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

define arm_aapcs_vfpcc void @test_fadd(ptr noalias nocapture readonly %A, float %B, ptr noalias nocapture %C, i32 %n) {
; CHECK-LABEL: test_fadd:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    cmp r2, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    bxlt lr
; CHECK-NEXT:  .LBB0_1: @ %vector.ph
; CHECK-NEXT:    vmov r3, s0
; CHECK-NEXT:  .LBB0_2: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q0, [r0], #16
; CHECK-NEXT:    subs r2, #4
; CHECK-NEXT:    vadd.f32 q0, q0, r3
; CHECK-NEXT:    vstrb.8 q0, [r1], #16
; CHECK-NEXT:    bne .LBB0_2
; CHECK-NEXT:  @ %bb.3: @ %for.cond.cleanup
; CHECK-NEXT:    bx lr
entry:
  %i = and i32 %n, 7
  %cmp = icmp eq i32 %i, 0
  tail call void @llvm.assume(i1 %cmp)
  %cmp18 = icmp sgt i32 %n, 0
  br i1 %cmp18, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %entry
  %broadcast.splatinsert10 = insertelement <4 x float> undef, float %B, i32 0
  %broadcast.splat11 = shufflevector <4 x float> %broadcast.splatinsert10, <4 x float> undef, <4 x i32> zeroinitializer
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %i1 = getelementptr inbounds float, ptr %A, i32 %index
  %wide.load = load <4 x float>, ptr %i1, align 4
  %i3 = fadd fast <4 x float> %wide.load, %broadcast.splat11
  %i4 = getelementptr inbounds float, ptr %C, i32 %index
  store <4 x float> %i3, ptr %i4, align 4
  %index.next = add i32 %index, 4
  %i6 = icmp eq i32 %index.next, %n
  br i1 %i6, label %for.cond.cleanup, label %vector.body

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

define arm_aapcs_vfpcc void @test_fadd_r(ptr noalias nocapture readonly %A, float %B, ptr noalias nocapture %C, i32 %n) {
; CHECK-LABEL: test_fadd_r:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    cmp r2, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    bxlt lr
; CHECK-NEXT:  .LBB1_1: @ %vector.ph
; CHECK-NEXT:    vmov r3, s0
; CHECK-NEXT:  .LBB1_2: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q0, [r0], #16
; CHECK-NEXT:    subs r2, #4
; CHECK-NEXT:    vadd.f32 q0, q0, r3
; CHECK-NEXT:    vstrb.8 q0, [r1], #16
; CHECK-NEXT:    bne .LBB1_2
; CHECK-NEXT:  @ %bb.3: @ %for.cond.cleanup
; CHECK-NEXT:    bx lr
entry:
  %i = and i32 %n, 7
  %cmp = icmp eq i32 %i, 0
  tail call void @llvm.assume(i1 %cmp)
  %cmp18 = icmp sgt i32 %n, 0
  br i1 %cmp18, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %entry
  %broadcast.splatinsert10 = insertelement <4 x float> undef, float %B, i32 0
  %broadcast.splat11 = shufflevector <4 x float> %broadcast.splatinsert10, <4 x float> undef, <4 x i32> zeroinitializer
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %i1 = getelementptr inbounds float, ptr %A, i32 %index
  %wide.load = load <4 x float>, ptr %i1, align 4
  %i3 = fadd fast <4 x float> %broadcast.splat11, %wide.load
  %i4 = getelementptr inbounds float, ptr %C, i32 %index
  store <4 x float> %i3, ptr %i4, align 4
  %index.next = add i32 %index, 4
  %i6 = icmp eq i32 %index.next, %n
  br i1 %i6, label %for.cond.cleanup, label %vector.body

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

define arm_aapcs_vfpcc void @test_fmul(ptr noalias nocapture readonly %A, float %B, ptr noalias nocapture %C, i32 %n) {
; CHECK-LABEL: test_fmul:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    cmp r2, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    bxlt lr
; CHECK-NEXT:  .LBB2_1: @ %vector.ph
; CHECK-NEXT:    vmov r3, s0
; CHECK-NEXT:  .LBB2_2: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q0, [r0], #16
; CHECK-NEXT:    subs r2, #4
; CHECK-NEXT:    vmul.f32 q0, q0, r3
; CHECK-NEXT:    vstrb.8 q0, [r1], #16
; CHECK-NEXT:    bne .LBB2_2
; CHECK-NEXT:  @ %bb.3: @ %for.cond.cleanup
; CHECK-NEXT:    bx lr
entry:
  %i = and i32 %n, 7
  %cmp = icmp eq i32 %i, 0
  tail call void @llvm.assume(i1 %cmp)
  %cmp18 = icmp sgt i32 %n, 0
  br i1 %cmp18, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %entry
  %broadcast.splatinsert10 = insertelement <4 x float> undef, float %B, i32 0
  %broadcast.splat11 = shufflevector <4 x float> %broadcast.splatinsert10, <4 x float> undef, <4 x i32> zeroinitializer
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %i1 = getelementptr inbounds float, ptr %A, i32 %index
  %wide.load = load <4 x float>, ptr %i1, align 4
  %i3 = fmul fast <4 x float> %wide.load, %broadcast.splat11
  %i4 = getelementptr inbounds float, ptr %C, i32 %index
  store <4 x float> %i3, ptr %i4, align 4
  %index.next = add i32 %index, 4
  %i6 = icmp eq i32 %index.next, %n
  br i1 %i6, label %for.cond.cleanup, label %vector.body

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

define arm_aapcs_vfpcc void @test_fmul_r(ptr noalias nocapture readonly %A, float %B, ptr noalias nocapture %C, i32 %n) {
; CHECK-LABEL: test_fmul_r:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    cmp r2, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    bxlt lr
; CHECK-NEXT:  .LBB3_1: @ %vector.ph
; CHECK-NEXT:    vmov r3, s0
; CHECK-NEXT:  .LBB3_2: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q0, [r0], #16
; CHECK-NEXT:    subs r2, #4
; CHECK-NEXT:    vmul.f32 q0, q0, r3
; CHECK-NEXT:    vstrb.8 q0, [r1], #16
; CHECK-NEXT:    bne .LBB3_2
; CHECK-NEXT:  @ %bb.3: @ %for.cond.cleanup
; CHECK-NEXT:    bx lr
entry:
  %i = and i32 %n, 7
  %cmp = icmp eq i32 %i, 0
  tail call void @llvm.assume(i1 %cmp)
  %cmp18 = icmp sgt i32 %n, 0
  br i1 %cmp18, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %entry
  %broadcast.splatinsert10 = insertelement <4 x float> undef, float %B, i32 0
  %broadcast.splat11 = shufflevector <4 x float> %broadcast.splatinsert10, <4 x float> undef, <4 x i32> zeroinitializer
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %i1 = getelementptr inbounds float, ptr %A, i32 %index
  %wide.load = load <4 x float>, ptr %i1, align 4
  %i3 = fmul fast <4 x float> %broadcast.splat11, %wide.load
  %i4 = getelementptr inbounds float, ptr %C, i32 %index
  store <4 x float> %i3, ptr %i4, align 4
  %index.next = add i32 %index, 4
  %i6 = icmp eq i32 %index.next, %n
  br i1 %i6, label %for.cond.cleanup, label %vector.body

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

define arm_aapcs_vfpcc void @test_fsub(ptr noalias nocapture readonly %A, float %B, ptr noalias nocapture %C, i32 %n) {
; CHECK-LABEL: test_fsub:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    cmp r2, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    bxlt lr
; CHECK-NEXT:  .LBB4_1: @ %vector.ph
; CHECK-NEXT:    vmov r3, s0
; CHECK-NEXT:  .LBB4_2: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q0, [r0], #16
; CHECK-NEXT:    subs r2, #4
; CHECK-NEXT:    vsub.f32 q0, q0, r3
; CHECK-NEXT:    vstrb.8 q0, [r1], #16
; CHECK-NEXT:    bne .LBB4_2
; CHECK-NEXT:  @ %bb.3: @ %for.cond.cleanup
; CHECK-NEXT:    bx lr
entry:
  %i = and i32 %n, 7
  %cmp = icmp eq i32 %i, 0
  tail call void @llvm.assume(i1 %cmp)
  %cmp18 = icmp sgt i32 %n, 0
  br i1 %cmp18, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %entry
  %broadcast.splatinsert10 = insertelement <4 x float> undef, float %B, i32 0
  %broadcast.splat11 = shufflevector <4 x float> %broadcast.splatinsert10, <4 x float> undef, <4 x i32> zeroinitializer
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %i1 = getelementptr inbounds float, ptr %A, i32 %index
  %wide.load = load <4 x float>, ptr %i1, align 4
  %i3 = fsub fast <4 x float> %wide.load, %broadcast.splat11
  %i4 = getelementptr inbounds float, ptr %C, i32 %index
  store <4 x float> %i3, ptr %i4, align 4
  %index.next = add i32 %index, 4
  %i6 = icmp eq i32 %index.next, %n
  br i1 %i6, label %for.cond.cleanup, label %vector.body

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

define arm_aapcs_vfpcc void @test_fsub_r(ptr noalias nocapture readonly %A, float %B, ptr noalias nocapture %C, i32 %n) {
; CHECK-LABEL: test_fsub_r:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    cmp r2, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    bxlt lr
; CHECK-NEXT:  .LBB5_1: @ %vector.ph
; CHECK-NEXT:    vmov r3, s0
; CHECK-NEXT:    vdup.32 q0, r3
; CHECK-NEXT:  .LBB5_2: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q1, [r0], #16
; CHECK-NEXT:    subs r2, #4
; CHECK-NEXT:    vsub.f32 q1, q0, q1
; CHECK-NEXT:    vstrb.8 q1, [r1], #16
; CHECK-NEXT:    bne .LBB5_2
; CHECK-NEXT:  @ %bb.3: @ %for.cond.cleanup
; CHECK-NEXT:    bx lr
entry:
  %i = and i32 %n, 7
  %cmp = icmp eq i32 %i, 0
  tail call void @llvm.assume(i1 %cmp)
  %cmp18 = icmp sgt i32 %n, 0
  br i1 %cmp18, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %entry
  %broadcast.splatinsert10 = insertelement <4 x float> undef, float %B, i32 0
  %broadcast.splat11 = shufflevector <4 x float> %broadcast.splatinsert10, <4 x float> undef, <4 x i32> zeroinitializer
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %i1 = getelementptr inbounds float, ptr %A, i32 %index
  %wide.load = load <4 x float>, ptr %i1, align 4
  %i3 = fsub fast <4 x float> %broadcast.splat11, %wide.load
  %i4 = getelementptr inbounds float, ptr %C, i32 %index
  store <4 x float> %i3, ptr %i4, align 4
  %index.next = add i32 %index, 4
  %i6 = icmp eq i32 %index.next, %n
  br i1 %i6, label %for.cond.cleanup, label %vector.body

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


define arm_aapcs_vfpcc void @test_fmas(ptr noalias nocapture readonly %A, ptr noalias nocapture readonly %B, float %C, ptr noalias nocapture %D, i32 %n) {
; CHECK-LABEL: test_fmas:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    cmp r3, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    bxlt lr
; CHECK-NEXT:  .LBB6_1: @ %vector.ph
; CHECK-NEXT:    vmov r12, s0
; CHECK-NEXT:  .LBB6_2: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q0, [r0], #16
; CHECK-NEXT:    vldrw.u32 q1, [r1], #16
; CHECK-NEXT:    subs r3, #4
; CHECK-NEXT:    vfmas.f32 q1, q0, r12
; CHECK-NEXT:    vstrb.8 q1, [r2], #16
; CHECK-NEXT:    bne .LBB6_2
; CHECK-NEXT:  @ %bb.3: @ %for.cond.cleanup
; CHECK-NEXT:    bx lr
entry:
  %i = and i32 %n, 7
  %cmp = icmp eq i32 %i, 0
  tail call void @llvm.assume(i1 %cmp)
  %cmp110 = icmp sgt i32 %n, 0
  br i1 %cmp110, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %entry
  %broadcast.splatinsert13 = insertelement <4 x float> undef, float %C, i32 0
  %broadcast.splat14 = shufflevector <4 x float> %broadcast.splatinsert13, <4 x float> undef, <4 x i32> zeroinitializer
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %i1 = getelementptr inbounds float, ptr %A, i32 %index
  %wide.load = load <4 x float>, ptr %i1, align 4
  %i3 = getelementptr inbounds float, ptr %B, i32 %index
  %wide.load12 = load <4 x float>, ptr %i3, align 4
  %i5 = fmul fast <4 x float> %wide.load12, %wide.load
  %i6 = fadd fast <4 x float> %i5, %broadcast.splat14
  %i7 = getelementptr inbounds float, ptr %D, i32 %index
  store <4 x float> %i6, ptr %i7, align 4
  %index.next = add i32 %index, 4
  %i9 = icmp eq i32 %index.next, %n
  br i1 %i9, label %for.cond.cleanup, label %vector.body

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

define arm_aapcs_vfpcc void @test_fmas_r(ptr noalias nocapture readonly %A, ptr noalias nocapture readonly %B, float %C, ptr noalias nocapture %D, i32 %n) {
; CHECK-LABEL: test_fmas_r:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    cmp r3, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    bxlt lr
; CHECK-NEXT:  .LBB7_1: @ %vector.ph
; CHECK-NEXT:    vmov r12, s0
; CHECK-NEXT:  .LBB7_2: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q0, [r0], #16
; CHECK-NEXT:    vldrw.u32 q1, [r1], #16
; CHECK-NEXT:    subs r3, #4
; CHECK-NEXT:    vfmas.f32 q1, q0, r12
; CHECK-NEXT:    vstrb.8 q1, [r2], #16
; CHECK-NEXT:    bne .LBB7_2
; CHECK-NEXT:  @ %bb.3: @ %for.cond.cleanup
; CHECK-NEXT:    bx lr
entry:
  %i = and i32 %n, 7
  %cmp = icmp eq i32 %i, 0
  tail call void @llvm.assume(i1 %cmp)
  %cmp110 = icmp sgt i32 %n, 0
  br i1 %cmp110, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %entry
  %broadcast.splatinsert13 = insertelement <4 x float> undef, float %C, i32 0
  %broadcast.splat14 = shufflevector <4 x float> %broadcast.splatinsert13, <4 x float> undef, <4 x i32> zeroinitializer
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %i1 = getelementptr inbounds float, ptr %A, i32 %index
  %wide.load = load <4 x float>, ptr %i1, align 4
  %i3 = getelementptr inbounds float, ptr %B, i32 %index
  %wide.load12 = load <4 x float>, ptr %i3, align 4
  %i5 = fmul fast <4 x float> %wide.load12, %wide.load
  %i6 = fadd fast <4 x float> %broadcast.splat14, %i5
  %i7 = getelementptr inbounds float, ptr %D, i32 %index
  store <4 x float> %i6, ptr %i7, align 4
  %index.next = add i32 %index, 4
  %i9 = icmp eq i32 %index.next, %n
  br i1 %i9, label %for.cond.cleanup, label %vector.body

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

define arm_aapcs_vfpcc void @test_fma(ptr noalias nocapture readonly %A, ptr noalias nocapture readonly %B, float %C, ptr noalias nocapture %D, i32 %n) {
; CHECK-LABEL: test_fma:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    cmp r3, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    bxlt lr
; CHECK-NEXT:  .LBB8_1: @ %vector.ph
; CHECK-NEXT:    vmov r12, s0
; CHECK-NEXT:  .LBB8_2: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q0, [r0], #16
; CHECK-NEXT:    vldrw.u32 q1, [r1], #16
; CHECK-NEXT:    subs r3, #4
; CHECK-NEXT:    vfma.f32 q1, q0, r12
; CHECK-NEXT:    vstrb.8 q1, [r2], #16
; CHECK-NEXT:    bne .LBB8_2
; CHECK-NEXT:  @ %bb.3: @ %for.cond.cleanup
; CHECK-NEXT:    bx lr
entry:
  %i = and i32 %n, 7
  %cmp = icmp eq i32 %i, 0
  tail call void @llvm.assume(i1 %cmp)
  %cmp110 = icmp sgt i32 %n, 0
  br i1 %cmp110, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %entry
  %broadcast.splatinsert12 = insertelement <4 x float> undef, float %C, i32 0
  %broadcast.splat13 = shufflevector <4 x float> %broadcast.splatinsert12, <4 x float> undef, <4 x i32> zeroinitializer
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %i1 = getelementptr inbounds float, ptr %A, i32 %index
  %wide.load = load <4 x float>, ptr %i1, align 4
  %i3 = fmul fast <4 x float> %wide.load, %broadcast.splat13
  %i4 = getelementptr inbounds float, ptr %B, i32 %index
  %wide.load14 = load <4 x float>, ptr %i4, align 4
  %i6 = fadd fast <4 x float> %i3, %wide.load14
  %i7 = getelementptr inbounds float, ptr %D, i32 %index
  store <4 x float> %i6, ptr %i7, align 4
  %index.next = add i32 %index, 4
  %i9 = icmp eq i32 %index.next, %n
  br i1 %i9, label %for.cond.cleanup, label %vector.body

