
; CHECK-NEXT:    vldrw.u32 q6, [r0, #176]
; CHECK-NEXT:    vldrw.u32 q0, [r0, #16]
; CHECK-NEXT:    vstmia r2, {d0, d1, d2, d3, d4, d5, d6, d7} @ 64-byte Spill
; CHECK-NEXT:    add r2, sp, #128
; CHECK-NEXT:    vldrw.u32 q2, [r0, #128]
; CHECK-NEXT:    vstmia r2, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Spill
; CHECK-NEXT:    add r2, sp, #64
; CHECK-NEXT:    vldrw.u32 q1, [r0, #64]
; CHECK-NEXT:    vldmia r2, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Reload
; CHECK-NEXT:    add r2, sp, #64
; CHECK-NEXT:    vmov q7, q6
; CHECK-NEXT:    vldrw.u32 q0, [r0]
; CHECK-NEXT:    vstmia r2, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Spill
; CHECK-NEXT:    add r2, sp, #128
; CHECK-NEXT:    vldrw.u32 q6, [r0, #160]
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vstmia sp, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Spill
; CHECK-NEXT:    vldmia r2, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Reload
; CHECK-NEXT:    add r2, sp, #128
; CHECK-NEXT:    vldrw.u32 q5, [r0, #112]
; CHECK-NEXT:    vstmia r2, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Spill
; CHECK-NEXT:    add r2, sp, #128
; CHECK-NEXT:    vldmia sp, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Reload
; CHECK-NEXT:    vldrw.u32 q5, [r0, #96]
; CHECK-NEXT:    vstmia sp, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Spill
; CHECK-NEXT:    vldmia r2, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Reload
; CHECK-NEXT:    add r2, sp, #128
; CHECK-NEXT:    vldrw.u32 q4, [r0, #48]
; CHECK-NEXT:    vstmia r2, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Spill
; CHECK-NEXT:    add r2, sp, #64
; CHECK-NEXT:    vldmia sp, {d8, d9, d10, d11, d12, d13, d14, d15} @ 64-byte Reload
; CHECK-NEXT:    vldrw.u32 q4, [r0, #32]
; CHECK-NEXT:    mov r0, r1
; CHECK-NEXT:    vst40.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst41.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst42.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst43.32 {q0, q1, q2, q3}, [r0]!
; CHECK-NEXT:    vldmia r2, {d0, d1, d2, d3, d4, d5, d6, d7} @ 64-byte Reload
; CHECK-NEXT:    add r2, sp, #128
; CHECK-NEXT:    vmov q7, q6
; CHECK-NEXT:    vst40.32 {q0, q1, q2, q3}, [r0]
; CHECK-NEXT:    vst41.32 {q0, q1, q2, q3}, [r0]
; CHECK-NEXT:    vst42.32 {q0, q1, q2, q3}, [r0]
; CHECK-NEXT:    vst43.32 {q0, q1, q2, q3}, [r0]
; CHECK-NEXT:    vldmia r2, {d0, d1, d2, d3, d4, d5, d6, d7} @ 64-byte Reload
; CHECK-NEXT:    add.w r0, r1, #192
; CHECK-NEXT:    adds r1, #128
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vst40.32 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst41.32 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst42.32 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst43.32 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst40.32 {q0, q1, q2, q3}, [r0]
; CHECK-NEXT:    vst41.32 {q0, q1, q2, q3}, [r0]
; CHECK-NEXT:    vst42.32 {q0, q1, q2, q3}, [r0]
; CHECK-NEXT:    vst43.32 {q0, q1, q2, q3}, [r0]
; CHECK-NEXT:    add sp, #216
; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    pop {r4, r5}
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <16 x float>, <16 x float>* %src, i32 0
  %l1 = load <16 x float>, <16 x float>* %s1, align 4
