PowerPC/SPE: Grab the emergency slot for the vreg(that was created by the eliminateFramePointer)

[long5hot]

Summary

For the SPE loads and store's (that requires 8 bit offest unlike others load and store instructions 16 bit offset)

PR : https://github.com/llvm/llvm-project/issues/62766

[jhibbits]

This looks just like D78669, which I've admittedly been very lazy on.

[umesh.kalappa0]

In D151711#4381449, @jhibbits wrote:

This looks just like D78669, which I've admittedly been very lazy on.

@jhibbits can we land this commit to master please ...

[jhibbits]

llvm/lib/Target/PowerPC/PPCFrameLowering.cpp

2304

Not a fan of including a lambda in a conditional expression. It makes the expression as a whole harder to read.

[long5hot]

removal of lambda expression from conditional statement, and updated existent test case.

[umesh.kalappa0]

@nemanjai ,Please can you help us to commit the changes .

[long5hot]

Removin Lambda expression.

[long5hot]

ping!

llvm/lib/Target/PowerPC/PPCFrameLowering.cpp

2304

got it!

[jhibbits]

Thanks for dropping the lambda, it's now cleaner than my original patch as well. I know @sfertile asked me for a simpler test case, but I don't think it's really feasible.

[DavidSpickett]

The asan test failure is unrelated to your change. You've got review approval so go ahead and land the change.

llvm/lib/Target/PowerPC/PPCFrameLowering.cpp

PPCFrameLowering::addScavengingSpillSlot(MachineFunction &MF,
   // large offset.
 
   // We need to have a scavenger spill slot for spills if the frame size is
   // large. In case there is no free register for large-offset addressing,
   // this slot is used for the necessary emergency spill. Also, we need the
   // slot for dynamic stack allocations.
 
   // The scavenger might be invoked if the frame offset does not fit into
-  // the 16-bit immediate. We don't know the complete frame size here
-  // because we've not yet computed callee-saved register spills or the
-  // needed alignment padding.
+  // the 16-bit immediate in case of not SPE and 8-bit in case of SPE.
+  // We don't know the complete frame size here because we've not yet computed
+  // callee-saved register spills or the needed alignment padding.
+
   unsigned StackSize = determineFrameLayout(MF, true);
   MachineFrameInfo &MFI = MF.getFrameInfo();
+  const PPCSubtarget &Subtarget = MF.getSubtarget<PPCSubtarget>();
+  bool NeedSpills = [&StackSize, &Subtarget]() {
+    if (Subtarget.hasSPE())
+      return !isInt<8>(StackSize);
+    else
+      return !isInt<16>(StackSize);
+  }();
+
   if (MFI.hasVarSizedObjects() || spillsCR(MF) || hasNonRISpills(MF) ||
-      (hasSpills(MF) && !isInt<16>(StackSize))) {
+      (hasSpills(MF) && NeedSpills)) {
     const TargetRegisterClass &GPRC = PPC::GPRCRegClass;
     const TargetRegisterClass &G8RC = PPC::G8RCRegClass;
     const TargetRegisterClass &RC = Subtarget.isPPC64() ? G8RC : GPRC;
     const TargetRegisterInfo &TRI = *Subtarget.getRegisterInfo();
     unsigned Size = TRI.getSpillSize(RC);
     Align Alignment = TRI.getSpillAlign(RC);
     RS->addScavengingFrameIndex(MFI.CreateStackObject(Size, Alignment, false));
 