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

define arm_aapcs_vfpcc void @test_fma_r(ptr noalias nocapture readonly %A, ptr noalias nocapture readonly %B, float %C, ptr noalias nocapture %D, i32 %n) {
; CHECK-LABEL: test_fma_r:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    cmp r3, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    bxlt lr
; CHECK-NEXT:  .LBB9_1: @ %vector.ph
; CHECK-NEXT:    vmov r12, s0
; CHECK-NEXT:  .LBB9_2: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q0, [r0], #16
; CHECK-NEXT:    vldrw.u32 q1, [r1], #16
; CHECK-NEXT:    subs r3, #4
; CHECK-NEXT:    vfma.f32 q1, q0, r12
; CHECK-NEXT:    vstrb.8 q1, [r2], #16
; CHECK-NEXT:    bne .LBB9_2
; CHECK-NEXT:  @ %bb.3: @ %for.cond.cleanup
; CHECK-NEXT:    bx lr
entry:
  %i = and i32 %n, 7
  %cmp = icmp eq i32 %i, 0
  tail call void @llvm.assume(i1 %cmp)
  %cmp110 = icmp sgt i32 %n, 0
  br i1 %cmp110, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %entry
  %broadcast.splatinsert12 = insertelement <4 x float> undef, float %C, i32 0
  %broadcast.splat13 = shufflevector <4 x float> %broadcast.splatinsert12, <4 x float> undef, <4 x i32> zeroinitializer
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %i1 = getelementptr inbounds float, ptr %A, i32 %index
  %wide.load = load <4 x float>, ptr %i1, align 4
  %i3 = fmul fast <4 x float> %broadcast.splat13, %wide.load
  %i4 = getelementptr inbounds float, ptr %B, i32 %index
  %wide.load14 = load <4 x float>, ptr %i4, align 4
  %i6 = fadd fast <4 x float> %i3, %wide.load14
  %i7 = getelementptr inbounds float, ptr %D, i32 %index
  store <4 x float> %i6, ptr %i7, align 4
  %index.next = add i32 %index, 4
  %i9 = icmp eq i32 %index.next, %n
  br i1 %i9, label %for.cond.cleanup, label %vector.body

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


define arm_aapcs_vfpcc void @test_fmss(ptr noalias nocapture readonly %A, ptr noalias nocapture readonly %B, float %C, ptr noalias nocapture %D, i32 %n) {
; CHECK-LABEL: test_fmss:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    cmp r3, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    bxlt lr
; CHECK-NEXT:  .LBB10_1: @ %vector.ph
; CHECK-NEXT:    vmov r12, s0
; CHECK-NEXT:    vdup.32 q0, r12
; CHECK-NEXT:    vneg.f32 q0, q0
; CHECK-NEXT:  .LBB10_2: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q1, [r0], #16
; CHECK-NEXT:    vldrw.u32 q2, [r1], #16
; CHECK-NEXT:    vmov q3, q0
; CHECK-NEXT:    subs r3, #4
; CHECK-NEXT:    vfma.f32 q3, q2, q1
; CHECK-NEXT:    vstrb.8 q3, [r2], #16
; CHECK-NEXT:    bne .LBB10_2
; CHECK-NEXT:  @ %bb.3: @ %for.cond.cleanup
; CHECK-NEXT:    bx lr
entry:
  %i = and i32 %n, 7
  %cmp = icmp eq i32 %i, 0
  tail call void @llvm.assume(i1 %cmp)
  %cmp110 = icmp sgt i32 %n, 0
  br i1 %cmp110, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %entry
  %broadcast.splatinsert13 = insertelement <4 x float> undef, float %C, i32 0
  %broadcast.splat14 = shufflevector <4 x float> %broadcast.splatinsert13, <4 x float> undef, <4 x i32> zeroinitializer
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %i1 = getelementptr inbounds float, ptr %A, i32 %index
  %wide.load = load <4 x float>, ptr %i1, align 4
  %i3 = getelementptr inbounds float, ptr %B, i32 %index
  %wide.load12 = load <4 x float>, ptr %i3, align 4
  %i5 = fmul fast <4 x float> %wide.load12, %wide.load
  %i6 = fsub fast <4 x float> %i5, %broadcast.splat14
  %i7 = getelementptr inbounds float, ptr %D, i32 %index
  store <4 x float> %i6, ptr %i7, align 4
  %index.next = add i32 %index, 4
  %i9 = icmp eq i32 %index.next, %n
  br i1 %i9, label %for.cond.cleanup, label %vector.body

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

define arm_aapcs_vfpcc void @test_fmss_r(ptr noalias nocapture readonly %A, ptr noalias nocapture readonly %B, float %C, ptr noalias nocapture %D, i32 %n) {
; CHECK-LABEL: test_fmss_r:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    cmp r3, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    bxlt lr
; CHECK-NEXT:  .LBB11_1: @ %vector.ph
; CHECK-NEXT:    vmov r12, s0
; CHECK-NEXT:    vdup.32 q0, r12
; CHECK-NEXT:  .LBB11_2: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q1, [r0], #16
; CHECK-NEXT:    vldrw.u32 q2, [r1], #16
; CHECK-NEXT:    vmov q3, q0
; CHECK-NEXT:    subs r3, #4
; CHECK-NEXT:    vfms.f32 q3, q2, q1
; CHECK-NEXT:    vstrb.8 q3, [r2], #16
; CHECK-NEXT:    bne .LBB11_2
; CHECK-NEXT:  @ %bb.3: @ %for.cond.cleanup
; CHECK-NEXT:    bx lr
entry:
  %i = and i32 %n, 7
  %cmp = icmp eq i32 %i, 0
  tail call void @llvm.assume(i1 %cmp)
  %cmp110 = icmp sgt i32 %n, 0
  br i1 %cmp110, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %entry
  %broadcast.splatinsert13 = insertelement <4 x float> undef, float %C, i32 0
  %broadcast.splat14 = shufflevector <4 x float> %broadcast.splatinsert13, <4 x float> undef, <4 x i32> zeroinitializer
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %i1 = getelementptr inbounds float, ptr %A, i32 %index
  %wide.load = load <4 x float>, ptr %i1, align 4
  %i3 = getelementptr inbounds float, ptr %B, i32 %index
  %wide.load12 = load <4 x float>, ptr %i3, align 4
  %i5 = fmul fast <4 x float> %wide.load12, %wide.load
  %i6 = fsub fast <4 x float> %broadcast.splat14, %i5
  %i7 = getelementptr inbounds float, ptr %D, i32 %index
  store <4 x float> %i6, ptr %i7, align 4
  %index.next = add i32 %index, 4
  %i9 = icmp eq i32 %index.next, %n
  br i1 %i9, label %for.cond.cleanup, label %vector.body

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

define arm_aapcs_vfpcc void @test_fms(ptr noalias nocapture readonly %A, ptr noalias nocapture readonly %B, float %C, ptr noalias nocapture %D, i32 %n) {
; CHECK-LABEL: test_fms:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    cmp r3, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    bxlt lr
; CHECK-NEXT:  .LBB12_1: @ %vector.ph
; CHECK-NEXT:    vmov r12, s0
; CHECK-NEXT:  .LBB12_2: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q0, [r1], #16
; CHECK-NEXT:    vldrw.u32 q1, [r0], #16
; CHECK-NEXT:    subs r3, #4
; CHECK-NEXT:    vneg.f32 q0, q0
; CHECK-NEXT:    vfma.f32 q0, q1, r12
; CHECK-NEXT:    vstrb.8 q0, [r2], #16
; CHECK-NEXT:    bne .LBB12_2
; CHECK-NEXT:  @ %bb.3: @ %for.cond.cleanup
; CHECK-NEXT:    bx lr
entry:
  %i = and i32 %n, 7
  %cmp = icmp eq i32 %i, 0
  tail call void @llvm.assume(i1 %cmp)
  %cmp110 = icmp sgt i32 %n, 0
  br i1 %cmp110, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %entry
  %broadcast.splatinsert12 = insertelement <4 x float> undef, float %C, i32 0
  %broadcast.splat13 = shufflevector <4 x float> %broadcast.splatinsert12, <4 x float> undef, <4 x i32> zeroinitializer
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %i1 = getelementptr inbounds float, ptr %A, i32 %index
  %wide.load = load <4 x float>, ptr %i1, align 4
  %i3 = fmul fast <4 x float> %wide.load, %broadcast.splat13
  %i4 = getelementptr inbounds float, ptr %B, i32 %index
  %wide.load14 = load <4 x float>, ptr %i4, align 4
  %i6 = fsub fast <4 x float> %i3, %wide.load14
  %i7 = getelementptr inbounds float, ptr %D, i32 %index
  store <4 x float> %i6, ptr %i7, align 4
  %index.next = add i32 %index, 4
  %i9 = icmp eq i32 %index.next, %n
  br i1 %i9, label %for.cond.cleanup, label %vector.body

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

define arm_aapcs_vfpcc void @test_fms_r(ptr noalias nocapture readonly %A, ptr noalias nocapture readonly %B, float %C, ptr noalias nocapture %D, i32 %n) {
; CHECK-LABEL: test_fms_r:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    cmp r3, #1
; CHECK-NEXT:    it lt
; CHECK-NEXT:    bxlt lr
; CHECK-NEXT:  .LBB13_1: @ %vector.ph
; CHECK-NEXT:    vmov r12, s0
; CHECK-NEXT:  .LBB13_2: @ %vector.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q0, [r1], #16
; CHECK-NEXT:    vldrw.u32 q1, [r0], #16
; CHECK-NEXT:    subs r3, #4
; CHECK-NEXT:    vneg.f32 q0, q0
; CHECK-NEXT:    vfma.f32 q0, q1, r12
; CHECK-NEXT:    vstrb.8 q0, [r2], #16
; CHECK-NEXT:    bne .LBB13_2
; CHECK-NEXT:  @ %bb.3: @ %for.cond.cleanup
; CHECK-NEXT:    bx lr
entry:
  %i = and i32 %n, 7
  %cmp = icmp eq i32 %i, 0
  tail call void @llvm.assume(i1 %cmp)
  %cmp110 = icmp sgt i32 %n, 0
  br i1 %cmp110, label %vector.ph, label %for.cond.cleanup

vector.ph:                                        ; preds = %entry
  %broadcast.splatinsert12 = insertelement <4 x float> undef, float %C, i32 0
  %broadcast.splat13 = shufflevector <4 x float> %broadcast.splatinsert12, <4 x float> undef, <4 x i32> zeroinitializer
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %vector.ph
  %index = phi i32 [ 0, %vector.ph ], [ %index.next, %vector.body ]
  %i1 = getelementptr inbounds float, ptr %A, i32 %index
  %wide.load = load <4 x float>, ptr %i1, align 4
  %i3 = fmul fast <4 x float> %broadcast.splat13, %wide.load
  %i4 = getelementptr inbounds float, ptr %B, i32 %index
  %wide.load14 = load <4 x float>, ptr %i4, align 4
  %i6 = fsub fast <4 x float> %i3, %wide.load14
  %i7 = getelementptr inbounds float, ptr %D, i32 %index
  store <4 x float> %i6, ptr %i7, align 4
  %index.next = add i32 %index, 4
  %i9 = icmp eq i32 %index.next, %n
  br i1 %i9, label %for.cond.cleanup, label %vector.body

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


define dso_local void @test_nested(ptr noalias nocapture %pInT1, ptr noalias nocapture readonly %pOutT1, ptr noalias nocapture readonly %pPRT_in, ptr noalias nocapture readnone %pPRT_pDst, i32 %numRows, i32 %numCols, i32 %l) local_unnamed_addr {
; CHECK-LABEL: test_nested:
; CHECK:       @ %bb.0: @ %for.body.us.preheader
; CHECK-NEXT:    .save {r4, r5, r6, lr}
; CHECK-NEXT:    push {r4, r5, r6, lr}
; CHECK-NEXT:    ldrd lr, r12, [sp, #16]
; CHECK-NEXT:    lsl.w r3, r12, #2
; CHECK-NEXT:  .LBB14_1: @ %for.body.us
; CHECK-NEXT:    @ =>This Loop Header: Depth=1
; CHECK-NEXT:    @ Child Loop BB14_2 Depth 2
; CHECK-NEXT:    ldr r4, [r1]
; CHECK-NEXT:    mov r5, r2
; CHECK-NEXT:    mov r6, r12
; CHECK-NEXT:    vdup.32 q0, r4
; CHECK-NEXT:    mov r4, r0
; CHECK-NEXT:  .LBB14_2: @ %vector.body
; CHECK-NEXT:    @ Parent Loop BB14_1 Depth=1
; CHECK-NEXT:    @ => This Inner Loop Header: Depth=2
; CHECK-NEXT:    vldrw.u32 q1, [r5], #16
; CHECK-NEXT:    vldrw.u32 q2, [r4]
; CHECK-NEXT:    subs r6, #4
; CHECK-NEXT:    vfms.f32 q2, q1, q0
; CHECK-NEXT:    vstrb.8 q2, [r4], #16
; CHECK-NEXT:    bne .LBB14_2
; CHECK-NEXT:  @ %bb.3: @ %for.cond6.for.end_crit_edge.us
; CHECK-NEXT:    @ in Loop: Header=BB14_1 Depth=1
; CHECK-NEXT:    add r0, r3
; CHECK-NEXT:    add r2, r3
; CHECK-NEXT:    adds r1, #4
; CHECK-NEXT:    le lr, .LBB14_1
; CHECK-NEXT:  @ %bb.4: @ %for.end14
; CHECK-NEXT:    pop {r4, r5, r6, pc}
for.body.us.preheader:
  %cmp = icmp sgt i32 %numRows, 0
  tail call void @llvm.assume(i1 %cmp)
  %cmp1 = icmp sgt i32 %numCols, 0
  tail call void @llvm.assume(i1 %cmp1)
  %rem = and i32 %numCols, 7
  %cmp2 = icmp eq i32 %rem, 0
  tail call void @llvm.assume(i1 %cmp2)
  %cmp3 = icmp slt i32 %l, %numCols
  tail call void @llvm.assume(i1 %cmp3)
  br label %for.body.us

for.body.us:                                      ; preds = %for.cond6.for.end_crit_edge.us, %for.body.us.preheader
  %pInT1.addr.038.us = phi ptr [ %scevgep40, %for.cond6.for.end_crit_edge.us ], [ %pInT1, %for.body.us.preheader ]
  %i.037.us = phi i32 [ %inc13.us, %for.cond6.for.end_crit_edge.us ], [ 0, %for.body.us.preheader ]
  %pOutT1.addr.036.us = phi ptr [ %incdec.ptr.us, %for.cond6.for.end_crit_edge.us ], [ %pOutT1, %for.body.us.preheader ]
  %pPRT_in.addr.035.us = phi ptr [ %scevgep, %for.cond6.for.end_crit_edge.us ], [ %pPRT_in, %for.body.us.preheader ]
  %scevgep = getelementptr float, ptr %pPRT_in.addr.035.us, i32 %numCols
  %i = load float, ptr %pOutT1.addr.036.us, align 4
  %broadcast.splatinsert47 = insertelement <4 x float> undef, float %i, i32 0
  %broadcast.splat48 = shufflevector <4 x float> %broadcast.splatinsert47, <4 x float> undef, <4 x i32> zeroinitializer
  br label %vector.body

vector.body:                                      ; preds = %vector.body, %for.body.us
  %index = phi i32 [ 0, %for.body.us ], [ %index.next, %vector.body ]
  %next.gep = getelementptr float, ptr %pInT1.addr.038.us, i32 %index
  %next.gep45 = getelementptr float, ptr %pPRT_in.addr.035.us, i32 %index
  %wide.load = load <4 x float>, ptr %next.gep, align 4
  %wide.load46 = load <4 x float>, ptr %next.gep45, align 4
  %i3 = fmul fast <4 x float> %wide.load46, %broadcast.splat48
  %i4 = fsub fast <4 x float> %wide.load, %i3
  store <4 x float> %i4, ptr %next.gep, align 4
  %index.next = add i32 %index, 4
  %i5 = icmp eq i32 %index.next, %numCols
  br i1 %i5, label %for.cond6.for.end_crit_edge.us, label %vector.body

for.cond6.for.end_crit_edge.us:                   ; preds = %vector.body
  %incdec.ptr.us = getelementptr inbounds float, ptr %pOutT1.addr.036.us, i32 1
  %scevgep40 = getelementptr float, ptr %pInT1.addr.038.us, i32 %numCols
  %inc13.us = add nuw nsw i32 %i.037.us, 1
  %exitcond41 = icmp eq i32 %inc13.us, %numRows
  br i1 %exitcond41, label %for.end14, label %for.body.us

for.end14:                                        ; preds = %for.cond6.for.end_crit_edge.us
  ret void
}