  %s2 = getelementptr <16 x float>, <16 x float>* %src, i32 1
  %l2 = load <16 x float>, <16 x float>* %s2, align 4
  %s3 = getelementptr <16 x float>, <16 x float>* %src, i32 2
  %l3 = load <16 x float>, <16 x float>* %s3, align 4
  %s4 = getelementptr <16 x float>, <16 x float>* %src, i32 3
  %l4 = load <16 x float>, <16 x float>* %s3, align 4
  %t1 = shufflevector <16 x float> %l1, <16 x float> %l2, <32 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, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
  %t2 = shufflevector <16 x float> %l3, <16 x float> %l4, <32 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, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
  %s = shufflevector <32 x float> %t1, <32 x float> %t2, <64 x i32> <i32 0, i32 16, i32 32, i32 48, i32 1, i32 17, i32 33, i32 49, i32 2, i32 18, i32 34, i32 50, i32 3, i32 19, i32 35, i32 51, i32 4, i32 20, i32 36, i32 52, i32 5, i32 21, i32 37, i32 53, i32 6, i32 22, i32 38, i32 54, i32 7, i32 23, i32 39, i32 55, i32 8, i32 24, i32 40, i32 56, i32 9, i32 25, i32 41, i32 57, i32 10, i32 26, i32 42, i32 58, i32 11, i32 27, i32 43, i32 59, i32 12, i32 28, i32 44, i32 60, i32 13, i32 29, i32 45, i32 61, i32 14, i32 30, i32 46, i32 62, i32 15, i32 31, i32 47, i32 63>
  store <64 x float> %s, <64 x float> *%dst, align 4
  ret void
}

define void @vst4_v4f32_align1(<4 x float> *%src, <16 x float> *%dst) {
; CHECK-LABEL: vst4_v4f32_align1:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldrw.u32 q2, [r0, #32]
; CHECK-NEXT:    vldrw.u32 q1, [r0, #16]
; CHECK-NEXT:    vldrw.u32 q0, [r0]
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vst40.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst41.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst42.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst43.32 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <4 x float>, <4 x float>* %src, i32 0
  %l1 = load <4 x float>, <4 x float>* %s1, align 4
  %s2 = getelementptr <4 x float>, <4 x float>* %src, i32 1
  %l2 = load <4 x float>, <4 x float>* %s2, align 4
  %s3 = getelementptr <4 x float>, <4 x float>* %src, i32 2
  %l3 = load <4 x float>, <4 x float>* %s3, align 4
  %s4 = getelementptr <4 x float>, <4 x float>* %src, i32 3
  %l4 = load <4 x float>, <4 x float>* %s3, align 4
  %t1 = shufflevector <4 x float> %l1, <4 x float> %l2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %t2 = shufflevector <4 x float> %l3, <4 x float> %l4, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %s = shufflevector <8 x float> %t1, <8 x float> %t2, <16 x i32> <i32 0, i32 4, i32 8, i32 12, i32 1, i32 5, i32 9, i32 13, i32 2, i32 6, i32 10, i32 14, i32 3, i32 7, i32 11, i32 15>
  store <16 x float> %s, <16 x float> *%dst, align 1
  ret void
}

; f16

define void @vst4_v2f16(<2 x half> *%src, <8 x half> *%dst) {
; CHECK-LABEL: vst4_v2f16:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldmia r0, {s4, s5}
; CHECK-NEXT:    vmov r2, s5
; CHECK-NEXT:    ldr r0, [r0, #8]
; CHECK-NEXT:    vmov r3, s4
; CHECK-NEXT:    vmovx.f16 s12, s4
; CHECK-NEXT:    vmov.16 q0[0], r3
; CHECK-NEXT:    vdup.32 q2, r0
; CHECK-NEXT:    vmov.16 q0[1], r2
; CHECK-NEXT:    vmov r0, s8