llvm/test/CodeGen/PowerPC/register-pressure.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -verify-machineinstrs < %s -mtriple=powerpc-unknown-linux-gnu \
+; RUN:          -mattr=+spe | FileCheck %s -check-prefixes=CHECK,SPE
+
+target datalayout = "E-m:e-p:32:32-Fn32-i64:64-n32"
+target triple = "ppc32"
+
+%struct.cmplx = type { double, double }
+
+; Function Attrs: noinline nounwind optnone uwtable
+define dso_local i32 @main() #0 {
+; CHECK-LABEL: main:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    mflr 0
+; CHECK-NEXT:    stwu 1, -48(1)
+; CHECK-NEXT:    stw 31, 44(1)
+; CHECK-NEXT:    stw 0, 52(1)
+; CHECK-NEXT:    .cfi_def_cfa_offset 48
+; CHECK-NEXT:    .cfi_offset r31, -4
+; CHECK-NEXT:    .cfi_offset lr, 4
+; CHECK-NEXT:    mr 31, 1
+; CHECK-NEXT:    .cfi_def_cfa_register r31
+; CHECK-NEXT:    li 3, 10
+; CHECK-NEXT:    stw 3, 40(31)
+; CHECK-NEXT:    li 3, 0
+; CHECK-NEXT:    stw 3, 28(31)
+; CHECK-NEXT:    lis 4, 16404
+; CHECK-NEXT:    stw 4, 24(31)
+; CHECK-NEXT:    stw 3, 36(31)
+; CHECK-NEXT:    lis 3, 16420
+; CHECK-NEXT:    stw 3, 32(31)
+; CHECK-NEXT:    lwz 3, 40(31)
+; CHECK-NEXT:    slwi 3, 3, 4
+; CHECK-NEXT:    bl malloc
+; CHECK-NEXT:    stw 3, 20(31)
+; CHECK-NEXT:    addi 7, 31, 24
+; CHECK-NEXT:    stw 7, 16(31)
+; CHECK-NEXT:    lwz 3, 20(31)
+; CHECK-NEXT:    stw 3, 12(31)
+; CHECK-NEXT:    lwz 5, 16(31)
+; CHECK-NEXT:    lwz 6, 12(31)
+; CHECK-NEXT:    li 3, 5
+; CHECK-NEXT:    li 4, 1
+; CHECK-NEXT:    li 8, 1
+; CHECK-NEXT:    bl pass11
+; CHECK-NEXT:    li 3, 0
+; CHECK-NEXT:    lwz 0, 52(1)
+; CHECK-NEXT:    lwz 31, 44(1)
+; CHECK-NEXT:    addi 1, 1, 48
+; CHECK-NEXT:    mtlr 0
+; CHECK-NEXT:    blr
+  %1 = alloca i32, align 4
+  %2 = alloca %struct.cmplx, align 8
+  %3 = alloca ptr, align 4
+  %4 = alloca ptr, align 4
+  %5 = alloca ptr, align 4
+  store i32 10, ptr %1, align 4
+  %6 = getelementptr inbounds %struct.cmplx, ptr %2, i32 0, i32 0
+  store double 5.000000e+00, ptr %6, align 8
+  %7 = getelementptr inbounds %struct.cmplx, ptr %2, i32 0, i32 1
+  store double 1.000000e+01, ptr %7, align 8
+  %8 = load i32, ptr %1, align 4
+  %9 = mul i32 %8, 16
+  %10 = call ptr @malloc(i32 noundef %9)
+  store ptr %10, ptr %3, align 4
+  store ptr %2, ptr %4, align 4
+  %11 = load ptr, ptr %3, align 4
+  store ptr %11, ptr %5, align 4
+  %12 = load ptr, ptr %4, align 4
+  %13 = load ptr, ptr %5, align 4
+  call void @pass11(i32 noundef 5, i32 noundef 1, ptr noundef %12, ptr noundef %13, ptr noundef %2, i32 noundef 1)
+  ret i32 0
+}
+
+declare dso_local ptr @malloc(i32 noundef) #1
+
+; Function Attrs: noinline nounwind optnone uwtable
+define internal void @pass11(i32 noundef %0, i32 noundef %1, ptr noalias noundef %2, ptr noalias noundef %3, ptr noalias noundef %4, i32 noundef %5) #0 {
+; CHECK-LABEL: pass11:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    stwu 1, -1040(1)
+; CHECK-NEXT:    stw 31, 1036(1)
+; CHECK-NEXT:    .cfi_def_cfa_offset 1040
+; CHECK-NEXT:    .cfi_offset r31, -4
+; CHECK-NEXT:    mr 31, 1
+; CHECK-NEXT:    .cfi_def_cfa_register r31
+; CHECK-NEXT:    .cfi_offset r15, -152
+; CHECK-NEXT:    .cfi_offset r16, -144
+; CHECK-NEXT:    .cfi_offset r17, -136
+; CHECK-NEXT:    .cfi_offset r18, -128
+; CHECK-NEXT:    .cfi_offset r19, -120
+; CHECK-NEXT:    .cfi_offset r20, -112
+; CHECK-NEXT:    .cfi_offset r21, -104
+; CHECK-NEXT:    .cfi_offset r22, -96
+; CHECK-NEXT:    .cfi_offset r23, -88
+; CHECK-NEXT:    .cfi_offset r24, -80
+; CHECK-NEXT:    .cfi_offset r25, -72
+; CHECK-NEXT:    .cfi_offset r26, -64
+; CHECK-NEXT:    .cfi_offset r27, -56
+; CHECK-NEXT:    .cfi_offset r28, -48
+; CHECK-NEXT:    .cfi_offset r29, -40
+; CHECK-NEXT:    .cfi_offset r30, -32
+; CHECK-NEXT:    li 9, 888
+; CHECK-NEXT:    evstddx 15, 31, 9 # 8-byte Folded Spill
+; CHECK-NEXT:    li 9, 896
+; CHECK-NEXT:    evstddx 16, 31, 9 # 8-byte Folded Spill
+; CHECK-NEXT:    li 9, 904
+; CHECK-NEXT:    evstddx 17, 31, 9 # 8-byte Folded Spill
+; CHECK-NEXT:    li 9, 912
+; CHECK-NEXT:    evstddx 18, 31, 9 # 8-byte Folded Spill
+; CHECK-NEXT:    li 9, 920
+; CHECK-NEXT:    evstddx 19, 31, 9 # 8-byte Folded Spill
+; CHECK-NEXT:    li 9, 928
+; CHECK-NEXT:    evstddx 20, 31, 9 # 8-byte Folded Spill
+; CHECK-NEXT:    li 9, 936
+; CHECK-NEXT:    evstddx 21, 31, 9 # 8-byte Folded Spill
+; CHECK-NEXT:    li 9, 944
+; CHECK-NEXT:    evstddx 22, 31, 9 # 8-byte Folded Spill
+; CHECK-NEXT:    li 9, 952
+; CHECK-NEXT:    evstddx 23, 31, 9 # 8-byte Folded Spill
+; CHECK-NEXT:    li 9, 960
+; CHECK-NEXT:    evstddx 24, 31, 9 # 8-byte Folded Spill
+; CHECK-NEXT:    li 9, 968
+; CHECK-NEXT:    evstddx 25, 31, 9 # 8-byte Folded Spill
+; CHECK-NEXT:    li 9, 976
+; CHECK-NEXT:    evstddx 26, 31, 9 # 8-byte Folded Spill
+; CHECK-NEXT:    li 9, 984
+; CHECK-NEXT:    evstddx 27, 31, 9 # 8-byte Folded Spill
+; CHECK-NEXT:    li 9, 992
+; CHECK-NEXT:    evstddx 28, 31, 9 # 8-byte Folded Spill
+; CHECK-NEXT:    li 9, 1000
+; CHECK-NEXT:    evstddx 29, 31, 9 # 8-byte Folded Spill
+; CHECK-NEXT:    li 9, 1008
+; CHECK-NEXT:    evstddx 30, 31, 9 # 8-byte Folded Spill
+; CHECK-NEXT:    stw 3, 884(31)
+; CHECK-NEXT:    stw 4, 880(31)
+; CHECK-NEXT:    stw 5, 876(31)
+; CHECK-NEXT:    stw 6, 872(31)
+; CHECK-NEXT:    stw 7, 868(31)
+; CHECK-NEXT:    stw 8, 864(31)
+; CHECK-NEXT:    li 3, 11
+; CHECK-NEXT:    stw 3, 860(31)
+; CHECK-NEXT:    lis 3, -30876
+; CHECK-NEXT:    ori 3, 3, 61626
+; CHECK-NEXT:    stw 3, 852(31)
+; CHECK-NEXT:    lis 3, 16362
+; CHECK-NEXT:    ori 3, 3, 60300
+; CHECK-NEXT:    stw 3, 848(31)
+; CHECK-NEXT:    lwz 3, 864(31)
+; CHECK-NEXT:    efdcfsi 3, 3
+; CHECK-NEXT:    li 4, .LCPI1_0@l
+; CHECK-NEXT:    lis 5, .LCPI1_0@ha
+; CHECK-NEXT:    evlddx 4, 5, 4
+; CHECK-NEXT:    efdmul 3, 3, 4
+; CHECK-NEXT:    li 4, 840
+; CHECK-NEXT:    evstddx 3, 31, 4
+; CHECK-NEXT:    lis 3, -9785
+; CHECK-NEXT:    ori 3, 3, 4790
+; CHECK-NEXT:    stw 3, 836(31)
+; CHECK-NEXT:    lis 3, 16346
+; CHECK-NEXT:    ori 3, 3, 38440
+; CHECK-NEXT:    stw 3, 832(31)
+; CHECK-NEXT:    lwz 3, 864(31)
+; CHECK-NEXT:    efdcfsi 3, 3
+; CHECK-NEXT:    li 4, .LCPI1_1@l
+; CHECK-NEXT:    lis 5, .LCPI1_1@ha
+; CHECK-NEXT:    evlddx 4, 5, 4
+; CHECK-NEXT:    efdmul 3, 3, 4
+; CHECK-NEXT:    li 4, 824
+; CHECK-NEXT:    evstddx 3, 31, 4
+; CHECK-NEXT:    lis 3, 25615
+; CHECK-NEXT:    ori 3, 3, 17627
+; CHECK-NEXT:    stw 3, 820(31)
+; CHECK-NEXT:    lis 3, -16446
+; CHECK-NEXT:    ori 3, 3, 14175
+; CHECK-NEXT:    stw 3, 816(31)
+; CHECK-NEXT:    lwz 3, 864(31)
+; CHECK-NEXT:    efdcfsi 3, 3
+; CHECK-NEXT:    li 4, .LCPI1_2@l
+; CHECK-NEXT:    lis 5, .LCPI1_2@ha
+; CHECK-NEXT:    evlddx 4, 5, 4
+; CHECK-NEXT:    efdmul 3, 3, 4
+; CHECK-NEXT:    li 4, 808
+; CHECK-NEXT:    evstddx 3, 31, 4
+; CHECK-NEXT:    lis 3, 32631
+; CHECK-NEXT:    ori 3, 3, 22663
+; CHECK-NEXT:    stw 3, 804(31)
+; CHECK-NEXT:    lis 3, -16412
+; CHECK-NEXT:    ori 3, 3, 62622
+; CHECK-NEXT:    stw 3, 800(31)
+; CHECK-NEXT:    lwz 3, 864(31)
+; CHECK-NEXT:    efdcfsi 3, 3
+; CHECK-NEXT:    li 4, .LCPI1_3@l
+; CHECK-NEXT:    lis 5, .LCPI1_3@ha
+; CHECK-NEXT:    evlddx 4, 5, 4
+; CHECK-NEXT:    efdmul 3, 3, 4
+; CHECK-NEXT:    li 4, 792
+; CHECK-NEXT:    evstddx 3, 31, 4
+; CHECK-NEXT:    lis 3, -25651
+; CHECK-NEXT:    ori 3, 3, 20567
+; CHECK-NEXT:    stw 3, 788(31)
+; CHECK-NEXT:    lis 3, -16402
+; CHECK-NEXT:    ori 3, 3, 46122
+; CHECK-NEXT:    stw 3, 784(31)
+; CHECK-NEXT:    lwz 3, 864(31)
+; CHECK-NEXT:    efdcfsi 3, 3
+; CHECK-NEXT:    li 4, .LCPI1_4@l
+; CHECK-NEXT:    lis 5, .LCPI1_4@ha
+; CHECK-NEXT:    evlddx 4, 5, 4
+; CHECK-NEXT:    efdmul 3, 3, 4
+; CHECK-NEXT:    li 4, 776
+; CHECK-NEXT:    evstddx 3, 31, 4
+; CHECK-NEXT:    li 3, 0
+; CHECK-NEXT:    stw 3, 772(31)
+; CHECK-NEXT:    b .LBB1_1
+; CHECK-NEXT:  .LBB1_1: # =>This Loop Header: Depth=1
+; CHECK-NEXT:    # Child Loop BB1_3 Depth 2
+; CHECK-NEXT:    lwz 3, 772(31)
+; CHECK-NEXT:    lwz 4, 880(31)
+; CHECK-NEXT:    cmplw 3, 4
+; CHECK-NEXT:    bge 0, .LBB1_8
+; CHECK-NEXT:    b .LBB1_2
+; CHECK-NEXT:  .LBB1_2:
+; CHECK-NEXT:    lwz 3, 876(31)
+; CHECK-NEXT:    lwz 4, 884(31)
+; CHECK-NEXT:    lwz 5, 772(31)
+; CHECK-NEXT:    mullw 4, 5, 4
+; CHECK-NEXT:    mulli 4, 4, 176
+; CHECK-NEXT:    lwzux 4, 3, 4
+; CHECK-NEXT:    stw 4, 752(31)
+; CHECK-NEXT:    lwz 4, 12(3)
+; CHECK-NEXT:    stw 4, 764(31)
+; CHECK-NEXT:    lwz 4, 8(3)
+; CHECK-NEXT:    stw 4, 760(31)
+; CHECK-NEXT:    lwz 3, 4(3)
+; CHECK-NEXT:    stw 3, 756(31)
+; CHECK-NEXT:    lwz 3, 876(31)
+; CHECK-NEXT:    lwz 4, 884(31)
+; CHECK-NEXT:    lwz 5, 772(31)
+; CHECK-NEXT:    mulli 5, 5, 11
+; CHECK-NEXT:    addi 6, 5, 1
+; CHECK-NEXT:    mullw 6, 4, 6
+; CHECK-NEXT:    slwi 6, 6, 4
+; CHECK-NEXT:    evlddx 6, 3, 6
+; CHECK-NEXT:    addi 5, 5, 10
+; CHECK-NEXT:    mullw 4, 4, 5
+; CHECK-NEXT:    slwi 4, 4, 4
+; CHECK-NEXT:    evlddx 3, 3, 4
+; CHECK-NEXT:    efdadd 3, 6, 3
+; CHECK-NEXT:    li 4, 736
+; CHECK-NEXT:    evstddx 3, 31, 4
+; CHECK-NEXT:    lwz 4, 876(31)
+; CHECK-NEXT:    lwz 5, 884(31)
+; CHECK-NEXT:    lwz 3, 772(31)
+; CHECK-NEXT:    mulli 6, 3, 11
+; CHECK-NEXT:    addi 3, 6, 1
+; CHECK-NEXT:    mullw 3, 5, 3
+; CHECK-NEXT:    slwi 3, 3, 4
+; CHECK-NEXT:    add 7, 4, 3
+; CHECK-NEXT:    li 3, 8
+; CHECK-NEXT:    evlddx 7, 7, 3
+; CHECK-NEXT:    addi 6, 6, 10
+; CHECK-NEXT:    mullw 5, 5, 6
+; CHECK-NEXT:    slwi 5, 5, 4
+; CHECK-NEXT:    add 4, 4, 5
+; CHECK-NEXT:    evlddx 4, 4, 3
+; CHECK-NEXT:    efdadd 4, 7, 4
+; CHECK-NEXT:    addi 5, 31, 736
+; CHECK-NEXT:    evstddx 4, 5, 3
+; CHECK-NEXT:    lwz 4, 876(31)
+; CHECK-NEXT:    lwz 6, 884(31)
+; CHECK-NEXT:    lwz 7, 772(31)
+; CHECK-NEXT:    mulli 7, 7, 11
+; CHECK-NEXT:    addi 8, 7, 1
+; CHECK-NEXT:    mullw 8, 6, 8
+; CHECK-NEXT:    slwi 8, 8, 4
+; CHECK-NEXT:    evlddx 8, 4, 8
+; CHECK-NEXT:    addi 7, 7, 10
+; CHECK-NEXT:    mullw 6, 6, 7
+; CHECK-NEXT:    slwi 6, 6, 4
+; CHECK-NEXT:    evlddx 4, 4, 6
+; CHECK-NEXT:    efdsub 4, 8, 4
+; CHECK-NEXT:    li 6, 592
+; CHECK-NEXT:    evstddx 4, 31, 6
+; CHECK-NEXT:    lwz 4, 876(31)
+; CHECK-NEXT:    lwz 6, 884(31)
+; CHECK-NEXT:    lwz 7, 772(31)
+; CHECK-NEXT:    mulli 7, 7, 11
+; CHECK-NEXT:    addi 8, 7, 1
+; CHECK-NEXT:    mullw 8, 6, 8
+; CHECK-NEXT:    slwi 8, 8, 4
+; CHECK-NEXT:    add 8, 4, 8
+; CHECK-NEXT:    evlddx 8, 8, 3
+; CHECK-NEXT:    addi 7, 7, 10
+; CHECK-NEXT:    mullw 6, 6, 7
+; CHECK-NEXT:    slwi 6, 6, 4
+; CHECK-NEXT:    add 4, 4, 6
+; CHECK-NEXT:    evlddx 4, 4, 3
+; CHECK-NEXT:    efdsub 6, 8, 4
+; CHECK-NEXT:    addi 4, 31, 592
+; CHECK-NEXT:    evstddx 6, 4, 3
+; CHECK-NEXT:    lwz 6, 876(31)
+; CHECK-NEXT:    lwz 7, 884(31)
+; CHECK-NEXT:    lwz 8, 772(31)
+; CHECK-NEXT:    mulli 8, 8, 11
+; CHECK-NEXT:    addi 9, 8, 2
+; CHECK-NEXT:    mullw 9, 7, 9
+; CHECK-NEXT:    slwi 9, 9, 4
+; CHECK-NEXT:    evlddx 9, 6, 9
+; CHECK-NEXT:    addi 8, 8, 9
+; CHECK-NEXT:    mullw 7, 7, 8
+; CHECK-NEXT:    slwi 7, 7, 4
+; CHECK-NEXT:    evlddx 6, 6, 7
+; CHECK-NEXT:    efdadd 6, 9, 6
+; CHECK-NEXT:    li 7, 720
+; CHECK-NEXT:    evstddx 6, 31, 7
+; CHECK-NEXT:    lwz 6, 876(31)
+; CHECK-NEXT:    lwz 7, 884(31)
+; CHECK-NEXT:    lwz 8, 772(31)
+; CHECK-NEXT:    mulli 8, 8, 11
+; CHECK-NEXT:    addi 9, 8, 2
+; CHECK-NEXT:    mullw 9, 7, 9
+; CHECK-NEXT:    slwi 9, 9, 4
+; CHECK-NEXT:    add 9, 6, 9
+; CHECK-NEXT:    evlddx 9, 9, 3
+; CHECK-NEXT:    addi 8, 8, 9
+; CHECK-NEXT:    mullw 7, 7, 8
+; CHECK-NEXT:    slwi 7, 7, 4
+; CHECK-NEXT:    add 6, 6, 7
+; CHECK-NEXT:    evlddx 6, 6, 3
+; CHECK-NEXT:    efdadd 6, 9, 6
+; CHECK-NEXT:    addi 7, 31, 720
+; CHECK-NEXT:    evstddx 6, 7, 3
+; CHECK-NEXT:    lwz 6, 876(31)
+; CHECK-NEXT:    lwz 8, 884(31)
+; CHECK-NEXT:    lwz 9, 772(31)
+; CHECK-NEXT:    mulli 9, 9, 11
+; CHECK-NEXT:    addi 10, 9, 2
+; CHECK-NEXT:    mullw 10, 8, 10
+; CHECK-NEXT:    slwi 10, 10, 4
+; CHECK-NEXT:    evlddx 10, 6, 10
+; CHECK-NEXT:    addi 9, 9, 9
+; CHECK-NEXT:    mullw 8, 8, 9
+; CHECK-NEXT:    slwi 8, 8, 4
+; CHECK-NEXT:    evlddx 6, 6, 8
+; CHECK-NEXT:    efdsub 6, 10, 6
+; CHECK-NEXT:    li 8, 608
+; CHECK-NEXT:    evstddx 6, 31, 8
+; CHECK-NEXT:    lwz 6, 876(31)
+; CHECK-NEXT:    lwz 8, 884(31)
+; CHECK-NEXT:    lwz 9, 772(31)
+; CHECK-NEXT:    mulli 9, 9, 11
+; CHECK-NEXT:    addi 10, 9, 2
+; CHECK-NEXT:    mullw 10, 8, 10
+; CHECK-NEXT:    slwi 10, 10, 4
+; CHECK-NEXT:    add 10, 6, 10
+; CHECK-NEXT:    evlddx 10, 10, 3
+; CHECK-NEXT:    addi 9, 9, 9
+; CHECK-NEXT:    mullw 8, 8, 9
+; CHECK-NEXT:    slwi 8, 8, 4
+; CHECK-NEXT:    add 6, 6, 8
+; CHECK-NEXT:    evlddx 6, 6, 3
+; CHECK-NEXT:    efdsub 8, 10, 6
+; CHECK-NEXT:    addi 6, 31, 608
+; CHECK-NEXT:    evstddx 8, 6, 3
+; CHECK-NEXT:    lwz 8, 876(31)
+; CHECK-NEXT:    lwz 9, 884(31)
+; CHECK-NEXT:    lwz 10, 772(31)
+; CHECK-NEXT:    mulli 10, 10, 11
+; CHECK-NEXT:    addi 11, 10, 3
+; CHECK-NEXT:    mullw 11, 9, 11
+; CHECK-NEXT:    slwi 11, 11, 4
+; CHECK-NEXT:    evlddx 11, 8, 11
+; CHECK-NEXT:    addi 10, 10, 8
+; CHECK-NEXT:    mullw 9, 9, 10
+; CHECK-NEXT:    slwi 9, 9, 4
+; CHECK-NEXT:    evlddx 8, 8, 9
+; CHECK-NEXT:    efdadd 8, 11, 8
+; CHECK-NEXT:    li 9, 704
+; CHECK-NEXT:    evstddx 8, 31, 9
+; CHECK-NEXT:    lwz 8, 876(31)
+; CHECK-NEXT:    lwz 9, 884(31)
+; CHECK-NEXT:    lwz 10, 772(31)
+; CHECK-NEXT:    mulli 10, 10, 11
+; CHECK-NEXT:    addi 11, 10, 3
+; CHECK-NEXT:    mullw 11, 9, 11
+; CHECK-NEXT:    slwi 11, 11, 4
+; CHECK-NEXT:    add 11, 8, 11
+; CHECK-NEXT:    evlddx 11, 11, 3
+; CHECK-NEXT:    addi 10, 10, 8
+; CHECK-NEXT:    mullw 9, 9, 10
+; CHECK-NEXT:    slwi 9, 9, 4
+; CHECK-NEXT:    add 8, 8, 9
+; CHECK-NEXT:    evlddx 8, 8, 3
+; CHECK-NEXT:    efdadd 8, 11, 8
+; CHECK-NEXT:    addi 9, 31, 704
+; CHECK-NEXT:    evstddx 8, 9, 3
+; CHECK-NEXT:    lwz 8, 876(31)
+; CHECK-NEXT:    lwz 10, 884(31)
+; CHECK-NEXT:    lwz 11, 772(31)
+; CHECK-NEXT:    mulli 11, 11, 11
+; CHECK-NEXT:    addi 12, 11, 3
+; CHECK-NEXT:    mullw 12, 10, 12
+; CHECK-NEXT:    slwi 12, 12, 4
+; CHECK-NEXT:    evlddx 12, 8, 12
+; CHECK-NEXT:    addi 11, 11, 8
+; CHECK-NEXT:    mullw 10, 10, 11
+; CHECK-NEXT:    slwi 10, 10, 4
+; CHECK-NEXT:    evlddx 8, 8, 10
+; CHECK-NEXT:    efdsub 8, 12, 8
+; CHECK-NEXT:    li 10, 624
+; CHECK-NEXT:    evstddx 8, 31, 10
+; CHECK-NEXT:    lwz 8, 876(31)
+; CHECK-NEXT:    lwz 10, 884(31)
+; CHECK-NEXT:    lwz 11, 772(31)
+; CHECK-NEXT:    mulli 11, 11, 11
+; CHECK-NEXT:    addi 12, 11, 3
+; CHECK-NEXT:    mullw 12, 10, 12
+; CHECK-NEXT:    slwi 12, 12, 4
+; CHECK-NEXT:    add 12, 8, 12
+; CHECK-NEXT:    evlddx 12, 12, 3
+; CHECK-NEXT:    addi 11, 11, 8
+; CHECK-NEXT:    mullw 10, 10, 11
+; CHECK-NEXT:    slwi 10, 10, 4
+; CHECK-NEXT:    add 8, 8, 10
+; CHECK-NEXT:    evlddx 8, 8, 3
+; CHECK-NEXT:    efdsub 10, 12, 8
+; CHECK-NEXT:    addi 8, 31, 624
+; CHECK-NEXT:    evstddx 10, 8, 3
+; CHECK-NEXT:    lwz 10, 876(31)
+; CHECK-NEXT:    lwz 11, 884(31)
+; CHECK-NEXT:    lwz 12, 772(31)
+; CHECK-NEXT:    mulli 12, 12, 11
+; CHECK-NEXT:    addi 0, 12, 4
+; CHECK-NEXT:    mullw 0, 11, 0
+; CHECK-NEXT:    slwi 0, 0, 4
+; CHECK-NEXT:    evlddx 0, 10, 0
+; CHECK-NEXT:    addi 12, 12, 7
+; CHECK-NEXT:    mullw 11, 11, 12
+; CHECK-NEXT:    slwi 11, 11, 4
+; CHECK-NEXT:    evlddx 10, 10, 11
+; CHECK-NEXT:    efdadd 10, 0, 10
+; CHECK-NEXT:    li 11, 688
+; CHECK-NEXT:    evstddx 10, 31, 11
+; CHECK-NEXT:    lwz 10, 876(31)
+; CHECK-NEXT:    lwz 11, 884(31)
+; CHECK-NEXT:    lwz 12, 772(31)
+; CHECK-NEXT:    mulli 12, 12, 11
+; CHECK-NEXT:    addi 0, 12, 4
+; CHECK-NEXT:    mullw 0, 11, 0
+; CHECK-NEXT:    slwi 0, 0, 4
+; CHECK-NEXT:    add 30, 10, 0
+; CHECK-NEXT:    evlddx 0, 30, 3
+; CHECK-NEXT:    addi 12, 12, 7
+; CHECK-NEXT:    mullw 11, 11, 12
+; CHECK-NEXT:    slwi 11, 11, 4
+; CHECK-NEXT:    add 10, 10, 11
+; CHECK-NEXT:    evlddx 10, 10, 3
+; CHECK-NEXT:    efdadd 10, 0, 10
+; CHECK-NEXT:    addi 11, 31, 688
+; CHECK-NEXT:    evstddx 10, 11, 3
+; CHECK-NEXT:    lwz 10, 876(31)
+; CHECK-NEXT:    lwz 12, 884(31)
+; CHECK-NEXT:    lwz 0, 772(31)
+; CHECK-NEXT:    mulli 30, 0, 11
+; CHECK-NEXT:    addi 0, 30, 4
+; CHECK-NEXT:    mullw 0, 12, 0
+; CHECK-NEXT:    slwi 0, 0, 4
+; CHECK-NEXT:    evlddx 0, 10, 0
+; CHECK-NEXT:    addi 30, 30, 7
+; CHECK-NEXT:    mullw 12, 12, 30
+; CHECK-NEXT:    slwi 12, 12, 4
+; CHECK-NEXT:    evlddx 10, 10, 12
+; CHECK-NEXT:    efdsub 10, 0, 10
+; CHECK-NEXT:    li 12, 640
+; CHECK-NEXT:    evstddx 10, 31, 12
+; CHECK-NEXT:    lwz 10, 876(31)
+; CHECK-NEXT:    lwz 12, 884(31)
+; CHECK-NEXT:    lwz 0, 772(31)
+; CHECK-NEXT:    mulli 30, 0, 11
+; CHECK-NEXT:    addi 0, 30, 4
+; CHECK-NEXT:    mullw 0, 12, 0
+; CHECK-NEXT:    slwi 0, 0, 4
+; CHECK-NEXT:    add 29, 10, 0
+; CHECK-NEXT:    evlddx 0, 29, 3
+; CHECK-NEXT:    addi 30, 30, 7
+; CHECK-NEXT:    mullw 12, 12, 30
+; CHECK-NEXT:    slwi 12, 12, 4
+; CHECK-NEXT:    add 10, 10, 12
+; CHECK-NEXT:    evlddx 10, 10, 3
+; CHECK-NEXT:    efdsub 12, 0, 10
+; CHECK-NEXT:    addi 10, 31, 640
+; CHECK-NEXT:    evstddx 12, 10, 3
+; CHECK-NEXT:    lwz 12, 876(31)
+; CHECK-NEXT:    lwz 0, 884(31)
+; CHECK-NEXT:    lwz 30, 772(31)
+; CHECK-NEXT:    mulli 30, 30, 11
+; CHECK-NEXT:    addi 29, 30, 5
+; CHECK-NEXT:    mullw 29, 0, 29
+; CHECK-NEXT:    slwi 29, 29, 4
+; CHECK-NEXT:    evlddx 29, 12, 29
+; CHECK-NEXT:    addi 30, 30, 6
+; CHECK-NEXT:    mullw 0, 0, 30
+; CHECK-NEXT:    slwi 0, 0, 4
+; CHECK-NEXT:    evlddx 12, 12, 0
+; CHECK-NEXT:    efdadd 12, 29, 12
+; CHECK-NEXT:    li 30, 672
+; CHECK-NEXT:    evstddx 12, 31, 30
+; CHECK-NEXT:    lwz 12, 876(31)
+; CHECK-NEXT:    lwz 0, 884(31)
+; CHECK-NEXT:    lwz 30, 772(31)
+; CHECK-NEXT:    mulli 30, 30, 11
+; CHECK-NEXT:    addi 29, 30, 5
+; CHECK-NEXT:    mullw 29, 0, 29
+; CHECK-NEXT:    slwi 29, 29, 4
+; CHECK-NEXT:    add 29, 12, 29
+; CHECK-NEXT:    evlddx 29, 29, 3
+; CHECK-NEXT:    addi 30, 30, 6
+; CHECK-NEXT:    mullw 0, 0, 30
+; CHECK-NEXT:    slwi 0, 0, 4
+; CHECK-NEXT:    add 12, 12, 0
+; CHECK-NEXT:    evlddx 12, 12, 3
+; CHECK-NEXT:    efdadd 12, 29, 12
+; CHECK-NEXT:    addi 30, 31, 672
+; CHECK-NEXT:    evstddx 12, 30, 3
+; CHECK-NEXT:    lwz 12, 876(31)
+; CHECK-NEXT:    lwz 0, 884(31)
+; CHECK-NEXT:    lwz 29, 772(31)
+; CHECK-NEXT:    mulli 29, 29, 11
+; CHECK-NEXT:    addi 28, 29, 5
+; CHECK-NEXT:    mullw 28, 0, 28
+; CHECK-NEXT:    slwi 28, 28, 4
+; CHECK-NEXT:    evlddx 28, 12, 28
+; CHECK-NEXT:    addi 29, 29, 6
+; CHECK-NEXT:    mullw 0, 0, 29
+; CHECK-NEXT:    slwi 0, 0, 4
+; CHECK-NEXT:    evlddx 12, 12, 0
+; CHECK-NEXT:    efdsub 12, 28, 12
+; CHECK-NEXT:    li 29, 656
+; CHECK-NEXT:    evstddx 12, 31, 29
+; CHECK-NEXT:    lwz 12, 876(31)
+; CHECK-NEXT:    lwz 0, 884(31)
+; CHECK-NEXT:    lwz 29, 772(31)
+; CHECK-NEXT:    mulli 29, 29, 11
+; CHECK-NEXT:    addi 28, 29, 5
+; CHECK-NEXT:    mullw 28, 0, 28
+; CHECK-NEXT:    slwi 28, 28, 4
+; CHECK-NEXT:    add 28, 12, 28
+; CHECK-NEXT:    evlddx 28, 28, 3
+; CHECK-NEXT:    addi 29, 29, 6
+; CHECK-NEXT:    mullw 0, 0, 29
+; CHECK-NEXT:    slwi 0, 0, 4
+; CHECK-NEXT:    add 12, 12, 0
+; CHECK-NEXT:    evlddx 12, 12, 3
+; CHECK-NEXT:    efdsub 0, 28, 12
+; CHECK-NEXT:    addi 12, 31, 656
+; CHECK-NEXT:    evstddx 0, 12, 3
+; CHECK-NEXT:    li 29, 752
+; CHECK-NEXT:    evlddx 0, 31, 29
+; CHECK-NEXT:    li 29, 736
+; CHECK-NEXT:    evlddx 29, 31, 29
+; CHECK-NEXT:    efdadd 0, 0, 29
+; CHECK-NEXT:    li 29, 720
+; CHECK-NEXT:    evlddx 29, 31, 29
+; CHECK-NEXT:    efdadd 0, 0, 29
+; CHECK-NEXT:    li 29, 704
+; CHECK-NEXT:    evlddx 29, 31, 29
+; CHECK-NEXT:    efdadd 0, 0, 29
+; CHECK-NEXT:    li 29, 688
+; CHECK-NEXT:    evlddx 29, 31, 29
+; CHECK-NEXT:    efdadd 0, 0, 29
+; CHECK-NEXT:    li 29, 672
+; CHECK-NEXT:    evlddx 29, 31, 29
+; CHECK-NEXT:    efdadd 0, 0, 29
+; CHECK-NEXT:    lwz 29, 872(31)
+; CHECK-NEXT:    lwz 28, 884(31)
+; CHECK-NEXT:    lwz 27, 772(31)
+; CHECK-NEXT:    mullw 28, 28, 27
+; CHECK-NEXT:    slwi 28, 28, 4
+; CHECK-NEXT:    evstddx 0, 29, 28
+; CHECK-NEXT:    addi 29, 31, 752
+; CHECK-NEXT:    evlddx 0, 29, 3
+; CHECK-NEXT:    evlddx 28, 5, 3
+; CHECK-NEXT:    efdadd 0, 0, 28
+; CHECK-NEXT:    evlddx 28, 7, 3
+; CHECK-NEXT:    efdadd 0, 0, 28
+; CHECK-NEXT:    evlddx 28, 9, 3
+; CHECK-NEXT:    efdadd 0, 0, 28
+; CHECK-NEXT:    evlddx 28, 11, 3
+; CHECK-NEXT:    efdadd 0, 0, 28
+; CHECK-NEXT:    evlddx 28, 30, 3
+; CHECK-NEXT:    efdadd 0, 0, 28
+; CHECK-NEXT:    lwz 28, 872(31)
+; CHECK-NEXT:    lwz 27, 884(31)
+; CHECK-NEXT:    lwz 26, 772(31)
+; CHECK-NEXT:    mullw 27, 27, 26
+; CHECK-NEXT:    slwi 27, 27, 4
+; CHECK-NEXT:    add 28, 28, 27
+; CHECK-NEXT:    evstddx 0, 28, 3
+; CHECK-NEXT:    li 28, 752
+; CHECK-NEXT:    evlddx 28, 31, 28
+; CHECK-NEXT:    li 27, 736
+; CHECK-NEXT:    evlddx 27, 31, 27
+; CHECK-NEXT:    li 0, .LCPI1_5@l
+; CHECK-NEXT:    lis 26, .LCPI1_5@ha
+; CHECK-NEXT:    evlddx 0, 26, 0
+; CHECK-NEXT:    efdmul 27, 27, 0
+; CHECK-NEXT:    efdadd 27, 27, 28
+; CHECK-NEXT:    li 28, 720
+; CHECK-NEXT:    evlddx 26, 31, 28
+; CHECK-NEXT:    li 28, .LCPI1_6@l
+; CHECK-NEXT:    lis 25, .LCPI1_6@ha
+; CHECK-NEXT:    evlddx 28, 25, 28
+; CHECK-NEXT:    efdmul 26, 26, 28
+; CHECK-NEXT:    efdadd 26, 26, 27
+; CHECK-NEXT:    li 27, 704
+; CHECK-NEXT:    evlddx 25, 31, 27
+; CHECK-NEXT:    li 27, .LCPI1_7@l
+; CHECK-NEXT:    lis 24, .LCPI1_7@ha
+; CHECK-NEXT:    evlddx 27, 24, 27
+; CHECK-NEXT:    efdmul 25, 25, 27
+; CHECK-NEXT:    efdadd 26, 25, 26
+; CHECK-NEXT:    li 25, 688
+; CHECK-NEXT:    evlddx 24, 31, 25
+; CHECK-NEXT:    li 25, .LCPI1_8@l
+; CHECK-NEXT:    lis 23, .LCPI1_8@ha
+; CHECK-NEXT:    evlddx 25, 23, 25
+; CHECK-NEXT:    efdmul 24, 24, 25
+; CHECK-NEXT:    efdadd 24, 24, 26
+; CHECK-NEXT:    li 26, 672
+; CHECK-NEXT:    evlddx 23, 31, 26
+; CHECK-NEXT:    li 26, .LCPI1_9@l
+; CHECK-NEXT:    lis 22, .LCPI1_9@ha
+; CHECK-NEXT:    evlddx 26, 22, 26
+; CHECK-NEXT:    efdmul 23, 23, 26
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 576
+; CHECK-NEXT:    evstddx 24, 31, 23
+; CHECK-NEXT:    evlddx 24, 29, 3
+; CHECK-NEXT:    evlddx 23, 5, 3
+; CHECK-NEXT:    efdmul 23, 23, 0
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 7, 3
+; CHECK-NEXT:    efdmul 23, 23, 28
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 9, 3
+; CHECK-NEXT:    efdmul 23, 23, 27
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 11, 3
+; CHECK-NEXT:    efdmul 23, 23, 25
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 30, 3
+; CHECK-NEXT:    efdmul 23, 23, 26
+; CHECK-NEXT:    efdadd 23, 23, 24
+; CHECK-NEXT:    addi 24, 31, 576
+; CHECK-NEXT:    evstddx 23, 24, 3
+; CHECK-NEXT:    li 23, 840
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    li 22, 592
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 824