%struct.arm_fir_instance_f32 = type { i16, ptr, ptr }
define void @arm_fir_f32_1_4_mve(ptr nocapture readonly %S, ptr nocapture readonly %pSrc, ptr %pDst, i32 %blockSize) {
; CHECK-LABEL: arm_fir_f32_1_4_mve:
; 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 #8
; CHECK-NEXT:    sub sp, #8
; CHECK-NEXT:    ldrh.w r10, [r0]
; CHECK-NEXT:    mov r11, r1
; CHECK-NEXT:    ldr.w r12, [r0, #4]
; CHECK-NEXT:    sub.w r1, r10, #1
; CHECK-NEXT:    cmp r1, #3
; CHECK-NEXT:    bhi .LBB15_6
; CHECK-NEXT:  @ %bb.1: @ %if.then
; CHECK-NEXT:    ldr r4, [r0, #8]
; CHECK-NEXT:    ldrd r7, r6, [r4]
; CHECK-NEXT:    ldrd r5, r8, [r4, #8]
; CHECK-NEXT:    add.w r4, r12, r1, lsl #2
; CHECK-NEXT:    lsrs r1, r3, #2
; CHECK-NEXT:    wls lr, r1, .LBB15_5
; CHECK-NEXT:  @ %bb.2: @ %while.body.lr.ph
; CHECK-NEXT:    bic r1, r3, #3
; CHECK-NEXT:    str r1, [sp] @ 4-byte Spill
; CHECK-NEXT:    add.w r9, r12, #4
; CHECK-NEXT:    add.w r1, r2, r1, lsl #2
; CHECK-NEXT:    str r1, [sp, #4] @ 4-byte Spill
; CHECK-NEXT:    mov r1, r11
; CHECK-NEXT:  .LBB15_3: @ %while.body
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q0, [r1], #16
; CHECK-NEXT:    vstrb.8 q0, [r4], #16
; CHECK-NEXT:    vldrw.u32 q0, [r9, #-4]
; CHECK-NEXT:    vldrw.u32 q1, [r9], #16
; CHECK-NEXT:    vmul.f32 q0, q0, r7
; CHECK-NEXT:    vldrw.u32 q2, [r9, #-8]
; CHECK-NEXT:    vfma.f32 q0, q1, r6
; CHECK-NEXT:    vldrw.u32 q1, [r9, #-12]
; CHECK-NEXT:    vfma.f32 q0, q1, r5
; CHECK-NEXT:    vfma.f32 q0, q2, r8
; CHECK-NEXT:    vstrb.8 q0, [r2], #16
; CHECK-NEXT:    le lr, .LBB15_3
; CHECK-NEXT:  @ %bb.4: @ %while.end.loopexit
; CHECK-NEXT:    ldr r1, [sp] @ 4-byte Reload
; CHECK-NEXT:    ldr r2, [sp, #4] @ 4-byte Reload
; CHECK-NEXT:    add.w r12, r12, r1, lsl #2
; CHECK-NEXT:    add.w r11, r11, r1, lsl #2
; CHECK-NEXT:  .LBB15_5: @ %while.end
; CHECK-NEXT:    and r1, r3, #3
; CHECK-NEXT:    vldrw.u32 q0, [r11]
; CHECK-NEXT:    vctp.32 r1
; CHECK-NEXT:    vpst
; CHECK-NEXT:    vstrwt.32 q0, [r4]
; CHECK-NEXT:    vldrw.u32 q0, [r12]
; CHECK-NEXT:    vldrw.u32 q1, [r12, #4]
; CHECK-NEXT:    vmul.f32 q0, q0, r7
; CHECK-NEXT:    vfma.f32 q0, q1, r6
; CHECK-NEXT:    vldrw.u32 q1, [r12, #8]
; CHECK-NEXT:    vfma.f32 q0, q1, r5
; CHECK-NEXT:    vldrw.u32 q1, [r12, #12]
; CHECK-NEXT:    vfma.f32 q0, q1, r8
; CHECK-NEXT:    vpst
; CHECK-NEXT:    vstrwt.32 q0, [r2]
; CHECK-NEXT:    ldr.w r12, [r0, #4]
; CHECK-NEXT:  .LBB15_6: @ %if.end
; CHECK-NEXT:    add.w r0, r12, r3, lsl #2
; CHECK-NEXT:    lsr.w r1, r10, #2
; CHECK-NEXT:    wls lr, r1, .LBB15_10
; CHECK-NEXT:  @ %bb.7: @ %while.body51.preheader
; CHECK-NEXT:    bic r2, r10, #3
; CHECK-NEXT:    adds r1, r2, r3
; CHECK-NEXT:    mov r3, r12
; CHECK-NEXT:    add.w r1, r12, r1, lsl #2
; CHECK-NEXT:  .LBB15_8: @ %while.body51
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q0, [r0], #16
; CHECK-NEXT:    vstrb.8 q0, [r3], #16
; CHECK-NEXT:    le lr, .LBB15_8
; CHECK-NEXT:  @ %bb.9: @ %while.end55.loopexit
; CHECK-NEXT:    add.w r12, r12, r2, lsl #2
; CHECK-NEXT:    mov r0, r1
; CHECK-NEXT:  .LBB15_10: @ %while.end55
; CHECK-NEXT:    ands r1, r10, #3
; CHECK-NEXT:    beq .LBB15_12
; CHECK-NEXT:  @ %bb.11: @ %if.then59
; CHECK-NEXT:    vldrw.u32 q0, [r0]
; CHECK-NEXT:    vctp.32 r1
; CHECK-NEXT:    vpst
; CHECK-NEXT:    vstrwt.32 q0, [r12]
; CHECK-NEXT:  .LBB15_12: @ %if.end61
; CHECK-NEXT:    add sp, #8
; CHECK-NEXT:    pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc}
entry:
  %pState1 = getelementptr inbounds %struct.arm_fir_instance_f32, ptr %S, i32 0, i32 1
  %i = load ptr, ptr %pState1, align 4
  %pCoeffs2 = getelementptr inbounds %struct.arm_fir_instance_f32, ptr %S, i32 0, i32 2
  %i1 = load ptr, ptr %pCoeffs2, align 4
  %numTaps3 = getelementptr inbounds %struct.arm_fir_instance_f32, ptr %S, i32 0, i32 0
  %i2 = load i16, ptr %numTaps3, align 4
  %conv = zext i16 %i2 to i32
  %sub = add nsw i32 %conv, -1
  %cmp = icmp ult i32 %sub, 4
  br i1 %cmp, label %if.then, label %if.end

if.then:                                          ; preds = %entry
  %arrayidx = getelementptr inbounds float, ptr %i, i32 %sub
  %incdec.ptr = getelementptr inbounds float, ptr %i1, i32 1
  %i3 = load float, ptr %i1, align 4
  %incdec.ptr6 = getelementptr inbounds float, ptr %i1, i32 2
  %i4 = load float, ptr %incdec.ptr, align 4
  %incdec.ptr7 = getelementptr inbounds float, ptr %i1, i32 3
  %i5 = load float, ptr %incdec.ptr6, align 4
  %i6 = load float, ptr %incdec.ptr7, align 4
  %shr = lshr i32 %blockSize, 2
  %cmp9146 = icmp eq i32 %shr, 0
  %.pre161 = insertelement <4 x float> undef, float %i3, i32 0
  %.pre162 = shufflevector <4 x float> %.pre161, <4 x float> undef, <4 x i32> zeroinitializer
  %.pre163 = insertelement <4 x float> undef, float %i4, i32 0
  %.pre164 = shufflevector <4 x float> %.pre163, <4 x float> undef, <4 x i32> zeroinitializer
  %.pre165 = insertelement <4 x float> undef, float %i5, i32 0
  %.pre166 = shufflevector <4 x float> %.pre165, <4 x float> undef, <4 x i32> zeroinitializer
  %.pre167 = insertelement <4 x float> undef, float %i6, i32 0
  %.pre168 = shufflevector <4 x float> %.pre167, <4 x float> undef, <4 x i32> zeroinitializer
  br i1 %cmp9146, label %while.end, label %while.body.lr.ph

while.body.lr.ph:                                 ; preds = %if.then
  %i7 = and i32 %blockSize, -4
  %scevgep158 = getelementptr float, ptr %pDst, i32 %i7
  br label %while.body

while.body:                                       ; preds = %while.body, %while.body.lr.ph
  %pStateCur.0151 = phi ptr [ %arrayidx, %while.body.lr.ph ], [ %add.ptr, %while.body ]
  %pSamples.0150 = phi ptr [ %i, %while.body.lr.ph ], [ %add.ptr24, %while.body ]
  %pOutput.0149 = phi ptr [ %pDst, %while.body.lr.ph ], [ %add.ptr23, %while.body ]
  %pTempSrc.0148 = phi ptr [ %pSrc, %while.body.lr.ph ], [ %add.ptr11, %while.body ]
  %blkCnt.0147 = phi i32 [ %shr, %while.body.lr.ph ], [ %dec, %while.body ]
  %i9 = load <4 x float>, ptr %pTempSrc.0148, align 4
  store <4 x float> %i9, ptr %pStateCur.0151, align 4
  %add.ptr = getelementptr inbounds float, ptr %pStateCur.0151, i32 4
  %add.ptr11 = getelementptr inbounds float, ptr %pTempSrc.0148, i32 4
  %i12 = load <4 x float>, ptr %pSamples.0150, align 4
  %i13 = fmul fast <4 x float> %i12, %.pre162
  %arrayidx12 = getelementptr inbounds float, ptr %pSamples.0150, i32 1
  %i15 = load <4 x float>, ptr %arrayidx12, align 4
  %mul = fmul fast <4 x float> %i15, %.pre164
  %add = fadd fast <4 x float> %mul, %i13
  %arrayidx13 = getelementptr inbounds float, ptr %pSamples.0150, i32 2
  %i17 = load <4 x float>, ptr %arrayidx13, align 4
  %mul16 = fmul fast <4 x float> %i17, %.pre166
  %add17 = fadd fast <4 x float> %add, %mul16
  %arrayidx18 = getelementptr inbounds float, ptr %pSamples.0150, i32 3
  %i19 = load <4 x float>, ptr %arrayidx18, align 4
  %mul21 = fmul fast <4 x float> %i19, %.pre168
  %add22 = fadd fast <4 x float> %add17, %mul21
  store <4 x float> %add22, ptr %pOutput.0149, align 4
  %add.ptr23 = getelementptr inbounds float, ptr %pOutput.0149, i32 4
  %add.ptr24 = getelementptr inbounds float, ptr %pSamples.0150, i32 4
  %dec = add nsw i32 %blkCnt.0147, -1
  %cmp9 = icmp eq i32 %dec, 0
  br i1 %cmp9, label %while.end.loopexit, label %while.body

while.end.loopexit:                               ; preds = %while.body
  %scevgep157 = getelementptr float, ptr %pSrc, i32 %i7
  %scevgep159 = getelementptr float, ptr %i, i32 %i7
  br label %while.end

while.end:                                        ; preds = %while.end.loopexit, %if.then
  %pTempSrc.0.lcssa = phi ptr [ %scevgep157, %while.end.loopexit ], [ %pSrc, %if.then ]
  %pOutput.0.lcssa = phi ptr [ %scevgep158, %while.end.loopexit ], [ %pDst, %if.then ]
  %pSamples.0.lcssa = phi ptr [ %scevgep159, %while.end.loopexit ], [ %i, %if.then ]
  %pStateCur.0.lcssa = phi ptr [ %add.ptr, %while.end.loopexit ], [ %arrayidx, %if.then ]
  %and = and i32 %blockSize, 3
  %i21 = tail call <4 x i1> @llvm.arm.mve.vctp32(i32 %and)
  %i23 = load <4 x float>, ptr %pTempSrc.0.lcssa, align 4
  tail call void @llvm.masked.store.v4f32.p0(<4 x float> %i23, ptr %pStateCur.0.lcssa, i32 4, <4 x i1> %i21)
  %i26 = load <4 x float>, ptr %pSamples.0.lcssa, align 4
  %i27 = fmul fast <4 x float> %i26, %.pre162
  %arrayidx29 = getelementptr inbounds float, ptr %pSamples.0.lcssa, i32 1
  %i29 = load <4 x float>, ptr %arrayidx29, align 4
  %mul32 = fmul fast <4 x float> %i29, %.pre164
  %add33 = fadd fast <4 x float> %mul32, %i27
  %arrayidx34 = getelementptr inbounds float, ptr %pSamples.0.lcssa, i32 2
  %i31 = load <4 x float>, ptr %arrayidx34, align 4
  %mul37 = fmul fast <4 x float> %i31, %.pre166
  %add38 = fadd fast <4 x float> %add33, %mul37
  %arrayidx39 = getelementptr inbounds float, ptr %pSamples.0.lcssa, i32 3
  %i33 = load <4 x float>, ptr %arrayidx39, align 4
  %mul42 = fmul fast <4 x float> %i33, %.pre168
  %add43 = fadd fast <4 x float> %add38, %mul42
  tail call void @llvm.masked.store.v4f32.p0(<4 x float> %add43, ptr %pOutput.0.lcssa, i32 4, <4 x i1> %i21)
  %.pre = load ptr, ptr %pState1, align 4
  br label %if.end

if.end:                                           ; preds = %while.end, %entry
  %i35 = phi ptr [ %.pre, %while.end ], [ %i, %entry ]
  %arrayidx45 = getelementptr inbounds float, ptr %i35, i32 %blockSize
  %shr47 = lshr i32 %conv, 2
  %cmp49141 = icmp eq i32 %shr47, 0
  br i1 %cmp49141, label %while.end55, label %while.body51.preheader

while.body51.preheader:                           ; preds = %if.end
  %i36 = and i32 %conv, 65532
  %i37 = add i32 %i36, %blockSize
  %scevgep = getelementptr float, ptr %i35, i32 %i37
  br label %while.body51

while.body51:                                     ; preds = %while.body51, %while.body51.preheader
  %pTempSrc.1144 = phi ptr [ %add.ptr52, %while.body51 ], [ %arrayidx45, %while.body51.preheader ]
  %pTempDest.0143 = phi ptr [ %add.ptr53, %while.body51 ], [ %i35, %while.body51.preheader ]
  %blkCnt.1142 = phi i32 [ %dec54, %while.body51 ], [ %shr47, %while.body51.preheader ]
  %i39 = load <4 x float>, ptr %pTempSrc.1144, align 4
  store <4 x float> %i39, ptr %pTempDest.0143, align 4
  %add.ptr52 = getelementptr inbounds float, ptr %pTempSrc.1144, i32 4
  %add.ptr53 = getelementptr inbounds float, ptr %pTempDest.0143, i32 4
  %dec54 = add nsw i32 %blkCnt.1142, -1
  %cmp49 = icmp eq i32 %dec54, 0
  br i1 %cmp49, label %while.end55.loopexit, label %while.body51