; CHECK-NEXT:    vmov.16 q0[2], r0
; CHECK-NEXT:    vmov r0, s9
; CHECK-NEXT:    vmov.16 q0[3], r0
; CHECK-NEXT:    vmov r0, s12
; CHECK-NEXT:    vmovx.f16 s4, s5
; CHECK-NEXT:    vmov.16 q0[4], r0
; CHECK-NEXT:    vmov r0, s4
; CHECK-NEXT:    vmovx.f16 s4, s8
; CHECK-NEXT:    vmov.16 q0[5], r0
; CHECK-NEXT:    vmov r0, s4
; CHECK-NEXT:    vmovx.f16 s4, s9
; CHECK-NEXT:    vmov.16 q0[6], r0
; CHECK-NEXT:    vmov r0, s4
; CHECK-NEXT:    vmov.16 q0[7], r0
; CHECK-NEXT:    vstrh.16 q0, [r1]
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <2 x half>, <2 x half>* %src, i32 0
  %l1 = load <2 x half>, <2 x half>* %s1, align 4
  %s2 = getelementptr <2 x half>, <2 x half>* %src, i32 1
  %l2 = load <2 x half>, <2 x half>* %s2, align 4
  %s3 = getelementptr <2 x half>, <2 x half>* %src, i32 2
  %l3 = load <2 x half>, <2 x half>* %s3, align 4
  %s4 = getelementptr <2 x half>, <2 x half>* %src, i32 3
  %l4 = load <2 x half>, <2 x half>* %s3, align 4
  %t1 = shufflevector <2 x half> %l1, <2 x half> %l2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %t2 = shufflevector <2 x half> %l3, <2 x half> %l4, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %s = shufflevector <4 x half> %t1, <4 x half> %t2, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 1, i32 3, i32 5, i32 7>
  store <8 x half> %s, <8 x half> *%dst, align 2
  ret void
}

define void @vst4_v4f16(<4 x half> *%src, <16 x half> *%dst) {
; CHECK-LABEL: vst4_v4f16:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, lr}
; CHECK-NEXT:    push {r4, lr}
; CHECK-NEXT:    ldrd lr, r12, [r0]
; CHECK-NEXT:    ldrd r3, r2, [r0, #8]
; CHECK-NEXT:    ldrd r4, r0, [r0, #16]
; CHECK-NEXT:    vmov q1[2], q1[0], lr, r3
; CHECK-NEXT:    vmov q1[3], q1[1], r12, r2
; CHECK-NEXT:    vmov q0[2], q0[0], r4, r4
; CHECK-NEXT:    vmov r3, s5
; CHECK-NEXT:    vmov q0[3], q0[1], r0, r0
; CHECK-NEXT:    vmov r2, s7
; CHECK-NEXT:    vmov.16 q2[0], r3
; CHECK-NEXT:    vmov.16 q2[1], r2
; CHECK-NEXT:    vmov r0, s1
; CHECK-NEXT:    vmov.16 q2[2], r0
; CHECK-NEXT:    vmov r0, s3
; CHECK-NEXT:    vmovx.f16 s12, s5
; CHECK-NEXT:    vmov.16 q2[3], r0
; CHECK-NEXT:    vmov r0, s12
; CHECK-NEXT:    vmovx.f16 s12, s7
; CHECK-NEXT:    vmov.16 q2[4], r0
; CHECK-NEXT:    vmov r0, s12
; CHECK-NEXT:    vmovx.f16 s12, s1
; CHECK-NEXT:    vmov.16 q2[5], r0
; CHECK-NEXT:    vmov r0, s12
; CHECK-NEXT:    vmovx.f16 s12, s3
; CHECK-NEXT:    vmov.16 q2[6], r0
; CHECK-NEXT:    vmov r0, s12
; CHECK-NEXT:    vmov.16 q2[7], r0
; CHECK-NEXT:    vmov r2, s4
; CHECK-NEXT:    vstrh.16 q2, [r1, #16]
; CHECK-NEXT:    vmov r0, s6
; CHECK-NEXT:    vmov.16 q2[0], r2
; CHECK-NEXT:    vmovx.f16 s12, s4
; CHECK-NEXT:    vmov.16 q2[1], r0
; CHECK-NEXT:    vmov r0, s0
; CHECK-NEXT:    vmov.16 q2[2], r0
; CHECK-NEXT:    vmov r0, s2
; CHECK-NEXT:    vmov.16 q2[3], r0
; CHECK-NEXT:    vmov r0, s12
; CHECK-NEXT:    vmovx.f16 s4, s6
; CHECK-NEXT:    vmov.16 q2[4], r0
; CHECK-NEXT:    vmov r0, s4
; CHECK-NEXT:    vmovx.f16 s4, s0
; CHECK-NEXT:    vmov.16 q2[5], r0
; CHECK-NEXT:    vmov r0, s4
; CHECK-NEXT:    vmovx.f16 s0, s2
; CHECK-NEXT:    vmov.16 q2[6], r0
; CHECK-NEXT:    vmov r0, s0
; CHECK-NEXT:    vmov.16 q2[7], r0
; CHECK-NEXT:    vstrh.16 q2, [r1]
; CHECK-NEXT:    pop {r4, pc}