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    li 20, 608
+; CHECK-NEXT:    evlddx 20, 31, 20
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdmul 23, 23, 22
+; CHECK-NEXT:    efdadd 23, 23, 21
+; CHECK-NEXT:    li 22, 808
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 624
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdadd 23, 22, 23
+; CHECK-NEXT:    li 22, 792
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 640
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdadd 23, 22, 23
+; CHECK-NEXT:    li 22, 776
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 656
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdadd 22, 22, 23
+; CHECK-NEXT:    addi 23, 31, 560
+; CHECK-NEXT:    evstddx 22, 23, 3
+; CHECK-NEXT:    li 22, 840
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    evlddx 21, 4, 3
+; CHECK-NEXT:    li 20, 824
+; CHECK-NEXT:    evlddx 20, 31, 20
+; CHECK-NEXT:    evlddx 19, 6, 3
+; CHECK-NEXT:    efdmul 20, 20, 19
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdadd 22, 22, 20
+; CHECK-NEXT:    li 21, 808
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 8, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdadd 22, 21, 22
+; CHECK-NEXT:    li 21, 792
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 10, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdadd 22, 21, 22
+; CHECK-NEXT:    li 21, 776
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 12, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdadd 22, 21, 22
+; CHECK-NEXT:    efdneg 22, 22
+; CHECK-NEXT:    li 21, 560
+; CHECK-NEXT:    evstddx 22, 31, 21
+; CHECK-NEXT:    li 22, 576
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 560
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdadd 22, 22, 21
+; CHECK-NEXT:    lwz 21, 872(31)
+; CHECK-NEXT:    lwz 20, 884(31)
+; CHECK-NEXT:    lwz 19, 772(31)
+; CHECK-NEXT:    lwz 18, 880(31)
+; CHECK-NEXT:    add 19, 19, 18
+; CHECK-NEXT:    mullw 20, 20, 19
+; CHECK-NEXT:    slwi 20, 20, 4
+; CHECK-NEXT:    evstddx 22, 21, 20
+; CHECK-NEXT:    evlddx 22, 24, 3
+; CHECK-NEXT:    evlddx 21, 23, 3
+; CHECK-NEXT:    efdadd 22, 22, 21
+; CHECK-NEXT:    lwz 21, 872(31)
+; CHECK-NEXT:    lwz 20, 884(31)
+; CHECK-NEXT:    lwz 19, 772(31)
+; CHECK-NEXT:    lwz 18, 880(31)
+; CHECK-NEXT:    add 19, 19, 18
+; CHECK-NEXT:    mullw 20, 20, 19
+; CHECK-NEXT:    slwi 20, 20, 4
+; CHECK-NEXT:    add 21, 21, 20
+; CHECK-NEXT:    evstddx 22, 21, 3
+; CHECK-NEXT:    li 22, 576
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 560
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdsub 22, 22, 21
+; CHECK-NEXT:    lwz 21, 872(31)
+; CHECK-NEXT:    lwz 20, 884(31)
+; CHECK-NEXT:    lwz 19, 772(31)
+; CHECK-NEXT:    lwz 18, 880(31)
+; CHECK-NEXT:    mulli 18, 18, 10
+; CHECK-NEXT:    add 19, 19, 18
+; CHECK-NEXT:    mullw 20, 20, 19
+; CHECK-NEXT:    slwi 20, 20, 4
+; CHECK-NEXT:    evstddx 22, 21, 20
+; CHECK-NEXT:    evlddx 24, 24, 3
+; CHECK-NEXT:    evlddx 23, 23, 3
+; CHECK-NEXT:    efdsub 24, 24, 23
+; CHECK-NEXT:    lwz 23, 872(31)
+; CHECK-NEXT:    lwz 22, 884(31)
+; CHECK-NEXT:    lwz 21, 772(31)
+; CHECK-NEXT:    lwz 20, 880(31)
+; CHECK-NEXT:    mulli 20, 20, 10
+; CHECK-NEXT:    add 21, 21, 20
+; CHECK-NEXT:    mullw 22, 22, 21
+; CHECK-NEXT:    slwi 22, 22, 4
+; CHECK-NEXT:    add 23, 23, 22
+; CHECK-NEXT:    evstddx 24, 23, 3
+; CHECK-NEXT:    li 24, 752
+; CHECK-NEXT:    evlddx 24, 31, 24
+; CHECK-NEXT:    li 23, 736
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 28
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 720
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 25
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 704
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 26
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 688
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 27
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 672
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 0
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 544
+; CHECK-NEXT:    evstddx 24, 31, 23
+; CHECK-NEXT:    evlddx 24, 29, 3
+; CHECK-NEXT:    evlddx 23, 5, 3
+; CHECK-NEXT:    efdmul 23, 23, 28
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 7, 3
+; CHECK-NEXT:    efdmul 23, 23, 25
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 9, 3
+; CHECK-NEXT:    efdmul 23, 23, 26
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 11, 3
+; CHECK-NEXT:    efdmul 23, 23, 27
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 30, 3
+; CHECK-NEXT:    efdmul 23, 23, 0
+; CHECK-NEXT:    efdadd 23, 23, 24
+; CHECK-NEXT:    addi 24, 31, 544
+; CHECK-NEXT:    evstddx 23, 24, 3
+; CHECK-NEXT:    li 23, 824
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    li 22, 592
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 792
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    li 20, 608
+; CHECK-NEXT:    evlddx 20, 31, 20
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdmul 23, 23, 22
+; CHECK-NEXT:    efdadd 23, 23, 21
+; CHECK-NEXT:    li 22, 776
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 624
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdsub 23, 23, 22
+; CHECK-NEXT:    li 22, 808
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 640
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdsub 23, 23, 22
+; CHECK-NEXT:    li 22, 840
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 656
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdsub 22, 23, 22
+; CHECK-NEXT:    addi 23, 31, 528
+; CHECK-NEXT:    evstddx 22, 23, 3
+; CHECK-NEXT:    li 22, 824
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    evlddx 21, 4, 3
+; CHECK-NEXT:    li 20, 792
+; CHECK-NEXT:    evlddx 20, 31, 20
+; CHECK-NEXT:    evlddx 19, 6, 3
+; CHECK-NEXT:    efdmul 20, 20, 19
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdadd 22, 22, 20
+; CHECK-NEXT:    li 21, 776
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 8, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdsub 22, 22, 21
+; CHECK-NEXT:    li 21, 808
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 10, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdsub 22, 22, 21
+; CHECK-NEXT:    li 21, 840
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 12, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdsub 22, 22, 21
+; CHECK-NEXT:    efdneg 22, 22
+; CHECK-NEXT:    li 21, 528
+; CHECK-NEXT:    evstddx 22, 31, 21
+; CHECK-NEXT:    li 22, 544
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 528
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdadd 22, 22, 21
+; CHECK-NEXT:    lwz 21, 872(31)
+; CHECK-NEXT:    lwz 20, 884(31)
+; CHECK-NEXT:    lwz 19, 772(31)
+; CHECK-NEXT:    lwz 18, 880(31)
+; CHECK-NEXT:    slwi 18, 18, 1
+; CHECK-NEXT:    add 19, 19, 18
+; CHECK-NEXT:    mullw 20, 20, 19
+; CHECK-NEXT:    slwi 20, 20, 4
+; CHECK-NEXT:    evstddx 22, 21, 20
+; CHECK-NEXT:    evlddx 22, 24, 3
+; CHECK-NEXT:    evlddx 21, 23, 3
+; CHECK-NEXT:    efdadd 22, 22, 21
+; CHECK-NEXT:    lwz 21, 872(31)
+; CHECK-NEXT:    lwz 20, 884(31)
+; CHECK-NEXT:    lwz 19, 772(31)
+; CHECK-NEXT:    lwz 18, 880(31)
+; CHECK-NEXT:    slwi 18, 18, 1
+; CHECK-NEXT:    add 19, 19, 18
+; CHECK-NEXT:    mullw 20, 20, 19
+; CHECK-NEXT:    slwi 20, 20, 4
+; CHECK-NEXT:    add 21, 21, 20
+; CHECK-NEXT:    evstddx 22, 21, 3
+; CHECK-NEXT:    li 22, 544
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 528
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdsub 22, 22, 21
+; CHECK-NEXT:    lwz 21, 872(31)
+; CHECK-NEXT:    lwz 20, 884(31)
+; CHECK-NEXT:    lwz 19, 772(31)
+; CHECK-NEXT:    lwz 18, 880(31)
+; CHECK-NEXT:    mulli 18, 18, 9
+; CHECK-NEXT:    add 19, 19, 18
+; CHECK-NEXT:    mullw 20, 20, 19
+; CHECK-NEXT:    slwi 20, 20, 4
+; CHECK-NEXT:    evstddx 22, 21, 20
+; CHECK-NEXT:    evlddx 24, 24, 3
+; CHECK-NEXT:    evlddx 23, 23, 3
+; CHECK-NEXT:    efdsub 24, 24, 23
+; CHECK-NEXT:    lwz 23, 872(31)
+; CHECK-NEXT:    lwz 22, 884(31)
+; CHECK-NEXT:    lwz 21, 772(31)
+; CHECK-NEXT:    lwz 20, 880(31)
+; CHECK-NEXT:    mulli 20, 20, 9
+; CHECK-NEXT:    add 21, 21, 20
+; CHECK-NEXT:    mullw 22, 22, 21
+; CHECK-NEXT:    slwi 22, 22, 4
+; CHECK-NEXT:    add 23, 23, 22
+; CHECK-NEXT:    evstddx 24, 23, 3
+; CHECK-NEXT:    li 24, 752
+; CHECK-NEXT:    evlddx 24, 31, 24
+; CHECK-NEXT:    li 23, 736
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 27
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 720
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 26
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 704
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 28
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 688
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 0
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 672
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 25
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 512
+; CHECK-NEXT:    evstddx 24, 31, 23
+; CHECK-NEXT:    evlddx 24, 29, 3
+; CHECK-NEXT:    evlddx 23, 5, 3
+; CHECK-NEXT:    efdmul 23, 23, 27
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 7, 3
+; CHECK-NEXT:    efdmul 23, 23, 26
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 9, 3
+; CHECK-NEXT:    efdmul 23, 23, 28
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 11, 3
+; CHECK-NEXT:    efdmul 23, 23, 0
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 30, 3
+; CHECK-NEXT:    efdmul 23, 23, 25
+; CHECK-NEXT:    efdadd 23, 23, 24
+; CHECK-NEXT:    addi 24, 31, 512
+; CHECK-NEXT:    evstddx 23, 24, 3
+; CHECK-NEXT:    li 23, 808
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    li 22, 592
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 776
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    li 20, 608
+; CHECK-NEXT:    evlddx 20, 31, 20
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdmul 23, 23, 22
+; CHECK-NEXT:    efdsub 23, 23, 21
+; CHECK-NEXT:    li 22, 824
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 624
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdsub 23, 23, 22
+; CHECK-NEXT:    li 22, 840
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 640
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdadd 23, 22, 23
+; CHECK-NEXT:    li 22, 792
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 656
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdadd 22, 22, 23
+; CHECK-NEXT:    addi 23, 31, 496
+; CHECK-NEXT:    evstddx 22, 23, 3
+; CHECK-NEXT:    li 22, 808
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    evlddx 21, 4, 3
+; CHECK-NEXT:    li 20, 776
+; CHECK-NEXT:    evlddx 20, 31, 20
+; CHECK-NEXT:    evlddx 19, 6, 3
+; CHECK-NEXT:    efdmul 20, 20, 19
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdsub 22, 22, 20
+; CHECK-NEXT:    li 21, 824
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 8, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdsub 22, 22, 21
+; CHECK-NEXT:    li 21, 840
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 10, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdadd 22, 21, 22
+; CHECK-NEXT:    li 21, 792
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 12, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdadd 22, 21, 22
+; CHECK-NEXT:    efdneg 22, 22
+; CHECK-NEXT:    li 21, 496
+; CHECK-NEXT:    evstddx 22, 31, 21
+; CHECK-NEXT:    li 22, 512
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 496
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdadd 22, 22, 21
+; CHECK-NEXT:    lwz 21, 872(31)
+; CHECK-NEXT:    lwz 20, 884(31)
+; CHECK-NEXT:    lwz 19, 772(31)
+; CHECK-NEXT:    lwz 18, 880(31)
+; CHECK-NEXT:    mulli 18, 18, 3
+; CHECK-NEXT:    add 19, 19, 18
+; CHECK-NEXT:    mullw 20, 20, 19
+; CHECK-NEXT:    slwi 20, 20, 4
+; CHECK-NEXT:    evstddx 22, 21, 20
+; CHECK-NEXT:    evlddx 22, 24, 3
+; CHECK-NEXT:    evlddx 21, 23, 3
+; CHECK-NEXT:    efdadd 22, 22, 21
+; CHECK-NEXT:    lwz 21, 872(31)
+; CHECK-NEXT:    lwz 20, 884(31)
+; CHECK-NEXT:    lwz 19, 772(31)
+; CHECK-NEXT:    lwz 18, 880(31)
+; CHECK-NEXT:    mulli 18, 18, 3
+; CHECK-NEXT:    add 19, 19, 18
+; CHECK-NEXT:    mullw 20, 20, 19
+; CHECK-NEXT:    slwi 20, 20, 4
+; CHECK-NEXT:    add 21, 21, 20
+; CHECK-NEXT:    evstddx 22, 21, 3
+; CHECK-NEXT:    li 22, 512
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 496
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdsub 22, 22, 21
+; CHECK-NEXT:    lwz 21, 872(31)
+; CHECK-NEXT:    lwz 20, 884(31)
+; CHECK-NEXT:    lwz 19, 772(31)
+; CHECK-NEXT:    lwz 18, 880(31)
+; CHECK-NEXT:    slwi 18, 18, 3
+; CHECK-NEXT:    add 19, 19, 18
+; CHECK-NEXT:    mullw 20, 20, 19
+; CHECK-NEXT:    slwi 20, 20, 4
+; CHECK-NEXT:    evstddx 22, 21, 20
+; CHECK-NEXT:    evlddx 24, 24, 3
+; CHECK-NEXT:    evlddx 23, 23, 3
+; CHECK-NEXT:    efdsub 24, 24, 23
+; CHECK-NEXT:    lwz 23, 872(31)
+; CHECK-NEXT:    lwz 22, 884(31)
+; CHECK-NEXT:    lwz 21, 772(31)
+; CHECK-NEXT:    lwz 20, 880(31)
+; CHECK-NEXT:    slwi 20, 20, 3
+; CHECK-NEXT:    add 21, 21, 20
+; CHECK-NEXT:    mullw 22, 22, 21
+; CHECK-NEXT:    slwi 22, 22, 4
+; CHECK-NEXT:    add 23, 23, 22
+; CHECK-NEXT:    evstddx 24, 23, 3
+; CHECK-NEXT:    li 24, 752
+; CHECK-NEXT:    evlddx 24, 31, 24
+; CHECK-NEXT:    li 23, 736
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 25
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 720
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 27
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 704
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 0
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 688
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 26
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 672
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 28
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 480
+; CHECK-NEXT:    evstddx 24, 31, 23
+; CHECK-NEXT:    evlddx 29, 29, 3
+; CHECK-NEXT:    evlddx 5, 5, 3
+; CHECK-NEXT:    efdmul 5, 5, 25
+; CHECK-NEXT:    efdadd 5, 5, 29
+; CHECK-NEXT:    evlddx 7, 7, 3
+; CHECK-NEXT:    efdmul 7, 7, 27
+; CHECK-NEXT:    efdadd 5, 7, 5
+; CHECK-NEXT:    evlddx 7, 9, 3
+; CHECK-NEXT:    efdmul 7, 7, 0
+; CHECK-NEXT:    efdadd 5, 7, 5
+; CHECK-NEXT:    evlddx 7, 11, 3
+; CHECK-NEXT:    efdmul 7, 7, 26
+; CHECK-NEXT:    efdadd 5, 7, 5
+; CHECK-NEXT:    evlddx 7, 30, 3
+; CHECK-NEXT:    efdmul 7, 7, 28
+; CHECK-NEXT:    efdadd 7, 7, 5
+; CHECK-NEXT:    addi 5, 31, 480
+; CHECK-NEXT:    evstddx 7, 5, 3
+; CHECK-NEXT:    li 7, 792
+; CHECK-NEXT:    evlddx 7, 31, 7
+; CHECK-NEXT:    li 9, 592
+; CHECK-NEXT:    evlddx 9, 31, 9
+; CHECK-NEXT:    li 11, 808
+; CHECK-NEXT:    evlddx 11, 31, 11
+; CHECK-NEXT:    li 30, 608
+; CHECK-NEXT:    evlddx 0, 31, 30
+; CHECK-NEXT:    efdmul 11, 11, 0
+; CHECK-NEXT:    efdmul 7, 7, 9
+; CHECK-NEXT:    efdsub 7, 7, 11
+; CHECK-NEXT:    li 9, 840
+; CHECK-NEXT:    evlddx 9, 31, 9
+; CHECK-NEXT:    li 11, 624
+; CHECK-NEXT:    evlddx 11, 31, 11
+; CHECK-NEXT:    efdmul 9, 9, 11
+; CHECK-NEXT:    efdadd 7, 9, 7
+; CHECK-NEXT:    li 9, 776
+; CHECK-NEXT:    evlddx 9, 31, 9
+; CHECK-NEXT:    li 11, 640
+; CHECK-NEXT:    evlddx 11, 31, 11
+; CHECK-NEXT:    efdmul 9, 9, 11
+; CHECK-NEXT:    efdadd 7, 9, 7
+; CHECK-NEXT:    li 9, 824
+; CHECK-NEXT:    evlddx 9, 31, 9
+; CHECK-NEXT:    li 11, 656
+; CHECK-NEXT:    evlddx 11, 31, 11
+; CHECK-NEXT:    efdmul 9, 9, 11
+; CHECK-NEXT:    efdsub 9, 7, 9
+; CHECK-NEXT:    addi 7, 31, 464
+; CHECK-NEXT:    evstddx 9, 7, 3
+; CHECK-NEXT:    li 9, 792
+; CHECK-NEXT:    evlddx 9, 31, 9
+; CHECK-NEXT:    evlddx 4, 4, 3
+; CHECK-NEXT:    li 11, 808
+; CHECK-NEXT:    evlddx 11, 31, 11
+; CHECK-NEXT:    evlddx 6, 6, 3
+; CHECK-NEXT:    efdmul 6, 11, 6
+; CHECK-NEXT:    efdmul 4, 9, 4
+; CHECK-NEXT:    efdsub 4, 4, 6
+; CHECK-NEXT:    li 6, 840
+; CHECK-NEXT:    evlddx 6, 31, 6
+; CHECK-NEXT:    evlddx 8, 8, 3
+; CHECK-NEXT:    efdmul 6, 6, 8
+; CHECK-NEXT:    efdadd 4, 6, 4
+; CHECK-NEXT:    li 6, 776
+; CHECK-NEXT:    evlddx 6, 31, 6
+; CHECK-NEXT:    evlddx 8, 10, 3
+; CHECK-NEXT:    efdmul 6, 6, 8
+; CHECK-NEXT:    efdadd 4, 6, 4
+; CHECK-NEXT:    li 6, 824
+; CHECK-NEXT:    evlddx 6, 31, 6
+; CHECK-NEXT:    evlddx 8, 12, 3
+; CHECK-NEXT:    efdmul 6, 6, 8
+; CHECK-NEXT:    efdsub 4, 4, 6
+; CHECK-NEXT:    efdneg 4, 4
+; CHECK-NEXT:    li 6, 464
+; CHECK-NEXT:    evstddx 4, 31, 6
+; CHECK-NEXT:    li 4, 480
+; CHECK-NEXT:    evlddx 4, 31, 4
+; CHECK-NEXT:    li 6, 464
+; CHECK-NEXT:    evlddx 6, 31, 6
+; CHECK-NEXT:    efdadd 4, 4, 6
+; CHECK-NEXT:    lwz 6, 872(31)
+; CHECK-NEXT:    lwz 8, 884(31)
+; CHECK-NEXT:    lwz 9, 772(31)
+; CHECK-NEXT:    lwz 10, 880(31)
+; CHECK-NEXT:    slwi 10, 10, 2
+; CHECK-NEXT:    add 9, 9, 10
+; CHECK-NEXT:    mullw 8, 8, 9
+; CHECK-NEXT:    slwi 8, 8, 4
+; CHECK-NEXT:    evstddx 4, 6, 8
+; CHECK-NEXT:    evlddx 4, 5, 3
+; CHECK-NEXT:    evlddx 6, 7, 3
+; CHECK-NEXT:    efdadd 4, 4, 6
+; CHECK-NEXT:    lwz 6, 872(31)
+; CHECK-NEXT:    lwz 8, 884(31)
+; CHECK-NEXT:    lwz 9, 772(31)
+; CHECK-NEXT:    lwz 10, 880(31)
+; CHECK-NEXT:    slwi 10, 10, 2
+; CHECK-NEXT:    add 9, 9, 10
+; CHECK-NEXT:    mullw 8, 8, 9
+; CHECK-NEXT:    slwi 8, 8, 4
+; CHECK-NEXT:    add 6, 6, 8
+; CHECK-NEXT:    evstddx 4, 6, 3
+; CHECK-NEXT:    li 4, 480
+; CHECK-NEXT:    evlddx 4, 31, 4
+; CHECK-NEXT:    li 6, 464
+; CHECK-NEXT:    evlddx 6, 31, 6
+; CHECK-NEXT:    efdsub 4, 4, 6
+; CHECK-NEXT:    lwz 6, 872(31)
+; CHECK-NEXT:    lwz 8, 884(31)
+; CHECK-NEXT:    lwz 9, 772(31)
+; CHECK-NEXT:    lwz 10, 880(31)
+; CHECK-NEXT:    mulli 10, 10, 7
+; CHECK-NEXT:    add 9, 9, 10
+; CHECK-NEXT:    mullw 8, 8, 9
+; CHECK-NEXT:    slwi 8, 8, 4
+; CHECK-NEXT:    evstddx 4, 6, 8
+; CHECK-NEXT:    evlddx 4, 5, 3
+; CHECK-NEXT:    evlddx 5, 7, 3
+; CHECK-NEXT:    efdsub 4, 4, 5
+; CHECK-NEXT:    lwz 5, 872(31)
+; CHECK-NEXT:    lwz 6, 884(31)
+; CHECK-NEXT:    lwz 7, 772(31)
+; CHECK-NEXT:    lwz 8, 880(31)
+; CHECK-NEXT:    mulli 8, 8, 7
+; CHECK-NEXT:    add 7, 7, 8
+; CHECK-NEXT:    mullw 6, 6, 7
+; CHECK-NEXT:    slwi 6, 6, 4
+; CHECK-NEXT:    add 5, 5, 6
+; CHECK-NEXT:    evstddx 4, 5, 3
+; CHECK-NEXT:    li 3, 1
+; CHECK-NEXT:    stw 3, 460(31)
+; CHECK-NEXT:    b .LBB1_3
+; CHECK-NEXT:  .LBB1_3: # Parent Loop BB1_1 Depth=1
+; CHECK-NEXT:    # => This Inner Loop Header: Depth=2
+; CHECK-NEXT:    lwz 3, 460(31)
+; CHECK-NEXT:    lwz 4, 884(31)
+; CHECK-NEXT:    cmplw 3, 4
+; CHECK-NEXT:    bge 0, .LBB1_6
+; CHECK-NEXT:    b .LBB1_4
+; CHECK-NEXT:  .LBB1_4:
+; CHECK-NEXT:    lwz 3, 876(31)
+; CHECK-NEXT:    lwz 4, 460(31)
+; CHECK-NEXT:    lwz 5, 884(31)
+; CHECK-NEXT:    lwz 6, 772(31)
+; CHECK-NEXT:    mullw 5, 6, 5
+; CHECK-NEXT:    mulli 5, 5, 11
+; CHECK-NEXT:    add 4, 4, 5
+; CHECK-NEXT:    slwi 4, 4, 4
+; CHECK-NEXT:    lwzux 4, 3, 4
+; CHECK-NEXT:    stw 4, 440(31)
+; CHECK-NEXT:    lwz 4, 12(3)
+; CHECK-NEXT:    stw 4, 452(31)
+; CHECK-NEXT:    lwz 4, 8(3)
+; CHECK-NEXT:    stw 4, 448(31)
+; CHECK-NEXT:    lwz 3, 4(3)
+; CHECK-NEXT:    stw 3, 444(31)
+; CHECK-NEXT:    lwz 3, 876(31)
+; CHECK-NEXT:    lwz 4, 460(31)
+; CHECK-NEXT:    lwz 5, 884(31)
+; CHECK-NEXT:    lwz 6, 772(31)
+; CHECK-NEXT:    mulli 6, 6, 11
+; CHECK-NEXT:    addi 7, 6, 1
+; CHECK-NEXT:    mullw 7, 5, 7
+; CHECK-NEXT:    add 7, 4, 7
+; CHECK-NEXT:    slwi 7, 7, 4
+; CHECK-NEXT:    evlddx 7, 3, 7
+; CHECK-NEXT:    addi 6, 6, 10
+; CHECK-NEXT:    mullw 5, 5, 6
+; CHECK-NEXT:    add 4, 4, 5
+; CHECK-NEXT:    slwi 4, 4, 4
+; CHECK-NEXT:    evlddx 3, 3, 4
+; CHECK-NEXT:    efdadd 3, 7, 3
+; CHECK-NEXT:    li 4, 424
+; CHECK-NEXT:    evstddx 3, 31, 4
+; CHECK-NEXT:    lwz 4, 876(31)
+; CHECK-NEXT:    lwz 5, 460(31)
+; CHECK-NEXT:    lwz 6, 884(31)
+; CHECK-NEXT:    lwz 3, 772(31)
+; CHECK-NEXT:    mulli 7, 3, 11
+; CHECK-NEXT:    addi 3, 7, 1
+; CHECK-NEXT:    mullw 3, 6, 3
+; CHECK-NEXT:    add 3, 5, 3
+; CHECK-NEXT:    slwi 3, 3, 4
+; CHECK-NEXT:    add 8, 4, 3
+; CHECK-NEXT:    li 3, 8
+; CHECK-NEXT:    evlddx 8, 8, 3
+; CHECK-NEXT:    addi 7, 7, 10
+; CHECK-NEXT:    mullw 6, 6, 7
+; CHECK-NEXT:    add 5, 5, 6
+; CHECK-NEXT:    slwi 5, 5, 4
+; CHECK-NEXT:    add 4, 4, 5
+; CHECK-NEXT:    evlddx 4, 4, 3
+; CHECK-NEXT:    efdadd 4, 8, 4
+; CHECK-NEXT:    addi 5, 31, 424
+; CHECK-NEXT:    evstddx 4, 5, 3
+; CHECK-NEXT:    lwz 4, 876(31)
+; CHECK-NEXT:    lwz 6, 460(31)
+; CHECK-NEXT:    lwz 7, 884(31)
+; CHECK-NEXT:    lwz 8, 772(31)
+; CHECK-NEXT:    mulli 8, 8, 11
+; CHECK-NEXT:    addi 9, 8, 1
+; CHECK-NEXT:    mullw 9, 7, 9
+; CHECK-NEXT:    add 9, 6, 9
+; CHECK-NEXT:    slwi 9, 9, 4
+; CHECK-NEXT:    evlddx 9, 4, 9
+; CHECK-NEXT:    addi 8, 8, 10
+; CHECK-NEXT:    mullw 7, 7, 8
+; CHECK-NEXT:    add 6, 6, 7
+; CHECK-NEXT:    slwi 6, 6, 4
+; CHECK-NEXT:    evlddx 4, 4, 6
+; CHECK-NEXT:    efdsub 4, 9, 4
+; CHECK-NEXT:    li 6, 280
+; CHECK-NEXT:    evstddx 4, 31, 6
+; CHECK-NEXT:    lwz 4, 876(31)
+; CHECK-NEXT:    lwz 6, 460(31)
+; CHECK-NEXT:    lwz 7, 884(31)
+; CHECK-NEXT:    lwz 8, 772(31)
+; CHECK-NEXT:    mulli 8, 8, 11
+; CHECK-NEXT:    addi 9, 8, 1
+; CHECK-NEXT:    mullw 9, 7, 9
+; CHECK-NEXT:    add 9, 6, 9
+; CHECK-NEXT:    slwi 9, 9, 4
+; CHECK-NEXT:    add 9, 4, 9
+; CHECK-NEXT:    evlddx 9, 9, 3
+; CHECK-NEXT:    addi 8, 8, 10
+; CHECK-NEXT:    mullw 7, 7, 8
+; CHECK-NEXT:    add 6, 6, 7
+; CHECK-NEXT:    slwi 6, 6, 4
+; CHECK-NEXT:    add 4, 4, 6
+; CHECK-NEXT:    evlddx 4, 4, 3
+; CHECK-NEXT:    efdsub 6, 9, 4
+; CHECK-NEXT:    addi 4, 31, 280
+; CHECK-NEXT:    evstddx 6, 4, 3
+; CHECK-NEXT:    lwz 6, 876(31)
+; CHECK-NEXT:    lwz 7, 460(31)
+; CHECK-NEXT:    lwz 8, 884(31)
+; CHECK-NEXT:    lwz 9, 772(31)
+; CHECK-NEXT:    mulli 9, 9, 11
+; CHECK-NEXT:    addi 10, 9, 2
+; CHECK-NEXT:    mullw 10, 8, 10
+; CHECK-NEXT:    add 10, 7, 10
+; CHECK-NEXT:    slwi 10, 10, 4
+; CHECK-NEXT:    evlddx 10, 6, 10
+; CHECK-NEXT:    addi 9, 9, 9
+; CHECK-NEXT:    mullw 8, 8, 9
+; CHECK-NEXT:    add 7, 7, 8
+; CHECK-NEXT:    slwi 7, 7, 4
+; CHECK-NEXT:    evlddx 6, 6, 7
+; CHECK-NEXT:    efdadd 6, 10, 6
+; CHECK-NEXT:    li 7, 408
+; CHECK-NEXT:    evstddx 6, 31, 7
+; CHECK-NEXT:    lwz 6, 876(31)
+; CHECK-NEXT:    lwz 7, 460(31)
+; CHECK-NEXT:    lwz 8, 884(31)
+; CHECK-NEXT:    lwz 9, 772(31)
+; CHECK-NEXT:    mulli 9, 9, 11
+; CHECK-NEXT:    addi 10, 9, 2
+; CHECK-NEXT:    mullw 10, 8, 10
+; CHECK-NEXT:    add 10, 7, 10
+; CHECK-NEXT:    slwi 10, 10, 4
+; CHECK-NEXT:    add 10, 6, 10
+; CHECK-NEXT:    evlddx 10, 10, 3
+; CHECK-NEXT:    addi 9, 9, 9
+; CHECK-NEXT:    mullw 8, 8, 9
+; CHECK-NEXT:    add 7, 7, 8
+; CHECK-NEXT:    slwi 7, 7, 4
+; CHECK-NEXT:    add 6, 6, 7
+; CHECK-NEXT:    evlddx 6, 6, 3
+; CHECK-NEXT:    efdadd 6, 10, 6