while.end55.loopexit:                             ; preds = %while.body51
  %scevgep156 = getelementptr float, ptr %i35, i32 %i36
  br label %while.end55

while.end55:                                      ; preds = %while.end55.loopexit, %if.end
  %pTempDest.0.lcssa = phi ptr [ %i35, %if.end ], [ %scevgep156, %while.end55.loopexit ]
  %pTempSrc.1.lcssa = phi ptr [ %arrayidx45, %if.end ], [ %scevgep, %while.end55.loopexit ]
  %and56 = and i32 %conv, 3
  %cmp57 = icmp eq i32 %and56, 0
  br i1 %cmp57, label %if.end61, label %if.then59

if.then59:                                        ; preds = %while.end55
  %i41 = tail call <4 x i1> @llvm.arm.mve.vctp32(i32 %and56)
  %i43 = load <4 x float>, ptr %pTempSrc.1.lcssa, align 4
  tail call void @llvm.masked.store.v4f32.p0(<4 x float> %i43, ptr %pTempDest.0.lcssa, i32 4, <4 x i1> %i41)
  br label %if.end61

if.end61:                                         ; preds = %if.then59, %while.end55
  ret void
}


define void @fir(ptr nocapture readonly %S, ptr nocapture readonly %pSrc, ptr nocapture %pDst, i32 %blockSize) {
; CHECK-LABEL: fir:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    cmp r3, #8
; CHECK-NEXT:    blo.w .LBB16_13
; CHECK-NEXT:  @ %bb.1: @ %if.then
; CHECK-NEXT:    lsrs.w r12, r3, #2
; CHECK-NEXT:    it eq
; CHECK-NEXT:    bxeq lr
; CHECK-NEXT:  .LBB16_2: @ %while.body.lr.ph
; 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}
; CHECK-NEXT:    vpush {d8, d9, d10, d11, d12, d13}
; CHECK-NEXT:    .pad #24
; CHECK-NEXT:    sub sp, #24
; CHECK-NEXT:    ldrh r6, [r0]
; CHECK-NEXT:    movs r4, #1
; CHECK-NEXT:    ldrd r7, r10, [r0, #4]
; CHECK-NEXT:    sub.w r0, r6, #8
; CHECK-NEXT:    add.w r3, r0, r0, lsr #29
; CHECK-NEXT:    and r0, r0, #7
; CHECK-NEXT:    asrs r5, r3, #3
; CHECK-NEXT:    cmp r5, #1
; CHECK-NEXT:    it gt
; CHECK-NEXT:    asrgt r4, r3, #3
; CHECK-NEXT:    add.w r3, r7, r6, lsl #2
; CHECK-NEXT:    sub.w r9, r3, #4
; CHECK-NEXT:    rsbs r3, r6, #0
; CHECK-NEXT:    str r4, [sp] @ 4-byte Spill
; CHECK-NEXT:    str r6, [sp, #8] @ 4-byte Spill
; CHECK-NEXT:    str r3, [sp, #4] @ 4-byte Spill
; CHECK-NEXT:    str r0, [sp, #12] @ 4-byte Spill
; CHECK-NEXT:    b .LBB16_6
; CHECK-NEXT:  .LBB16_3: @ %while.end.loopexit
; CHECK-NEXT:    @ in Loop: Header=BB16_6 Depth=1
; CHECK-NEXT:    ldr r0, [sp, #12] @ 4-byte Reload
; CHECK-NEXT:    add.w r7, r7, r0, lsl #2
; CHECK-NEXT:    b .LBB16_5
; CHECK-NEXT:  .LBB16_4: @ %for.end
; CHECK-NEXT:    @ in Loop: Header=BB16_6 Depth=1
; CHECK-NEXT:    ldr r1, [sp, #20] @ 4-byte Reload
; CHECK-NEXT:    ldrd r0, r9, [sp, #12] @ 8-byte Folded Reload
; CHECK-NEXT:    wls lr, r0, .LBB16_5
; CHECK-NEXT:    b .LBB16_10
; CHECK-NEXT:  .LBB16_5: @ %while.end
; CHECK-NEXT:    @ in Loop: Header=BB16_6 Depth=1
; CHECK-NEXT:    ldr r0, [sp, #4] @ 4-byte Reload
; CHECK-NEXT:    subs.w r12, r12, #1
; CHECK-NEXT:    vstrb.8 q0, [r2], #16
; CHECK-NEXT:    add.w r0, r7, r0, lsl #2
; CHECK-NEXT:    add.w r7, r0, #16
; CHECK-NEXT:    beq .LBB16_12
; CHECK-NEXT:  .LBB16_6: @ %while.body
; CHECK-NEXT:    @ =>This Loop Header: Depth=1
; CHECK-NEXT:    @ Child Loop BB16_8 Depth 2
; CHECK-NEXT:    @ Child Loop BB16_11 Depth 2
; CHECK-NEXT:    add.w lr, r10, #8
; CHECK-NEXT:    vldrw.u32 q0, [r1], #16
; CHECK-NEXT:    ldrd r3, r4, [r10]
; CHECK-NEXT:    ldm.w lr, {r0, r5, r6, lr}
; CHECK-NEXT:    ldrd r11, r8, [r10, #24]
; CHECK-NEXT:    vstrb.8 q0, [r9], #16
; CHECK-NEXT:    vldrw.u32 q0, [r7], #32
; CHECK-NEXT:    str r1, [sp, #20] @ 4-byte Spill
; CHECK-NEXT:    str.w r9, [sp, #16] @ 4-byte Spill
; CHECK-NEXT:    vldrw.u32 q1, [r7, #-28]
; CHECK-NEXT:    vmul.f32 q0, q0, r3
; CHECK-NEXT:    vldrw.u32 q6, [r7, #-24]
; CHECK-NEXT:    vldrw.u32 q4, [r7, #-20]
; CHECK-NEXT:    vfma.f32 q0, q1, r4
; CHECK-NEXT:    vldrw.u32 q5, [r7, #-16]
; CHECK-NEXT:    vfma.f32 q0, q6, r0
; CHECK-NEXT:    vldrw.u32 q2, [r7, #-12]
; CHECK-NEXT:    vfma.f32 q0, q4, r5
; CHECK-NEXT:    vldrw.u32 q3, [r7, #-8]
; CHECK-NEXT:    vfma.f32 q0, q5, r6
; CHECK-NEXT:    vldrw.u32 q1, [r7, #-4]
; CHECK-NEXT:    vfma.f32 q0, q2, lr
; CHECK-NEXT:    ldr r0, [sp, #8] @ 4-byte Reload
; CHECK-NEXT:    vfma.f32 q0, q3, r11
; CHECK-NEXT:    vfma.f32 q0, q1, r8
; CHECK-NEXT:    cmp r0, #16
; CHECK-NEXT:    blo .LBB16_9
; CHECK-NEXT:  @ %bb.7: @ %for.body.preheader
; CHECK-NEXT:    @ in Loop: Header=BB16_6 Depth=1
; CHECK-NEXT:    ldr r0, [sp] @ 4-byte Reload
; CHECK-NEXT:    add.w r4, r10, #32
; CHECK-NEXT:    dls lr, r0
; CHECK-NEXT:  .LBB16_8: @ %for.body
; CHECK-NEXT:    @ Parent Loop BB16_6 Depth=1
; CHECK-NEXT:    @ => This Inner Loop Header: Depth=2
; CHECK-NEXT:    ldm.w r4, {r0, r3, r5, r6, r8, r11}
; CHECK-NEXT:    vldrw.u32 q1, [r7], #32
; CHECK-NEXT:    vldrw.u32 q6, [r7, #-24]
; CHECK-NEXT:    vldrw.u32 q4, [r7, #-20]
; CHECK-NEXT:    vfma.f32 q0, q1, r0
; CHECK-NEXT:    vldrw.u32 q1, [r7, #-28]
; CHECK-NEXT:    vldrw.u32 q5, [r7, #-16]
; CHECK-NEXT:    vldrw.u32 q2, [r7, #-12]
; CHECK-NEXT:    vfma.f32 q0, q1, r3
; CHECK-NEXT:    ldrd r9, r1, [r4, #24]
; CHECK-NEXT:    vfma.f32 q0, q6, r5
; CHECK-NEXT:    vldrw.u32 q3, [r7, #-8]
; CHECK-NEXT:    vfma.f32 q0, q4, r6
; CHECK-NEXT:    vldrw.u32 q1, [r7, #-4]
; CHECK-NEXT:    vfma.f32 q0, q5, r8
; CHECK-NEXT:    adds r4, #32
; CHECK-NEXT:    vfma.f32 q0, q2, r11
; CHECK-NEXT:    vfma.f32 q0, q3, r9
; CHECK-NEXT:    vfma.f32 q0, q1, r1
; CHECK-NEXT:    le lr, .LBB16_8
; CHECK-NEXT:    b .LBB16_4
; CHECK-NEXT:  .LBB16_9: @ in Loop: Header=BB16_6 Depth=1
; CHECK-NEXT:    add.w r4, r10, #32
; CHECK-NEXT:    b .LBB16_4
; CHECK-NEXT:  .LBB16_10: @ %while.body76.preheader
; CHECK-NEXT:    @ in Loop: Header=BB16_6 Depth=1
; CHECK-NEXT:    mov r3, r7
; CHECK-NEXT:  .LBB16_11: @ %while.body76
; CHECK-NEXT:    @ Parent Loop BB16_6 Depth=1
; CHECK-NEXT:    @ => This Inner Loop Header: Depth=2
; CHECK-NEXT:    ldr r0, [r4], #4
; CHECK-NEXT:    vldrw.u32 q1, [r3], #4
; CHECK-NEXT:    vfma.f32 q0, q1, r0
; CHECK-NEXT:    le lr, .LBB16_11
; CHECK-NEXT:    b .LBB16_3
; CHECK-NEXT:  .LBB16_12:
; CHECK-NEXT:    add sp, #24
; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13}
; CHECK-NEXT:    add sp, #4
; CHECK-NEXT:    pop.w {r4, r5, r6, r7, r8, r9, r10, r11, lr}
; CHECK-NEXT:  .LBB16_13: @ %if.end
; CHECK-NEXT:    bx lr
entry:
  %pState1 = getelementptr inbounds %struct.arm_fir_instance_f32, ptr %S, i32 0, i32 1
  %i = load ptr, ptr %pState1, align 4
  %pCoeffs2 = getelementptr inbounds %struct.arm_fir_instance_f32, ptr %S, i32 0, i32 2
  %i1 = load ptr, ptr %pCoeffs2, align 4
  %numTaps3 = getelementptr inbounds %struct.arm_fir_instance_f32, ptr %S, i32 0, i32 0
  %i2 = load i16, ptr %numTaps3, align 4
  %conv = zext i16 %i2 to i32
  %cmp = icmp ugt i32 %blockSize, 7
  br i1 %cmp, label %if.then, label %if.end

if.then:                                          ; preds = %entry
  %shr = lshr i32 %blockSize, 2
  %cmp5217 = icmp eq i32 %shr, 0
  br i1 %cmp5217, label %if.end, label %while.body.lr.ph

while.body.lr.ph:                                 ; preds = %if.then
  %sub = add nsw i32 %conv, -1
  %arrayidx = getelementptr inbounds float, ptr %i, i32 %sub
  %incdec.ptr = getelementptr inbounds float, ptr %i1, i32 1
  %incdec.ptr7 = getelementptr inbounds float, ptr %i1, i32 2
  %incdec.ptr8 = getelementptr inbounds float, ptr %i1, i32 3
  %incdec.ptr9 = getelementptr inbounds float, ptr %i1, i32 4
  %incdec.ptr10 = getelementptr inbounds float, ptr %i1, i32 5
  %incdec.ptr11 = getelementptr inbounds float, ptr %i1, i32 6
  %incdec.ptr12 = getelementptr inbounds float, ptr %i1, i32 7
  %sub37 = add nsw i32 %conv, -8
  %div = sdiv i32 %sub37, 8
  %pCoeffsCur.0199 = getelementptr inbounds float, ptr %i1, i32 8
  %cmp38201 = icmp ugt i16 %i2, 15
  %and = and i32 %sub37, 7
  %cmp74210 = icmp eq i32 %and, 0
  %idx.neg = sub nsw i32 0, %conv
  %i3 = icmp sgt i32 %div, 1
  %smax = select i1 %i3, i32 %div, i32 1
  br label %while.body

while.body:                                       ; preds = %while.end, %while.body.lr.ph
  %blkCnt.0222 = phi i32 [ %shr, %while.body.lr.ph ], [ %dec84, %while.end ]
  %pStateCur.0221 = phi ptr [ %arrayidx, %while.body.lr.ph ], [ %add.ptr, %while.end ]
  %pSamples.0220 = phi ptr [ %i, %while.body.lr.ph ], [ %add.ptr83, %while.end ]
  %pTempSrc.0219 = phi ptr [ %pSrc, %while.body.lr.ph ], [ %add.ptr14, %while.end ]
  %pOutput.0218 = phi ptr [ %pDst, %while.body.lr.ph ], [ %add.ptr81, %while.end ]
  %i4 = load float, ptr %i1, align 4
  %i5 = load float, ptr %incdec.ptr, align 4
  %i6 = load float, ptr %incdec.ptr7, align 4
  %i7 = load float, ptr %incdec.ptr8, align 4
  %i8 = load float, ptr %incdec.ptr9, align 4
  %i9 = load float, ptr %incdec.ptr10, align 4
  %i10 = load float, ptr %incdec.ptr11, align 4
  %i11 = load float, ptr %incdec.ptr12, align 4
  %i13 = load <4 x float>, ptr %pTempSrc.0219, align 4
  store <4 x float> %i13, ptr %pStateCur.0221, align 4
  %add.ptr = getelementptr inbounds float, ptr %pStateCur.0221, i32 4
  %add.ptr14 = getelementptr inbounds float, ptr %pTempSrc.0219, i32 4
  %i16 = load <4 x float>, ptr %pSamples.0220, align 4
  %.splatinsert = insertelement <4 x float> undef, float %i4, i32 0
  %.splat = shufflevector <4 x float> %.splatinsert, <4 x float> undef, <4 x i32> zeroinitializer
  %i17 = fmul fast <4 x float> %i16, %.splat
  %arrayidx15 = getelementptr inbounds float, ptr %pSamples.0220, i32 1
  %i19 = load <4 x float>, ptr %arrayidx15, align 4
  %.splatinsert16 = insertelement <4 x float> undef, float %i5, i32 0
  %.splat17 = shufflevector <4 x float> %.splatinsert16, <4 x float> undef, <4 x i32> zeroinitializer
  %i20 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i19, <4 x float> %.splat17, <4 x float> %i17)
  %arrayidx18 = getelementptr inbounds float, ptr %pSamples.0220, i32 2
  %i22 = load <4 x float>, ptr %arrayidx18, align 4
  %.splatinsert19 = insertelement <4 x float> undef, float %i6, i32 0
  %.splat20 = shufflevector <4 x float> %.splatinsert19, <4 x float> undef, <4 x i32> zeroinitializer
  %i23 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i22, <4 x float> %.splat20, <4 x float> %i20)
  %arrayidx21 = getelementptr inbounds float, ptr %pSamples.0220, i32 3
  %i25 = load <4 x float>, ptr %arrayidx21, align 4
  %.splatinsert22 = insertelement <4 x float> undef, float %i7, i32 0
  %.splat23 = shufflevector <4 x float> %.splatinsert22, <4 x float> undef, <4 x i32> zeroinitializer
  %i26 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i25, <4 x float> %.splat23, <4 x float> %i23)
  %arrayidx24 = getelementptr inbounds float, ptr %pSamples.0220, i32 4
  %i28 = load <4 x float>, ptr %arrayidx24, align 4
  %.splatinsert25 = insertelement <4 x float> undef, float %i8, i32 0
  %.splat26 = shufflevector <4 x float> %.splatinsert25, <4 x float> undef, <4 x i32> zeroinitializer
  %i29 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i28, <4 x float> %.splat26, <4 x float> %i26)
  %arrayidx27 = getelementptr inbounds float, ptr %pSamples.0220, i32 5
  %i31 = load <4 x float>, ptr %arrayidx27, align 4
  %.splatinsert28 = insertelement <4 x float> undef, float %i9, i32 0
  %.splat29 = shufflevector <4 x float> %.splatinsert28, <4 x float> undef, <4 x i32> zeroinitializer
  %i32 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i31, <4 x float> %.splat29, <4 x float> %i29)
  %arrayidx30 = getelementptr inbounds float, ptr %pSamples.0220, i32 6
  %i34 = load <4 x float>, ptr %arrayidx30, align 4
  %.splatinsert31 = insertelement <4 x float> undef, float %i10, i32 0
  %.splat32 = shufflevector <4 x float> %.splatinsert31, <4 x float> undef, <4 x i32> zeroinitializer
  %i35 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i34, <4 x float> %.splat32, <4 x float> %i32)
  %arrayidx33 = getelementptr inbounds float, ptr %pSamples.0220, i32 7
  %i37 = load <4 x float>, ptr %arrayidx33, align 4
  %.splatinsert34 = insertelement <4 x float> undef, float %i11, i32 0
  %.splat35 = shufflevector <4 x float> %.splatinsert34, <4 x float> undef, <4 x i32> zeroinitializer
  %i38 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i37, <4 x float> %.splat35, <4 x float> %i35)
  %pSamples.1200 = getelementptr inbounds float, ptr %pSamples.0220, i32 8
  br i1 %cmp38201, label %for.body, label %for.end