entry:
  %s1 = getelementptr <4 x half>, <4 x half>* %src, i32 0
  %l1 = load <4 x half>, <4 x half>* %s1, align 4
  %s2 = getelementptr <4 x half>, <4 x half>* %src, i32 1
  %l2 = load <4 x half>, <4 x half>* %s2, align 4
  %s3 = getelementptr <4 x half>, <4 x half>* %src, i32 2
  %l3 = load <4 x half>, <4 x half>* %s3, align 4
  %s4 = getelementptr <4 x half>, <4 x half>* %src, i32 3
  %l4 = load <4 x half>, <4 x half>* %s3, align 4
  %t1 = shufflevector <4 x half> %l1, <4 x half> %l2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %t2 = shufflevector <4 x half> %l3, <4 x half> %l4, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %s = shufflevector <8 x half> %t1, <8 x half> %t2, <16 x i32> <i32 0, i32 4, i32 8, i32 12, i32 1, i32 5, i32 9, i32 13, i32 2, i32 6, i32 10, i32 14, i32 3, i32 7, i32 11, i32 15>
  store <16 x half> %s, <16 x half> *%dst, align 2
  ret void
}

define void @vst4_v8f16(<8 x half> *%src, <32 x half> *%dst) {
; CHECK-LABEL: vst4_v8f16:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldrw.u32 q2, [r0, #32]
; CHECK-NEXT:    vldrw.u32 q1, [r0, #16]
; CHECK-NEXT:    vldrw.u32 q0, [r0]
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vst40.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst41.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst42.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst43.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <8 x half>, <8 x half>* %src, i32 0
  %l1 = load <8 x half>, <8 x half>* %s1, align 4
  %s2 = getelementptr <8 x half>, <8 x half>* %src, i32 1
  %l2 = load <8 x half>, <8 x half>* %s2, align 4
  %s3 = getelementptr <8 x half>, <8 x half>* %src, i32 2
  %l3 = load <8 x half>, <8 x half>* %s3, align 4
  %s4 = getelementptr <8 x half>, <8 x half>* %src, i32 3
  %l4 = load <8 x half>, <8 x half>* %s3, align 4
  %t1 = shufflevector <8 x half> %l1, <8 x half> %l2, <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>
  %t2 = shufflevector <8 x half> %l3, <8 x half> %l4, <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>
  %s = shufflevector <16 x half> %t1, <16 x half> %t2, <32 x i32> <i32 0, i32 8, i32 16, i32 24, i32 1, i32 9, i32 17, i32 25, i32 2, i32 10, i32 18, i32 26, i32 3, i32 11, i32 19, i32 27, i32 4, i32 12, i32 20, i32 28, i32 5, i32 13, i32 21, i32 29, i32 6, i32 14, i32 22, i32 30, i32 7, i32 15, i32 23, i32 31>
  store <32 x half> %s, <32 x half> *%dst, align 2
  ret void
}

define void @vst4_v16f16(<16 x half> *%src, <64 x half> *%dst) {
; CHECK-LABEL: vst4_v16f16:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .vsave {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    vpush {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    vldrw.u32 q6, [r0, #64]
; CHECK-NEXT:    vldrw.u32 q2, [r0, #80]
; CHECK-NEXT:    vldrw.u32 q5, [r0, #32]
; CHECK-NEXT:    vldrw.u32 q1, [r0, #48]
; CHECK-NEXT:    vldrw.u32 q4, [r0]
; CHECK-NEXT:    vldrw.u32 q0, [r0, #16]
; CHECK-NEXT:    vmov q7, q6
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vst40.16 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst41.16 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst42.16 {q4, q5, q6, q7}, [r1]
; CHECK-NEXT:    vst43.16 {q4, q5, q6, q7}, [r1]!