+; CHECK-NEXT:    addi 7, 31, 408
+; CHECK-NEXT:    evstddx 6, 7, 3
+; CHECK-NEXT:    lwz 6, 876(31)
+; CHECK-NEXT:    lwz 8, 460(31)
+; CHECK-NEXT:    lwz 9, 884(31)
+; CHECK-NEXT:    lwz 10, 772(31)
+; CHECK-NEXT:    mulli 10, 10, 11
+; CHECK-NEXT:    addi 11, 10, 2
+; CHECK-NEXT:    mullw 11, 9, 11
+; CHECK-NEXT:    add 11, 8, 11
+; CHECK-NEXT:    slwi 11, 11, 4
+; CHECK-NEXT:    evlddx 11, 6, 11
+; CHECK-NEXT:    addi 10, 10, 9
+; CHECK-NEXT:    mullw 9, 9, 10
+; CHECK-NEXT:    add 8, 8, 9
+; CHECK-NEXT:    slwi 8, 8, 4
+; CHECK-NEXT:    evlddx 6, 6, 8
+; CHECK-NEXT:    efdsub 6, 11, 6
+; CHECK-NEXT:    li 8, 296
+; CHECK-NEXT:    evstddx 6, 31, 8
+; CHECK-NEXT:    lwz 6, 876(31)
+; CHECK-NEXT:    lwz 8, 460(31)
+; CHECK-NEXT:    lwz 9, 884(31)
+; CHECK-NEXT:    lwz 10, 772(31)
+; CHECK-NEXT:    mulli 10, 10, 11
+; CHECK-NEXT:    addi 11, 10, 2
+; CHECK-NEXT:    mullw 11, 9, 11
+; CHECK-NEXT:    add 11, 8, 11
+; CHECK-NEXT:    slwi 11, 11, 4
+; CHECK-NEXT:    add 11, 6, 11
+; CHECK-NEXT:    evlddx 11, 11, 3
+; CHECK-NEXT:    addi 10, 10, 9
+; CHECK-NEXT:    mullw 9, 9, 10
+; CHECK-NEXT:    add 8, 8, 9
+; CHECK-NEXT:    slwi 8, 8, 4
+; CHECK-NEXT:    add 6, 6, 8
+; CHECK-NEXT:    evlddx 6, 6, 3
+; CHECK-NEXT:    efdsub 8, 11, 6
+; CHECK-NEXT:    addi 6, 31, 296
+; CHECK-NEXT:    evstddx 8, 6, 3
+; CHECK-NEXT:    lwz 8, 876(31)
+; CHECK-NEXT:    lwz 9, 460(31)
+; CHECK-NEXT:    lwz 10, 884(31)
+; CHECK-NEXT:    lwz 11, 772(31)
+; CHECK-NEXT:    mulli 11, 11, 11
+; CHECK-NEXT:    addi 12, 11, 3
+; CHECK-NEXT:    mullw 12, 10, 12
+; CHECK-NEXT:    add 12, 9, 12
+; CHECK-NEXT:    slwi 12, 12, 4
+; CHECK-NEXT:    evlddx 12, 8, 12
+; CHECK-NEXT:    addi 11, 11, 8
+; CHECK-NEXT:    mullw 10, 10, 11
+; CHECK-NEXT:    add 9, 9, 10
+; CHECK-NEXT:    slwi 9, 9, 4
+; CHECK-NEXT:    evlddx 8, 8, 9
+; CHECK-NEXT:    efdadd 8, 12, 8
+; CHECK-NEXT:    li 9, 392
+; CHECK-NEXT:    evstddx 8, 31, 9
+; CHECK-NEXT:    lwz 8, 876(31)
+; CHECK-NEXT:    lwz 9, 460(31)
+; CHECK-NEXT:    lwz 10, 884(31)
+; CHECK-NEXT:    lwz 11, 772(31)
+; CHECK-NEXT:    mulli 11, 11, 11
+; CHECK-NEXT:    addi 12, 11, 3
+; CHECK-NEXT:    mullw 12, 10, 12
+; CHECK-NEXT:    add 12, 9, 12
+; CHECK-NEXT:    slwi 12, 12, 4
+; CHECK-NEXT:    add 12, 8, 12
+; CHECK-NEXT:    evlddx 12, 12, 3
+; CHECK-NEXT:    addi 11, 11, 8
+; CHECK-NEXT:    mullw 10, 10, 11
+; CHECK-NEXT:    add 9, 9, 10
+; CHECK-NEXT:    slwi 9, 9, 4
+; CHECK-NEXT:    add 8, 8, 9
+; CHECK-NEXT:    evlddx 8, 8, 3
+; CHECK-NEXT:    efdadd 8, 12, 8
+; CHECK-NEXT:    addi 9, 31, 392
+; CHECK-NEXT:    evstddx 8, 9, 3
+; CHECK-NEXT:    lwz 8, 876(31)
+; CHECK-NEXT:    lwz 10, 460(31)
+; CHECK-NEXT:    lwz 11, 884(31)
+; CHECK-NEXT:    lwz 12, 772(31)
+; CHECK-NEXT:    mulli 12, 12, 11
+; CHECK-NEXT:    addi 0, 12, 3
+; CHECK-NEXT:    mullw 0, 11, 0
+; CHECK-NEXT:    add 0, 10, 0
+; CHECK-NEXT:    slwi 0, 0, 4
+; CHECK-NEXT:    evlddx 0, 8, 0
+; CHECK-NEXT:    addi 12, 12, 8
+; CHECK-NEXT:    mullw 11, 11, 12
+; CHECK-NEXT:    add 10, 10, 11
+; CHECK-NEXT:    slwi 10, 10, 4
+; CHECK-NEXT:    evlddx 8, 8, 10
+; CHECK-NEXT:    efdsub 8, 0, 8
+; CHECK-NEXT:    li 10, 312
+; CHECK-NEXT:    evstddx 8, 31, 10
+; CHECK-NEXT:    lwz 8, 876(31)
+; CHECK-NEXT:    lwz 10, 460(31)
+; CHECK-NEXT:    lwz 11, 884(31)
+; CHECK-NEXT:    lwz 12, 772(31)
+; CHECK-NEXT:    mulli 12, 12, 11
+; CHECK-NEXT:    addi 0, 12, 3
+; CHECK-NEXT:    mullw 0, 11, 0
+; CHECK-NEXT:    add 0, 10, 0
+; CHECK-NEXT:    slwi 0, 0, 4
+; CHECK-NEXT:    add 30, 8, 0
+; CHECK-NEXT:    evlddx 0, 30, 3
+; CHECK-NEXT:    addi 12, 12, 8
+; CHECK-NEXT:    mullw 11, 11, 12
+; CHECK-NEXT:    add 10, 10, 11
+; CHECK-NEXT:    slwi 10, 10, 4
+; CHECK-NEXT:    add 8, 8, 10
+; CHECK-NEXT:    evlddx 8, 8, 3
+; CHECK-NEXT:    efdsub 10, 0, 8
+; CHECK-NEXT:    addi 8, 31, 312
+; CHECK-NEXT:    evstddx 10, 8, 3
+; CHECK-NEXT:    lwz 10, 876(31)
+; CHECK-NEXT:    lwz 11, 460(31)
+; CHECK-NEXT:    lwz 12, 884(31)
+; CHECK-NEXT:    lwz 0, 772(31)
+; CHECK-NEXT:    mulli 30, 0, 11
+; CHECK-NEXT:    addi 0, 30, 4
+; CHECK-NEXT:    mullw 0, 12, 0
+; CHECK-NEXT:    add 0, 11, 0
+; CHECK-NEXT:    slwi 0, 0, 4
+; CHECK-NEXT:    evlddx 0, 10, 0
+; CHECK-NEXT:    addi 30, 30, 7
+; CHECK-NEXT:    mullw 12, 12, 30
+; CHECK-NEXT:    add 11, 11, 12
+; CHECK-NEXT:    slwi 11, 11, 4
+; CHECK-NEXT:    evlddx 10, 10, 11
+; CHECK-NEXT:    efdadd 10, 0, 10
+; CHECK-NEXT:    li 11, 376
+; CHECK-NEXT:    evstddx 10, 31, 11
+; CHECK-NEXT:    lwz 10, 876(31)
+; CHECK-NEXT:    lwz 11, 460(31)
+; CHECK-NEXT:    lwz 12, 884(31)
+; CHECK-NEXT:    lwz 0, 772(31)
+; CHECK-NEXT:    mulli 30, 0, 11
+; CHECK-NEXT:    addi 0, 30, 4
+; CHECK-NEXT:    mullw 0, 12, 0
+; CHECK-NEXT:    add 0, 11, 0
+; CHECK-NEXT:    slwi 0, 0, 4
+; CHECK-NEXT:    add 29, 10, 0
+; CHECK-NEXT:    evlddx 0, 29, 3
+; CHECK-NEXT:    addi 30, 30, 7
+; CHECK-NEXT:    mullw 12, 12, 30
+; CHECK-NEXT:    add 11, 11, 12
+; CHECK-NEXT:    slwi 11, 11, 4
+; CHECK-NEXT:    add 10, 10, 11
+; CHECK-NEXT:    evlddx 10, 10, 3
+; CHECK-NEXT:    efdadd 10, 0, 10
+; CHECK-NEXT:    addi 11, 31, 376
+; CHECK-NEXT:    evstddx 10, 11, 3
+; CHECK-NEXT:    lwz 10, 876(31)
+; CHECK-NEXT:    lwz 12, 460(31)
+; CHECK-NEXT:    lwz 0, 884(31)
+; CHECK-NEXT:    lwz 30, 772(31)
+; CHECK-NEXT:    mulli 30, 30, 11
+; CHECK-NEXT:    addi 29, 30, 4
+; CHECK-NEXT:    mullw 29, 0, 29
+; CHECK-NEXT:    add 29, 12, 29
+; CHECK-NEXT:    slwi 29, 29, 4
+; CHECK-NEXT:    evlddx 29, 10, 29
+; CHECK-NEXT:    addi 30, 30, 7
+; CHECK-NEXT:    mullw 0, 0, 30
+; CHECK-NEXT:    add 12, 12, 0
+; CHECK-NEXT:    slwi 12, 12, 4
+; CHECK-NEXT:    evlddx 10, 10, 12
+; CHECK-NEXT:    efdsub 10, 29, 10
+; CHECK-NEXT:    li 12, 328
+; CHECK-NEXT:    evstddx 10, 31, 12
+; CHECK-NEXT:    lwz 10, 876(31)
+; CHECK-NEXT:    lwz 12, 460(31)
+; CHECK-NEXT:    lwz 0, 884(31)
+; CHECK-NEXT:    lwz 30, 772(31)
+; CHECK-NEXT:    mulli 30, 30, 11
+; CHECK-NEXT:    addi 29, 30, 4
+; CHECK-NEXT:    mullw 29, 0, 29
+; CHECK-NEXT:    add 29, 12, 29
+; CHECK-NEXT:    slwi 29, 29, 4
+; CHECK-NEXT:    add 29, 10, 29
+; CHECK-NEXT:    evlddx 29, 29, 3
+; CHECK-NEXT:    addi 30, 30, 7
+; CHECK-NEXT:    mullw 0, 0, 30
+; CHECK-NEXT:    add 12, 12, 0
+; CHECK-NEXT:    slwi 12, 12, 4
+; CHECK-NEXT:    add 10, 10, 12
+; CHECK-NEXT:    evlddx 10, 10, 3
+; CHECK-NEXT:    efdsub 12, 29, 10
+; CHECK-NEXT:    addi 10, 31, 328
+; CHECK-NEXT:    evstddx 12, 10, 3
+; CHECK-NEXT:    lwz 12, 876(31)
+; CHECK-NEXT:    lwz 0, 460(31)
+; CHECK-NEXT:    lwz 30, 884(31)
+; CHECK-NEXT:    lwz 29, 772(31)
+; CHECK-NEXT:    mulli 29, 29, 11
+; CHECK-NEXT:    addi 28, 29, 5
+; CHECK-NEXT:    mullw 28, 30, 28
+; CHECK-NEXT:    add 28, 0, 28
+; CHECK-NEXT:    slwi 28, 28, 4
+; CHECK-NEXT:    evlddx 28, 12, 28
+; CHECK-NEXT:    addi 29, 29, 6
+; CHECK-NEXT:    mullw 30, 30, 29
+; CHECK-NEXT:    add 0, 0, 30
+; CHECK-NEXT:    slwi 0, 0, 4
+; CHECK-NEXT:    evlddx 12, 12, 0
+; CHECK-NEXT:    efdadd 12, 28, 12
+; CHECK-NEXT:    li 30, 360
+; CHECK-NEXT:    evstddx 12, 31, 30
+; CHECK-NEXT:    lwz 12, 876(31)
+; CHECK-NEXT:    lwz 0, 460(31)
+; CHECK-NEXT:    lwz 30, 884(31)
+; CHECK-NEXT:    lwz 29, 772(31)
+; CHECK-NEXT:    mulli 29, 29, 11
+; CHECK-NEXT:    addi 28, 29, 5
+; CHECK-NEXT:    mullw 28, 30, 28
+; CHECK-NEXT:    add 28, 0, 28
+; CHECK-NEXT:    slwi 28, 28, 4
+; CHECK-NEXT:    add 28, 12, 28
+; CHECK-NEXT:    evlddx 28, 28, 3
+; CHECK-NEXT:    addi 29, 29, 6
+; CHECK-NEXT:    mullw 30, 30, 29
+; CHECK-NEXT:    add 0, 0, 30
+; CHECK-NEXT:    slwi 0, 0, 4
+; CHECK-NEXT:    add 12, 12, 0
+; CHECK-NEXT:    evlddx 12, 12, 3
+; CHECK-NEXT:    efdadd 12, 28, 12
+; CHECK-NEXT:    addi 30, 31, 360
+; CHECK-NEXT:    evstddx 12, 30, 3
+; CHECK-NEXT:    lwz 12, 876(31)
+; CHECK-NEXT:    lwz 0, 460(31)
+; CHECK-NEXT:    lwz 29, 884(31)
+; CHECK-NEXT:    lwz 28, 772(31)
+; CHECK-NEXT:    mulli 28, 28, 11
+; CHECK-NEXT:    addi 27, 28, 5
+; CHECK-NEXT:    mullw 27, 29, 27
+; CHECK-NEXT:    add 27, 0, 27
+; CHECK-NEXT:    slwi 27, 27, 4
+; CHECK-NEXT:    evlddx 27, 12, 27
+; CHECK-NEXT:    addi 28, 28, 6
+; CHECK-NEXT:    mullw 29, 29, 28
+; CHECK-NEXT:    add 0, 0, 29
+; CHECK-NEXT:    slwi 0, 0, 4
+; CHECK-NEXT:    evlddx 12, 12, 0
+; CHECK-NEXT:    efdsub 12, 27, 12
+; CHECK-NEXT:    li 29, 344
+; CHECK-NEXT:    evstddx 12, 31, 29
+; CHECK-NEXT:    lwz 12, 876(31)
+; CHECK-NEXT:    lwz 0, 460(31)
+; CHECK-NEXT:    lwz 29, 884(31)
+; CHECK-NEXT:    lwz 28, 772(31)
+; CHECK-NEXT:    mulli 28, 28, 11
+; CHECK-NEXT:    addi 27, 28, 5
+; CHECK-NEXT:    mullw 27, 29, 27
+; CHECK-NEXT:    add 27, 0, 27
+; CHECK-NEXT:    slwi 27, 27, 4
+; CHECK-NEXT:    add 27, 12, 27
+; CHECK-NEXT:    evlddx 27, 27, 3
+; CHECK-NEXT:    addi 28, 28, 6
+; CHECK-NEXT:    mullw 29, 29, 28
+; CHECK-NEXT:    add 0, 0, 29
+; CHECK-NEXT:    slwi 0, 0, 4
+; CHECK-NEXT:    add 12, 12, 0
+; CHECK-NEXT:    evlddx 12, 12, 3
+; CHECK-NEXT:    efdsub 0, 27, 12
+; CHECK-NEXT:    addi 12, 31, 344
+; CHECK-NEXT:    evstddx 0, 12, 3
+; CHECK-NEXT:    li 29, 440
+; CHECK-NEXT:    evlddx 0, 31, 29
+; CHECK-NEXT:    li 29, 424
+; CHECK-NEXT:    evlddx 29, 31, 29
+; CHECK-NEXT:    efdadd 0, 0, 29
+; CHECK-NEXT:    li 29, 408
+; CHECK-NEXT:    evlddx 29, 31, 29
+; CHECK-NEXT:    efdadd 0, 0, 29
+; CHECK-NEXT:    li 29, 392
+; CHECK-NEXT:    evlddx 29, 31, 29
+; CHECK-NEXT:    efdadd 0, 0, 29
+; CHECK-NEXT:    li 29, 376
+; CHECK-NEXT:    evlddx 29, 31, 29
+; CHECK-NEXT:    efdadd 0, 0, 29
+; CHECK-NEXT:    li 29, 360
+; CHECK-NEXT:    evlddx 29, 31, 29
+; CHECK-NEXT:    efdadd 0, 0, 29
+; CHECK-NEXT:    lwz 29, 872(31)
+; CHECK-NEXT:    lwz 28, 460(31)
+; CHECK-NEXT:    lwz 27, 884(31)
+; CHECK-NEXT:    lwz 26, 772(31)
+; CHECK-NEXT:    mullw 27, 27, 26
+; CHECK-NEXT:    add 28, 28, 27
+; CHECK-NEXT:    slwi 28, 28, 4
+; CHECK-NEXT:    evstddx 0, 29, 28
+; CHECK-NEXT:    addi 29, 31, 440
+; CHECK-NEXT:    evlddx 0, 29, 3
+; CHECK-NEXT:    evlddx 28, 5, 3
+; CHECK-NEXT:    efdadd 0, 0, 28
+; CHECK-NEXT:    evlddx 28, 7, 3
+; CHECK-NEXT:    efdadd 0, 0, 28
+; CHECK-NEXT:    evlddx 28, 9, 3
+; CHECK-NEXT:    efdadd 0, 0, 28
+; CHECK-NEXT:    evlddx 28, 11, 3
+; CHECK-NEXT:    efdadd 0, 0, 28
+; CHECK-NEXT:    evlddx 28, 30, 3
+; CHECK-NEXT:    efdadd 0, 0, 28
+; CHECK-NEXT:    lwz 28, 872(31)
+; CHECK-NEXT:    lwz 27, 460(31)
+; CHECK-NEXT:    lwz 26, 884(31)
+; CHECK-NEXT:    lwz 25, 772(31)
+; CHECK-NEXT:    mullw 26, 26, 25
+; CHECK-NEXT:    add 27, 27, 26
+; CHECK-NEXT:    slwi 27, 27, 4
+; CHECK-NEXT:    add 28, 28, 27
+; CHECK-NEXT:    evstddx 0, 28, 3
+; CHECK-NEXT:    li 28, 440
+; CHECK-NEXT:    evlddx 28, 31, 28
+; CHECK-NEXT:    li 27, 424
+; CHECK-NEXT:    evlddx 27, 31, 27
+; CHECK-NEXT:    li 0, .LCPI1_5@l
+; CHECK-NEXT:    lis 26, .LCPI1_5@ha
+; CHECK-NEXT:    evlddx 0, 26, 0
+; CHECK-NEXT:    efdmul 27, 27, 0
+; CHECK-NEXT:    efdadd 27, 27, 28
+; CHECK-NEXT:    li 28, 408
+; CHECK-NEXT:    evlddx 26, 31, 28
+; CHECK-NEXT:    li 28, .LCPI1_6@l
+; CHECK-NEXT:    lis 25, .LCPI1_6@ha
+; CHECK-NEXT:    evlddx 28, 25, 28
+; CHECK-NEXT:    efdmul 26, 26, 28
+; CHECK-NEXT:    efdadd 26, 26, 27
+; CHECK-NEXT:    li 27, 392
+; CHECK-NEXT:    evlddx 25, 31, 27
+; CHECK-NEXT:    li 27, .LCPI1_7@l
+; CHECK-NEXT:    lis 24, .LCPI1_7@ha
+; CHECK-NEXT:    evlddx 27, 24, 27
+; CHECK-NEXT:    efdmul 25, 25, 27
+; CHECK-NEXT:    efdadd 26, 25, 26
+; CHECK-NEXT:    li 25, 376
+; CHECK-NEXT:    evlddx 24, 31, 25
+; CHECK-NEXT:    li 25, .LCPI1_8@l
+; CHECK-NEXT:    lis 23, .LCPI1_8@ha
+; CHECK-NEXT:    evlddx 25, 23, 25
+; CHECK-NEXT:    efdmul 24, 24, 25
+; CHECK-NEXT:    efdadd 24, 24, 26
+; CHECK-NEXT:    li 26, 360
+; CHECK-NEXT:    evlddx 23, 31, 26
+; CHECK-NEXT:    li 26, .LCPI1_9@l
+; CHECK-NEXT:    lis 22, .LCPI1_9@ha
+; CHECK-NEXT:    evlddx 26, 22, 26
+; CHECK-NEXT:    efdmul 23, 23, 26
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evstdd 24, 232(31)
+; CHECK-NEXT:    evlddx 24, 29, 3
+; CHECK-NEXT:    evlddx 23, 5, 3
+; CHECK-NEXT:    efdmul 23, 23, 0
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 7, 3
+; CHECK-NEXT:    efdmul 23, 23, 28
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 9, 3
+; CHECK-NEXT:    efdmul 23, 23, 27
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 11, 3
+; CHECK-NEXT:    efdmul 23, 23, 25
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 30, 3
+; CHECK-NEXT:    efdmul 23, 23, 26
+; CHECK-NEXT:    efdadd 23, 23, 24
+; CHECK-NEXT:    addi 24, 31, 232
+; CHECK-NEXT:    evstddx 23, 24, 3
+; CHECK-NEXT:    li 23, 840
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    li 22, 280
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 824
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    li 20, 296
+; CHECK-NEXT:    evlddx 20, 31, 20
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdmul 23, 23, 22
+; CHECK-NEXT:    efdadd 23, 23, 21
+; CHECK-NEXT:    li 22, 808
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 312
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdadd 23, 22, 23
+; CHECK-NEXT:    li 22, 792
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 328
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdadd 23, 22, 23
+; CHECK-NEXT:    li 22, 776
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 344
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdadd 23, 22, 23
+; CHECK-NEXT:    addi 22, 31, 216
+; CHECK-NEXT:    evstddx 23, 22, 3
+; CHECK-NEXT:    li 23, 840
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    evlddx 21, 4, 3
+; CHECK-NEXT:    li 20, 824
+; CHECK-NEXT:    evlddx 20, 31, 20
+; CHECK-NEXT:    evlddx 19, 6, 3
+; CHECK-NEXT:    efdmul 20, 20, 19
+; CHECK-NEXT:    efdmul 23, 23, 21
+; CHECK-NEXT:    efdadd 23, 23, 20
+; CHECK-NEXT:    li 21, 808
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 8, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdadd 23, 21, 23
+; CHECK-NEXT:    li 21, 792
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 10, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdadd 23, 21, 23
+; CHECK-NEXT:    li 21, 776
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 12, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdadd 23, 21, 23
+; CHECK-NEXT:    efdneg 23, 23
+; CHECK-NEXT:    evstdd 23, 216(31)
+; CHECK-NEXT:    evldd 23, 232(31)
+; CHECK-NEXT:    evldd 21, 216(31)
+; CHECK-NEXT:    efdadd 23, 23, 21
+; CHECK-NEXT:    li 21, 264
+; CHECK-NEXT:    evstddx 23, 31, 21
+; CHECK-NEXT:    evlddx 23, 24, 3
+; CHECK-NEXT:    evlddx 21, 22, 3
+; CHECK-NEXT:    efdadd 23, 23, 21
+; CHECK-NEXT:    addi 21, 31, 264
+; CHECK-NEXT:    evstddx 23, 21, 3
+; CHECK-NEXT:    evldd 23, 232(31)
+; CHECK-NEXT:    evldd 20, 216(31)
+; CHECK-NEXT:    efdsub 23, 23, 20
+; CHECK-NEXT:    evstdd 23, 248(31)
+; CHECK-NEXT:    evlddx 24, 24, 3
+; CHECK-NEXT:    evlddx 23, 22, 3
+; CHECK-NEXT:    efdsub 23, 24, 23
+; CHECK-NEXT:    addi 24, 31, 248
+; CHECK-NEXT:    evstddx 23, 24, 3
+; CHECK-NEXT:    lwz 23, 868(31)
+; CHECK-NEXT:    lwz 22, 460(31)
+; CHECK-NEXT:    slwi 20, 22, 4
+; CHECK-NEXT:    lwz 19, 884(31)
+; CHECK-NEXT:    add 23, 20, 23
+; CHECK-NEXT:    li 20, -16
+; CHECK-NEXT:    evlddx 18, 23, 20
+; CHECK-NEXT:    li 17, 264
+; CHECK-NEXT:    evlddx 17, 31, 17
+; CHECK-NEXT:    lwz 16, 864(31)
+; CHECK-NEXT:    efdcfsi 16, 16
+; CHECK-NEXT:    li 15, -8
+; CHECK-NEXT:    evlddx 23, 23, 15
+; CHECK-NEXT:    efdmul 23, 16, 23
+; CHECK-NEXT:    evlddx 16, 21, 3
+; CHECK-NEXT:    efdmul 23, 23, 16
+; CHECK-NEXT:    efdmul 18, 18, 17
+; CHECK-NEXT:    efdsub 23, 18, 23
+; CHECK-NEXT:    lwz 18, 872(31)
+; CHECK-NEXT:    lwz 17, 772(31)
+; CHECK-NEXT:    lwz 16, 880(31)
+; CHECK-NEXT:    add 17, 17, 16
+; CHECK-NEXT:    mullw 19, 19, 17
+; CHECK-NEXT:    add 22, 22, 19
+; CHECK-NEXT:    slwi 22, 22, 4
+; CHECK-NEXT:    evstddx 23, 18, 22
+; CHECK-NEXT:    lwz 23, 868(31)
+; CHECK-NEXT:    lwz 22, 460(31)
+; CHECK-NEXT:    slwi 19, 22, 4
+; CHECK-NEXT:    lwz 18, 884(31)
+; CHECK-NEXT:    add 23, 19, 23
+; CHECK-NEXT:    evlddx 20, 23, 20
+; CHECK-NEXT:    evlddx 21, 21, 3
+; CHECK-NEXT:    lwz 19, 864(31)
+; CHECK-NEXT:    efdcfsi 19, 19
+; CHECK-NEXT:    evlddx 23, 23, 15
+; CHECK-NEXT:    efdmul 23, 19, 23
+; CHECK-NEXT:    li 19, 264
+; CHECK-NEXT:    evlddx 19, 31, 19
+; CHECK-NEXT:    efdmul 23, 23, 19
+; CHECK-NEXT:    efdmul 21, 20, 21
+; CHECK-NEXT:    efdadd 23, 21, 23
+; CHECK-NEXT:    lwz 21, 872(31)
+; CHECK-NEXT:    lwz 20, 772(31)
+; CHECK-NEXT:    lwz 19, 880(31)
+; CHECK-NEXT:    add 20, 20, 19
+; CHECK-NEXT:    mullw 20, 18, 20
+; CHECK-NEXT:    add 22, 22, 20
+; CHECK-NEXT:    slwi 22, 22, 4
+; CHECK-NEXT:    add 22, 21, 22
+; CHECK-NEXT:    evstddx 23, 22, 3
+; CHECK-NEXT:    lwz 23, 868(31)
+; CHECK-NEXT:    lwz 22, 460(31)
+; CHECK-NEXT:    lwz 21, 884(31)
+; CHECK-NEXT:    mulli 20, 21, 9
+; CHECK-NEXT:    add 20, 20, 22
+; CHECK-NEXT:    slwi 20, 20, 4
+; CHECK-NEXT:    add 23, 20, 23
+; CHECK-NEXT:    li 20, -160
+; CHECK-NEXT:    evlddx 19, 23, 20
+; CHECK-NEXT:    evldd 18, 248(31)
+; CHECK-NEXT:    lwz 17, 864(31)
+; CHECK-NEXT:    efdcfsi 17, 17
+; CHECK-NEXT:    li 16, -152
+; CHECK-NEXT:    evlddx 23, 23, 16
+; CHECK-NEXT:    efdmul 23, 17, 23
+; CHECK-NEXT:    evlddx 17, 24, 3
+; CHECK-NEXT:    efdmul 23, 23, 17
+; CHECK-NEXT:    efdmul 19, 19, 18
+; CHECK-NEXT:    efdsub 23, 19, 23
+; CHECK-NEXT:    lwz 19, 872(31)
+; CHECK-NEXT:    lwz 18, 772(31)
+; CHECK-NEXT:    lwz 17, 880(31)
+; CHECK-NEXT:    mulli 17, 17, 10
+; CHECK-NEXT:    add 18, 18, 17
+; CHECK-NEXT:    mullw 21, 21, 18
+; CHECK-NEXT:    add 22, 22, 21
+; CHECK-NEXT:    slwi 22, 22, 4
+; CHECK-NEXT:    evstddx 23, 19, 22
+; CHECK-NEXT:    lwz 23, 868(31)
+; CHECK-NEXT:    lwz 22, 460(31)
+; CHECK-NEXT:    lwz 21, 884(31)
+; CHECK-NEXT:    mulli 19, 21, 9
+; CHECK-NEXT:    add 19, 19, 22
+; CHECK-NEXT:    slwi 19, 19, 4
+; CHECK-NEXT:    add 23, 19, 23
+; CHECK-NEXT:    evlddx 20, 23, 20
+; CHECK-NEXT:    evlddx 24, 24, 3
+; CHECK-NEXT:    lwz 19, 864(31)
+; CHECK-NEXT:    efdcfsi 19, 19
+; CHECK-NEXT:    evlddx 23, 23, 16
+; CHECK-NEXT:    efdmul 23, 19, 23
+; CHECK-NEXT:    evldd 19, 248(31)
+; CHECK-NEXT:    efdmul 23, 23, 19
+; CHECK-NEXT:    efdmul 24, 20, 24
+; CHECK-NEXT:    efdadd 24, 24, 23
+; CHECK-NEXT:    lwz 23, 872(31)
+; CHECK-NEXT:    lwz 20, 772(31)
+; CHECK-NEXT:    lwz 19, 880(31)
+; CHECK-NEXT:    mulli 19, 19, 10
+; CHECK-NEXT:    add 20, 20, 19
+; CHECK-NEXT:    mullw 21, 21, 20
+; CHECK-NEXT:    add 22, 22, 21
+; CHECK-NEXT:    slwi 22, 22, 4
+; CHECK-NEXT:    add 23, 23, 22
+; CHECK-NEXT:    evstddx 24, 23, 3
+; CHECK-NEXT:    li 24, 440
+; CHECK-NEXT:    evlddx 24, 31, 24
+; CHECK-NEXT:    li 23, 424
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 28
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 408
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 25
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 392
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 26
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 376
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 27
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 360
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 0
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evstdd 24, 168(31)
+; CHECK-NEXT:    evlddx 24, 29, 3
+; CHECK-NEXT:    evlddx 23, 5, 3
+; CHECK-NEXT:    efdmul 23, 23, 28
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 7, 3
+; CHECK-NEXT:    efdmul 23, 23, 25
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 9, 3
+; CHECK-NEXT:    efdmul 23, 23, 26
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 11, 3
+; CHECK-NEXT:    efdmul 23, 23, 27
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 30, 3
+; CHECK-NEXT:    efdmul 23, 23, 0
+; CHECK-NEXT:    efdadd 23, 23, 24
+; CHECK-NEXT:    addi 24, 31, 168
+; CHECK-NEXT:    evstddx 23, 24, 3
+; CHECK-NEXT:    li 23, 824
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    li 22, 280
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 792
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    li 20, 296
+; CHECK-NEXT:    evlddx 20, 31, 20
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdmul 23, 23, 22
+; CHECK-NEXT:    efdadd 23, 23, 21
+; CHECK-NEXT:    li 22, 776
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 312
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdsub 23, 23, 22
+; CHECK-NEXT:    li 22, 808
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 328
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdsub 23, 23, 22
+; CHECK-NEXT:    li 22, 840
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 344
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdsub 23, 23, 22
+; CHECK-NEXT:    addi 22, 31, 152
+; CHECK-NEXT:    evstddx 23, 22, 3
+; CHECK-NEXT:    li 23, 824
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    evlddx 21, 4, 3
+; CHECK-NEXT:    li 20, 792
+; CHECK-NEXT:    evlddx 20, 31, 20
+; CHECK-NEXT:    evlddx 19, 6, 3
+; CHECK-NEXT:    efdmul 20, 20, 19
+; CHECK-NEXT:    efdmul 23, 23, 21
+; CHECK-NEXT:    efdadd 23, 23, 20
+; CHECK-NEXT:    li 21, 776
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 8, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdsub 23, 23, 21
+; CHECK-NEXT:    li 21, 808
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 10, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdsub 23, 23, 21
+; CHECK-NEXT:    li 21, 840
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 12, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdsub 23, 23, 21
+; CHECK-NEXT:    efdneg 23, 23
+; CHECK-NEXT:    evstdd 23, 152(31)
+; CHECK-NEXT:    evldd 23, 168(31)
+; CHECK-NEXT:    evldd 21, 152(31)
+; CHECK-NEXT:    efdadd 23, 23, 21
+; CHECK-NEXT:    evstdd 23, 200(31)
+; CHECK-NEXT:    evlddx 23, 24, 3
+; CHECK-NEXT:    evlddx 21, 22, 3
+; CHECK-NEXT:    efdadd 23, 23, 21
+; CHECK-NEXT:    addi 21, 31, 200
+; CHECK-NEXT:    evstddx 23, 21, 3
+; CHECK-NEXT:    evldd 23, 168(31)
+; CHECK-NEXT:    evldd 20, 152(31)
+; CHECK-NEXT:    efdsub 23, 23, 20
+; CHECK-NEXT:    evstdd 23, 184(31)
+; CHECK-NEXT:    evlddx 24, 24, 3
+; CHECK-NEXT:    evlddx 23, 22, 3
+; CHECK-NEXT:    efdsub 23, 24, 23
+; CHECK-NEXT:    addi 24, 31, 184
+; CHECK-NEXT:    evstddx 23, 24, 3
+; CHECK-NEXT:    lwz 23, 868(31)
+; CHECK-NEXT:    lwz 22, 460(31)
+; CHECK-NEXT:    lwz 20, 884(31)
+; CHECK-NEXT:    add 19, 20, 22
+; CHECK-NEXT:    slwi 19, 19, 4
+; CHECK-NEXT:    add 23, 19, 23
+; CHECK-NEXT:    li 19, -32
+; CHECK-NEXT:    evlddx 18, 23, 19
+; CHECK-NEXT:    evldd 17, 200(31)
+; CHECK-NEXT:    lwz 16, 864(31)
+; CHECK-NEXT:    efdcfsi 16, 16
+; CHECK-NEXT:    li 15, -24
+; CHECK-NEXT:    evlddx 23, 23, 15