for.body:                                         ; preds = %for.body, %while.body
  %pSamples.1207 = phi ptr [ %pSamples.1, %for.body ], [ %pSamples.1200, %while.body ]
  %pCoeffsCur.0206 = phi ptr [ %pCoeffsCur.0, %for.body ], [ %pCoeffsCur.0199, %while.body ]
  %.pn205 = phi ptr [ %pCoeffsCur.0206, %for.body ], [ %i1, %while.body ]
  %i.0204 = phi i32 [ %inc, %for.body ], [ 0, %while.body ]
  %vecAcc0.0203 = phi <4 x float> [ %i70, %for.body ], [ %i38, %while.body ]
  %pSamples.0.pn202 = phi ptr [ %pSamples.1207, %for.body ], [ %pSamples.0220, %while.body ]
  %incdec.ptr40 = getelementptr inbounds float, ptr %.pn205, i32 9
  %i39 = load float, ptr %pCoeffsCur.0206, align 4
  %incdec.ptr41 = getelementptr inbounds float, ptr %.pn205, i32 10
  %i40 = load float, ptr %incdec.ptr40, align 4
  %incdec.ptr42 = getelementptr inbounds float, ptr %.pn205, i32 11
  %i41 = load float, ptr %incdec.ptr41, align 4
  %incdec.ptr43 = getelementptr inbounds float, ptr %.pn205, i32 12
  %i42 = load float, ptr %incdec.ptr42, align 4
  %incdec.ptr44 = getelementptr inbounds float, ptr %.pn205, i32 13
  %i43 = load float, ptr %incdec.ptr43, align 4
  %incdec.ptr45 = getelementptr inbounds float, ptr %.pn205, i32 14
  %i44 = load float, ptr %incdec.ptr44, align 4
  %incdec.ptr46 = getelementptr inbounds float, ptr %.pn205, i32 15
  %i45 = load float, ptr %incdec.ptr45, align 4
  %i46 = load float, ptr %incdec.ptr46, align 4
  %i48 = load <4 x float>, ptr %pSamples.1207, align 4
  %.splatinsert48 = insertelement <4 x float> undef, float %i39, i32 0
  %.splat49 = shufflevector <4 x float> %.splatinsert48, <4 x float> undef, <4 x i32> zeroinitializer
  %i49 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i48, <4 x float> %.splat49, <4 x float> %vecAcc0.0203)
  %arrayidx50 = getelementptr inbounds float, ptr %pSamples.0.pn202, i32 9
  %i51 = load <4 x float>, ptr %arrayidx50, align 4
  %.splatinsert51 = insertelement <4 x float> undef, float %i40, i32 0
  %.splat52 = shufflevector <4 x float> %.splatinsert51, <4 x float> undef, <4 x i32> zeroinitializer
  %i52 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i51, <4 x float> %.splat52, <4 x float> %i49)
  %arrayidx53 = getelementptr inbounds float, ptr %pSamples.0.pn202, i32 10
  %i54 = load <4 x float>, ptr %arrayidx53, align 4
  %.splatinsert54 = insertelement <4 x float> undef, float %i41, i32 0
  %.splat55 = shufflevector <4 x float> %.splatinsert54, <4 x float> undef, <4 x i32> zeroinitializer
  %i55 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i54, <4 x float> %.splat55, <4 x float> %i52)
  %arrayidx56 = getelementptr inbounds float, ptr %pSamples.0.pn202, i32 11
  %i57 = load <4 x float>, ptr %arrayidx56, align 4
  %.splatinsert57 = insertelement <4 x float> undef, float %i42, i32 0
  %.splat58 = shufflevector <4 x float> %.splatinsert57, <4 x float> undef, <4 x i32> zeroinitializer
  %i58 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i57, <4 x float> %.splat58, <4 x float> %i55)
  %arrayidx59 = getelementptr inbounds float, ptr %pSamples.0.pn202, i32 12
  %i60 = load <4 x float>, ptr %arrayidx59, align 4
  %.splatinsert60 = insertelement <4 x float> undef, float %i43, i32 0
  %.splat61 = shufflevector <4 x float> %.splatinsert60, <4 x float> undef, <4 x i32> zeroinitializer
  %i61 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i60, <4 x float> %.splat61, <4 x float> %i58)
  %arrayidx62 = getelementptr inbounds float, ptr %pSamples.0.pn202, i32 13
  %i63 = load <4 x float>, ptr %arrayidx62, align 4
  %.splatinsert63 = insertelement <4 x float> undef, float %i44, i32 0
  %.splat64 = shufflevector <4 x float> %.splatinsert63, <4 x float> undef, <4 x i32> zeroinitializer
  %i64 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i63, <4 x float> %.splat64, <4 x float> %i61)
  %arrayidx65 = getelementptr inbounds float, ptr %pSamples.0.pn202, i32 14
  %i66 = load <4 x float>, ptr %arrayidx65, align 4
  %.splatinsert66 = insertelement <4 x float> undef, float %i45, i32 0
  %.splat67 = shufflevector <4 x float> %.splatinsert66, <4 x float> undef, <4 x i32> zeroinitializer
  %i67 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i66, <4 x float> %.splat67, <4 x float> %i64)
  %arrayidx68 = getelementptr inbounds float, ptr %pSamples.0.pn202, i32 15
  %i69 = load <4 x float>, ptr %arrayidx68, align 4
  %.splatinsert69 = insertelement <4 x float> undef, float %i46, i32 0
  %.splat70 = shufflevector <4 x float> %.splatinsert69, <4 x float> undef, <4 x i32> zeroinitializer
  %i70 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i69, <4 x float> %.splat70, <4 x float> %i67)
  %inc = add nuw nsw i32 %i.0204, 1
  %pCoeffsCur.0 = getelementptr inbounds float, ptr %pCoeffsCur.0206, i32 8
  %pSamples.1 = getelementptr inbounds float, ptr %pSamples.1207, i32 8
  %exitcond = icmp eq i32 %inc, %smax
  br i1 %exitcond, label %for.end, label %for.body

for.end:                                          ; preds = %for.body, %while.body
  %vecAcc0.0.lcssa = phi <4 x float> [ %i38, %while.body ], [ %i70, %for.body ]
  %pCoeffsCur.0.lcssa = phi ptr [ %pCoeffsCur.0199, %while.body ], [ %pCoeffsCur.0, %for.body ]
  %pSamples.1.lcssa = phi ptr [ %pSamples.1200, %while.body ], [ %pSamples.1, %for.body ]
  br i1 %cmp74210, label %while.end, label %while.body76

while.body76:                                     ; preds = %while.body76, %for.end
  %pCoeffsCur.1214 = phi ptr [ %incdec.ptr77, %while.body76 ], [ %pCoeffsCur.0.lcssa, %for.end ]
  %vecAcc0.1213 = phi <4 x float> [ %i74, %while.body76 ], [ %vecAcc0.0.lcssa, %for.end ]
  %numCnt.0212 = phi i32 [ %dec, %while.body76 ], [ %and, %for.end ]
  %pSamples.2211 = phi ptr [ %incdec.ptr80, %while.body76 ], [ %pSamples.1.lcssa, %for.end ]
  %incdec.ptr77 = getelementptr inbounds float, ptr %pCoeffsCur.1214, i32 1
  %i71 = load float, ptr %pCoeffsCur.1214, align 4
  %i73 = load <4 x float>, ptr %pSamples.2211, align 4
  %.splatinsert78 = insertelement <4 x float> undef, float %i71, i32 0
  %.splat79 = shufflevector <4 x float> %.splatinsert78, <4 x float> undef, <4 x i32> zeroinitializer
  %i74 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i73, <4 x float> %.splat79, <4 x float> %vecAcc0.1213)
  %incdec.ptr80 = getelementptr inbounds float, ptr %pSamples.2211, i32 1
  %dec = add nsw i32 %numCnt.0212, -1
  %cmp74 = icmp sgt i32 %numCnt.0212, 1
  br i1 %cmp74, label %while.body76, label %while.end.loopexit

while.end.loopexit:                               ; preds = %while.body76
  %scevgep = getelementptr float, ptr %pSamples.1.lcssa, i32 %and
  br label %while.end

while.end:                                        ; preds = %while.end.loopexit, %for.end
  %pSamples.2.lcssa = phi ptr [ %pSamples.1.lcssa, %for.end ], [ %scevgep, %while.end.loopexit ]
  %vecAcc0.1.lcssa = phi <4 x float> [ %vecAcc0.0.lcssa, %for.end ], [ %i74, %while.end.loopexit ]
  store <4 x float> %vecAcc0.1.lcssa, ptr %pOutput.0218, align 4
  %add.ptr81 = getelementptr inbounds float, ptr %pOutput.0218, i32 4
  %add.ptr82 = getelementptr inbounds float, ptr %pSamples.2.lcssa, i32 4
  %add.ptr83 = getelementptr inbounds float, ptr %add.ptr82, i32 %idx.neg
  %dec84 = add nsw i32 %blkCnt.0222, -1
  %cmp5 = icmp eq i32 %dec84, 0
  br i1 %cmp5, label %if.end, label %while.body

if.end:                                           ; preds = %while.end, %if.then, %entry
  ret void
}

%struct.arm_biquad_cascade_stereo_df2T_instance_f32 = type { i8, ptr, ptr }
define arm_aapcs_vfpcc void @arm_biquad_cascade_stereo_df2T_f32(ptr nocapture readonly %arg, ptr %arg1, ptr %arg2, i32 %arg3) {
; CHECK-LABEL: arm_biquad_cascade_stereo_df2T_f32:
; CHECK:       @ %bb.0: @ %bb
; CHECK-NEXT:    .save {r4, r5, r6, r7, r8, lr}
; CHECK-NEXT:    push.w {r4, r5, r6, r7, r8, lr}
; CHECK-NEXT:    .vsave {d8, d9, d10, d11}
; CHECK-NEXT:    vpush {d8, d9, d10, d11}
; CHECK-NEXT:    .pad #24
; CHECK-NEXT:    sub sp, #24
; CHECK-NEXT:    mov r8, r3
; CHECK-NEXT:    ldrb.w r12, [r0]
; CHECK-NEXT:    ldrd r0, r3, [r0, #4]
; CHECK-NEXT:    movs r4, #0
; CHECK-NEXT:    cmp.w r8, #0
; CHECK-NEXT:    strd r4, r4, [sp, #16]
; CHECK-NEXT:    beq .LBB17_5
; CHECK-NEXT:  @ %bb.1:
; CHECK-NEXT:    movs r5, #2
; CHECK-NEXT:    viwdup.u32 q0, r4, r5, #1
; CHECK-NEXT:    mov r4, sp
; CHECK-NEXT:  .LBB17_2: @ %bb29
; CHECK-NEXT:    @ =>This Loop Header: Depth=1
; CHECK-NEXT:    @ Child Loop BB17_3 Depth 2
; CHECK-NEXT:    ldrd r5, r7, [r3]
; CHECK-NEXT:    vldrw.u32 q1, [r0]
; CHECK-NEXT:    ldr r6, [r3, #12]
; CHECK-NEXT:    vldr s8, [r3, #8]
; CHECK-NEXT:    vstrw.32 q1, [r4]
; CHECK-NEXT:    vdup.32 q1, r7
; CHECK-NEXT:    vldr s12, [r3, #16]
; CHECK-NEXT:    vmov.f32 s6, s8
; CHECK-NEXT:    dls lr, r8
; CHECK-NEXT:    vmov.f32 s7, s8
; CHECK-NEXT:    vdup.32 q2, r6
; CHECK-NEXT:    vmov.f32 s10, s12
; CHECK-NEXT:    mov r7, r2
; CHECK-NEXT:    vmov.f32 s11, s12
; CHECK-NEXT:  .LBB17_3: @ %bb55
; CHECK-NEXT:    @ Parent Loop BB17_2 Depth=1
; CHECK-NEXT:    @ => This Inner Loop Header: Depth=2
; CHECK-NEXT:    vldrw.u32 q4, [r1, q0, uxtw #2]
; CHECK-NEXT:    vldrw.u32 q5, [r4, q0, uxtw #2]
; CHECK-NEXT:    vldrw.u32 q3, [sp, #8]
; CHECK-NEXT:    adds r1, #8
; CHECK-NEXT:    vfma.f32 q5, q4, r5
; CHECK-NEXT:    vfma.f32 q3, q5, q2
; CHECK-NEXT:    vstmia r7!, {s20, s21}
; CHECK-NEXT:    vfma.f32 q3, q4, q1
; CHECK-NEXT:    vstrw.32 q3, [r4]
; CHECK-NEXT:    le lr, .LBB17_3
; CHECK-NEXT:  @ %bb.4: @ %bb75
; CHECK-NEXT:    @ in Loop: Header=BB17_2 Depth=1
; CHECK-NEXT:    adds r3, #20
; CHECK-NEXT:    subs.w r12, r12, #1
; CHECK-NEXT:    vstrb.8 q3, [r0], #16
; CHECK-NEXT:    mov r1, r2
; CHECK-NEXT:    bne .LBB17_2
; CHECK-NEXT:    b .LBB17_7
; CHECK-NEXT:  .LBB17_5: @ %bb21.preheader
; CHECK-NEXT:    dls lr, r12
; CHECK-NEXT:    mov r1, sp
; CHECK-NEXT:  .LBB17_6: @ %bb21
; CHECK-NEXT:    @ =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    vldrw.u32 q0, [r0], #16
; CHECK-NEXT:    vstrw.32 q0, [r1]
; CHECK-NEXT:    le lr, .LBB17_6
; CHECK-NEXT:  .LBB17_7: @ %bb80
; CHECK-NEXT:    add sp, #24
; CHECK-NEXT:    vpop {d8, d9, d10, d11}
; CHECK-NEXT:    pop.w {r4, r5, r6, r7, r8, pc}
bb:
  %i = alloca [6 x float], align 4
  %i4 = getelementptr inbounds %struct.arm_biquad_cascade_stereo_df2T_instance_f32, ptr %arg, i32 0, i32 1
  %i5 = load ptr, ptr %i4, align 4
  %i6 = getelementptr inbounds %struct.arm_biquad_cascade_stereo_df2T_instance_f32, ptr %arg, i32 0, i32 2
  %i7 = load ptr, ptr %i6, align 4
  %i8 = getelementptr inbounds %struct.arm_biquad_cascade_stereo_df2T_instance_f32, ptr %arg, i32 0, i32 0
  %i9 = load i8, ptr %i8, align 4
  %i10 = zext i8 %i9 to i32
  call void @llvm.lifetime.start.p0(i64 24, ptr nonnull %i)
  %i12 = tail call { <4 x i32>, i32 } @llvm.arm.mve.viwdup.v4i32(i32 0, i32 2, i32 1)
  %i13 = extractvalue { <4 x i32>, i32 } %i12, 0
  %i14 = getelementptr inbounds [6 x float], ptr %i, i32 0, i32 4
  store float 0.000000e+00, ptr %i14, align 4
  %i15 = getelementptr inbounds [6 x float], ptr %i, i32 0, i32 5
  store float 0.000000e+00, ptr %i15, align 4
  %i17 = icmp eq i32 %arg3, 0
  %i19 = getelementptr inbounds [6 x float], ptr %i, i32 0, i32 2
  br i1 %i17, label %bb21, label %bb29