; CHECK-NEXT:    vst40.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst41.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst42.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst43.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <16 x half>, <16 x half>* %src, i32 0
  %l1 = load <16 x half>, <16 x half>* %s1, align 4
  %s2 = getelementptr <16 x half>, <16 x half>* %src, i32 1
  %l2 = load <16 x half>, <16 x half>* %s2, align 4
  %s3 = getelementptr <16 x half>, <16 x half>* %src, i32 2
  %l3 = load <16 x half>, <16 x half>* %s3, align 4
  %s4 = getelementptr <16 x half>, <16 x half>* %src, i32 3
  %l4 = load <16 x half>, <16 x half>* %s3, align 4
  %t1 = shufflevector <16 x half> %l1, <16 x half> %l2, <32 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, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
  %t2 = shufflevector <16 x half> %l3, <16 x half> %l4, <32 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, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
  %s = shufflevector <32 x half> %t1, <32 x half> %t2, <64 x i32> <i32 0, i32 16, i32 32, i32 48, i32 1, i32 17, i32 33, i32 49, i32 2, i32 18, i32 34, i32 50, i32 3, i32 19, i32 35, i32 51, i32 4, i32 20, i32 36, i32 52, i32 5, i32 21, i32 37, i32 53, i32 6, i32 22, i32 38, i32 54, i32 7, i32 23, i32 39, i32 55, i32 8, i32 24, i32 40, i32 56, i32 9, i32 25, i32 41, i32 57, i32 10, i32 26, i32 42, i32 58, i32 11, i32 27, i32 43, i32 59, i32 12, i32 28, i32 44, i32 60, i32 13, i32 29, i32 45, i32 61, i32 14, i32 30, i32 46, i32 62, i32 15, i32 31, i32 47, i32 63>
  store <64 x half> %s, <64 x half> *%dst, align 2
  ret void
}

define void @vst4_v8f16_align1(<8 x half> *%src, <32 x half> *%dst) {
; CHECK-LABEL: vst4_v8f16_align1:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldrw.u32 q2, [r0, #32]
; CHECK-NEXT:    vldrw.u32 q1, [r0, #16]
; CHECK-NEXT:    vldrw.u32 q0, [r0]
; CHECK-NEXT:    vmov q3, q2
; CHECK-NEXT:    vst40.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst41.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst42.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    vst43.16 {q0, q1, q2, q3}, [r1]
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <8 x half>, <8 x half>* %src, i32 0
  %l1 = load <8 x half>, <8 x half>* %s1, align 4
  %s2 = getelementptr <8 x half>, <8 x half>* %src, i32 1
  %l2 = load <8 x half>, <8 x half>* %s2, align 4
  %s3 = getelementptr <8 x half>, <8 x half>* %src, i32 2
  %l3 = load <8 x half>, <8 x half>* %s3, align 4
  %s4 = getelementptr <8 x half>, <8 x half>* %src, i32 3
  %l4 = load <8 x half>, <8 x half>* %s3, align 4
  %t1 = shufflevector <8 x half> %l1, <8 x half> %l2, <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>
  %t2 = shufflevector <8 x half> %l3, <8 x half> %l4, <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>
  %s = shufflevector <16 x half> %t1, <16 x half> %t2, <32 x i32> <i32 0, i32 8, i32 16, i32 24, i32 1, i32 9, i32 17, i32 25, i32 2, i32 10, i32 18, i32 26, i32 3, i32 11, i32 19, i32 27, i32 4, i32 12, i32 20, i32 28, i32 5, i32 13, i32 21, i32 29, i32 6, i32 14, i32 22, i32 30, i32 7, i32 15, i32 23, i32 31>
  store <32 x half> %s, <32 x half> *%dst, align 1
  ret void
}

; f64

define void @vst4_v2f64(<2 x double> *%src, <8 x double> *%dst) {
; CHECK-LABEL: vst4_v2f64:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vldrw.u32 q1, [r0]
; CHECK-NEXT:    vldrw.u32 q0, [r0, #16]
; CHECK-NEXT:    vldrw.u32 q2, [r0, #32]
; CHECK-NEXT:    vmov.f64 d6, d2