+; CHECK-NEXT:    efdmul 23, 16, 23
+; CHECK-NEXT:    evlddx 16, 21, 3
+; CHECK-NEXT:    efdmul 23, 23, 16
+; CHECK-NEXT:    efdmul 18, 18, 17
+; CHECK-NEXT:    efdsub 23, 18, 23
+; CHECK-NEXT:    lwz 18, 872(31)
+; CHECK-NEXT:    lwz 17, 772(31)
+; CHECK-NEXT:    lwz 16, 880(31)
+; CHECK-NEXT:    slwi 16, 16, 1
+; CHECK-NEXT:    add 17, 17, 16
+; CHECK-NEXT:    mullw 20, 20, 17
+; CHECK-NEXT:    add 22, 22, 20
+; CHECK-NEXT:    slwi 22, 22, 4
+; CHECK-NEXT:    evstddx 23, 18, 22
+; CHECK-NEXT:    lwz 23, 868(31)
+; CHECK-NEXT:    lwz 22, 460(31)
+; CHECK-NEXT:    lwz 20, 884(31)
+; CHECK-NEXT:    add 18, 20, 22
+; CHECK-NEXT:    slwi 18, 18, 4
+; CHECK-NEXT:    add 23, 18, 23
+; CHECK-NEXT:    evlddx 19, 23, 19
+; CHECK-NEXT:    evlddx 21, 21, 3
+; CHECK-NEXT:    lwz 18, 864(31)
+; CHECK-NEXT:    efdcfsi 18, 18
+; CHECK-NEXT:    evlddx 23, 23, 15
+; CHECK-NEXT:    efdmul 23, 18, 23
+; CHECK-NEXT:    evldd 18, 200(31)
+; CHECK-NEXT:    efdmul 23, 23, 18
+; CHECK-NEXT:    efdmul 21, 19, 21
+; CHECK-NEXT:    efdadd 23, 21, 23
+; CHECK-NEXT:    lwz 21, 872(31)
+; CHECK-NEXT:    lwz 19, 772(31)
+; CHECK-NEXT:    lwz 18, 880(31)
+; CHECK-NEXT:    slwi 18, 18, 1
+; CHECK-NEXT:    add 19, 19, 18
+; CHECK-NEXT:    mullw 20, 20, 19
+; CHECK-NEXT:    add 22, 22, 20
+; CHECK-NEXT:    slwi 22, 22, 4
+; CHECK-NEXT:    add 22, 21, 22
+; CHECK-NEXT:    evstddx 23, 22, 3
+; CHECK-NEXT:    lwz 23, 868(31)
+; CHECK-NEXT:    lwz 22, 460(31)
+; CHECK-NEXT:    lwz 21, 884(31)
+; CHECK-NEXT:    slwi 20, 21, 3
+; CHECK-NEXT:    add 20, 20, 22
+; CHECK-NEXT:    slwi 20, 20, 4
+; CHECK-NEXT:    add 23, 20, 23
+; CHECK-NEXT:    li 20, -144
+; CHECK-NEXT:    evlddx 19, 23, 20
+; CHECK-NEXT:    evldd 18, 184(31)
+; CHECK-NEXT:    lwz 17, 864(31)
+; CHECK-NEXT:    efdcfsi 17, 17
+; CHECK-NEXT:    li 16, -136
+; CHECK-NEXT:    evlddx 23, 23, 16
+; CHECK-NEXT:    efdmul 23, 17, 23
+; CHECK-NEXT:    evlddx 17, 24, 3
+; CHECK-NEXT:    efdmul 23, 23, 17
+; CHECK-NEXT:    efdmul 19, 19, 18
+; CHECK-NEXT:    efdsub 23, 19, 23
+; CHECK-NEXT:    lwz 19, 872(31)
+; CHECK-NEXT:    lwz 18, 772(31)
+; CHECK-NEXT:    lwz 17, 880(31)
+; CHECK-NEXT:    mulli 17, 17, 9
+; CHECK-NEXT:    add 18, 18, 17
+; CHECK-NEXT:    mullw 21, 21, 18
+; CHECK-NEXT:    add 22, 22, 21
+; CHECK-NEXT:    slwi 22, 22, 4
+; CHECK-NEXT:    evstddx 23, 19, 22
+; CHECK-NEXT:    lwz 23, 868(31)
+; CHECK-NEXT:    lwz 22, 460(31)
+; CHECK-NEXT:    lwz 21, 884(31)
+; CHECK-NEXT:    slwi 19, 21, 3
+; CHECK-NEXT:    add 19, 19, 22
+; CHECK-NEXT:    slwi 19, 19, 4
+; CHECK-NEXT:    add 23, 19, 23
+; CHECK-NEXT:    evlddx 20, 23, 20
+; CHECK-NEXT:    evlddx 24, 24, 3
+; CHECK-NEXT:    lwz 19, 864(31)
+; CHECK-NEXT:    efdcfsi 19, 19
+; CHECK-NEXT:    evlddx 23, 23, 16
+; CHECK-NEXT:    efdmul 23, 19, 23
+; CHECK-NEXT:    evldd 19, 184(31)
+; CHECK-NEXT:    efdmul 23, 23, 19
+; CHECK-NEXT:    efdmul 24, 20, 24
+; CHECK-NEXT:    efdadd 24, 24, 23
+; CHECK-NEXT:    lwz 23, 872(31)
+; CHECK-NEXT:    lwz 20, 772(31)
+; CHECK-NEXT:    lwz 19, 880(31)
+; CHECK-NEXT:    mulli 19, 19, 9
+; CHECK-NEXT:    add 20, 20, 19
+; CHECK-NEXT:    mullw 21, 21, 20
+; CHECK-NEXT:    add 22, 22, 21
+; CHECK-NEXT:    slwi 22, 22, 4
+; CHECK-NEXT:    add 23, 23, 22
+; CHECK-NEXT:    evstddx 24, 23, 3
+; CHECK-NEXT:    li 24, 440
+; CHECK-NEXT:    evlddx 24, 31, 24
+; CHECK-NEXT:    li 23, 424
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 27
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 408
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 26
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 392
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 28
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 376
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 0
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 360
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 25
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evstdd 24, 104(31)
+; CHECK-NEXT:    evlddx 24, 29, 3
+; CHECK-NEXT:    evlddx 23, 5, 3
+; CHECK-NEXT:    efdmul 23, 23, 27
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 7, 3
+; CHECK-NEXT:    efdmul 23, 23, 26
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 9, 3
+; CHECK-NEXT:    efdmul 23, 23, 28
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 11, 3
+; CHECK-NEXT:    efdmul 23, 23, 0
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evlddx 23, 30, 3
+; CHECK-NEXT:    efdmul 23, 23, 25
+; CHECK-NEXT:    efdadd 23, 23, 24
+; CHECK-NEXT:    addi 24, 31, 104
+; CHECK-NEXT:    evstddx 23, 24, 3
+; CHECK-NEXT:    li 23, 808
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    li 22, 280
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 776
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    li 20, 296
+; CHECK-NEXT:    evlddx 20, 31, 20
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdmul 23, 23, 22
+; CHECK-NEXT:    efdsub 23, 23, 21
+; CHECK-NEXT:    li 22, 824
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 312
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdsub 23, 23, 22
+; CHECK-NEXT:    li 22, 840
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 328
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdadd 23, 22, 23
+; CHECK-NEXT:    li 22, 792
+; CHECK-NEXT:    evlddx 22, 31, 22
+; CHECK-NEXT:    li 21, 344
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    efdmul 22, 22, 21
+; CHECK-NEXT:    efdadd 23, 22, 23
+; CHECK-NEXT:    addi 22, 31, 88
+; CHECK-NEXT:    evstddx 23, 22, 3
+; CHECK-NEXT:    li 23, 808
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    evlddx 21, 4, 3
+; CHECK-NEXT:    li 20, 776
+; CHECK-NEXT:    evlddx 20, 31, 20
+; CHECK-NEXT:    evlddx 19, 6, 3
+; CHECK-NEXT:    efdmul 20, 20, 19
+; CHECK-NEXT:    efdmul 23, 23, 21
+; CHECK-NEXT:    efdsub 23, 23, 20
+; CHECK-NEXT:    li 21, 824
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 8, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdsub 23, 23, 21
+; CHECK-NEXT:    li 21, 840
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 10, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdadd 23, 21, 23
+; CHECK-NEXT:    li 21, 792
+; CHECK-NEXT:    evlddx 21, 31, 21
+; CHECK-NEXT:    evlddx 20, 12, 3
+; CHECK-NEXT:    efdmul 21, 21, 20
+; CHECK-NEXT:    efdadd 23, 21, 23
+; CHECK-NEXT:    efdneg 23, 23
+; CHECK-NEXT:    evstdd 23, 88(31)
+; CHECK-NEXT:    evldd 23, 104(31)
+; CHECK-NEXT:    evldd 21, 88(31)
+; CHECK-NEXT:    efdadd 23, 23, 21
+; CHECK-NEXT:    evstdd 23, 136(31)
+; CHECK-NEXT:    evlddx 23, 24, 3
+; CHECK-NEXT:    evlddx 21, 22, 3
+; CHECK-NEXT:    efdadd 23, 23, 21
+; CHECK-NEXT:    addi 21, 31, 136
+; CHECK-NEXT:    evstddx 23, 21, 3
+; CHECK-NEXT:    evldd 23, 104(31)
+; CHECK-NEXT:    evldd 20, 88(31)
+; CHECK-NEXT:    efdsub 23, 23, 20
+; CHECK-NEXT:    evstdd 23, 120(31)
+; CHECK-NEXT:    evlddx 24, 24, 3
+; CHECK-NEXT:    evlddx 23, 22, 3
+; CHECK-NEXT:    efdsub 23, 24, 23
+; CHECK-NEXT:    addi 24, 31, 120
+; CHECK-NEXT:    evstddx 23, 24, 3
+; CHECK-NEXT:    lwz 23, 868(31)
+; CHECK-NEXT:    lwz 22, 460(31)
+; CHECK-NEXT:    lwz 20, 884(31)
+; CHECK-NEXT:    slwi 19, 20, 1
+; CHECK-NEXT:    add 19, 19, 22
+; CHECK-NEXT:    slwi 19, 19, 4
+; CHECK-NEXT:    add 23, 19, 23
+; CHECK-NEXT:    li 19, -48
+; CHECK-NEXT:    evlddx 18, 23, 19
+; CHECK-NEXT:    evldd 17, 136(31)
+; CHECK-NEXT:    lwz 16, 864(31)
+; CHECK-NEXT:    efdcfsi 16, 16
+; CHECK-NEXT:    li 15, -40
+; CHECK-NEXT:    evlddx 23, 23, 15
+; CHECK-NEXT:    efdmul 23, 16, 23
+; CHECK-NEXT:    evlddx 16, 21, 3
+; CHECK-NEXT:    efdmul 23, 23, 16
+; CHECK-NEXT:    efdmul 18, 18, 17
+; CHECK-NEXT:    efdsub 23, 18, 23
+; CHECK-NEXT:    lwz 18, 872(31)
+; CHECK-NEXT:    lwz 17, 772(31)
+; CHECK-NEXT:    lwz 16, 880(31)
+; CHECK-NEXT:    mulli 16, 16, 3
+; CHECK-NEXT:    add 17, 17, 16
+; CHECK-NEXT:    mullw 20, 20, 17
+; CHECK-NEXT:    add 22, 22, 20
+; CHECK-NEXT:    slwi 22, 22, 4
+; CHECK-NEXT:    evstddx 23, 18, 22
+; CHECK-NEXT:    lwz 23, 868(31)
+; CHECK-NEXT:    lwz 22, 460(31)
+; CHECK-NEXT:    lwz 20, 884(31)
+; CHECK-NEXT:    slwi 18, 20, 1
+; CHECK-NEXT:    add 18, 18, 22
+; CHECK-NEXT:    slwi 18, 18, 4
+; CHECK-NEXT:    add 23, 18, 23
+; CHECK-NEXT:    evlddx 19, 23, 19
+; CHECK-NEXT:    evlddx 21, 21, 3
+; CHECK-NEXT:    lwz 18, 864(31)
+; CHECK-NEXT:    efdcfsi 18, 18
+; CHECK-NEXT:    evlddx 23, 23, 15
+; CHECK-NEXT:    efdmul 23, 18, 23
+; CHECK-NEXT:    evldd 18, 136(31)
+; CHECK-NEXT:    efdmul 23, 23, 18
+; CHECK-NEXT:    efdmul 21, 19, 21
+; CHECK-NEXT:    efdadd 23, 21, 23
+; CHECK-NEXT:    lwz 21, 872(31)
+; CHECK-NEXT:    lwz 19, 772(31)
+; CHECK-NEXT:    lwz 18, 880(31)
+; CHECK-NEXT:    mulli 18, 18, 3
+; CHECK-NEXT:    add 19, 19, 18
+; CHECK-NEXT:    mullw 20, 20, 19
+; CHECK-NEXT:    add 22, 22, 20
+; CHECK-NEXT:    slwi 22, 22, 4
+; CHECK-NEXT:    add 22, 21, 22
+; CHECK-NEXT:    evstddx 23, 22, 3
+; CHECK-NEXT:    lwz 23, 868(31)
+; CHECK-NEXT:    lwz 22, 460(31)
+; CHECK-NEXT:    lwz 21, 884(31)
+; CHECK-NEXT:    mulli 20, 21, 7
+; CHECK-NEXT:    add 20, 20, 22
+; CHECK-NEXT:    slwi 20, 20, 4
+; CHECK-NEXT:    add 23, 20, 23
+; CHECK-NEXT:    li 20, -128
+; CHECK-NEXT:    evlddx 19, 23, 20
+; CHECK-NEXT:    evldd 18, 120(31)
+; CHECK-NEXT:    lwz 17, 864(31)
+; CHECK-NEXT:    efdcfsi 17, 17
+; CHECK-NEXT:    li 16, -120
+; CHECK-NEXT:    evlddx 23, 23, 16
+; CHECK-NEXT:    efdmul 23, 17, 23
+; CHECK-NEXT:    evlddx 17, 24, 3
+; CHECK-NEXT:    efdmul 23, 23, 17
+; CHECK-NEXT:    efdmul 19, 19, 18
+; CHECK-NEXT:    efdsub 23, 19, 23
+; CHECK-NEXT:    lwz 19, 872(31)
+; CHECK-NEXT:    lwz 18, 772(31)
+; CHECK-NEXT:    lwz 17, 880(31)
+; CHECK-NEXT:    slwi 17, 17, 3
+; CHECK-NEXT:    add 18, 18, 17
+; CHECK-NEXT:    mullw 21, 21, 18
+; CHECK-NEXT:    add 22, 22, 21
+; CHECK-NEXT:    slwi 22, 22, 4
+; CHECK-NEXT:    evstddx 23, 19, 22
+; CHECK-NEXT:    lwz 23, 868(31)
+; CHECK-NEXT:    lwz 22, 460(31)
+; CHECK-NEXT:    lwz 21, 884(31)
+; CHECK-NEXT:    mulli 19, 21, 7
+; CHECK-NEXT:    add 19, 19, 22
+; CHECK-NEXT:    slwi 19, 19, 4
+; CHECK-NEXT:    add 23, 19, 23
+; CHECK-NEXT:    evlddx 20, 23, 20
+; CHECK-NEXT:    evlddx 24, 24, 3
+; CHECK-NEXT:    lwz 19, 864(31)
+; CHECK-NEXT:    efdcfsi 19, 19
+; CHECK-NEXT:    evlddx 23, 23, 16
+; CHECK-NEXT:    efdmul 23, 19, 23
+; CHECK-NEXT:    evldd 19, 120(31)
+; CHECK-NEXT:    efdmul 23, 23, 19
+; CHECK-NEXT:    efdmul 24, 20, 24
+; CHECK-NEXT:    efdadd 24, 24, 23
+; CHECK-NEXT:    lwz 23, 872(31)
+; CHECK-NEXT:    lwz 20, 772(31)
+; CHECK-NEXT:    lwz 19, 880(31)
+; CHECK-NEXT:    slwi 19, 19, 3
+; CHECK-NEXT:    add 20, 20, 19
+; CHECK-NEXT:    mullw 21, 21, 20
+; CHECK-NEXT:    add 22, 22, 21
+; CHECK-NEXT:    slwi 22, 22, 4
+; CHECK-NEXT:    add 23, 23, 22
+; CHECK-NEXT:    evstddx 24, 23, 3
+; CHECK-NEXT:    li 24, 440
+; CHECK-NEXT:    evlddx 24, 31, 24
+; CHECK-NEXT:    li 23, 424
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 25
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 408
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 27
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 392
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 0
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 376
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 26
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    li 23, 360
+; CHECK-NEXT:    evlddx 23, 31, 23
+; CHECK-NEXT:    efdmul 23, 23, 28
+; CHECK-NEXT:    efdadd 24, 23, 24
+; CHECK-NEXT:    evstdd 24, 40(31)
+; CHECK-NEXT:    evlddx 29, 29, 3
+; CHECK-NEXT:    evlddx 5, 5, 3
+; CHECK-NEXT:    efdmul 5, 5, 25
+; CHECK-NEXT:    efdadd 5, 5, 29
+; CHECK-NEXT:    evlddx 7, 7, 3
+; CHECK-NEXT:    efdmul 7, 7, 27
+; CHECK-NEXT:    efdadd 5, 7, 5
+; CHECK-NEXT:    evlddx 7, 9, 3
+; CHECK-NEXT:    efdmul 7, 7, 0
+; CHECK-NEXT:    efdadd 5, 7, 5
+; CHECK-NEXT:    evlddx 7, 11, 3
+; CHECK-NEXT:    efdmul 7, 7, 26
+; CHECK-NEXT:    efdadd 5, 7, 5
+; CHECK-NEXT:    evlddx 7, 30, 3
+; CHECK-NEXT:    efdmul 7, 7, 28
+; CHECK-NEXT:    efdadd 7, 7, 5
+; CHECK-NEXT:    addi 5, 31, 40
+; CHECK-NEXT:    evstddx 7, 5, 3
+; CHECK-NEXT:    li 7, 792
+; CHECK-NEXT:    evlddx 7, 31, 7
+; CHECK-NEXT:    li 9, 280
+; CHECK-NEXT:    evlddx 9, 31, 9
+; CHECK-NEXT:    li 11, 808
+; CHECK-NEXT:    evlddx 11, 31, 11
+; CHECK-NEXT:    li 30, 296
+; CHECK-NEXT:    evlddx 0, 31, 30
+; CHECK-NEXT:    efdmul 11, 11, 0
+; CHECK-NEXT:    efdmul 7, 7, 9
+; CHECK-NEXT:    efdsub 7, 7, 11
+; CHECK-NEXT:    li 9, 840
+; CHECK-NEXT:    evlddx 9, 31, 9
+; CHECK-NEXT:    li 11, 312
+; CHECK-NEXT:    evlddx 11, 31, 11
+; CHECK-NEXT:    efdmul 9, 9, 11
+; CHECK-NEXT:    efdadd 7, 9, 7
+; CHECK-NEXT:    li 9, 776
+; CHECK-NEXT:    evlddx 9, 31, 9
+; CHECK-NEXT:    li 11, 328
+; CHECK-NEXT:    evlddx 11, 31, 11
+; CHECK-NEXT:    efdmul 9, 9, 11
+; CHECK-NEXT:    efdadd 7, 9, 7
+; CHECK-NEXT:    li 9, 824
+; CHECK-NEXT:    evlddx 9, 31, 9
+; CHECK-NEXT:    li 11, 344
+; CHECK-NEXT:    evlddx 11, 31, 11
+; CHECK-NEXT:    efdmul 9, 9, 11
+; CHECK-NEXT:    efdsub 7, 7, 9
+; CHECK-NEXT:    addi 9, 31, 24
+; CHECK-NEXT:    evstddx 7, 9, 3
+; CHECK-NEXT:    li 7, 792
+; CHECK-NEXT:    evlddx 7, 31, 7
+; CHECK-NEXT:    evlddx 4, 4, 3
+; CHECK-NEXT:    li 11, 808
+; CHECK-NEXT:    evlddx 11, 31, 11
+; CHECK-NEXT:    evlddx 6, 6, 3
+; CHECK-NEXT:    efdmul 6, 11, 6
+; CHECK-NEXT:    efdmul 4, 7, 4
+; CHECK-NEXT:    efdsub 4, 4, 6
+; CHECK-NEXT:    li 6, 840
+; CHECK-NEXT:    evlddx 6, 31, 6
+; CHECK-NEXT:    evlddx 7, 8, 3
+; CHECK-NEXT:    efdmul 6, 6, 7
+; CHECK-NEXT:    efdadd 4, 6, 4
+; CHECK-NEXT:    li 6, 776
+; CHECK-NEXT:    evlddx 6, 31, 6
+; CHECK-NEXT:    evlddx 7, 10, 3
+; CHECK-NEXT:    efdmul 6, 6, 7
+; CHECK-NEXT:    efdadd 4, 6, 4
+; CHECK-NEXT:    li 6, 824
+; CHECK-NEXT:    evlddx 6, 31, 6
+; CHECK-NEXT:    evlddx 7, 12, 3
+; CHECK-NEXT:    efdmul 6, 6, 7
+; CHECK-NEXT:    efdsub 4, 4, 6
+; CHECK-NEXT:    efdneg 4, 4
+; CHECK-NEXT:    evstdd 4, 24(31)
+; CHECK-NEXT:    evldd 4, 40(31)
+; CHECK-NEXT:    evldd 6, 24(31)
+; CHECK-NEXT:    efdadd 4, 4, 6
+; CHECK-NEXT:    evstdd 4, 72(31)
+; CHECK-NEXT:    evlddx 4, 5, 3
+; CHECK-NEXT:    evlddx 6, 9, 3
+; CHECK-NEXT:    efdadd 4, 4, 6
+; CHECK-NEXT:    addi 6, 31, 72
+; CHECK-NEXT:    evstddx 4, 6, 3
+; CHECK-NEXT:    evldd 4, 40(31)
+; CHECK-NEXT:    evldd 7, 24(31)
+; CHECK-NEXT:    efdsub 4, 4, 7
+; CHECK-NEXT:    evstdd 4, 56(31)
+; CHECK-NEXT:    evlddx 4, 5, 3
+; CHECK-NEXT:    evlddx 5, 9, 3
+; CHECK-NEXT:    efdsub 5, 4, 5
+; CHECK-NEXT:    addi 4, 31, 56
+; CHECK-NEXT:    evstddx 5, 4, 3
+; CHECK-NEXT:    lwz 5, 868(31)
+; CHECK-NEXT:    lwz 7, 460(31)
+; CHECK-NEXT:    lwz 8, 884(31)
+; CHECK-NEXT:    mulli 9, 8, 3
+; CHECK-NEXT:    add 9, 9, 7
+; CHECK-NEXT:    slwi 9, 9, 4
+; CHECK-NEXT:    add 5, 9, 5
+; CHECK-NEXT:    li 9, -64
+; CHECK-NEXT:    evlddx 10, 5, 9
+; CHECK-NEXT:    evldd 11, 72(31)
+; CHECK-NEXT:    lwz 12, 864(31)
+; CHECK-NEXT:    efdcfsi 12, 12
+; CHECK-NEXT:    li 0, -56
+; CHECK-NEXT:    evlddx 5, 5, 0
+; CHECK-NEXT:    efdmul 5, 12, 5
+; CHECK-NEXT:    evlddx 12, 6, 3
+; CHECK-NEXT:    efdmul 5, 5, 12
+; CHECK-NEXT:    efdmul 10, 10, 11
+; CHECK-NEXT:    efdsub 5, 10, 5
+; CHECK-NEXT:    lwz 10, 872(31)
+; CHECK-NEXT:    lwz 11, 772(31)
+; CHECK-NEXT:    lwz 12, 880(31)
+; CHECK-NEXT:    slwi 12, 12, 2
+; CHECK-NEXT:    add 11, 11, 12
+; CHECK-NEXT:    mullw 8, 8, 11
+; CHECK-NEXT:    add 7, 7, 8
+; CHECK-NEXT:    slwi 7, 7, 4
+; CHECK-NEXT:    evstddx 5, 10, 7
+; CHECK-NEXT:    lwz 5, 868(31)
+; CHECK-NEXT:    lwz 7, 460(31)
+; CHECK-NEXT:    lwz 8, 884(31)
+; CHECK-NEXT:    mulli 10, 8, 3
+; CHECK-NEXT:    add 10, 10, 7
+; CHECK-NEXT:    slwi 10, 10, 4
+; CHECK-NEXT:    add 5, 10, 5
+; CHECK-NEXT:    evlddx 9, 5, 9
+; CHECK-NEXT:    evlddx 6, 6, 3
+; CHECK-NEXT:    lwz 10, 864(31)
+; CHECK-NEXT:    efdcfsi 10, 10
+; CHECK-NEXT:    evlddx 5, 5, 0
+; CHECK-NEXT:    efdmul 5, 10, 5
+; CHECK-NEXT:    evldd 10, 72(31)
+; CHECK-NEXT:    efdmul 5, 5, 10
+; CHECK-NEXT:    efdmul 6, 9, 6
+; CHECK-NEXT:    efdadd 5, 6, 5
+; CHECK-NEXT:    lwz 6, 872(31)
+; CHECK-NEXT:    lwz 9, 772(31)
+; CHECK-NEXT:    lwz 10, 880(31)
+; CHECK-NEXT:    slwi 10, 10, 2
+; CHECK-NEXT:    add 9, 9, 10
+; CHECK-NEXT:    mullw 8, 8, 9
+; CHECK-NEXT:    add 7, 7, 8
+; CHECK-NEXT:    slwi 7, 7, 4
+; CHECK-NEXT:    add 6, 6, 7
+; CHECK-NEXT:    evstddx 5, 6, 3
+; CHECK-NEXT:    lwz 5, 868(31)
+; CHECK-NEXT:    lwz 6, 460(31)
+; CHECK-NEXT:    lwz 7, 884(31)
+; CHECK-NEXT:    mulli 8, 7, 6
+; CHECK-NEXT:    add 8, 8, 6
+; CHECK-NEXT:    slwi 8, 8, 4
+; CHECK-NEXT:    add 5, 8, 5
+; CHECK-NEXT:    li 8, -112
+; CHECK-NEXT:    evlddx 9, 5, 8
+; CHECK-NEXT:    evldd 10, 56(31)
+; CHECK-NEXT:    lwz 11, 864(31)
+; CHECK-NEXT:    efdcfsi 11, 11
+; CHECK-NEXT:    li 12, -104
+; CHECK-NEXT:    evlddx 5, 5, 12
+; CHECK-NEXT:    efdmul 5, 11, 5
+; CHECK-NEXT:    evlddx 11, 4, 3
+; CHECK-NEXT:    efdmul 5, 5, 11
+; CHECK-NEXT:    efdmul 9, 9, 10
+; CHECK-NEXT:    efdsub 5, 9, 5
+; CHECK-NEXT:    lwz 9, 872(31)
+; CHECK-NEXT:    lwz 10, 772(31)
+; CHECK-NEXT:    lwz 11, 880(31)
+; CHECK-NEXT:    mulli 11, 11, 7
+; CHECK-NEXT:    add 10, 10, 11
+; CHECK-NEXT:    mullw 7, 7, 10
+; CHECK-NEXT:    add 6, 6, 7
+; CHECK-NEXT:    slwi 6, 6, 4
+; CHECK-NEXT:    evstddx 5, 9, 6
+; CHECK-NEXT:    lwz 5, 868(31)
+; CHECK-NEXT:    lwz 6, 460(31)
+; CHECK-NEXT:    lwz 7, 884(31)
+; CHECK-NEXT:    mulli 9, 7, 6
+; CHECK-NEXT:    add 9, 9, 6
+; CHECK-NEXT:    slwi 9, 9, 4
+; CHECK-NEXT:    add 5, 9, 5
+; CHECK-NEXT:    evlddx 8, 5, 8
+; CHECK-NEXT:    evlddx 4, 4, 3
+; CHECK-NEXT:    lwz 9, 864(31)
+; CHECK-NEXT:    efdcfsi 9, 9
+; CHECK-NEXT:    evlddx 5, 5, 12
+; CHECK-NEXT:    efdmul 5, 9, 5
+; CHECK-NEXT:    evldd 9, 56(31)
+; CHECK-NEXT:    efdmul 5, 5, 9
+; CHECK-NEXT:    efdmul 4, 8, 4
+; CHECK-NEXT:    efdadd 4, 4, 5
+; CHECK-NEXT:    lwz 5, 872(31)
+; CHECK-NEXT:    lwz 8, 772(31)
+; CHECK-NEXT:    lwz 9, 880(31)
+; CHECK-NEXT:    mulli 9, 9, 7
+; CHECK-NEXT:    add 8, 8, 9
+; CHECK-NEXT:    mullw 7, 7, 8
+; CHECK-NEXT:    add 6, 6, 7
+; CHECK-NEXT:    slwi 6, 6, 4
+; CHECK-NEXT:    add 5, 5, 6
+; CHECK-NEXT:    evstddx 4, 5, 3
+; CHECK-NEXT:    b .LBB1_5
+; CHECK-NEXT:  .LBB1_5:
+; CHECK-NEXT:    lwz 3, 460(31)
+; CHECK-NEXT:    addi 3, 3, 1
+; CHECK-NEXT:    stw 3, 460(31)
+; CHECK-NEXT:    b .LBB1_3
+; CHECK-NEXT:  .LBB1_6:
+; CHECK-NEXT:    b .LBB1_7
+; CHECK-NEXT:  .LBB1_7:
+; CHECK-NEXT:    lwz 3, 772(31)
+; CHECK-NEXT:    addi 3, 3, 1
+; CHECK-NEXT:    stw 3, 772(31)
+; CHECK-NEXT:    b .LBB1_1
+; CHECK-NEXT:  .LBB1_8:
+; CHECK-NEXT:    li 3, 1008
+; CHECK-NEXT:    evlddx 30, 31, 3 # 8-byte Folded Reload
+; CHECK-NEXT:    li 3, 1000
+; CHECK-NEXT:    evlddx 29, 31, 3 # 8-byte Folded Reload
+; CHECK-NEXT:    li 3, 992
+; CHECK-NEXT:    evlddx 28, 31, 3 # 8-byte Folded Reload
+; CHECK-NEXT:    li 3, 984
+; CHECK-NEXT:    evlddx 27, 31, 3 # 8-byte Folded Reload
+; CHECK-NEXT:    li 3, 976
+; CHECK-NEXT:    evlddx 26, 31, 3 # 8-byte Folded Reload
+; CHECK-NEXT:    li 3, 968
+; CHECK-NEXT:    evlddx 25, 31, 3 # 8-byte Folded Reload
+; CHECK-NEXT:    li 3, 960
+; CHECK-NEXT:    evlddx 24, 31, 3 # 8-byte Folded Reload
+; CHECK-NEXT:    li 3, 952
+; CHECK-NEXT:    evlddx 23, 31, 3 # 8-byte Folded Reload
+; CHECK-NEXT:    li 3, 944
+; CHECK-NEXT:    evlddx 22, 31, 3 # 8-byte Folded Reload
+; CHECK-NEXT:    li 3, 936
+; CHECK-NEXT:    evlddx 21, 31, 3 # 8-byte Folded Reload
+; CHECK-NEXT:    li 3, 928
+; CHECK-NEXT:    evlddx 20, 31, 3 # 8-byte Folded Reload
+; CHECK-NEXT:    li 3, 920
+; CHECK-NEXT:    evlddx 19, 31, 3 # 8-byte Folded Reload
+; CHECK-NEXT:    li 3, 912
+; CHECK-NEXT:    evlddx 18, 31, 3 # 8-byte Folded Reload
+; CHECK-NEXT:    li 3, 904
+; CHECK-NEXT:    evlddx 17, 31, 3 # 8-byte Folded Reload
+; CHECK-NEXT:    li 3, 896
+; CHECK-NEXT:    evlddx 16, 31, 3 # 8-byte Folded Reload
+; CHECK-NEXT:    li 3, 888
+; CHECK-NEXT:    evlddx 15, 31, 3 # 8-byte Folded Reload
+; CHECK-NEXT:    lwz 31, 1036(1)
+; CHECK-NEXT:    addi 1, 1, 1040
+; CHECK-NEXT:    blr
+  %7 = alloca i32, align 4
+  %8 = alloca i32, align 4
+  %9 = alloca ptr, align 4
+  %10 = alloca ptr, align 4
+  %11 = alloca ptr, align 4
+  %12 = alloca i32, align 4
+  %13 = alloca i32, align 4
+  %14 = alloca double, align 8
+  %15 = alloca double, align 8
+  %16 = alloca double, align 8
+  %17 = alloca double, align 8
+  %18 = alloca double, align 8
+  %19 = alloca double, align 8
+  %20 = alloca double, align 8
+  %21 = alloca double, align 8
+  %22 = alloca double, align 8
+  %23 = alloca double, align 8
+  %24 = alloca i32, align 4
+  %25 = alloca %struct.cmplx, align 8
+  %26 = alloca %struct.cmplx, align 8
+  %27 = alloca %struct.cmplx, align 8
+  %28 = alloca %struct.cmplx, align 8
+  %29 = alloca %struct.cmplx, align 8
+  %30 = alloca %struct.cmplx, align 8
+  %31 = alloca %struct.cmplx, align 8
+  %32 = alloca %struct.cmplx, align 8
+  %33 = alloca %struct.cmplx, align 8
+  %34 = alloca %struct.cmplx, align 8
+  %35 = alloca %struct.cmplx, align 8
+  %36 = alloca %struct.cmplx, align 8
+  %37 = alloca %struct.cmplx, align 8
+  %38 = alloca %struct.cmplx, align 8
+  %39 = alloca %struct.cmplx, align 8
+  %40 = alloca %struct.cmplx, align 8
+  %41 = alloca %struct.cmplx, align 8
+  %42 = alloca %struct.cmplx, align 8
+  %43 = alloca %struct.cmplx, align 8
+  %44 = alloca i32, align 4
+  %45 = alloca %struct.cmplx, align 8
+  %46 = alloca %struct.cmplx, align 8
+  %47 = alloca %struct.cmplx, align 8
+  %48 = alloca %struct.cmplx, align 8
+  %49 = alloca %struct.cmplx, align 8
+  %50 = alloca %struct.cmplx, align 8
+  %51 = alloca %struct.cmplx, align 8
+  %52 = alloca %struct.cmplx, align 8
+  %53 = alloca %struct.cmplx, align 8
+  %54 = alloca %struct.cmplx, align 8
+  %55 = alloca %struct.cmplx, align 8
+  %56 = alloca %struct.cmplx, align 8
+  %57 = alloca %struct.cmplx, align 8
+  %58 = alloca %struct.cmplx, align 8
+  %59 = alloca %struct.cmplx, align 8
+  %60 = alloca %struct.cmplx, align 8
+  %61 = alloca %struct.cmplx, align 8
+  %62 = alloca %struct.cmplx, align 8
+  %63 = alloca %struct.cmplx, align 8
+  %64 = alloca %struct.cmplx, align 8
+  %65 = alloca %struct.cmplx, align 8
+  %66 = alloca %struct.cmplx, align 8
+  %67 = alloca %struct.cmplx, align 8
+  %68 = alloca %struct.cmplx, align 8
+  %69 = alloca %struct.cmplx, align 8
+  %70 = alloca %struct.cmplx, align 8
+  %71 = alloca %struct.cmplx, align 8
+  store i32 %0, ptr %7, align 4
+  store i32 %1, ptr %8, align 4
+  store ptr %2, ptr %9, align 4
+  store ptr %3, ptr %10, align 4
+  store ptr %4, ptr %11, align 4
+  store i32 %5, ptr %12, align 4
+  store i32 11, ptr %13, align 4