bb21:                                             ; preds = %bb21, %bb
  %i22 = phi i32 [ %i27, %bb21 ], [ %i10, %bb ]
  %i23 = phi ptr [ %i26, %bb21 ], [ %i5, %bb ]
  %i25 = load <4 x float>, ptr %i23, align 8
  store <4 x float> %i25, ptr %i, align 4
  %i26 = getelementptr inbounds float, ptr %i23, i32 4
  %i27 = add i32 %i22, -1
  %i28 = icmp eq i32 %i27, 0
  br i1 %i28, label %bb80, label %bb21

bb29:                                             ; preds = %bb75, %bb
  %i30 = phi i32 [ %i78, %bb75 ], [ %i10, %bb ]
  %i31 = phi ptr [ %i76, %bb75 ], [ %i7, %bb ]
  %i32 = phi ptr [ %i77, %bb75 ], [ %i5, %bb ]
  %i33 = phi ptr [ %arg2, %bb75 ], [ %arg1, %bb ]
  %i34 = getelementptr inbounds float, ptr %i31, i32 1
  %i35 = load float, ptr %i31, align 4
  %i36 = getelementptr inbounds float, ptr %i31, i32 2
  %i37 = load float, ptr %i34, align 4
  %i38 = getelementptr inbounds float, ptr %i31, i32 3
  %i39 = load float, ptr %i36, align 4
  %i40 = getelementptr inbounds float, ptr %i31, i32 4
  %i41 = load float, ptr %i38, align 4
  %i42 = load float, ptr %i40, align 4
  %i43 = insertelement <4 x float> undef, float %i41, i32 0
  %i44 = shufflevector <4 x float> %i43, <4 x float> undef, <4 x i32> <i32 0, i32 0, i32 undef, i32 undef>
  %i45 = insertelement <4 x float> %i44, float %i42, i32 2
  %i46 = insertelement <4 x float> %i45, float %i42, i32 3
  %i47 = insertelement <4 x float> undef, float %i37, i32 0
  %i48 = shufflevector <4 x float> %i47, <4 x float> undef, <4 x i32> <i32 0, i32 0, i32 undef, i32 undef>
  %i49 = insertelement <4 x float> %i48, float %i39, i32 2
  %i50 = insertelement <4 x float> %i49, float %i39, i32 3
  %i52 = load <4 x float>, ptr %i32, align 8
  store <4 x float> %i52, ptr %i, align 4
  %i53 = insertelement <4 x float> undef, float %i35, i32 0
  %i54 = shufflevector <4 x float> %i53, <4 x float> undef, <4 x i32> zeroinitializer
  br label %bb55

bb55:                                             ; preds = %bb55, %bb29
  %i56 = phi ptr [ %i33, %bb29 ], [ %i72, %bb55 ]
  %i57 = phi ptr [ %arg2, %bb29 ], [ %i68, %bb55 ]
  %i58 = phi i32 [ %arg3, %bb29 ], [ %i73, %bb55 ]
  %i59 = call <4 x i32> @llvm.arm.mve.vldr.gather.offset.v4i32.p0.v4i32(ptr nonnull %i, <4 x i32> %i13, i32 32, i32 2, i32 1)
  %i60 = bitcast <4 x i32> %i59 to <4 x float>
  %i62 = call <4 x i32> @llvm.arm.mve.vldr.gather.offset.v4i32.p0.v4i32(ptr %i56, <4 x i32> %i13, i32 32, i32 2, i32 1)
  %i63 = bitcast <4 x i32> %i62 to <4 x float>
  %i64 = call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i63, <4 x float> %i54, <4 x float> %i60)
  %i65 = extractelement <4 x float> %i64, i32 0
  %i66 = getelementptr inbounds float, ptr %i57, i32 1
  store float %i65, ptr %i57, align 4
  %i67 = extractelement <4 x float> %i64, i32 1
  %i68 = getelementptr inbounds float, ptr %i57, i32 2
  store float %i67, ptr %i66, align 4
  %i69 = load <4 x float>, ptr %i19, align 4
  %i70 = call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i64, <4 x float> %i46, <4 x float> %i69)
  %i71 = call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i63, <4 x float> %i50, <4 x float> %i70)
  store <4 x float> %i71, ptr %i, align 4
  %i72 = getelementptr inbounds float, ptr %i56, i32 2
  %i73 = add i32 %i58, -1
  %i74 = icmp eq i32 %i73, 0
  br i1 %i74, label %bb75, label %bb55

bb75:                                             ; preds = %bb55
  %i76 = getelementptr inbounds float, ptr %i31, i32 5
  store <4 x float> %i71, ptr %i32, align 4
  %i77 = getelementptr inbounds float, ptr %i32, i32 4
  %i78 = add i32 %i30, -1
  %i79 = icmp eq i32 %i78, 0
  br i1 %i79, label %bb80, label %bb29

bb80:                                             ; preds = %bb75, %bb21
  call void @llvm.lifetime.end.p0(i64 24, ptr nonnull %i)
  ret void
}

define arm_aapcs_vfpcc void @fms(ptr nocapture readonly %pSrc1, ptr nocapture readonly %pSrc2, ptr nocapture readonly %pSrc3, ptr nocapture %pDst, i32 %N, i32 %M) {
; CHECK-LABEL: fms:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, r5, r7, lr}
; CHECK-NEXT:    push {r4, r5, r7, lr}
; CHECK-NEXT:    ldr r4, [sp, #16]
; CHECK-NEXT:    lsrs r5, r4, #2
; CHECK-NEXT:    beq .LBB18_5
; CHECK-NEXT:  @ %bb.1: @ %do.body.preheader
; CHECK-NEXT:    ldr.w r12, [sp, #20]
; CHECK-NEXT:  .LBB18_2: @ %do.body
; CHECK-NEXT:    @ =>This Loop Header: Depth=1
; CHECK-NEXT:    @ Child Loop BB18_3 Depth 2
; CHECK-NEXT:    ldr r4, [r2]
; CHECK-NEXT:    dls lr, r5
; CHECK-NEXT:    vdup.32 q0, r4
; CHECK-NEXT:  .LBB18_3: @ %while.body
; CHECK-NEXT:    @ Parent Loop BB18_2 Depth=1
; CHECK-NEXT:    @ => This Inner Loop Header: Depth=2
; CHECK-NEXT:    vldrw.u32 q1, [r1], #16
; CHECK-NEXT:    vldrw.u32 q2, [r0], #16
; CHECK-NEXT:    vfms.f32 q2, q1, q0
; CHECK-NEXT:    vstrb.8 q2, [r3], #16
; CHECK-NEXT:    le lr, .LBB18_3
; CHECK-NEXT:  @ %bb.4: @ %while.end
; CHECK-NEXT:    @ in Loop: Header=BB18_2 Depth=1
; CHECK-NEXT:    adds r2, #4
; CHECK-NEXT:    subs.w r12, r12, #1
; CHECK-NEXT:    bne .LBB18_2
; CHECK-NEXT:  .LBB18_5: @ %do.end
; CHECK-NEXT:    pop {r4, r5, r7, pc}
entry:
  %shr = lshr i32 %N, 2
  %cmp15 = icmp eq i32 %shr, 0
  br i1 %cmp15, label %do.end, label %do.body

do.body:                                          ; preds = %while.end, %entry
  %pDst.addr.0 = phi ptr [ %add.ptr2, %while.end ], [ %pDst, %entry ]
  %M.addr.0 = phi i32 [ %dec3, %while.end ], [ %M, %entry ]
  %pSrc3.addr.0 = phi ptr [ %incdec.ptr, %while.end ], [ %pSrc3, %entry ]
  %pSrc2.addr.0 = phi ptr [ %add.ptr1, %while.end ], [ %pSrc2, %entry ]
  %pSrc1.addr.0 = phi ptr [ %add.ptr, %while.end ], [ %pSrc1, %entry ]
  %i = load float, ptr %pSrc3.addr.0, align 4
  %.splatinsert = insertelement <4 x float> undef, float %i, i32 0
  %.splat = shufflevector <4 x float> %.splatinsert, <4 x float> undef, <4 x i32> zeroinitializer
  br label %while.body

while.body:                                       ; preds = %while.body, %do.body
  %pSrc1.addr.119 = phi ptr [ %pSrc1.addr.0, %do.body ], [ %add.ptr, %while.body ]
  %pSrc2.addr.118 = phi ptr [ %pSrc2.addr.0, %do.body ], [ %add.ptr1, %while.body ]
  %blkCnt.017 = phi i32 [ %shr, %do.body ], [ %dec, %while.body ]
  %pDst.addr.116 = phi ptr [ %pDst.addr.0, %do.body ], [ %add.ptr2, %while.body ]
  %i2 = load <4 x float>, ptr %pSrc1.addr.119, align 4
  %i4 = load <4 x float>, ptr %pSrc2.addr.118, align 4
  %i5 = fneg fast <4 x float> %i4
  %i6 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %.splat, <4 x float> %i5, <4 x float> %i2)
  store <4 x float> %i6, ptr %pDst.addr.116, align 4
  %add.ptr = getelementptr inbounds float, ptr %pSrc1.addr.119, i32 4
  %add.ptr1 = getelementptr inbounds float, ptr %pSrc2.addr.118, i32 4
  %add.ptr2 = getelementptr inbounds float, ptr %pDst.addr.116, i32 4
  %dec = add nsw i32 %blkCnt.017, -1
  %cmp = icmp eq i32 %dec, 0
  br i1 %cmp, label %while.end, label %while.body

while.end:                                        ; preds = %while.body
  %incdec.ptr = getelementptr inbounds float, ptr %pSrc3.addr.0, i32 1
  %dec3 = add i32 %M.addr.0, -1
  %cmp4 = icmp eq i32 %dec3, 0
  br i1 %cmp4, label %do.end, label %do.body

do.end:                                           ; preds = %while.end, %entry
  ret void
}


%struct.arm_biquad_casd_df1_inst_f32 = type { i32, ptr, ptr }
define arm_aapcs_vfpcc void @arm_biquad_cascade_df1_f32(ptr nocapture readonly %S, ptr nocapture readonly %pSrc, ptr nocapture %pDst, i32 %blockSize) {
; CHECK-LABEL: arm_biquad_cascade_df1_f32:
; 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 #16
; CHECK-NEXT:    sub sp, #16
; CHECK-NEXT:    ldrd r7, r9, [r0]
; CHECK-NEXT:    and r6, r3, #3
; CHECK-NEXT:    ldr r0, [r0, #8]
; CHECK-NEXT:    lsrs r3, r3, #2
; CHECK-NEXT:    @ implicit-def: $r12
; CHECK-NEXT:    str r6, [sp, #4] @ 4-byte Spill
; CHECK-NEXT:    str r3, [sp] @ 4-byte Spill
; CHECK-NEXT:    str r2, [sp, #8] @ 4-byte Spill
; CHECK-NEXT:    b .LBB19_3
; CHECK-NEXT:  .LBB19_1: @ in Loop: Header=BB19_3 Depth=1
; CHECK-NEXT:    mov r3, r8
; CHECK-NEXT:    mov r2, r5
; CHECK-NEXT:    mov r4, r11
; CHECK-NEXT:    mov r8, r10
; CHECK-NEXT:  .LBB19_2: @ %if.end69
; CHECK-NEXT:    @ in Loop: Header=BB19_3 Depth=1
; CHECK-NEXT:    ldr r7, [sp, #12] @ 4-byte Reload
; CHECK-NEXT:    adds r0, #128
; CHECK-NEXT:    strd r2, r4, [r9]
; CHECK-NEXT:    ldr r2, [sp, #8] @ 4-byte Reload
; CHECK-NEXT:    subs r7, #1
; CHECK-NEXT:    strd r3, r8, [r9, #8]
; CHECK-NEXT:    add.w r9, r9, #16
; CHECK-NEXT:    mov r1, r2
; CHECK-NEXT:    beq.w .LBB19_13
; CHECK-NEXT:  .LBB19_3: @ %do.body
; CHECK-NEXT:    @ =>This Loop Header: Depth=1
; CHECK-NEXT:    @ Child Loop BB19_5 Depth 2
; CHECK-NEXT:    ldrd r5, r11, [r9]
; CHECK-NEXT:    mov r6, r2
; CHECK-NEXT:    ldrd r8, r10, [r9, #8]
; CHECK-NEXT:    ldr r2, [sp] @ 4-byte Reload
; CHECK-NEXT:    str r7, [sp, #12] @ 4-byte Spill
; CHECK-NEXT:    wls lr, r2, .LBB19_6
; CHECK-NEXT:  @ %bb.4: @ %while.body.lr.ph
; CHECK-NEXT:    @ in Loop: Header=BB19_3 Depth=1
; CHECK-NEXT:    ldr r6, [sp, #8] @ 4-byte Reload
; CHECK-NEXT:    mov r4, r11
; CHECK-NEXT:    mov r3, r5
; CHECK-NEXT:  .LBB19_5: @ %while.body
; CHECK-NEXT:    @ Parent Loop BB19_3 Depth=1
; CHECK-NEXT:    @ => This Inner Loop Header: Depth=2
; CHECK-NEXT:    ldr r5, [r1, #12]
; CHECK-NEXT:    vldrw.u32 q2, [r0]
; CHECK-NEXT:    vldrw.u32 q6, [r0, #16]
; CHECK-NEXT:    ldm.w r1, {r2, r7, r11}
; CHECK-NEXT:    vmul.f32 q2, q2, r5
; CHECK-NEXT:    vldrw.u32 q7, [r0, #32]
; CHECK-NEXT:    vfma.f32 q2, q6, r11
; CHECK-NEXT:    vldrw.u32 q4, [r0, #48]
; CHECK-NEXT:    vfma.f32 q2, q7, r7
; CHECK-NEXT:    vldrw.u32 q5, [r0, #64]
; CHECK-NEXT:    vfma.f32 q2, q4, r2
; CHECK-NEXT:    vldrw.u32 q3, [r0, #80]
; CHECK-NEXT:    vfma.f32 q2, q5, r3
; CHECK-NEXT:    vldrw.u32 q1, [r0, #96]
; CHECK-NEXT:    vfma.f32 q2, q3, r4
; CHECK-NEXT:    vldrw.u32 q0, [r0, #112]
; CHECK-NEXT:    vfma.f32 q2, q1, r8
; CHECK-NEXT:    adds r1, #16
; CHECK-NEXT:    vfma.f32 q2, q0, r10
; CHECK-NEXT:    mov r4, r11
; CHECK-NEXT:    vmov r10, r8, d5
; CHECK-NEXT:    vstrb.8 q2, [r6], #16
; CHECK-NEXT:    mov r3, r5
; CHECK-NEXT:    mov r12, r5
; CHECK-NEXT:    le lr, .LBB19_5
; CHECK-NEXT:  .LBB19_6: @ %while.end
; CHECK-NEXT:    @ in Loop: Header=BB19_3 Depth=1
; CHECK-NEXT:    ldr r3, [sp, #4] @ 4-byte Reload
; CHECK-NEXT:    cmp r3, #0
; CHECK-NEXT:    beq .LBB19_1
; CHECK-NEXT:  @ %bb.7: @ %if.then
; CHECK-NEXT:    @ in Loop: Header=BB19_3 Depth=1
; CHECK-NEXT:    ldrd lr, r4, [r1]
; CHECK-NEXT:    vldrw.u32 q0, [r0]
; CHECK-NEXT:    ldrd r2, r1, [r1, #8]
; CHECK-NEXT:    vldrw.u32 q6, [r0, #16]
; CHECK-NEXT:    vldrw.u32 q7, [r0, #32]
; CHECK-NEXT:    vldrw.u32 q4, [r0, #48]
; CHECK-NEXT:    vmul.f32 q0, q0, r1
; CHECK-NEXT:    vldrw.u32 q5, [r0, #64]
; CHECK-NEXT:    vfma.f32 q0, q6, r2
; CHECK-NEXT:    vldrw.u32 q3, [r0, #80]
; CHECK-NEXT:    vfma.f32 q0, q7, r4
; CHECK-NEXT:    vldrw.u32 q2, [r0, #96]
; CHECK-NEXT:    vfma.f32 q0, q4, lr
; CHECK-NEXT:    vldrw.u32 q1, [r0, #112]
; CHECK-NEXT:    vfma.f32 q0, q5, r5
; CHECK-NEXT:    cmp r3, #1
; CHECK-NEXT:    vfma.f32 q0, q3, r11
; CHECK-NEXT:    vfma.f32 q0, q2, r8
; CHECK-NEXT:    vfma.f32 q0, q1, r10
; CHECK-NEXT:    vmov r5, s0
; CHECK-NEXT:    bne .LBB19_9
; CHECK-NEXT:  @ %bb.8: @ %if.then58
; CHECK-NEXT:    @ in Loop: Header=BB19_3 Depth=1
; CHECK-NEXT:    str r5, [r6]
; CHECK-NEXT:    mov r2, lr
; CHECK-NEXT:    mov r4, r12
; CHECK-NEXT:    mov r3, r5
; CHECK-NEXT:    b .LBB19_12
; CHECK-NEXT:  .LBB19_9: @ %if.else
; CHECK-NEXT:    @ in Loop: Header=BB19_3 Depth=1
; CHECK-NEXT:    vmov r8, s1
; CHECK-NEXT:    cmp r3, #2
; CHECK-NEXT:    vstr s1, [r6, #4]
; CHECK-NEXT:    str r5, [r6]
; CHECK-NEXT:    bne .LBB19_11
; CHECK-NEXT:  @ %bb.10: @ in Loop: Header=BB19_3 Depth=1
; CHECK-NEXT:    mov r2, r4
; CHECK-NEXT:    mov r3, r8
; CHECK-NEXT:    mov r4, lr
; CHECK-NEXT:    mov r8, r5
; CHECK-NEXT:    b .LBB19_12
; CHECK-NEXT:  .LBB19_11: @ %if.else64
; CHECK-NEXT:    @ in Loop: Header=BB19_3 Depth=1
; CHECK-NEXT:    vmov r3, s2
; CHECK-NEXT:    vstr s2, [r6, #8]
; CHECK-NEXT:  .LBB19_12: @ %if.end69
; CHECK-NEXT:    @ in Loop: Header=BB19_3 Depth=1
; CHECK-NEXT:    mov r12, r1
; CHECK-NEXT:    b .LBB19_2
; CHECK-NEXT:  .LBB19_13: @ %do.end
; CHECK-NEXT:    add sp, #16
; 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}
entry:
  %pState1 = getelementptr inbounds %struct.arm_biquad_casd_df1_inst_f32, ptr %S, i32 0, i32 1
  %i = load ptr, ptr %pState1, align 4
  %pCoeffs2 = getelementptr inbounds %struct.arm_biquad_casd_df1_inst_f32, ptr %S, i32 0, i32 2
  %i1 = load ptr, ptr %pCoeffs2, align 4
  %numStages = getelementptr inbounds %struct.arm_biquad_casd_df1_inst_f32, ptr %S, i32 0, i32 0
  %i2 = load i32, ptr %numStages, align 4
  %shr = lshr i32 %blockSize, 2
  %cmp201 = icmp eq i32 %shr, 0
  %and = and i32 %blockSize, 3
  %tobool = icmp eq i32 %and, 0
  %cmp57 = icmp eq i32 %and, 1
  %cmp60 = icmp eq i32 %and, 2
  br label %do.body