; CHECK-NEXT:    vmov.f64 d7, d0
; CHECK-NEXT:    vmov.f64 d0, d3
; CHECK-NEXT:    vstrw.32 q3, [r1]
; CHECK-NEXT:    vmov.f64 d2, d4
; CHECK-NEXT:    vstrw.32 q0, [r1, #32]
; CHECK-NEXT:    vmov.f64 d3, d4
; CHECK-NEXT:    vmov.f64 d4, d5
; CHECK-NEXT:    vstrw.32 q1, [r1, #16]
; CHECK-NEXT:    vstrw.32 q2, [r1, #48]
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <2 x double>, <2 x double>* %src, i32 0
  %l1 = load <2 x double>, <2 x double>* %s1, align 4
  %s2 = getelementptr <2 x double>, <2 x double>* %src, i32 1
  %l2 = load <2 x double>, <2 x double>* %s2, align 4
  %s3 = getelementptr <2 x double>, <2 x double>* %src, i32 2
  %l3 = load <2 x double>, <2 x double>* %s3, align 4
  %s4 = getelementptr <2 x double>, <2 x double>* %src, i32 3
  %l4 = load <2 x double>, <2 x double>* %s3, align 4
  %t1 = shufflevector <2 x double> %l1, <2 x double> %l2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %t2 = shufflevector <2 x double> %l3, <2 x double> %l4, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %s = shufflevector <4 x double> %t1, <4 x double> %t2, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 1, i32 3, i32 5, i32 7>
  store <8 x double> %s, <8 x double> *%dst, align 8
  ret void
}

define void @vst4_v4f64(<4 x double> *%src, <16 x double> *%dst) {
; CHECK-LABEL: vst4_v4f64:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .vsave {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    vpush {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    vldrw.u32 q5, [r0]
; CHECK-NEXT:    vldrw.u32 q0, [r0, #32]
; CHECK-NEXT:    vldrw.u32 q6, [r0, #16]
; CHECK-NEXT:    vldrw.u32 q4, [r0, #80]
; CHECK-NEXT:    vmov.f64 d4, d10
; CHECK-NEXT:    vldrw.u32 q3, [r0, #48]
; CHECK-NEXT:    vldrw.u32 q1, [r0, #64]
; CHECK-NEXT:    vmov.f64 d5, d0
; CHECK-NEXT:    vmov.f64 d0, d11
; CHECK-NEXT:    vstrw.32 q2, [r1]
; CHECK-NEXT:    vmov.f64 d10, d12
; CHECK-NEXT:    vstrw.32 q0, [r1, #32]
; CHECK-NEXT:    vmov.f64 d11, d6
; CHECK-NEXT:    vmov.f64 d6, d13
; CHECK-NEXT:    vstrw.32 q5, [r1, #64]
; CHECK-NEXT:    vmov.f64 d12, d2
; CHECK-NEXT:    vstrw.32 q3, [r1, #96]
; CHECK-NEXT:    vmov.f64 d14, d8
; CHECK-NEXT:    vmov.f64 d13, d2
; CHECK-NEXT:    vmov.f64 d15, d8
; CHECK-NEXT:    vstrw.32 q6, [r1, #16]
; CHECK-NEXT:    vmov.f64 d2, d3
; CHECK-NEXT:    vstrw.32 q7, [r1, #80]
; CHECK-NEXT:    vmov.f64 d8, d9
; CHECK-NEXT:    vstrw.32 q1, [r1, #48]
; CHECK-NEXT:    vstrw.32 q4, [r1, #112]
; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13, d14, d15}
; CHECK-NEXT:    bx lr
entry:
  %s1 = getelementptr <4 x double>, <4 x double>* %src, i32 0
  %l1 = load <4 x double>, <4 x double>* %s1, align 4
  %s2 = getelementptr <4 x double>, <4 x double>* %src, i32 1
  %l2 = load <4 x double>, <4 x double>* %s2, align 4
  %s3 = getelementptr <4 x double>, <4 x double>* %src, i32 2
  %l3 = load <4 x double>, <4 x double>* %s3, align 4
  %s4 = getelementptr <4 x double>, <4 x double>* %src, i32 3
  %l4 = load <4 x double>, <4 x double>* %s3, align 4
  %t1 = shufflevector <4 x double> %l1, <4 x double> %l2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %t2 = shufflevector <4 x double> %l3, <4 x double> %l4, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %s = shufflevector <8 x double> %t1, <8 x double> %t2, <16 x i32> <i32 0, i32 4, i32 8, i32 12, i32 1, i32 5, i32 9, i32 13, i32 2, i32 6, i32 10, i32 14, i32 3, i32 7, i32 11, i32 15>
  store <16 x double> %s, <16 x double> *%dst, align 8
  ret void
}