+  store double 0x3FEAEB8C8764F0BA, ptr %14, align 8
+  %72 = load i32, ptr %12, align 4
+  %73 = sitofp i32 %72 to double
+  %74 = fmul double %73, 0x3FE14CEDF8BB580B
+  store double %74, ptr %15, align 8
+  store double 0x3FDA9628D9C712B6, ptr %16, align 8
+  %75 = load i32, ptr %12, align 4
+  %76 = sitofp i32 %75 to double
+  %77 = fmul double %76, 0x3FED1BB48EEE2C13
+  store double %77, ptr %17, align 8
+  store double 0xBFC2375F640F44DB, ptr %18, align 8
+  %78 = load i32, ptr %12, align 4
+  %79 = sitofp i32 %78 to double
+  %80 = fmul double %79, 0x3FEFAC9E043842EF
+  store double %80, ptr %19, align 8
+  store double 0xBFE4F49E7F775887, ptr %20, align 8
+  %81 = load i32, ptr %12, align 4
+  %82 = sitofp i32 %81 to double
+  %83 = fmul double %82, 0x3FE82F19BB3A28A1
+  store double %83, ptr %21, align 8
+  store double 0xBFEEB42A9BCD5057, ptr %22, align 8
+  %84 = load i32, ptr %12, align 4
+  %85 = sitofp i32 %84 to double
+  %86 = fmul double %85, 0x3FD207E7FD768DBF
+  store double %86, ptr %23, align 8
+  store i32 0, ptr %24, align 4
+  br label %87
+
+87:                                               ; preds = %2792, %6
+  %88 = load i32, ptr %24, align 4
+  %89 = load i32, ptr %8, align 4
+  %90 = icmp ult i32 %88, %89
+  br i1 %90, label %91, label %2795
+
+91:                                               ; preds = %87
+  %92 = load ptr, ptr %9, align 4
+  %93 = load i32, ptr %7, align 4
+  %94 = load i32, ptr %24, align 4
+  %95 = mul i32 11, %94
+  %96 = add i32 0, %95
+  %97 = mul i32 %93, %96
+  %98 = add i32 0, %97
+  %99 = getelementptr inbounds %struct.cmplx, ptr %92, i32 %98
+  call void @llvm.memcpy.p0.p0.i32(ptr align 8 %25, ptr align 8 %99, i32 16, i1 false)
+  %100 = load ptr, ptr %9, align 4
+  %101 = load i32, ptr %7, align 4
+  %102 = load i32, ptr %24, align 4
+  %103 = mul i32 11, %102
+  %104 = add i32 1, %103
+  %105 = mul i32 %101, %104
+  %106 = add i32 0, %105
+  %107 = getelementptr inbounds %struct.cmplx, ptr %100, i32 %106
+  %108 = getelementptr inbounds %struct.cmplx, ptr %107, i32 0, i32 0
+  %109 = load double, ptr %108, align 8
+  %110 = load ptr, ptr %9, align 4
+  %111 = load i32, ptr %7, align 4
+  %112 = load i32, ptr %24, align 4
+  %113 = mul i32 11, %112
+  %114 = add i32 10, %113
+  %115 = mul i32 %111, %114
+  %116 = add i32 0, %115
+  %117 = getelementptr inbounds %struct.cmplx, ptr %110, i32 %116
+  %118 = getelementptr inbounds %struct.cmplx, ptr %117, i32 0, i32 0
+  %119 = load double, ptr %118, align 8
+  %120 = fadd double %109, %119
+  %121 = getelementptr inbounds %struct.cmplx, ptr %26, i32 0, i32 0
+  store double %120, ptr %121, align 8
+  %122 = load ptr, ptr %9, align 4
+  %123 = load i32, ptr %7, align 4
+  %124 = load i32, ptr %24, align 4
+  %125 = mul i32 11, %124
+  %126 = add i32 1, %125
+  %127 = mul i32 %123, %126
+  %128 = add i32 0, %127
+  %129 = getelementptr inbounds %struct.cmplx, ptr %122, i32 %128
+  %130 = getelementptr inbounds %struct.cmplx, ptr %129, i32 0, i32 1
+  %131 = load double, ptr %130, align 8
+  %132 = load ptr, ptr %9, align 4
+  %133 = load i32, ptr %7, align 4
+  %134 = load i32, ptr %24, align 4
+  %135 = mul i32 11, %134
+  %136 = add i32 10, %135
+  %137 = mul i32 %133, %136
+  %138 = add i32 0, %137
+  %139 = getelementptr inbounds %struct.cmplx, ptr %132, i32 %138
+  %140 = getelementptr inbounds %struct.cmplx, ptr %139, i32 0, i32 1
+  %141 = load double, ptr %140, align 8
+  %142 = fadd double %131, %141
+  %143 = getelementptr inbounds %struct.cmplx, ptr %26, i32 0, i32 1
+  store double %142, ptr %143, align 8
+  %144 = load ptr, ptr %9, align 4
+  %145 = load i32, ptr %7, align 4
+  %146 = load i32, ptr %24, align 4
+  %147 = mul i32 11, %146
+  %148 = add i32 1, %147
+  %149 = mul i32 %145, %148
+  %150 = add i32 0, %149
+  %151 = getelementptr inbounds %struct.cmplx, ptr %144, i32 %150
+  %152 = getelementptr inbounds %struct.cmplx, ptr %151, i32 0, i32 0
+  %153 = load double, ptr %152, align 8
+  %154 = load ptr, ptr %9, align 4
+  %155 = load i32, ptr %7, align 4
+  %156 = load i32, ptr %24, align 4
+  %157 = mul i32 11, %156
+  %158 = add i32 10, %157
+  %159 = mul i32 %155, %158
+  %160 = add i32 0, %159
+  %161 = getelementptr inbounds %struct.cmplx, ptr %154, i32 %160
+  %162 = getelementptr inbounds %struct.cmplx, ptr %161, i32 0, i32 0
+  %163 = load double, ptr %162, align 8
+  %164 = fsub double %153, %163
+  %165 = getelementptr inbounds %struct.cmplx, ptr %35, i32 0, i32 0
+  store double %164, ptr %165, align 8
+  %166 = load ptr, ptr %9, align 4
+  %167 = load i32, ptr %7, align 4
+  %168 = load i32, ptr %24, align 4
+  %169 = mul i32 11, %168
+  %170 = add i32 1, %169
+  %171 = mul i32 %167, %170
+  %172 = add i32 0, %171
+  %173 = getelementptr inbounds %struct.cmplx, ptr %166, i32 %172
+  %174 = getelementptr inbounds %struct.cmplx, ptr %173, i32 0, i32 1
+  %175 = load double, ptr %174, align 8
+  %176 = load ptr, ptr %9, align 4
+  %177 = load i32, ptr %7, align 4
+  %178 = load i32, ptr %24, align 4
+  %179 = mul i32 11, %178
+  %180 = add i32 10, %179
+  %181 = mul i32 %177, %180
+  %182 = add i32 0, %181
+  %183 = getelementptr inbounds %struct.cmplx, ptr %176, i32 %182
+  %184 = getelementptr inbounds %struct.cmplx, ptr %183, i32 0, i32 1
+  %185 = load double, ptr %184, align 8
+  %186 = fsub double %175, %185
+  %187 = getelementptr inbounds %struct.cmplx, ptr %35, i32 0, i32 1
+  store double %186, ptr %187, align 8
+  %188 = load ptr, ptr %9, align 4
+  %189 = load i32, ptr %7, align 4
+  %190 = load i32, ptr %24, align 4
+  %191 = mul i32 11, %190
+  %192 = add i32 2, %191
+  %193 = mul i32 %189, %192
+  %194 = add i32 0, %193
+  %195 = getelementptr inbounds %struct.cmplx, ptr %188, i32 %194
+  %196 = getelementptr inbounds %struct.cmplx, ptr %195, i32 0, i32 0
+  %197 = load double, ptr %196, align 8
+  %198 = load ptr, ptr %9, align 4
+  %199 = load i32, ptr %7, align 4
+  %200 = load i32, ptr %24, align 4
+  %201 = mul i32 11, %200
+  %202 = add i32 9, %201
+  %203 = mul i32 %199, %202
+  %204 = add i32 0, %203
+  %205 = getelementptr inbounds %struct.cmplx, ptr %198, i32 %204
+  %206 = getelementptr inbounds %struct.cmplx, ptr %205, i32 0, i32 0
+  %207 = load double, ptr %206, align 8
+  %208 = fadd double %197, %207
+  %209 = getelementptr inbounds %struct.cmplx, ptr %27, i32 0, i32 0
+  store double %208, ptr %209, align 8
+  %210 = load ptr, ptr %9, align 4
+  %211 = load i32, ptr %7, align 4
+  %212 = load i32, ptr %24, align 4
+  %213 = mul i32 11, %212
+  %214 = add i32 2, %213
+  %215 = mul i32 %211, %214
+  %216 = add i32 0, %215
+  %217 = getelementptr inbounds %struct.cmplx, ptr %210, i32 %216
+  %218 = getelementptr inbounds %struct.cmplx, ptr %217, i32 0, i32 1
+  %219 = load double, ptr %218, align 8
+  %220 = load ptr, ptr %9, align 4
+  %221 = load i32, ptr %7, align 4
+  %222 = load i32, ptr %24, align 4
+  %223 = mul i32 11, %222
+  %224 = add i32 9, %223
+  %225 = mul i32 %221, %224
+  %226 = add i32 0, %225
+  %227 = getelementptr inbounds %struct.cmplx, ptr %220, i32 %226
+  %228 = getelementptr inbounds %struct.cmplx, ptr %227, i32 0, i32 1
+  %229 = load double, ptr %228, align 8
+  %230 = fadd double %219, %229
+  %231 = getelementptr inbounds %struct.cmplx, ptr %27, i32 0, i32 1
+  store double %230, ptr %231, align 8
+  %232 = load ptr, ptr %9, align 4
+  %233 = load i32, ptr %7, align 4
+  %234 = load i32, ptr %24, align 4
+  %235 = mul i32 11, %234
+  %236 = add i32 2, %235
+  %237 = mul i32 %233, %236
+  %238 = add i32 0, %237
+  %239 = getelementptr inbounds %struct.cmplx, ptr %232, i32 %238
+  %240 = getelementptr inbounds %struct.cmplx, ptr %239, i32 0, i32 0
+  %241 = load double, ptr %240, align 8
+  %242 = load ptr, ptr %9, align 4
+  %243 = load i32, ptr %7, align 4
+  %244 = load i32, ptr %24, align 4
+  %245 = mul i32 11, %244
+  %246 = add i32 9, %245
+  %247 = mul i32 %243, %246
+  %248 = add i32 0, %247
+  %249 = getelementptr inbounds %struct.cmplx, ptr %242, i32 %248
+  %250 = getelementptr inbounds %struct.cmplx, ptr %249, i32 0, i32 0
+  %251 = load double, ptr %250, align 8
+  %252 = fsub double %241, %251
+  %253 = getelementptr inbounds %struct.cmplx, ptr %34, i32 0, i32 0
+  store double %252, ptr %253, align 8
+  %254 = load ptr, ptr %9, align 4
+  %255 = load i32, ptr %7, align 4
+  %256 = load i32, ptr %24, align 4
+  %257 = mul i32 11, %256
+  %258 = add i32 2, %257
+  %259 = mul i32 %255, %258
+  %260 = add i32 0, %259
+  %261 = getelementptr inbounds %struct.cmplx, ptr %254, i32 %260
+  %262 = getelementptr inbounds %struct.cmplx, ptr %261, i32 0, i32 1
+  %263 = load double, ptr %262, align 8
+  %264 = load ptr, ptr %9, align 4
+  %265 = load i32, ptr %7, align 4
+  %266 = load i32, ptr %24, align 4
+  %267 = mul i32 11, %266
+  %268 = add i32 9, %267
+  %269 = mul i32 %265, %268
+  %270 = add i32 0, %269
+  %271 = getelementptr inbounds %struct.cmplx, ptr %264, i32 %270
+  %272 = getelementptr inbounds %struct.cmplx, ptr %271, i32 0, i32 1
+  %273 = load double, ptr %272, align 8
+  %274 = fsub double %263, %273
+  %275 = getelementptr inbounds %struct.cmplx, ptr %34, i32 0, i32 1
+  store double %274, ptr %275, align 8
+  %276 = load ptr, ptr %9, align 4
+  %277 = load i32, ptr %7, align 4
+  %278 = load i32, ptr %24, align 4
+  %279 = mul i32 11, %278
+  %280 = add i32 3, %279
+  %281 = mul i32 %277, %280
+  %282 = add i32 0, %281
+  %283 = getelementptr inbounds %struct.cmplx, ptr %276, i32 %282
+  %284 = getelementptr inbounds %struct.cmplx, ptr %283, i32 0, i32 0
+  %285 = load double, ptr %284, align 8
+  %286 = load ptr, ptr %9, align 4
+  %287 = load i32, ptr %7, align 4
+  %288 = load i32, ptr %24, align 4
+  %289 = mul i32 11, %288
+  %290 = add i32 8, %289
+  %291 = mul i32 %287, %290
+  %292 = add i32 0, %291
+  %293 = getelementptr inbounds %struct.cmplx, ptr %286, i32 %292
+  %294 = getelementptr inbounds %struct.cmplx, ptr %293, i32 0, i32 0
+  %295 = load double, ptr %294, align 8
+  %296 = fadd double %285, %295
+  %297 = getelementptr inbounds %struct.cmplx, ptr %28, i32 0, i32 0
+  store double %296, ptr %297, align 8
+  %298 = load ptr, ptr %9, align 4
+  %299 = load i32, ptr %7, align 4
+  %300 = load i32, ptr %24, align 4
+  %301 = mul i32 11, %300
+  %302 = add i32 3, %301
+  %303 = mul i32 %299, %302
+  %304 = add i32 0, %303
+  %305 = getelementptr inbounds %struct.cmplx, ptr %298, i32 %304
+  %306 = getelementptr inbounds %struct.cmplx, ptr %305, i32 0, i32 1
+  %307 = load double, ptr %306, align 8
+  %308 = load ptr, ptr %9, align 4
+  %309 = load i32, ptr %7, align 4
+  %310 = load i32, ptr %24, align 4
+  %311 = mul i32 11, %310
+  %312 = add i32 8, %311
+  %313 = mul i32 %309, %312
+  %314 = add i32 0, %313
+  %315 = getelementptr inbounds %struct.cmplx, ptr %308, i32 %314
+  %316 = getelementptr inbounds %struct.cmplx, ptr %315, i32 0, i32 1
+  %317 = load double, ptr %316, align 8
+  %318 = fadd double %307, %317
+  %319 = getelementptr inbounds %struct.cmplx, ptr %28, i32 0, i32 1
+  store double %318, ptr %319, align 8
+  %320 = load ptr, ptr %9, align 4
+  %321 = load i32, ptr %7, align 4
+  %322 = load i32, ptr %24, align 4
+  %323 = mul i32 11, %322
+  %324 = add i32 3, %323
+  %325 = mul i32 %321, %324
+  %326 = add i32 0, %325
+  %327 = getelementptr inbounds %struct.cmplx, ptr %320, i32 %326
+  %328 = getelementptr inbounds %struct.cmplx, ptr %327, i32 0, i32 0
+  %329 = load double, ptr %328, align 8
+  %330 = load ptr, ptr %9, align 4
+  %331 = load i32, ptr %7, align 4
+  %332 = load i32, ptr %24, align 4
+  %333 = mul i32 11, %332
+  %334 = add i32 8, %333
+  %335 = mul i32 %331, %334
+  %336 = add i32 0, %335
+  %337 = getelementptr inbounds %struct.cmplx, ptr %330, i32 %336
+  %338 = getelementptr inbounds %struct.cmplx, ptr %337, i32 0, i32 0
+  %339 = load double, ptr %338, align 8
+  %340 = fsub double %329, %339
+  %341 = getelementptr inbounds %struct.cmplx, ptr %33, i32 0, i32 0
+  store double %340, ptr %341, align 8
+  %342 = load ptr, ptr %9, align 4
+  %343 = load i32, ptr %7, align 4
+  %344 = load i32, ptr %24, align 4
+  %345 = mul i32 11, %344
+  %346 = add i32 3, %345
+  %347 = mul i32 %343, %346
+  %348 = add i32 0, %347
+  %349 = getelementptr inbounds %struct.cmplx, ptr %342, i32 %348
+  %350 = getelementptr inbounds %struct.cmplx, ptr %349, i32 0, i32 1
+  %351 = load double, ptr %350, align 8
+  %352 = load ptr, ptr %9, align 4
+  %353 = load i32, ptr %7, align 4
+  %354 = load i32, ptr %24, align 4
+  %355 = mul i32 11, %354
+  %356 = add i32 8, %355
+  %357 = mul i32 %353, %356
+  %358 = add i32 0, %357
+  %359 = getelementptr inbounds %struct.cmplx, ptr %352, i32 %358
+  %360 = getelementptr inbounds %struct.cmplx, ptr %359, i32 0, i32 1
+  %361 = load double, ptr %360, align 8
+  %362 = fsub double %351, %361
+  %363 = getelementptr inbounds %struct.cmplx, ptr %33, i32 0, i32 1
+  store double %362, ptr %363, align 8
+  %364 = load ptr, ptr %9, align 4
+  %365 = load i32, ptr %7, align 4
+  %366 = load i32, ptr %24, align 4
+  %367 = mul i32 11, %366
+  %368 = add i32 4, %367
+  %369 = mul i32 %365, %368
+  %370 = add i32 0, %369
+  %371 = getelementptr inbounds %struct.cmplx, ptr %364, i32 %370
+  %372 = getelementptr inbounds %struct.cmplx, ptr %371, i32 0, i32 0
+  %373 = load double, ptr %372, align 8
+  %374 = load ptr, ptr %9, align 4
+  %375 = load i32, ptr %7, align 4
+  %376 = load i32, ptr %24, align 4
+  %377 = mul i32 11, %376
+  %378 = add i32 7, %377
+  %379 = mul i32 %375, %378
+  %380 = add i32 0, %379
+  %381 = getelementptr inbounds %struct.cmplx, ptr %374, i32 %380
+  %382 = getelementptr inbounds %struct.cmplx, ptr %381, i32 0, i32 0
+  %383 = load double, ptr %382, align 8
+  %384 = fadd double %373, %383
+  %385 = getelementptr inbounds %struct.cmplx, ptr %29, i32 0, i32 0
+  store double %384, ptr %385, align 8
+  %386 = load ptr, ptr %9, align 4
+  %387 = load i32, ptr %7, align 4
+  %388 = load i32, ptr %24, align 4
+  %389 = mul i32 11, %388
+  %390 = add i32 4, %389
+  %391 = mul i32 %387, %390
+  %392 = add i32 0, %391
+  %393 = getelementptr inbounds %struct.cmplx, ptr %386, i32 %392
+  %394 = getelementptr inbounds %struct.cmplx, ptr %393, i32 0, i32 1
+  %395 = load double, ptr %394, align 8
+  %396 = load ptr, ptr %9, align 4
+  %397 = load i32, ptr %7, align 4
+  %398 = load i32, ptr %24, align 4
+  %399 = mul i32 11, %398
+  %400 = add i32 7, %399
+  %401 = mul i32 %397, %400
+  %402 = add i32 0, %401
+  %403 = getelementptr inbounds %struct.cmplx, ptr %396, i32 %402
+  %404 = getelementptr inbounds %struct.cmplx, ptr %403, i32 0, i32 1
+  %405 = load double, ptr %404, align 8
+  %406 = fadd double %395, %405
+  %407 = getelementptr inbounds %struct.cmplx, ptr %29, i32 0, i32 1
+  store double %406, ptr %407, align 8
+  %408 = load ptr, ptr %9, align 4
+  %409 = load i32, ptr %7, align 4
+  %410 = load i32, ptr %24, align 4
+  %411 = mul i32 11, %410
+  %412 = add i32 4, %411
+  %413 = mul i32 %409, %412
+  %414 = add i32 0, %413
+  %415 = getelementptr inbounds %struct.cmplx, ptr %408, i32 %414
+  %416 = getelementptr inbounds %struct.cmplx, ptr %415, i32 0, i32 0
+  %417 = load double, ptr %416, align 8
+  %418 = load ptr, ptr %9, align 4
+  %419 = load i32, ptr %7, align 4
+  %420 = load i32, ptr %24, align 4
+  %421 = mul i32 11, %420
+  %422 = add i32 7, %421
+  %423 = mul i32 %419, %422
+  %424 = add i32 0, %423
+  %425 = getelementptr inbounds %struct.cmplx, ptr %418, i32 %424
+  %426 = getelementptr inbounds %struct.cmplx, ptr %425, i32 0, i32 0
+  %427 = load double, ptr %426, align 8
+  %428 = fsub double %417, %427
+  %429 = getelementptr inbounds %struct.cmplx, ptr %32, i32 0, i32 0
+  store double %428, ptr %429, align 8
+  %430 = load ptr, ptr %9, align 4
+  %431 = load i32, ptr %7, align 4
+  %432 = load i32, ptr %24, align 4
+  %433 = mul i32 11, %432
+  %434 = add i32 4, %433
+  %435 = mul i32 %431, %434
+  %436 = add i32 0, %435
+  %437 = getelementptr inbounds %struct.cmplx, ptr %430, i32 %436
+  %438 = getelementptr inbounds %struct.cmplx, ptr %437, i32 0, i32 1
+  %439 = load double, ptr %438, align 8
+  %440 = load ptr, ptr %9, align 4
+  %441 = load i32, ptr %7, align 4
+  %442 = load i32, ptr %24, align 4
+  %443 = mul i32 11, %442
+  %444 = add i32 7, %443
+  %445 = mul i32 %441, %444
+  %446 = add i32 0, %445
+  %447 = getelementptr inbounds %struct.cmplx, ptr %440, i32 %446
+  %448 = getelementptr inbounds %struct.cmplx, ptr %447, i32 0, i32 1
+  %449 = load double, ptr %448, align 8
+  %450 = fsub double %439, %449
+  %451 = getelementptr inbounds %struct.cmplx, ptr %32, i32 0, i32 1
+  store double %450, ptr %451, align 8
+  %452 = load ptr, ptr %9, align 4
+  %453 = load i32, ptr %7, align 4
+  %454 = load i32, ptr %24, align 4
+  %455 = mul i32 11, %454
+  %456 = add i32 5, %455
+  %457 = mul i32 %453, %456
+  %458 = add i32 0, %457
+  %459 = getelementptr inbounds %struct.cmplx, ptr %452, i32 %458
+  %460 = getelementptr inbounds %struct.cmplx, ptr %459, i32 0, i32 0
+  %461 = load double, ptr %460, align 8
+  %462 = load ptr, ptr %9, align 4
+  %463 = load i32, ptr %7, align 4
+  %464 = load i32, ptr %24, align 4
+  %465 = mul i32 11, %464
+  %466 = add i32 6, %465
+  %467 = mul i32 %463, %466
+  %468 = add i32 0, %467
+  %469 = getelementptr inbounds %struct.cmplx, ptr %462, i32 %468
+  %470 = getelementptr inbounds %struct.cmplx, ptr %469, i32 0, i32 0
+  %471 = load double, ptr %470, align 8
+  %472 = fadd double %461, %471
+  %473 = getelementptr inbounds %struct.cmplx, ptr %30, i32 0, i32 0
+  store double %472, ptr %473, align 8
+  %474 = load ptr, ptr %9, align 4
+  %475 = load i32, ptr %7, align 4
+  %476 = load i32, ptr %24, align 4
+  %477 = mul i32 11, %476
+  %478 = add i32 5, %477
+  %479 = mul i32 %475, %478
+  %480 = add i32 0, %479
+  %481 = getelementptr inbounds %struct.cmplx, ptr %474, i32 %480
+  %482 = getelementptr inbounds %struct.cmplx, ptr %481, i32 0, i32 1
+  %483 = load double, ptr %482, align 8
+  %484 = load ptr, ptr %9, align 4
+  %485 = load i32, ptr %7, align 4
+  %486 = load i32, ptr %24, align 4
+  %487 = mul i32 11, %486
+  %488 = add i32 6, %487
+  %489 = mul i32 %485, %488
+  %490 = add i32 0, %489
+  %491 = getelementptr inbounds %struct.cmplx, ptr %484, i32 %490
+  %492 = getelementptr inbounds %struct.cmplx, ptr %491, i32 0, i32 1
+  %493 = load double, ptr %492, align 8
+  %494 = fadd double %483, %493
+  %495 = getelementptr inbounds %struct.cmplx, ptr %30, i32 0, i32 1
+  store double %494, ptr %495, align 8
+  %496 = load ptr, ptr %9, align 4
+  %497 = load i32, ptr %7, align 4
+  %498 = load i32, ptr %24, align 4
+  %499 = mul i32 11, %498
+  %500 = add i32 5, %499
+  %501 = mul i32 %497, %500
+  %502 = add i32 0, %501
+  %503 = getelementptr inbounds %struct.cmplx, ptr %496, i32 %502
+  %504 = getelementptr inbounds %struct.cmplx, ptr %503, i32 0, i32 0
+  %505 = load double, ptr %504, align 8
+  %506 = load ptr, ptr %9, align 4
+  %507 = load i32, ptr %7, align 4
+  %508 = load i32, ptr %24, align 4
+  %509 = mul i32 11, %508
+  %510 = add i32 6, %509
+  %511 = mul i32 %507, %510
+  %512 = add i32 0, %511
+  %513 = getelementptr inbounds %struct.cmplx, ptr %506, i32 %512
+  %514 = getelementptr inbounds %struct.cmplx, ptr %513, i32 0, i32 0
+  %515 = load double, ptr %514, align 8
+  %516 = fsub double %505, %515
+  %517 = getelementptr inbounds %struct.cmplx, ptr %31, i32 0, i32 0
+  store double %516, ptr %517, align 8
+  %518 = load ptr, ptr %9, align 4
+  %519 = load i32, ptr %7, align 4
+  %520 = load i32, ptr %24, align 4
+  %521 = mul i32 11, %520
+  %522 = add i32 5, %521
+  %523 = mul i32 %519, %522
+  %524 = add i32 0, %523
+  %525 = getelementptr inbounds %struct.cmplx, ptr %518, i32 %524
+  %526 = getelementptr inbounds %struct.cmplx, ptr %525, i32 0, i32 1
+  %527 = load double, ptr %526, align 8
+  %528 = load ptr, ptr %9, align 4
+  %529 = load i32, ptr %7, align 4
+  %530 = load i32, ptr %24, align 4
+  %531 = mul i32 11, %530
+  %532 = add i32 6, %531
+  %533 = mul i32 %529, %532
+  %534 = add i32 0, %533
+  %535 = getelementptr inbounds %struct.cmplx, ptr %528, i32 %534
+  %536 = getelementptr inbounds %struct.cmplx, ptr %535, i32 0, i32 1
+  %537 = load double, ptr %536, align 8
+  %538 = fsub double %527, %537
+  %539 = getelementptr inbounds %struct.cmplx, ptr %31, i32 0, i32 1
+  store double %538, ptr %539, align 8
+  %540 = getelementptr inbounds %struct.cmplx, ptr %25, i32 0, i32 0
+  %541 = load double, ptr %540, align 8
+  %542 = getelementptr inbounds %struct.cmplx, ptr %26, i32 0, i32 0
+  %543 = load double, ptr %542, align 8
+  %544 = fadd double %541, %543
+  %545 = getelementptr inbounds %struct.cmplx, ptr %27, i32 0, i32 0
+  %546 = load double, ptr %545, align 8
+  %547 = fadd double %544, %546
+  %548 = getelementptr inbounds %struct.cmplx, ptr %28, i32 0, i32 0
+  %549 = load double, ptr %548, align 8
+  %550 = fadd double %547, %549
+  %551 = getelementptr inbounds %struct.cmplx, ptr %29, i32 0, i32 0
+  %552 = load double, ptr %551, align 8
+  %553 = fadd double %550, %552
+  %554 = getelementptr inbounds %struct.cmplx, ptr %30, i32 0, i32 0
+  %555 = load double, ptr %554, align 8
+  %556 = fadd double %553, %555
+  %557 = load ptr, ptr %10, align 4
+  %558 = load i32, ptr %7, align 4
+  %559 = load i32, ptr %24, align 4
+  %560 = load i32, ptr %8, align 4
+  %561 = mul i32 %560, 0
+  %562 = add i32 %559, %561
+  %563 = mul i32 %558, %562
+  %564 = add i32 0, %563
+  %565 = getelementptr inbounds %struct.cmplx, ptr %557, i32 %564
+  %566 = getelementptr inbounds %struct.cmplx, ptr %565, i32 0, i32 0
+  store double %556, ptr %566, align 8
+  %567 = getelementptr inbounds %struct.cmplx, ptr %25, i32 0, i32 1
+  %568 = load double, ptr %567, align 8
+  %569 = getelementptr inbounds %struct.cmplx, ptr %26, i32 0, i32 1
+  %570 = load double, ptr %569, align 8
+  %571 = fadd double %568, %570
+  %572 = getelementptr inbounds %struct.cmplx, ptr %27, i32 0, i32 1
+  %573 = load double, ptr %572, align 8
+  %574 = fadd double %571, %573
+  %575 = getelementptr inbounds %struct.cmplx, ptr %28, i32 0, i32 1
+  %576 = load double, ptr %575, align 8
+  %577 = fadd double %574, %576
+  %578 = getelementptr inbounds %struct.cmplx, ptr %29, i32 0, i32 1
+  %579 = load double, ptr %578, align 8
+  %580 = fadd double %577, %579
+  %581 = getelementptr inbounds %struct.cmplx, ptr %30, i32 0, i32 1
+  %582 = load double, ptr %581, align 8
+  %583 = fadd double %580, %582
+  %584 = load ptr, ptr %10, align 4
+  %585 = load i32, ptr %7, align 4