do.body:                                          ; preds = %if.end69, %entry
  %pState.0 = phi ptr [ %i, %entry ], [ %incdec.ptr73, %if.end69 ]
  %pCoeffs.0 = phi ptr [ %i1, %entry ], [ %add.ptr74, %if.end69 ]
  %pIn.0 = phi ptr [ %pSrc, %entry ], [ %pDst, %if.end69 ]
  %X3.0 = phi float [ undef, %entry ], [ %X3.2, %if.end69 ]
  %stage.0 = phi i32 [ %i2, %entry ], [ %dec75, %if.end69 ]
  %i3 = load float, ptr %pState.0, align 4
  %arrayidx3 = getelementptr inbounds float, ptr %pState.0, i32 1
  %i4 = load float, ptr %arrayidx3, align 4
  %arrayidx4 = getelementptr inbounds float, ptr %pState.0, i32 2
  %i5 = load float, ptr %arrayidx4, align 4
  %arrayidx5 = getelementptr inbounds float, ptr %pState.0, i32 3
  %i6 = load float, ptr %arrayidx5, align 4
  br i1 %cmp201, label %while.end, label %while.body.lr.ph

while.body.lr.ph:                                 ; preds = %do.body
  %arrayidx9 = getelementptr inbounds float, ptr %pCoeffs.0, i32 4
  %arrayidx12 = getelementptr inbounds float, ptr %pCoeffs.0, i32 8
  %arrayidx15 = getelementptr inbounds float, ptr %pCoeffs.0, i32 12
  %arrayidx18 = getelementptr inbounds float, ptr %pCoeffs.0, i32 16
  %arrayidx21 = getelementptr inbounds float, ptr %pCoeffs.0, i32 20
  %arrayidx24 = getelementptr inbounds float, ptr %pCoeffs.0, i32 24
  %arrayidx27 = getelementptr inbounds float, ptr %pCoeffs.0, i32 28
  br label %while.body

while.body:                                       ; preds = %while.body, %while.body.lr.ph
  %sample.0208 = phi i32 [ %shr, %while.body.lr.ph ], [ %dec, %while.body ]
  %pIn.1207 = phi ptr [ %pIn.0, %while.body.lr.ph ], [ %incdec.ptr8, %while.body ]
  %pOut.1206 = phi ptr [ %pDst, %while.body.lr.ph ], [ %add.ptr, %while.body ]
  %Yn2.0205 = phi float [ %i6, %while.body.lr.ph ], [ %i37, %while.body ]
  %Yn1.0204 = phi float [ %i5, %while.body.lr.ph ], [ %i36, %while.body ]
  %Xn2.0203 = phi float [ %i4, %while.body.lr.ph ], [ %i17, %while.body ]
  %Xn1.0202 = phi float [ %i3, %while.body.lr.ph ], [ %i18, %while.body ]
  %incdec.ptr = getelementptr inbounds float, ptr %pIn.1207, i32 1
  %i15 = load float, ptr %pIn.1207, align 4
  %incdec.ptr6 = getelementptr inbounds float, ptr %pIn.1207, i32 2
  %i16 = load float, ptr %incdec.ptr, align 4
  %incdec.ptr7 = getelementptr inbounds float, ptr %pIn.1207, i32 3
  %i17 = load float, ptr %incdec.ptr6, align 4
  %incdec.ptr8 = getelementptr inbounds float, ptr %pIn.1207, i32 4
  %i18 = load float, ptr %incdec.ptr7, align 4
  %i19 = load <4 x float>, ptr %pCoeffs.0, align 4
  %.splatinsert = insertelement <4 x float> undef, float %i18, i32 0
  %.splat = shufflevector <4 x float> %.splatinsert, <4 x float> undef, <4 x i32> zeroinitializer
  %i20 = fmul fast <4 x float> %.splat, %i19
  %i21 = load <4 x float>, ptr %arrayidx9, align 4
  %.splatinsert10 = insertelement <4 x float> undef, float %i17, i32 0
  %.splat11 = shufflevector <4 x float> %.splatinsert10, <4 x float> undef, <4 x i32> zeroinitializer
  %i22 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i21, <4 x float> %.splat11, <4 x float> %i20)
  %i23 = load <4 x float>, ptr %arrayidx12, align 4
  %.splatinsert13 = insertelement <4 x float> undef, float %i16, i32 0
  %.splat14 = shufflevector <4 x float> %.splatinsert13, <4 x float> undef, <4 x i32> zeroinitializer
  %i24 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i23, <4 x float> %.splat14, <4 x float> %i22)
  %i25 = load <4 x float>, ptr %arrayidx15, align 4
  %.splatinsert16 = insertelement <4 x float> undef, float %i15, i32 0
  %.splat17 = shufflevector <4 x float> %.splatinsert16, <4 x float> undef, <4 x i32> zeroinitializer
  %i26 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i25, <4 x float> %.splat17, <4 x float> %i24)
  %i27 = load <4 x float>, ptr %arrayidx18, align 4
  %.splatinsert19 = insertelement <4 x float> undef, float %Xn1.0202, i32 0
  %.splat20 = shufflevector <4 x float> %.splatinsert19, <4 x float> undef, <4 x i32> zeroinitializer
  %i28 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i27, <4 x float> %.splat20, <4 x float> %i26)
  %i29 = load <4 x float>, ptr %arrayidx21, align 4
  %.splatinsert22 = insertelement <4 x float> undef, float %Xn2.0203, i32 0
  %.splat23 = shufflevector <4 x float> %.splatinsert22, <4 x float> undef, <4 x i32> zeroinitializer
  %i30 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i29, <4 x float> %.splat23, <4 x float> %i28)
  %i31 = load <4 x float>, ptr %arrayidx24, align 4
  %.splatinsert25 = insertelement <4 x float> undef, float %Yn1.0204, i32 0
  %.splat26 = shufflevector <4 x float> %.splatinsert25, <4 x float> undef, <4 x i32> zeroinitializer
  %i32 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i31, <4 x float> %.splat26, <4 x float> %i30)
  %i33 = load <4 x float>, ptr %arrayidx27, align 4
  %.splatinsert28 = insertelement <4 x float> undef, float %Yn2.0205, i32 0
  %.splat29 = shufflevector <4 x float> %.splatinsert28, <4 x float> undef, <4 x i32> zeroinitializer
  %i34 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i33, <4 x float> %.splat29, <4 x float> %i32)
  store <4 x float> %i34, ptr %pOut.1206, align 4
  %add.ptr = getelementptr inbounds float, ptr %pOut.1206, i32 4
  %i36 = extractelement <4 x float> %i34, i32 3
  %i37 = extractelement <4 x float> %i34, i32 2
  %dec = add nsw i32 %sample.0208, -1
  %cmp = icmp eq i32 %dec, 0
  br i1 %cmp, label %while.end, label %while.body

while.end:                                        ; preds = %while.body, %do.body
  %Xn1.0.lcssa = phi float [ %i3, %do.body ], [ %i18, %while.body ]
  %Xn2.0.lcssa = phi float [ %i4, %do.body ], [ %i17, %while.body ]
  %Yn1.0.lcssa = phi float [ %i5, %do.body ], [ %i36, %while.body ]
  %Yn2.0.lcssa = phi float [ %i6, %do.body ], [ %i37, %while.body ]
  %pOut.1.lcssa = phi ptr [ %pDst, %do.body ], [ %add.ptr, %while.body ]
  %pIn.1.lcssa = phi ptr [ %pIn.0, %do.body ], [ %incdec.ptr8, %while.body ]
  %X3.1.lcssa = phi float [ %X3.0, %do.body ], [ %i18, %while.body ]
  br i1 %tobool, label %if.end69, label %if.then

if.then:                                          ; preds = %while.end
  %incdec.ptr30 = getelementptr inbounds float, ptr %pIn.1.lcssa, i32 1
  %i38 = load float, ptr %pIn.1.lcssa, align 4
  %incdec.ptr31 = getelementptr inbounds float, ptr %pIn.1.lcssa, i32 2
  %i39 = load float, ptr %incdec.ptr30, align 4
  %incdec.ptr32 = getelementptr inbounds float, ptr %pIn.1.lcssa, i32 3
  %i40 = load float, ptr %incdec.ptr31, align 4
  %i41 = load float, ptr %incdec.ptr32, align 4
  %i43 = load <4 x float>, ptr %pCoeffs.0, align 4
  %.splatinsert34 = insertelement <4 x float> undef, float %i41, i32 0
  %.splat35 = shufflevector <4 x float> %.splatinsert34, <4 x float> undef, <4 x i32> zeroinitializer
  %i44 = fmul fast <4 x float> %.splat35, %i43
  %arrayidx36 = getelementptr inbounds float, ptr %pCoeffs.0, i32 4
  %i46 = load <4 x float>, ptr %arrayidx36, align 4
  %.splatinsert37 = insertelement <4 x float> undef, float %i40, i32 0
  %.splat38 = shufflevector <4 x float> %.splatinsert37, <4 x float> undef, <4 x i32> zeroinitializer
  %i47 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i46, <4 x float> %.splat38, <4 x float> %i44)
  %arrayidx39 = getelementptr inbounds float, ptr %pCoeffs.0, i32 8
  %i49 = load <4 x float>, ptr %arrayidx39, align 4
  %.splatinsert40 = insertelement <4 x float> undef, float %i39, i32 0
  %.splat41 = shufflevector <4 x float> %.splatinsert40, <4 x float> undef, <4 x i32> zeroinitializer
  %i50 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i49, <4 x float> %.splat41, <4 x float> %i47)
  %arrayidx42 = getelementptr inbounds float, ptr %pCoeffs.0, i32 12
  %i52 = load <4 x float>, ptr %arrayidx42, align 4
  %.splatinsert43 = insertelement <4 x float> undef, float %i38, i32 0
  %.splat44 = shufflevector <4 x float> %.splatinsert43, <4 x float> undef, <4 x i32> zeroinitializer
  %i53 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i52, <4 x float> %.splat44, <4 x float> %i50)
  %arrayidx45 = getelementptr inbounds float, ptr %pCoeffs.0, i32 16
  %i55 = load <4 x float>, ptr %arrayidx45, align 4
  %.splatinsert46 = insertelement <4 x float> undef, float %Xn1.0.lcssa, i32 0
  %.splat47 = shufflevector <4 x float> %.splatinsert46, <4 x float> undef, <4 x i32> zeroinitializer
  %i56 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i55, <4 x float> %.splat47, <4 x float> %i53)
  %arrayidx48 = getelementptr inbounds float, ptr %pCoeffs.0, i32 20
  %i58 = load <4 x float>, ptr %arrayidx48, align 4
  %.splatinsert49 = insertelement <4 x float> undef, float %Xn2.0.lcssa, i32 0
  %.splat50 = shufflevector <4 x float> %.splatinsert49, <4 x float> undef, <4 x i32> zeroinitializer
  %i59 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i58, <4 x float> %.splat50, <4 x float> %i56)
  %arrayidx51 = getelementptr inbounds float, ptr %pCoeffs.0, i32 24
  %i61 = load <4 x float>, ptr %arrayidx51, align 4
  %.splatinsert52 = insertelement <4 x float> undef, float %Yn1.0.lcssa, i32 0
  %.splat53 = shufflevector <4 x float> %.splatinsert52, <4 x float> undef, <4 x i32> zeroinitializer
  %i62 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i61, <4 x float> %.splat53, <4 x float> %i59)
  %arrayidx54 = getelementptr inbounds float, ptr %pCoeffs.0, i32 28
  %i64 = load <4 x float>, ptr %arrayidx54, align 4
  %.splatinsert55 = insertelement <4 x float> undef, float %Yn2.0.lcssa, i32 0
  %.splat56 = shufflevector <4 x float> %.splatinsert55, <4 x float> undef, <4 x i32> zeroinitializer
  %i65 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i64, <4 x float> %.splat56, <4 x float> %i62)
  %i66 = extractelement <4 x float> %i65, i32 0
  br i1 %cmp57, label %if.then58, label %if.else

if.then58:                                        ; preds = %if.then
  store float %i66, ptr %pOut.1.lcssa, align 4
  br label %if.end69

if.else:                                          ; preds = %if.then
  %incdec.ptr62 = getelementptr inbounds float, ptr %pOut.1.lcssa, i32 1
  store float %i66, ptr %pOut.1.lcssa, align 4
  %i67 = extractelement <4 x float> %i65, i32 1
  store float %i67, ptr %incdec.ptr62, align 4
  br i1 %cmp60, label %if.end69, label %if.else64

if.else64:                                        ; preds = %if.else
  %incdec.ptr63 = getelementptr inbounds float, ptr %pOut.1.lcssa, i32 2
  %i68 = extractelement <4 x float> %i65, i32 2
  store float %i68, ptr %incdec.ptr63, align 4
  br label %if.end69

if.end69:                                         ; preds = %if.else64, %if.else, %if.then58, %while.end
  %Xn1.1 = phi float [ %i38, %if.then58 ], [ %i40, %if.else64 ], [ %Xn1.0.lcssa, %while.end ], [ %i39, %if.else ]
  %Xn2.1 = phi float [ %X3.1.lcssa, %if.then58 ], [ %i39, %if.else64 ], [ %Xn2.0.lcssa, %while.end ], [ %i38, %if.else ]
  %Yn1.1 = phi float [ %i66, %if.then58 ], [ %i68, %if.else64 ], [ %Yn1.0.lcssa, %while.end ], [ %i67, %if.else ]
  %Yn2.1 = phi float [ %Yn1.0.lcssa, %if.then58 ], [ %i67, %if.else64 ], [ %Yn2.0.lcssa, %while.end ], [ %i66, %if.else ]
  %X3.2 = phi float [ %i41, %if.then58 ], [ %i41, %if.else64 ], [ %X3.1.lcssa, %while.end ], [ %i41, %if.else ]
  store float %Xn1.1, ptr %pState.0, align 4
  store float %Xn2.1, ptr %arrayidx3, align 4
  store float %Yn1.1, ptr %arrayidx4, align 4
  %incdec.ptr73 = getelementptr inbounds float, ptr %pState.0, i32 4
  store float %Yn2.1, ptr %arrayidx5, align 4
  %add.ptr74 = getelementptr inbounds float, ptr %pCoeffs.0, i32 32
  %dec75 = add i32 %stage.0, -1
  %cmp76 = icmp eq i32 %dec75, 0
  br i1 %cmp76, label %do.end, label %do.body

do.end:                                           ; preds = %if.end69
  ret void
}


%struct.arm_biquad_cascade_df2T_instance_f32 = type { i8, ptr, ptr }
define void @arm_biquad_cascade_df2T_f32(ptr nocapture readonly %S, ptr nocapture readonly %pSrc, ptr nocapture %pDst, i32 %blockSize) {
; CHECK-LABEL: arm_biquad_cascade_df2T_f32:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r4, r5, r6, r7, r8, lr}
; CHECK-NEXT:    push.w {r4, r5, r6, r7, r8, lr}
; CHECK-NEXT:    .vsave {d8, d9, d10, d11, d12, d13}
; CHECK-NEXT:    vpush {d8, d9, d10, d11, d12, d13}
; CHECK-NEXT:    ldrd r6, r12, [r0, #4]
; CHECK-NEXT:    lsr.w r8, r3, #1
; CHECK-NEXT:    ldrb r0, [r0]
; CHECK-NEXT:    vldr s0, .LCPI20_0
; CHECK-NEXT:    b .LBB20_3
; CHECK-NEXT:  .LBB20_1: @ %if.else
; CHECK-NEXT:    @ in Loop: Header=BB20_3 Depth=1
; CHECK-NEXT:    vmov.f32 s14, s13
; CHECK-NEXT:    vstr s12, [r6]
; CHECK-NEXT:  .LBB20_2: @ %if.end
; CHECK-NEXT:    @ in Loop: Header=BB20_3 Depth=1
; CHECK-NEXT:    vstr s14, [r6, #4]
; CHECK-NEXT:    add.w r12, r12, #20
; CHECK-NEXT:    adds r6, #8
; CHECK-NEXT:    subs r0, #1
; CHECK-NEXT:    mov r1, r2
; CHECK-NEXT:    beq .LBB20_8
; CHECK-NEXT:  .LBB20_3: @ %do.body
; CHECK-NEXT:    @ =>This Loop Header: Depth=1
; CHECK-NEXT:    @ Child Loop BB20_5 Depth 2
; CHECK-NEXT:    vldrw.u32 q2, [r12]
; CHECK-NEXT:    movs r5, #0
; CHECK-NEXT:    vmov q4, q2
; CHECK-NEXT:    vshlc q4, r5, #32
; CHECK-NEXT:    vldrw.u32 q1, [r12, #8]
; CHECK-NEXT:    vmov q5, q1
; CHECK-NEXT:    vshlc q5, r5, #32
; CHECK-NEXT:    vldrw.u32 q3, [r6]
; CHECK-NEXT:    vmov.f32 s14, s0
; CHECK-NEXT:    mov r5, r2
; CHECK-NEXT:    vmov.f32 s15, s0
; CHECK-NEXT:    wls lr, r8, .LBB20_6
; CHECK-NEXT:  @ %bb.4: @ %while.body.preheader
; CHECK-NEXT:    @ in Loop: Header=BB20_3 Depth=1
; CHECK-NEXT:    vmov q6, q3
; CHECK-NEXT:    mov r5, r2
; CHECK-NEXT:  .LBB20_5: @ %while.body
; CHECK-NEXT:    @ Parent Loop BB20_3 Depth=1
; CHECK-NEXT:    @ => This Inner Loop Header: Depth=2
; CHECK-NEXT:    ldrd r7, r4, [r1], #8
; CHECK-NEXT:    vfma.f32 q6, q2, r7
; CHECK-NEXT:    vmov r7, s24
; CHECK-NEXT:    vmov q3, q6
; CHECK-NEXT:    vfma.f32 q3, q1, r7
; CHECK-NEXT:    vstr s24, [r5]
; CHECK-NEXT:    vmov.f32 s15, s0
; CHECK-NEXT:    vfma.f32 q3, q4, r4
; CHECK-NEXT:    vmov r4, s13
; CHECK-NEXT:    vstr s13, [r5, #4]
; CHECK-NEXT:    vfma.f32 q3, q5, r4
; CHECK-NEXT:    adds r5, #8
; CHECK-NEXT:    vmov.f32 s12, s14
; CHECK-NEXT:    vmov.f32 s13, s15
; CHECK-NEXT:    vmov.f32 s14, s0
; CHECK-NEXT:    vmov q6, q3
; CHECK-NEXT:    le lr, .LBB20_5
; CHECK-NEXT:  .LBB20_6: @ %while.end
; CHECK-NEXT:    @ in Loop: Header=BB20_3 Depth=1
; CHECK-NEXT:    lsls r7, r3, #31
; CHECK-NEXT:    beq .LBB20_1
; CHECK-NEXT:  @ %bb.7: @ %if.then
; CHECK-NEXT:    @ in Loop: Header=BB20_3 Depth=1
; CHECK-NEXT:    ldr r1, [r1]
; CHECK-NEXT:    vfma.f32 q3, q2, r1
; CHECK-NEXT:    vmov r1, s12
; CHECK-NEXT:    vstr s12, [r5]
; CHECK-NEXT:    vfma.f32 q3, q1, r1
; CHECK-NEXT:    vstr s13, [r6]
; CHECK-NEXT:    b .LBB20_2
; CHECK-NEXT:  .LBB20_8: @ %do.end
; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13}
; CHECK-NEXT:    pop.w {r4, r5, r6, r7, r8, pc}
; CHECK-NEXT:    .p2align 2
; CHECK-NEXT:  @ %bb.9:
; CHECK-NEXT:  .LCPI20_0:
; CHECK-NEXT:    .long 0x00000000 @ float 0
entry:
  %pState1 = getelementptr inbounds %struct.arm_biquad_cascade_df2T_instance_f32, ptr %S, i32 0, i32 1
  %i = load ptr, ptr %pState1, align 4
  %numStages = getelementptr inbounds %struct.arm_biquad_cascade_df2T_instance_f32, ptr %S, i32 0, i32 0
  %i1 = load i8, ptr %numStages, align 4
  %conv = zext i8 %i1 to i32
  %pCoeffs = getelementptr inbounds %struct.arm_biquad_cascade_df2T_instance_f32, ptr %S, i32 0, i32 2
  %i2 = load ptr, ptr %pCoeffs, align 4
  %div = lshr i32 %blockSize, 1
  %cmp.not90 = icmp eq i32 %div, 0
  %and = and i32 %blockSize, 1
  %tobool.not = icmp eq i32 %and, 0
  br label %do.body

do.body:                                          ; preds = %if.end, %entry
  %stage.0 = phi i32 [ %conv, %entry ], [ %dec23, %if.end ]
  %pCurCoeffs.0 = phi ptr [ %i2, %entry ], [ %add.ptr2, %if.end ]
  %pState.0 = phi ptr [ %i, %entry ], [ %pState.1, %if.end ]
  %pIn.0 = phi ptr [ %pSrc, %entry ], [ %pDst, %if.end ]
  %i4 = load <4 x float>, ptr %pCurCoeffs.0, align 4
  %add.ptr = getelementptr inbounds float, ptr %pCurCoeffs.0, i32 2
  %i6 = load <4 x float>, ptr %add.ptr, align 4
  %add.ptr2 = getelementptr inbounds float, ptr %pCurCoeffs.0, i32 5
  %i8 = load <4 x float>, ptr %pState.0, align 8
  %i9 = shufflevector <4 x float> %i8, <4 x float> <float poison, float poison, float 0.000000e+00, float 0.000000e+00>, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
  %i10 = bitcast <4 x float> %i4 to <4 x i32>
  %i11 = tail call { i32, <4 x i32> } @llvm.arm.mve.vshlc.v4i32(<4 x i32> %i10, i32 0, i32 32)
  %i12 = extractvalue { i32, <4 x i32> } %i11, 0
  %i13 = extractvalue { i32, <4 x i32> } %i11, 1
  %i14 = bitcast <4 x i32> %i13 to <4 x float>
  %i15 = bitcast <4 x float> %i6 to <4 x i32>
  %i16 = tail call { i32, <4 x i32> } @llvm.arm.mve.vshlc.v4i32(<4 x i32> %i15, i32 %i12, i32 32)
  %i17 = extractvalue { i32, <4 x i32> } %i16, 1
  %i18 = bitcast <4 x i32> %i17 to <4 x float>
  br i1 %cmp.not90, label %while.end, label %while.body

while.body:                                       ; preds = %while.body, %do.body
  %pIn.194 = phi ptr [ %incdec.ptr4, %while.body ], [ %pIn.0, %do.body ]
  %state.093 = phi <4 x float> [ %i30, %while.body ], [ %i9, %do.body ]
  %pOut.192 = phi ptr [ %incdec.ptr12, %while.body ], [ %pDst, %do.body ]
  %sample.091 = phi i32 [ %dec, %while.body ], [ %div, %do.body ]
  %incdec.ptr = getelementptr inbounds float, ptr %pIn.194, i32 1
  %i19 = load float, ptr %pIn.194, align 4
  %incdec.ptr4 = getelementptr inbounds float, ptr %pIn.194, i32 2
  %i20 = load float, ptr %incdec.ptr, align 4
  %.splatinsert = insertelement <4 x float> poison, float %i19, i32 0
  %.splat = shufflevector <4 x float> %.splatinsert, <4 x float> poison, <4 x i32> zeroinitializer
  %i21 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i4, <4 x float> %.splat, <4 x float> %state.093)
  %i22 = extractelement <4 x float> %i21, i32 0
  %.splat6 = shufflevector <4 x float> %i21, <4 x float> poison, <4 x i32> zeroinitializer
  %i23 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i6, <4 x float> %.splat6, <4 x float> %i21)
  %i24 = insertelement <4 x float> %i23, float 0.000000e+00, i32 3
  %.splatinsert7 = insertelement <4 x float> poison, float %i20, i32 0
  %.splat8 = shufflevector <4 x float> %.splatinsert7, <4 x float> poison, <4 x i32> zeroinitializer
  %i25 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i14, <4 x float> %.splat8, <4 x float> %i24)
  %i26 = extractelement <4 x float> %i25, i32 1
  %.splat10 = shufflevector <4 x float> %i25, <4 x float> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
  %i27 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i18, <4 x float> %.splat10, <4 x float> %i25)
  %i28 = shufflevector <4 x float> %i27, <4 x float> undef, <4 x i32> <i32 2, i32 undef, i32 undef, i32 3>
  %i29 = insertelement <4 x float> %i28, float 0.000000e+00, i32 2
  %i30 = shufflevector <4 x float> %i29, <4 x float> %i27, <4 x i32> <i32 0, i32 7, i32 2, i32 3>
  %incdec.ptr11 = getelementptr inbounds float, ptr %pOut.192, i32 1
  store float %i22, ptr %pOut.192, align 4
  %incdec.ptr12 = getelementptr inbounds float, ptr %pOut.192, i32 2
  store float %i26, ptr %incdec.ptr11, align 4
  %dec = add nsw i32 %sample.091, -1
  %cmp.not = icmp eq i32 %dec, 0
  br i1 %cmp.not, label %while.end, label %while.body

while.end:                                        ; preds = %while.body, %do.body
  %pOut.1.lcssa = phi ptr [ %pDst, %do.body ], [ %incdec.ptr12, %while.body ]
  %state.0.lcssa = phi <4 x float> [ %i9, %do.body ], [ %i30, %while.body ]
  %pIn.1.lcssa = phi ptr [ %pIn.0, %do.body ], [ %incdec.ptr4, %while.body ]
  br i1 %tobool.not, label %if.else, label %if.then

if.then:                                          ; preds = %while.end
  %i31 = load float, ptr %pIn.1.lcssa, align 4
  %.splatinsert14 = insertelement <4 x float> poison, float %i31, i32 0
  %.splat15 = shufflevector <4 x float> %.splatinsert14, <4 x float> poison, <4 x i32> zeroinitializer
  %i32 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i4, <4 x float> %.splat15, <4 x float> %state.0.lcssa)
  %i33 = extractelement <4 x float> %i32, i32 0
  %.splat17 = shufflevector <4 x float> %i32, <4 x float> poison, <4 x i32> zeroinitializer
  %i34 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %i6, <4 x float> %.splat17, <4 x float> %i32)
  store float %i33, ptr %pOut.1.lcssa, align 4
  %i35 = extractelement <4 x float> %i34, i32 1
  store float %i35, ptr %pState.0, align 4
  %i36 = extractelement <4 x float> %i34, i32 2
  br label %if.end

if.else:                                          ; preds = %while.end
  %i37 = extractelement <4 x float> %state.0.lcssa, i32 0
  store float %i37, ptr %pState.0, align 4
  %i38 = extractelement <4 x float> %state.0.lcssa, i32 1
  br label %if.end

if.end:                                           ; preds = %if.else, %if.then
  %.sink = phi float [ %i38, %if.else ], [ %i36, %if.then ]
  %i39 = getelementptr inbounds float, ptr %pState.0, i32 1
  store float %.sink, ptr %i39, align 4
  %pState.1 = getelementptr inbounds float, ptr %pState.0, i32 2
  %dec23 = add i32 %stage.0, -1
  %cmp24.not = icmp eq i32 %dec23, 0
  br i1 %cmp24.not, label %do.end, label %do.body

do.end:                                           ; preds = %if.end
  ret void
}

define arm_aapcs_vfpcc float @vecAddAcrossF32Mve(<4 x float> %in) {
; CHECK-LABEL: vecAddAcrossF32Mve:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vadd.f32 s0, s0, s1
; CHECK-NEXT:    vadd.f32 s0, s0, s2
; CHECK-NEXT:    vadd.f32 s0, s0, s3
; CHECK-NEXT:    bx lr
entry:
  %i = extractelement <4 x float> %in, i32 0
  %i1 = extractelement <4 x float> %in, i32 1
  %add = fadd fast float %i, %i1
  %i2 = extractelement <4 x float> %in, i32 2
  %add1 = fadd fast float %add, %i2
  %i3 = extractelement <4 x float> %in, i32 3
  %add2 = fadd fast float %add1, %i3
  ret float %add2
}


declare { i32, <4 x i32> } @llvm.arm.mve.vshlc.v4i32(<4 x i32>, i32, i32) #1
declare void @llvm.lifetime.end.p0(i64 immarg, ptr nocapture)
declare void @llvm.lifetime.start.p0(i64 immarg, ptr nocapture)
declare { <4 x i32>, i32 } @llvm.arm.mve.viwdup.v4i32(i32, i32, i32)
declare <4 x i32> @llvm.arm.mve.vldr.gather.offset.v4i32.p0.v4i32(ptr, <4 x i32>, i32, i32, i32)
declare void @llvm.assume(i1)
declare <4 x i1> @llvm.arm.mve.vctp32(i32)
declare <4 x float> @llvm.fma.v4f32(<4 x float>, <4 x float>, <4 x float>)
declare void @llvm.masked.store.v4f32.p0(<4 x float>, ptr, i32 immarg, <4 x i1>)