+  %586 = load i32, ptr %24, align 4
+  %587 = load i32, ptr %8, align 4
+  %588 = mul i32 %587, 0
+  %589 = add i32 %586, %588
+  %590 = mul i32 %585, %589
+  %591 = add i32 0, %590
+  %592 = getelementptr inbounds %struct.cmplx, ptr %584, i32 %591
+  %593 = getelementptr inbounds %struct.cmplx, ptr %592, i32 0, i32 1
+  store double %583, ptr %593, align 8
+  %594 = getelementptr inbounds %struct.cmplx, ptr %25, i32 0, i32 0
+  %595 = load double, ptr %594, align 8
+  %596 = getelementptr inbounds %struct.cmplx, ptr %26, i32 0, i32 0
+  %597 = load double, ptr %596, align 8
+  %598 = call double @llvm.fmuladd.f64(double 0x3FEAEB8C8764F0BA, double %597, double %595)
+  %599 = getelementptr inbounds %struct.cmplx, ptr %27, i32 0, i32 0
+  %600 = load double, ptr %599, align 8
+  %601 = call double @llvm.fmuladd.f64(double 0x3FDA9628D9C712B6, double %600, double %598)
+  %602 = getelementptr inbounds %struct.cmplx, ptr %28, i32 0, i32 0
+  %603 = load double, ptr %602, align 8
+  %604 = call double @llvm.fmuladd.f64(double 0xBFC2375F640F44DB, double %603, double %601)
+  %605 = getelementptr inbounds %struct.cmplx, ptr %29, i32 0, i32 0
+  %606 = load double, ptr %605, align 8
+  %607 = call double @llvm.fmuladd.f64(double 0xBFE4F49E7F775887, double %606, double %604)
+  %608 = getelementptr inbounds %struct.cmplx, ptr %30, i32 0, i32 0
+  %609 = load double, ptr %608, align 8
+  %610 = call double @llvm.fmuladd.f64(double 0xBFEEB42A9BCD5057, double %609, double %607)
+  %611 = getelementptr inbounds %struct.cmplx, ptr %36, i32 0, i32 0
+  store double %610, ptr %611, align 8
+  %612 = getelementptr inbounds %struct.cmplx, ptr %25, i32 0, i32 1
+  %613 = load double, ptr %612, align 8
+  %614 = getelementptr inbounds %struct.cmplx, ptr %26, i32 0, i32 1
+  %615 = load double, ptr %614, align 8
+  %616 = call double @llvm.fmuladd.f64(double 0x3FEAEB8C8764F0BA, double %615, double %613)
+  %617 = getelementptr inbounds %struct.cmplx, ptr %27, i32 0, i32 1
+  %618 = load double, ptr %617, align 8
+  %619 = call double @llvm.fmuladd.f64(double 0x3FDA9628D9C712B6, double %618, double %616)
+  %620 = getelementptr inbounds %struct.cmplx, ptr %28, i32 0, i32 1
+  %621 = load double, ptr %620, align 8
+  %622 = call double @llvm.fmuladd.f64(double 0xBFC2375F640F44DB, double %621, double %619)
+  %623 = getelementptr inbounds %struct.cmplx, ptr %29, i32 0, i32 1
+  %624 = load double, ptr %623, align 8
+  %625 = call double @llvm.fmuladd.f64(double 0xBFE4F49E7F775887, double %624, double %622)
+  %626 = getelementptr inbounds %struct.cmplx, ptr %30, i32 0, i32 1
+  %627 = load double, ptr %626, align 8
+  %628 = call double @llvm.fmuladd.f64(double 0xBFEEB42A9BCD5057, double %627, double %625)
+  %629 = getelementptr inbounds %struct.cmplx, ptr %36, i32 0, i32 1
+  store double %628, ptr %629, align 8
+  %630 = load double, ptr %15, align 8
+  %631 = getelementptr inbounds %struct.cmplx, ptr %35, i32 0, i32 0
+  %632 = load double, ptr %631, align 8
+  %633 = load double, ptr %17, align 8
+  %634 = getelementptr inbounds %struct.cmplx, ptr %34, i32 0, i32 0
+  %635 = load double, ptr %634, align 8
+  %636 = fmul double %633, %635
+  %637 = call double @llvm.fmuladd.f64(double %630, double %632, double %636)
+  %638 = load double, ptr %19, align 8
+  %639 = getelementptr inbounds %struct.cmplx, ptr %33, i32 0, i32 0
+  %640 = load double, ptr %639, align 8
+  %641 = call double @llvm.fmuladd.f64(double %638, double %640, double %637)
+  %642 = load double, ptr %21, align 8
+  %643 = getelementptr inbounds %struct.cmplx, ptr %32, i32 0, i32 0
+  %644 = load double, ptr %643, align 8
+  %645 = call double @llvm.fmuladd.f64(double %642, double %644, double %641)
+  %646 = load double, ptr %23, align 8
+  %647 = getelementptr inbounds %struct.cmplx, ptr %31, i32 0, i32 0
+  %648 = load double, ptr %647, align 8
+  %649 = call double @llvm.fmuladd.f64(double %646, double %648, double %645)
+  %650 = getelementptr inbounds %struct.cmplx, ptr %37, i32 0, i32 1
+  store double %649, ptr %650, align 8
+  %651 = load double, ptr %15, align 8
+  %652 = getelementptr inbounds %struct.cmplx, ptr %35, i32 0, i32 1
+  %653 = load double, ptr %652, align 8
+  %654 = load double, ptr %17, align 8
+  %655 = getelementptr inbounds %struct.cmplx, ptr %34, i32 0, i32 1
+  %656 = load double, ptr %655, align 8
+  %657 = fmul double %654, %656
+  %658 = call double @llvm.fmuladd.f64(double %651, double %653, double %657)
+  %659 = load double, ptr %19, align 8
+  %660 = getelementptr inbounds %struct.cmplx, ptr %33, i32 0, i32 1
+  %661 = load double, ptr %660, align 8
+  %662 = call double @llvm.fmuladd.f64(double %659, double %661, double %658)
+  %663 = load double, ptr %21, align 8
+  %664 = getelementptr inbounds %struct.cmplx, ptr %32, i32 0, i32 1
+  %665 = load double, ptr %664, align 8
+  %666 = call double @llvm.fmuladd.f64(double %663, double %665, double %662)
+  %667 = load double, ptr %23, align 8
+  %668 = getelementptr inbounds %struct.cmplx, ptr %31, i32 0, i32 1
+  %669 = load double, ptr %668, align 8
+  %670 = call double @llvm.fmuladd.f64(double %667, double %669, double %666)
+  %671 = fneg double %670
+  %672 = getelementptr inbounds %struct.cmplx, ptr %37, i32 0, i32 0
+  store double %671, ptr %672, align 8
+  %673 = getelementptr inbounds %struct.cmplx, ptr %36, i32 0, i32 0
+  %674 = load double, ptr %673, align 8
+  %675 = getelementptr inbounds %struct.cmplx, ptr %37, i32 0, i32 0
+  %676 = load double, ptr %675, align 8
+  %677 = fadd double %674, %676
+  %678 = load ptr, ptr %10, align 4
+  %679 = load i32, ptr %7, align 4
+  %680 = load i32, ptr %24, align 4
+  %681 = load i32, ptr %8, align 4
+  %682 = mul i32 %681, 1
+  %683 = add i32 %680, %682
+  %684 = mul i32 %679, %683
+  %685 = add i32 0, %684
+  %686 = getelementptr inbounds %struct.cmplx, ptr %678, i32 %685
+  %687 = getelementptr inbounds %struct.cmplx, ptr %686, i32 0, i32 0
+  store double %677, ptr %687, align 8
+  %688 = getelementptr inbounds %struct.cmplx, ptr %36, i32 0, i32 1
+  %689 = load double, ptr %688, align 8
+  %690 = getelementptr inbounds %struct.cmplx, ptr %37, i32 0, i32 1
+  %691 = load double, ptr %690, align 8
+  %692 = fadd double %689, %691
+  %693 = load ptr, ptr %10, align 4
+  %694 = load i32, ptr %7, align 4
+  %695 = load i32, ptr %24, align 4
+  %696 = load i32, ptr %8, align 4
+  %697 = mul i32 %696, 1
+  %698 = add i32 %695, %697
+  %699 = mul i32 %694, %698
+  %700 = add i32 0, %699
+  %701 = getelementptr inbounds %struct.cmplx, ptr %693, i32 %700
+  %702 = getelementptr inbounds %struct.cmplx, ptr %701, i32 0, i32 1
+  store double %692, ptr %702, align 8
+  %703 = getelementptr inbounds %struct.cmplx, ptr %36, i32 0, i32 0
+  %704 = load double, ptr %703, align 8
+  %705 = getelementptr inbounds %struct.cmplx, ptr %37, i32 0, i32 0
+  %706 = load double, ptr %705, align 8
+  %707 = fsub double %704, %706
+  %708 = load ptr, ptr %10, align 4
+  %709 = load i32, ptr %7, align 4
+  %710 = load i32, ptr %24, align 4
+  %711 = load i32, ptr %8, align 4
+  %712 = mul i32 %711, 10
+  %713 = add i32 %710, %712
+  %714 = mul i32 %709, %713
+  %715 = add i32 0, %714
+  %716 = getelementptr inbounds %struct.cmplx, ptr %708, i32 %715
+  %717 = getelementptr inbounds %struct.cmplx, ptr %716, i32 0, i32 0
+  store double %707, ptr %717, align 8
+  %718 = getelementptr inbounds %struct.cmplx, ptr %36, i32 0, i32 1
+  %719 = load double, ptr %718, align 8
+  %720 = getelementptr inbounds %struct.cmplx, ptr %37, i32 0, i32 1
+  %721 = load double, ptr %720, align 8
+  %722 = fsub double %719, %721
+  %723 = load ptr, ptr %10, align 4
+  %724 = load i32, ptr %7, align 4
+  %725 = load i32, ptr %24, align 4
+  %726 = load i32, ptr %8, align 4
+  %727 = mul i32 %726, 10
+  %728 = add i32 %725, %727
+  %729 = mul i32 %724, %728
+  %730 = add i32 0, %729
+  %731 = getelementptr inbounds %struct.cmplx, ptr %723, i32 %730
+  %732 = getelementptr inbounds %struct.cmplx, ptr %731, i32 0, i32 1
+  store double %722, ptr %732, align 8
+  %733 = getelementptr inbounds %struct.cmplx, ptr %25, i32 0, i32 0
+  %734 = load double, ptr %733, align 8
+  %735 = getelementptr inbounds %struct.cmplx, ptr %26, i32 0, i32 0
+  %736 = load double, ptr %735, align 8
+  %737 = call double @llvm.fmuladd.f64(double 0x3FDA9628D9C712B6, double %736, double %734)
+  %738 = getelementptr inbounds %struct.cmplx, ptr %27, i32 0, i32 0
+  %739 = load double, ptr %738, align 8
+  %740 = call double @llvm.fmuladd.f64(double 0xBFE4F49E7F775887, double %739, double %737)
+  %741 = getelementptr inbounds %struct.cmplx, ptr %28, i32 0, i32 0
+  %742 = load double, ptr %741, align 8
+  %743 = call double @llvm.fmuladd.f64(double 0xBFEEB42A9BCD5057, double %742, double %740)
+  %744 = getelementptr inbounds %struct.cmplx, ptr %29, i32 0, i32 0
+  %745 = load double, ptr %744, align 8
+  %746 = call double @llvm.fmuladd.f64(double 0xBFC2375F640F44DB, double %745, double %743)
+  %747 = getelementptr inbounds %struct.cmplx, ptr %30, i32 0, i32 0
+  %748 = load double, ptr %747, align 8
+  %749 = call double @llvm.fmuladd.f64(double 0x3FEAEB8C8764F0BA, double %748, double %746)
+  %750 = getelementptr inbounds %struct.cmplx, ptr %38, i32 0, i32 0
+  store double %749, ptr %750, align 8
+  %751 = getelementptr inbounds %struct.cmplx, ptr %25, i32 0, i32 1
+  %752 = load double, ptr %751, align 8
+  %753 = getelementptr inbounds %struct.cmplx, ptr %26, i32 0, i32 1
+  %754 = load double, ptr %753, align 8
+  %755 = call double @llvm.fmuladd.f64(double 0x3FDA9628D9C712B6, double %754, double %752)
+  %756 = getelementptr inbounds %struct.cmplx, ptr %27, i32 0, i32 1
+  %757 = load double, ptr %756, align 8
+  %758 = call double @llvm.fmuladd.f64(double 0xBFE4F49E7F775887, double %757, double %755)
+  %759 = getelementptr inbounds %struct.cmplx, ptr %28, i32 0, i32 1
+  %760 = load double, ptr %759, align 8
+  %761 = call double @llvm.fmuladd.f64(double 0xBFEEB42A9BCD5057, double %760, double %758)
+  %762 = getelementptr inbounds %struct.cmplx, ptr %29, i32 0, i32 1
+  %763 = load double, ptr %762, align 8
+  %764 = call double @llvm.fmuladd.f64(double 0xBFC2375F640F44DB, double %763, double %761)
+  %765 = getelementptr inbounds %struct.cmplx, ptr %30, i32 0, i32 1
+  %766 = load double, ptr %765, align 8
+  %767 = call double @llvm.fmuladd.f64(double 0x3FEAEB8C8764F0BA, double %766, double %764)
+  %768 = getelementptr inbounds %struct.cmplx, ptr %38, i32 0, i32 1
+  store double %767, ptr %768, align 8
+  %769 = load double, ptr %17, align 8
+  %770 = getelementptr inbounds %struct.cmplx, ptr %35, i32 0, i32 0
+  %771 = load double, ptr %770, align 8
+  %772 = load double, ptr %21, align 8
+  %773 = getelementptr inbounds %struct.cmplx, ptr %34, i32 0, i32 0
+  %774 = load double, ptr %773, align 8
+  %775 = fmul double %772, %774
+  %776 = call double @llvm.fmuladd.f64(double %769, double %771, double %775)
+  %777 = load double, ptr %23, align 8
+  %778 = getelementptr inbounds %struct.cmplx, ptr %33, i32 0, i32 0
+  %779 = load double, ptr %778, align 8
+  %780 = fneg double %777
+  %781 = call double @llvm.fmuladd.f64(double %780, double %779, double %776)
+  %782 = load double, ptr %19, align 8
+  %783 = getelementptr inbounds %struct.cmplx, ptr %32, i32 0, i32 0
+  %784 = load double, ptr %783, align 8
+  %785 = fneg double %782
+  %786 = call double @llvm.fmuladd.f64(double %785, double %784, double %781)
+  %787 = load double, ptr %15, align 8
+  %788 = getelementptr inbounds %struct.cmplx, ptr %31, i32 0, i32 0
+  %789 = load double, ptr %788, align 8
+  %790 = fneg double %787
+  %791 = call double @llvm.fmuladd.f64(double %790, double %789, double %786)
+  %792 = getelementptr inbounds %struct.cmplx, ptr %39, i32 0, i32 1
+  store double %791, ptr %792, align 8
+  %793 = load double, ptr %17, align 8
+  %794 = getelementptr inbounds %struct.cmplx, ptr %35, i32 0, i32 1
+  %795 = load double, ptr %794, align 8
+  %796 = load double, ptr %21, align 8
+  %797 = getelementptr inbounds %struct.cmplx, ptr %34, i32 0, i32 1
+  %798 = load double, ptr %797, align 8
+  %799 = fmul double %796, %798
+  %800 = call double @llvm.fmuladd.f64(double %793, double %795, double %799)
+  %801 = load double, ptr %23, align 8
+  %802 = getelementptr inbounds %struct.cmplx, ptr %33, i32 0, i32 1
+  %803 = load double, ptr %802, align 8
+  %804 = fneg double %801
+  %805 = call double @llvm.fmuladd.f64(double %804, double %803, double %800)
+  %806 = load double, ptr %19, align 8
+  %807 = getelementptr inbounds %struct.cmplx, ptr %32, i32 0, i32 1
+  %808 = load double, ptr %807, align 8
+  %809 = fneg double %806
+  %810 = call double @llvm.fmuladd.f64(double %809, double %808, double %805)
+  %811 = load double, ptr %15, align 8
+  %812 = getelementptr inbounds %struct.cmplx, ptr %31, i32 0, i32 1
+  %813 = load double, ptr %812, align 8
+  %814 = fneg double %811
+  %815 = call double @llvm.fmuladd.f64(double %814, double %813, double %810)
+  %816 = fneg double %815
+  %817 = getelementptr inbounds %struct.cmplx, ptr %39, i32 0, i32 0
+  store double %816, ptr %817, align 8
+  %818 = getelementptr inbounds %struct.cmplx, ptr %38, i32 0, i32 0
+  %819 = load double, ptr %818, align 8
+  %820 = getelementptr inbounds %struct.cmplx, ptr %39, i32 0, i32 0
+  %821 = load double, ptr %820, align 8
+  %822 = fadd double %819, %821
+  %823 = load ptr, ptr %10, align 4
+  %824 = load i32, ptr %7, align 4
+  %825 = load i32, ptr %24, align 4
+  %826 = load i32, ptr %8, align 4
+  %827 = mul i32 %826, 2
+  %828 = add i32 %825, %827
+  %829 = mul i32 %824, %828
+  %830 = add i32 0, %829
+  %831 = getelementptr inbounds %struct.cmplx, ptr %823, i32 %830
+  %832 = getelementptr inbounds %struct.cmplx, ptr %831, i32 0, i32 0
+  store double %822, ptr %832, align 8
+  %833 = getelementptr inbounds %struct.cmplx, ptr %38, i32 0, i32 1
+  %834 = load double, ptr %833, align 8
+  %835 = getelementptr inbounds %struct.cmplx, ptr %39, i32 0, i32 1
+  %836 = load double, ptr %835, align 8
+  %837 = fadd double %834, %836
+  %838 = load ptr, ptr %10, align 4
+  %839 = load i32, ptr %7, align 4
+  %840 = load i32, ptr %24, align 4
+  %841 = load i32, ptr %8, align 4
+  %842 = mul i32 %841, 2
+  %843 = add i32 %840, %842
+  %844 = mul i32 %839, %843
+  %845 = add i32 0, %844
+  %846 = getelementptr inbounds %struct.cmplx, ptr %838, i32 %845
+  %847 = getelementptr inbounds %struct.cmplx, ptr %846, i32 0, i32 1
+  store double %837, ptr %847, align 8
+  %848 = getelementptr inbounds %struct.cmplx, ptr %38, i32 0, i32 0
+  %849 = load double, ptr %848, align 8
+  %850 = getelementptr inbounds %struct.cmplx, ptr %39, i32 0, i32 0
+  %851 = load double, ptr %850, align 8
+  %852 = fsub double %849, %851
+  %853 = load ptr, ptr %10, align 4
+  %854 = load i32, ptr %7, align 4
+  %855 = load i32, ptr %24, align 4
+  %856 = load i32, ptr %8, align 4
+  %857 = mul i32 %856, 9
+  %858 = add i32 %855, %857
+  %859 = mul i32 %854, %858
+  %860 = add i32 0, %859
+  %861 = getelementptr inbounds %struct.cmplx, ptr %853, i32 %860
+  %862 = getelementptr inbounds %struct.cmplx, ptr %861, i32 0, i32 0
+  store double %852, ptr %862, align 8
+  %863 = getelementptr inbounds %struct.cmplx, ptr %38, i32 0, i32 1
+  %864 = load double, ptr %863, align 8
+  %865 = getelementptr inbounds %struct.cmplx, ptr %39, i32 0, i32 1
+  %866 = load double, ptr %865, align 8
+  %867 = fsub double %864, %866
+  %868 = load ptr, ptr %10, align 4
+  %869 = load i32, ptr %7, align 4
+  %870 = load i32, ptr %24, align 4
+  %871 = load i32, ptr %8, align 4
+  %872 = mul i32 %871, 9
+  %873 = add i32 %870, %872
+  %874 = mul i32 %869, %873
+  %875 = add i32 0, %874
+  %876 = getelementptr inbounds %struct.cmplx, ptr %868, i32 %875
+  %877 = getelementptr inbounds %struct.cmplx, ptr %876, i32 0, i32 1
+  store double %867, ptr %877, align 8
+  %878 = getelementptr inbounds %struct.cmplx, ptr %25, i32 0, i32 0
+  %879 = load double, ptr %878, align 8
+  %880 = getelementptr inbounds %struct.cmplx, ptr %26, i32 0, i32 0
+  %881 = load double, ptr %880, align 8
+  %882 = call double @llvm.fmuladd.f64(double 0xBFC2375F640F44DB, double %881, double %879)
+  %883 = getelementptr inbounds %struct.cmplx, ptr %27, i32 0, i32 0
+  %884 = load double, ptr %883, align 8
+  %885 = call double @llvm.fmuladd.f64(double 0xBFEEB42A9BCD5057, double %884, double %882)
+  %886 = getelementptr inbounds %struct.cmplx, ptr %28, i32 0, i32 0
+  %887 = load double, ptr %886, align 8
+  %888 = call double @llvm.fmuladd.f64(double 0x3FDA9628D9C712B6, double %887, double %885)
+  %889 = getelementptr inbounds %struct.cmplx, ptr %29, i32 0, i32 0
+  %890 = load double, ptr %889, align 8
+  %891 = call double @llvm.fmuladd.f64(double 0x3FEAEB8C8764F0BA, double %890, double %888)
+  %892 = getelementptr inbounds %struct.cmplx, ptr %30, i32 0, i32 0
+  %893 = load double, ptr %892, align 8
+  %894 = call double @llvm.fmuladd.f64(double 0xBFE4F49E7F775887, double %893, double %891)
+  %895 = getelementptr inbounds %struct.cmplx, ptr %40, i32 0, i32 0
+  store double %894, ptr %895, align 8
+  %896 = getelementptr inbounds %struct.cmplx, ptr %25, i32 0, i32 1
+  %897 = load double, ptr %896, align 8
+  %898 = getelementptr inbounds %struct.cmplx, ptr %26, i32 0, i32 1
+  %899 = load double, ptr %898, align 8
+  %900 = call double @llvm.fmuladd.f64(double 0xBFC2375F640F44DB, double %899, double %897)
+  %901 = getelementptr inbounds %struct.cmplx, ptr %27, i32 0, i32 1
+  %902 = load double, ptr %901, align 8
+  %903 = call double @llvm.fmuladd.f64(double 0xBFEEB42A9BCD5057, double %902, double %900)
+  %904 = getelementptr inbounds %struct.cmplx, ptr %28, i32 0, i32 1
+  %905 = load double, ptr %904, align 8
+  %906 = call double @llvm.fmuladd.f64(double 0x3FDA9628D9C712B6, double %905, double %903)
+  %907 = getelementptr inbounds %struct.cmplx, ptr %29, i32 0, i32 1
+  %908 = load double, ptr %907, align 8
+  %909 = call double @llvm.fmuladd.f64(double 0x3FEAEB8C8764F0BA, double %908, double %906)
+  %910 = getelementptr inbounds %struct.cmplx, ptr %30, i32 0, i32 1
+  %911 = load double, ptr %910, align 8
+  %912 = call double @llvm.fmuladd.f64(double 0xBFE4F49E7F775887, double %911, double %909)
+  %913 = getelementptr inbounds %struct.cmplx, ptr %40, i32 0, i32 1
+  store double %912, ptr %913, align 8
+  %914 = load double, ptr %19, align 8
+  %915 = getelementptr inbounds %struct.cmplx, ptr %35, i32 0, i32 0
+  %916 = load double, ptr %915, align 8
+  %917 = load double, ptr %23, align 8
+  %918 = getelementptr inbounds %struct.cmplx, ptr %34, i32 0, i32 0
+  %919 = load double, ptr %918, align 8
+  %920 = fmul double %917, %919
+  %921 = fneg double %920
+  %922 = call double @llvm.fmuladd.f64(double %914, double %916, double %921)
+  %923 = load double, ptr %17, align 8
+  %924 = getelementptr inbounds %struct.cmplx, ptr %33, i32 0, i32 0
+  %925 = load double, ptr %924, align 8
+  %926 = fneg double %923
+  %927 = call double @llvm.fmuladd.f64(double %926, double %925, double %922)
+  %928 = load double, ptr %15, align 8
+  %929 = getelementptr inbounds %struct.cmplx, ptr %32, i32 0, i32 0
+  %930 = load double, ptr %929, align 8
+  %931 = call double @llvm.fmuladd.f64(double %928, double %930, double %927)
+  %932 = load double, ptr %21, align 8
+  %933 = getelementptr inbounds %struct.cmplx, ptr %31, i32 0, i32 0
+  %934 = load double, ptr %933, align 8
+  %935 = call double @llvm.fmuladd.f64(double %932, double %934, double %931)
+  %936 = getelementptr inbounds %struct.cmplx, ptr %41, i32 0, i32 1
+  store double %935, ptr %936, align 8
+  %937 = load double, ptr %19, align 8
+  %938 = getelementptr inbounds %struct.cmplx, ptr %35, i32 0, i32 1
+  %939 = load double, ptr %938, align 8
+  %940 = load double, ptr %23, align 8
+  %941 = getelementptr inbounds %struct.cmplx, ptr %34, i32 0, i32 1
+  %942 = load double, ptr %941, align 8
+  %943 = fmul double %940, %942
+  %944 = fneg double %943
+  %945 = call double @llvm.fmuladd.f64(double %937, double %939, double %944)
+  %946 = load double, ptr %17, align 8
+  %947 = getelementptr inbounds %struct.cmplx, ptr %33, i32 0, i32 1
+  %948 = load double, ptr %947, align 8
+  %949 = fneg double %946
+  %950 = call double @llvm.fmuladd.f64(double %949, double %948, double %945)
+  %951 = load double, ptr %15, align 8
+  %952 = getelementptr inbounds %struct.cmplx, ptr %32, i32 0, i32 1
+  %953 = load double, ptr %952, align 8
+  %954 = call double @llvm.fmuladd.f64(double %951, double %953, double %950)
+  %955 = load double, ptr %21, align 8
+  %956 = getelementptr inbounds %struct.cmplx, ptr %31, i32 0, i32 1
+  %957 = load double, ptr %956, align 8
+  %958 = call double @llvm.fmuladd.f64(double %955, double %957, double %954)
+  %959 = fneg double %958
+  %960 = getelementptr inbounds %struct.cmplx, ptr %41, i32 0, i32 0
+  store double %959, ptr %960, align 8
+  %961 = getelementptr inbounds %struct.cmplx, ptr %40, i32 0, i32 0
+  %962 = load double, ptr %961, align 8
+  %963 = getelementptr inbounds %struct.cmplx, ptr %41, i32 0, i32 0
+  %964 = load double, ptr %963, align 8
+  %965 = fadd double %962, %964
+  %966 = load ptr, ptr %10, align 4
+  %967 = load i32, ptr %7, align 4
+  %968 = load i32, ptr %24, align 4
+  %969 = load i32, ptr %8, align 4
+  %970 = mul i32 %969, 3
+  %971 = add i32 %968, %970
+  %972 = mul i32 %967, %971
+  %973 = add i32 0, %972
+  %974 = getelementptr inbounds %struct.cmplx, ptr %966, i32 %973
+  %975 = getelementptr inbounds %struct.cmplx, ptr %974, i32 0, i32 0
+  store double %965, ptr %975, align 8
+  %976 = getelementptr inbounds %struct.cmplx, ptr %40, i32 0, i32 1
+  %977 = load double, ptr %976, align 8
+  %978 = getelementptr inbounds %struct.cmplx, ptr %41, i32 0, i32 1
+  %979 = load double, ptr %978, align 8
+  %980 = fadd double %977, %979
+  %981 = load ptr, ptr %10, align 4
+  %982 = load i32, ptr %7, align 4
+  %983 = load i32, ptr %24, align 4
+  %984 = load i32, ptr %8, align 4
+  %985 = mul i32 %984, 3
+  %986 = add i32 %983, %985
+  %987 = mul i32 %982, %986
+  %988 = add i32 0, %987
+  %989 = getelementptr inbounds %struct.cmplx, ptr %981, i32 %988
+  %990 = getelementptr inbounds %struct.cmplx, ptr %989, i32 0, i32 1
+  store double %980, ptr %990, align 8
+  %991 = getelementptr inbounds %struct.cmplx, ptr %40, i32 0, i32 0
+  %992 = load double, ptr %991, align 8
+  %993 = getelementptr inbounds %struct.cmplx, ptr %41, i32 0, i32 0
+  %994 = load double, ptr %993, align 8
+  %995 = fsub double %992, %994
+  %996 = load ptr, ptr %10, align 4
+  %997 = load i32, ptr %7, align 4
+  %998 = load i32, ptr %24, align 4
+  %999 = load i32, ptr %8, align 4
+  %1000 = mul i32 %999, 8
+  %1001 = add i32 %998, %1000
+  %1002 = mul i32 %997, %1001
+  %1003 = add i32 0, %1002
+  %1004 = getelementptr inbounds %struct.cmplx, ptr %996, i32 %1003
+  %1005 = getelementptr inbounds %struct.cmplx, ptr %1004, i32 0, i32 0
+  store double %995, ptr %1005, align 8
+  %1006 = getelementptr inbounds %struct.cmplx, ptr %40, i32 0, i32 1
+  %1007 = load double, ptr %1006, align 8
+  %1008 = getelementptr inbounds %struct.cmplx, ptr %41, i32 0, i32 1
+  %1009 = load double, ptr %1008, align 8
+  %1010 = fsub double %1007, %1009
+  %1011 = load ptr, ptr %10, align 4
+  %1012 = load i32, ptr %7, align 4
+  %1013 = load i32, ptr %24, align 4
+  %1014 = load i32, ptr %8, align 4
+  %1015 = mul i32 %1014, 8
+  %1016 = add i32 %1013, %1015
+  %1017 = mul i32 %1012, %1016
+  %1018 = add i32 0, %1017
+  %1019 = getelementptr inbounds %struct.cmplx, ptr %1011, i32 %1018
+  %1020 = getelementptr inbounds %struct.cmplx, ptr %1019, i32 0, i32 1
+  store double %1010, ptr %1020, align 8
+  %1021 = getelementptr inbounds %struct.cmplx, ptr %25, i32 0, i32 0
+  %1022 = load double, ptr %1021, align 8
+  %1023 = getelementptr inbounds %struct.cmplx, ptr %26, i32 0, i32 0
+  %1024 = load double, ptr %1023, align 8
+  %1025 = call double @llvm.fmuladd.f64(double 0xBFE4F49E7F775887, double %1024, double %1022)
+  %1026 = getelementptr inbounds %struct.cmplx, ptr %27, i32 0, i32 0
+  %1027 = load double, ptr %1026, align 8
+  %1028 = call double @llvm.fmuladd.f64(double 0xBFC2375F640F44DB, double %1027, double %1025)
+  %1029 = getelementptr inbounds %struct.cmplx, ptr %28, i32 0, i32 0
+  %1030 = load double, ptr %1029, align 8
+  %1031 = call double @llvm.fmuladd.f64(double 0x3FEAEB8C8764F0BA, double %1030, double %1028)
+  %1032 = getelementptr inbounds %struct.cmplx, ptr %29, i32 0, i32 0
+  %1033 = load double, ptr %1032, align 8
+  %1034 = call double @llvm.fmuladd.f64(double 0xBFEEB42A9BCD5057, double %1033, double %1031)
+  %1035 = getelementptr inbounds %struct.cmplx, ptr %30, i32 0, i32 0
+  %1036 = load double, ptr %1035, align 8
+  %1037 = call double @llvm.fmuladd.f64(double 0x3FDA9628D9C712B6, double %1036, double %1034)
+  %1038 = getelementptr inbounds %struct.cmplx, ptr %42, i32 0, i32 0
+  store double %1037, ptr %1038, align 8
+  %1039 = getelementptr inbounds %struct.cmplx, ptr %25, i32 0, i32 1
+  %1040 = load double, ptr %1039, align 8
+  %1041 = getelementptr inbounds %struct.cmplx, ptr %26, i32 0, i32 1
+  %1042 = load double, ptr %1041, align 8
+  %1043 = call double @llvm.fmuladd.f64(double 0xBFE4F49E7F775887, double %1042, double %1040)
+  %1044 = getelementptr inbounds %struct.cmplx, ptr %27, i32 0, i32 1
+  %1045 = load double, ptr %1044, align 8
+  %1046 = call double @llvm.fmuladd.f64(double 0xBFC2375F640F44DB, double %1045, double %1043)
+  %1047 = getelementptr inbounds %struct.cmplx, ptr %28, i32 0, i32 1
+  %1048 = load double, ptr %1047, align 8
+  %1049 = call double @llvm.fmuladd.f64(double 0x3FEAEB8C8764F0BA, double %1048, double %1046)
+  %1050 = getelementptr inbounds %struct.cmplx, ptr %29, i32 0, i32 1
+  %1051 = load double, ptr %1050, align 8
+  %1052 = call double @llvm.fmuladd.f64(double 0xBFEEB42A9BCD5057, double %1051, double %1049)
+  %1053 = getelementptr inbounds %struct.cmplx, ptr %30, i32 0, i32 1
+  %1054 = load double, ptr %1053, align 8
+  %1055 = call double @llvm.fmuladd.f64(double 0x3FDA9628D9C712B6, double %1054, double %1052)
+  %1056 = getelementptr inbounds %struct.cmplx, ptr %42, i32 0, i32 1
+  store double %1055, ptr %1056, align 8
+  %1057 = load double, ptr %21, align 8
+  %1058 = getelementptr inbounds %struct.cmplx, ptr %35, i32 0, i32 0
+  %1059 = load double, ptr %1058, align 8
+  %1060 = load double, ptr %19, align 8
+  %1061 = getelementptr inbounds %struct.cmplx, ptr %34, i32 0, i32 0
+  %1062 = load double, ptr %1061, align 8
+  %1063 = fmul double %1060, %1062
+  %1064 = fneg double %1063
+  %1065 = call double @llvm.fmuladd.f64(double %1057, double %1059, double %1064)
+  %1066 = load double, ptr %15, align 8
+  %1067 = getelementptr inbounds %struct.cmplx, ptr %33, i32 0, i32 0
+  %1068 = load double, ptr %1067, align 8
+  %1069 = call double @llvm.fmuladd.f64(double %1066, double %1068, double %1065)
+  %1070 = load double, ptr %23, align 8
+  %1071 = getelementptr inbounds %struct.cmplx, ptr %32, i32 0, i32 0
+  %1072 = load double, ptr %1071, align 8
+  %1073 = call double @llvm.fmuladd.f64(double %1070, double %1072, double %1069)
+  %1074 = load double, ptr %17, align 8
+  %1075 = getelementptr inbounds %struct.cmplx, ptr %31, i32 0, i32 0
+  %1076 = load double, ptr %1075, align 8
+  %1077 = fneg double %1074
+  %1078 = call double @llvm.fmuladd.f64(double %1077, double %1076, double %1073)
+  %1079 = getelementptr inbounds %struct.cmplx, ptr %43, i32 0, i32 1
+  store double %1078, ptr %1079, align 8
+  %1080 = load double, ptr %21, align 8
+  %1081 = getelementptr inbounds %struct.cmplx, ptr %35, i32 0, i32 1
+  %1082 = load double, ptr %1081, align 8
+  %1083 = load double, ptr %19, align 8
+  %1084 = getelementptr inbounds %struct.cmplx, ptr %34, i32 0, i32 1
+  %1085 = load double, ptr %1084, align 8
+  %1086 = fmul double %1083, %1085
+  %1087 = fneg double %1086
+  %1088 = call double @llvm.fmuladd.f64(double %1080, double %1082, double %1087)
+  %1089 = load double, ptr %15, align 8
+  %1090 = getelementptr inbounds %struct.cmplx, ptr %33, i32 0, i32 1
+  %1091 = load double, ptr %1090, align 8
+  %1092 = call double @llvm.fmuladd.f64(double %1089, double %1091, double %1088)
+  %1093 = load double, ptr %23, align 8
+  %1094 = getelementptr inbounds %struct.cmplx, ptr %32, i32 0, i32 1
+  %1095 = load double, ptr %1094, align 8
+  %1096 = call double @llvm.fmuladd.f64(double %1093, double %1095, double %1092)
+  %1097 = load double, ptr %17, align 8
+  %1098 = getelementptr inbounds %struct.cmplx, ptr %31, i32 0, i32 1
+  %1099 = load double, ptr %1098, align 8
+  %1100 = fneg double %1097
+  %1101 = call double @llvm.fmuladd.f64(double %1100, double %1099, double %1096)
+  %1102 = fneg double %1101
+  %1103 = getelementptr inbounds %struct.cmplx, ptr %43, i32 0, i32 0
+  store double %1102, ptr %1103, align 8
+  %1104 = getelementptr inbounds %struct.cmplx, ptr %42, i32 0, i32 0
+  %1105 = load double, ptr %1104, align 8
+  %1106 = getelementptr inbounds %struct.cmplx, ptr %43, i32 0, i32 0
+  %1107 = load double, ptr %1106, align 8
+  %1108 = fadd double %1105, %1107
+  %1109 = load ptr, ptr %10, align 4
+  %1110 = load i32, ptr %7, align 4
+  %1111 = load i32, ptr %24, align 4
+  %1112 = load i32, ptr %8, align 4
+  %1113 = mul i32 %1112, 4
+  %1114 = add i32 %1111, %1113
+  %1115 = mul i32 %1110, %1114
+  %1116 = add i32 0, %1115
+  %1117 = getelementptr inbounds %struct.cmplx, ptr %1109, i32 %1116
+  %1118 = getelementptr inbounds %struct.cmplx, ptr %1117, i32 0, i32 0
+  store double %1108, ptr %1118, align 8
+  %1119 = getelementptr inbounds %struct.cmplx, ptr %42, i32 0, i32 1
+  %1120 = load double, ptr %1119, align 8
+  %1121 = getelementptr inbounds %struct.cmplx, ptr %43, i32 0, i32 1
+  %1122 = load double, ptr %1121, align 8
+  %1123 = fadd double %1120, %1122
+  %1124 = load ptr, ptr %10, align 4
+  %1125 = load i32, ptr %7, align 4
+  %1126 = load i32, ptr %24, align 4
+  %1127 = load i32, ptr %8, align 4
+  %1128 = mul i32 %1127, 4
+  %1129 = add i32 %1126, %1128
+  %1130 = mul i32 %1125, %1129
+  %1131 = add i32 0, %1130
+  %1132 = getelementptr inbounds %struct.cmplx, ptr %1124, i32 %1131
+  %1133 = getelementptr inbounds %struct.cmplx, ptr %1132, i32 0, i32 1
+  store double %1123, ptr %1133, align 8
+  %1134 = getelementptr inbounds %struct.cmplx, ptr %42, i32 0, i32 0
+  %1135 = load double, ptr %1134, align 8
+  %1136 = getelementptr inbounds %struct.cmplx, ptr %43, i32 0, i32 0
+  %1137 = load double, ptr %1136, align 8
+  %1138 = fsub double %1135, %1137
+  %1139 = load ptr, ptr %10, align 4
+  %1140 = load i32, ptr %7, align 4
+  %1141 = load i32, ptr %24, align 4
+  %1142 = load i32, ptr %8, align 4
+  %1143 = mul i32 %1142, 7
+  %1144 = add i32 %1141, %1143
+  %1145 = mul i32 %1140, %1144
+  %1146 = add i32 0, %1145
+  %1147 = getelementptr inbounds %struct.cmplx, ptr %1139, i32 %1146
+  %1148 = getelementptr inbounds %struct.cmplx, ptr %1147, i32 0, i32 0
+  store double %1138, ptr %1148, align 8
+  %1149 = getelementptr inbounds %struct.cmplx, ptr %42, i32 0, i32 1
+  %1150 = load double, ptr %1149, align 8
+  %1151 = getelementptr inbounds %struct.cmplx, ptr %43, i32 0, i32 1
+  %1152 = load double, ptr %1151, align 8
+  %1153 = fsub double %1150, %1152
+  %1154 = load ptr, ptr %10, align 4
+  %1155 = load i32, ptr %7, align 4
+  %1156 = load i32, ptr %24, align 4
+  %1157 = load i32, ptr %8, align 4
+  %1158 = mul i32 %1157, 7
+  %1159 = add i32 %1156, %1158
+  %1160 = mul i32 %1155, %1159
+  %1161 = add i32 0, %1160
+  %1162 = getelementptr inbounds %struct.cmplx, ptr %1154, i32 %1161
+  %1163 = getelementptr inbounds %struct.cmplx, ptr %1162, i32 0, i32 1
+  store double %1153, ptr %1163, align 8
+  store i32 1, ptr %44, align 4
+  br label %1164
+
+1164:                                             ; preds = %2788, %91
+  %1165 = load i32, ptr %44, align 4
+  %1166 = load i32, ptr %7, align 4
+  %1167 = icmp ult i32 %1165, %1166
+  br i1 %1167, label %1168, label %2791
+
+1168:                                             ; preds = %1164
+  %1169 = load ptr, ptr %9, align 4
+  %1170 = load i32, ptr %44, align 4
+  %1171 = load i32, ptr %7, align 4
+  %1172 = load i32, ptr %24, align 4
+  %1173 = mul i32 11, %1172
+  %1174 = add i32 0, %1173
+  %1175 = mul i32 %1171, %1174
+  %1176 = add i32 %1170, %1175
+  %1177 = getelementptr inbounds %struct.cmplx, ptr %1169, i32 %1176
+  call void @llvm.memcpy.p0.p0.i32(ptr align 8 %45, ptr align 8 %1177, i32 16, i1 false)
+  %1178 = load ptr, ptr %9, align 4
+  %1179 = load i32, ptr %44, align 4
+  %1180 = load i32, ptr %7, align 4
+  %1181 = load i32, ptr %24, align 4
+  %1182 = mul i32 11, %1181
+  %1183 = add i32 1, %1182
+  %1184 = mul i32 %1180, %1183
+  %1185 = add i32 %1179, %1184
+  %1186 = getelementptr inbounds %struct.cmplx, ptr %1178, i32 %1185
+  %1187 = getelementptr inbounds %struct.cmplx, ptr %1186, i32 0, i32 0
+  %1188 = load double, ptr %1187, align 8
+  %1189 = load ptr, ptr %9, align 4
+  %1190 = load i32, ptr %44, align 4
+  %1191 = load i32, ptr %7, align 4
+  %1192 = load i32, ptr %24, align 4
+  %1193 = mul i32 11, %1192
+  %1194 = add i32 10, %1193
+  %1195 = mul i32 %1191, %1194
+  %1196 = add i32 %1190, %1195
+  %1197 = getelementptr inbounds %struct.cmplx, ptr %1189, i32 %1196
+  %1198 = getelementptr inbounds %struct.cmplx, ptr %1197, i32 0, i32 0
+  %1199 = load double, ptr %1198, align 8
+  %1200 = fadd double %1188, %1199
+  %1201 = getelementptr inbounds %struct.cmplx, ptr %46, i32 0, i32 0
+  store double %1200, ptr %1201, align 8
+  %1202 = load ptr, ptr %9, align 4
+  %1203 = load i32, ptr %44, align 4
+  %1204 = load i32, ptr %7, align 4
+  %1205 = load i32, ptr %24, align 4
+  %1206 = mul i32 11, %1205
+  %1207 = add i32 1, %1206
+  %1208 = mul i32 %1204, %1207
+  %1209 = add i32 %1203, %1208
+  %1210 = getelementptr inbounds %struct.cmplx, ptr %1202, i32 %1209
+  %1211 = getelementptr inbounds %struct.cmplx, ptr %1210, i32 0, i32 1
+  %1212 = load double, ptr %1211, align 8
+  %1213 = load ptr, ptr %9, align 4
+  %1214 = load i32, ptr %44, align 4
+  %1215 = load i32, ptr %7, align 4
+  %1216 = load i32, ptr %24, align 4
+  %1217 = mul i32 11, %1216
+  %1218 = add i32 10, %1217
+  %1219 = mul i32 %1215, %1218
+  %1220 = add i32 %1214, %1219
+  %1221 = getelementptr inbounds %struct.cmplx, ptr %1213, i32 %1220
+  %1222 = getelementptr inbounds %struct.cmplx, ptr %1221, i32 0, i32 1
+  %1223 = load double, ptr %1222, align 8
+  %1224 = fadd double %1212, %1223
+  %1225 = getelementptr inbounds %struct.cmplx, ptr %46, i32 0, i32 1
+  store double %1224, ptr %1225, align 8
+  %1226 = load ptr, ptr %9, align 4
+  %1227 = load i32, ptr %44, align 4
+  %1228 = load i32, ptr %7, align 4
+  %1229 = load i32, ptr %24, align 4
+  %1230 = mul i32 11, %1229
+  %1231 = add i32 1, %1230
+  %1232 = mul i32 %1228, %1231
+  %1233 = add i32 %1227, %1232
+  %1234 = getelementptr inbounds %struct.cmplx, ptr %1226, i32 %1233
+  %1235 = getelementptr inbounds %struct.cmplx, ptr %1234, i32 0, i32 0
+  %1236 = load double, ptr %1235, align 8
+  %1237 = load ptr, ptr %9, align 4
+  %1238 = load i32, ptr %44, align 4
+  %1239 = load i32, ptr %7, align 4
+  %1240 = load i32, ptr %24, align 4
+  %1241 = mul i32 11, %1240
+  %1242 = add i32 10, %1241
+  %1243 = mul i32 %1239, %1242
+  %1244 = add i32 %1238, %1243
+  %1245 = getelementptr inbounds %struct.cmplx, ptr %1237, i32 %1244
+  %1246 = getelementptr inbounds %struct.cmplx, ptr %1245, i32 0, i32 0
+  %1247 = load double, ptr %1246, align 8
+  %1248 = fsub double %1236, %1247
+  %1249 = getelementptr inbounds %struct.cmplx, ptr %55, i32 0, i32 0
+  store double %1248, ptr %1249, align 8
+  %1250 = load ptr, ptr %9, align 4
+  %1251 = load i32, ptr %44, align 4
+  %1252 = load i32, ptr %7, align 4
+  %1253 = load i32, ptr %24, align 4
+  %1254 = mul i32 11, %1253
+  %1255 = add i32 1, %1254
+  %1256 = mul i32 %1252, %1255
+  %1257 = add i32 %1251, %1256
+  %1258 = getelementptr inbounds %struct.cmplx, ptr %1250, i32 %1257
+  %1259 = getelementptr inbounds %struct.cmplx, ptr %1258, i32 0, i32 1
+  %1260 = load double, ptr %1259, align 8
+  %1261 = load ptr, ptr %9, align 4
+  %1262 = load i32, ptr %44, align 4
+  %1263 = load i32, ptr %7, align 4
+  %1264 = load i32, ptr %24, align 4
+  %1265 = mul i32 11, %1264
+  %1266 = add i32 10, %1265
+  %1267 = mul i32 %1263, %1266
+  %1268 = add i32 %1262, %1267
+  %1269 = getelementptr inbounds %struct.cmplx, ptr %1261, i32 %1268
+  %1270 = getelementptr inbounds %struct.cmplx, ptr %1269, i32 0, i32 1
+  %1271 = load double, ptr %1270, align 8
+  %1272 = fsub double %1260, %1271
+  %1273 = getelementptr inbounds %struct.cmplx, ptr %55, i32 0, i32 1
+  store double %1272, ptr %1273, align 8
+  %1274 = load ptr, ptr %9, align 4
+  %1275 = load i32, ptr %44, align 4
+  %1276 = load i32, ptr %7, align 4
+  %1277 = load i32, ptr %24, align 4
+  %1278 = mul i32 11, %1277
+  %1279 = add i32 2, %1278
+  %1280 = mul i32 %1276, %1279
+  %1281 = add i32 %1275, %1280
+  %1282 = getelementptr inbounds %struct.cmplx, ptr %1274, i32 %1281
+  %1283 = getelementptr inbounds %struct.cmplx, ptr %1282, i32 0, i32 0
+  %1284 = load double, ptr %1283, align 8
+  %1285 = load ptr, ptr %9, align 4
+  %1286 = load i32, ptr %44, align 4
+  %1287 = load i32, ptr %7, align 4
+  %1288 = load i32, ptr %24, align 4
+  %1289 = mul i32 11, %1288
+  %1290 = add i32 9, %1289
+  %1291 = mul i32 %1287, %1290
+  %1292 = add i32 %1286, %1291
+  %1293 = getelementptr inbounds %struct.cmplx, ptr %1285, i32 %1292
+  %1294 = getelementptr inbounds %struct.cmplx, ptr %1293, i32 0, i32 0
+  %1295 = load double, ptr %1294, align 8
+  %1296 = fadd double %1284, %1295
+  %1297 = getelementptr inbounds %struct.cmplx, ptr %47, i32 0, i32 0
+  store double %1296, ptr %1297, align 8
+  %1298 = load ptr, ptr %9, align 4
+  %1299 = load i32, ptr %44, align 4
+  %1300 = load i32, ptr %7, align 4
+  %1301 = load i32, ptr %24, align 4
+  %1302 = mul i32 11, %1301
+  %1303 = add i32 2, %1302
+  %1304 = mul i32 %1300, %1303
+  %1305 = add i32 %1299, %1304
+  %1306 = getelementptr inbounds %struct.cmplx, ptr %1298, i32 %1305
+  %1307 = getelementptr inbounds %struct.cmplx, ptr %1306, i32 0, i32 1
+  %1308 = load double, ptr %1307, align 8
+  %1309 = load ptr, ptr %9, align 4
+  %1310 = load i32, ptr %44, align 4
+  %1311 = load i32, ptr %7, align 4
+  %1312 = load i32, ptr %24, align 4
+  %1313 = mul i32 11, %1312
+  %1314 = add i32 9, %1313
+  %1315 = mul i32 %1311, %1314
+  %1316 = add i32 %1310, %1315
+  %1317 = getelementptr inbounds %struct.cmplx, ptr %1309, i32 %1316
+  %1318 = getelementptr inbounds %struct.cmplx, ptr %1317, i32 0, i32 1
+  %1319 = load double, ptr %1318, align 8
+  %1320 = fadd double %1308, %1319
+  %1321 = getelementptr inbounds %struct.cmplx, ptr %47, i32 0, i32 1
+  store double %1320, ptr %1321, align 8
+  %1322 = load ptr, ptr %9, align 4
+  %1323 = load i32, ptr %44, align 4
+  %1324 = load i32, ptr %7, align 4
+  %1325 = load i32, ptr %24, align 4
+  %1326 = mul i32 11, %1325
+  %1327 = add i32 2, %1326
+  %1328 = mul i32 %1324, %1327
+  %1329 = add i32 %1323, %1328
+  %1330 = getelementptr inbounds %struct.cmplx, ptr %1322, i32 %1329
+  %1331 = getelementptr inbounds %struct.cmplx, ptr %1330, i32 0, i32 0
+  %1332 = load double, ptr %1331, align 8
+  %1333 = load ptr, ptr %9, align 4
+  %1334 = load i32, ptr %44, align 4
+  %1335 = load i32, ptr %7, align 4
+  %1336 = load i32, ptr %24, align 4
+  %1337 = mul i32 11, %1336
+  %1338 = add i32 9, %1337
+  %1339 = mul i32 %1335, %1338
+  %1340 = add i32 %1334, %1339
+  %1341 = getelementptr inbounds %struct.cmplx, ptr %1333, i32 %1340
+  %1342 = getelementptr inbounds %struct.cmplx, ptr %1341, i32 0, i32 0
+  %1343 = load double, ptr %1342, align 8
+  %1344 = fsub double %1332, %1343
+  %1345 = getelementptr inbounds %struct.cmplx, ptr %54, i32 0, i32 0
+  store double %1344, ptr %1345, align 8
+  %1346 = load ptr, ptr %9, align 4
+  %1347 = load i32, ptr %44, align 4
+  %1348 = load i32, ptr %7, align 4
+  %1349 = load i32, ptr %24, align 4
+  %1350 = mul i32 11, %1349
+  %1351 = add i32 2, %1350
+  %1352 = mul i32 %1348, %1351
+  %1353 = add i32 %1347, %1352
+  %1354 = getelementptr inbounds %struct.cmplx, ptr %1346, i32 %1353
+  %1355 = getelementptr inbounds %struct.cmplx, ptr %1354, i32 0, i32 1
+  %1356 = load double, ptr %1355, align 8
+  %1357 = load ptr, ptr %9, align 4
+  %1358 = load i32, ptr %44, align 4
+  %1359 = load i32, ptr %7, align 4
+  %1360 = load i32, ptr %24, align 4
+  %1361 = mul i32 11, %1360
+  %1362 = add i32 9, %1361
+  %1363 = mul i32 %1359, %1362
+  %1364 = add i32 %1358, %1363
+  %1365 = getelementptr inbounds %struct.cmplx, ptr %1357, i32 %1364
+  %1366 = getelementptr inbounds %struct.cmplx, ptr %1365, i32 0, i32 1
+  %1367 = load double, ptr %1366, align 8
+  %1368 = fsub double %1356, %1367
+  %1369 = getelementptr inbounds %struct.cmplx, ptr %54, i32 0, i32 1
+  store double %1368, ptr %1369, align 8
+  %1370 = load ptr, ptr %9, align 4
+  %1371 = load i32, ptr %44, align 4
+  %1372 = load i32, ptr %7, align 4
+  %1373 = load i32, ptr %24, align 4
+  %1374 = mul i32 11, %1373
+  %1375 = add i32 3, %1374
+  %1376 = mul i32 %1372, %1375
+  %1377 = add i32 %1371, %1376
+  %1378 = getelementptr inbounds %struct.cmplx, ptr %1370, i32 %1377
+  %1379 = getelementptr inbounds %struct.cmplx, ptr %1378, i32 0, i32 0
+  %1380 = load double, ptr %1379, align 8
+  %1381 = load ptr, ptr %9, align 4
+  %1382 = load i32, ptr %44, align 4
+  %1383 = load i32, ptr %7, align 4
+  %1384 = load i32, ptr %24, align 4
+  %1385 = mul i32 11, %1384
+  %1386 = add i32 8, %1385
+  %1387 = mul i32 %1383, %1386
+  %1388 = add i32 %1382, %1387
+  %1389 = getelementptr inbounds %struct.cmplx, ptr %1381, i32 %1388
+  %1390 = getelementptr inbounds %struct.cmplx, ptr %1389, i32 0, i32 0
+  %1391 = load double, ptr %1390, align 8
+  %1392 = fadd double %1380, %1391
+  %1393 = getelementptr inbounds %struct.cmplx, ptr %48, i32 0, i32 0
+  store double %1392, ptr %1393, align 8
+  %1394 = load ptr, ptr %9, align 4
+  %1395 = load i32, ptr %44, align 4
+  %1396 = load i32, ptr %7, align 4
+  %1397 = load i32, ptr %24, align 4
+  %1398 = mul i32 11, %1397
+  %1399 = add i32 3, %1398
+  %1400 = mul i32 %1396, %1399
+  %1401 = add i32 %1395, %1400
+  %1402 = getelementptr inbounds %struct.cmplx, ptr %1394, i32 %1401
+  %1403 = getelementptr inbounds %struct.cmplx, ptr %1402, i32 0, i32 1
+  %1404 = load double, ptr %1403, align 8
+  %1405 = load ptr, ptr %9, align 4
+  %1406 = load i32, ptr %44, align 4
+  %1407 = load i32, ptr %7, align 4
+  %1408 = load i32, ptr %24, align 4
+  %1409 = mul i32 11, %1408
+  %1410 = add i32 8, %1409
+  %1411 = mul i32 %1407, %1410
+  %1412 = add i32 %1406, %1411
+  %1413 = getelementptr inbounds %struct.cmplx, ptr %1405, i32 %1412
+  %1414 = getelementptr inbounds %struct.cmplx, ptr %1413, i32 0, i32 1
+  %1415 = load double, ptr %1414, align 8
+  %1416 = fadd double %1404, %1415
+  %1417 = getelementptr inbounds %struct.cmplx, ptr %48, i32 0, i32 1
+  store double %1416, ptr %1417, align 8
+  %1418 = load ptr, ptr %9, align 4
+  %1419 = load i32, ptr %44, align 4
+  %1420 = load i32, ptr %7, align 4
+  %1421 = load i32, ptr %24, align 4
+  %1422 = mul i32 11, %1421
+  %1423 = add i32 3, %1422
+  %1424 = mul i32 %1420, %1423
+  %1425 = add i32 %1419, %1424
+  %1426 = getelementptr inbounds %struct.cmplx, ptr %1418, i32 %1425
+  %1427 = getelementptr inbounds %struct.cmplx, ptr %1426, i32 0, i32 0
+  %1428 = load double, ptr %1427, align 8
+  %1429 = load ptr, ptr %9, align 4
+  %1430 = load i32, ptr %44, align 4
+  %1431 = load i32, ptr %7, align 4
+  %1432 = load i32, ptr %24, align 4
+  %1433 = mul i32 11, %1432
+  %1434 = add i32 8, %1433
+  %1435 = mul i32 %1431, %1434
+  %1436 = add i32 %1430, %1435
+  %1437 = getelementptr inbounds %struct.cmplx, ptr %1429, i32 %1436
+  %1438 = getelementptr inbounds %struct.cmplx, ptr %1437, i32 0, i32 0
+  %1439 = load double, ptr %1438, align 8
+  %1440 = fsub double %1428, %1439
+  %1441 = getelementptr inbounds %struct.cmplx, ptr %53, i32 0, i32 0
+  store double %1440, ptr %1441, align 8
+  %1442 = load ptr, ptr %9, align 4
+  %1443 = load i32, ptr %44, align 4
+  %1444 = load i32, ptr %7, align 4
+  %1445 = load i32, ptr %24, align 4
+  %1446 = mul i32 11, %1445
+  %1447 = add i32 3, %1446
+  %1448 = mul i32 %1444, %1447
+  %1449 = add i32 %1443, %1448
+  %1450 = getelementptr inbounds %struct.cmplx, ptr %1442, i32 %1449
+  %1451 = getelementptr inbounds %struct.cmplx, ptr %1450, i32 0, i32 1
+  %1452 = load double, ptr %1451, align 8
+  %1453 = load ptr, ptr %9, align 4
+  %1454 = load i32, ptr %44, align 4
+  %1455 = load i32, ptr %7, align 4
+  %1456 = load i32, ptr %24, align 4
+  %1457 = mul i32 11, %1456
+  %1458 = add i32 8, %1457
+  %1459 = mul i32 %1455, %1458
+  %1460 = add i32 %1454, %1459
+  %1461 = getelementptr inbounds %struct.cmplx, ptr %1453, i32 %1460
+  %1462 = getelementptr inbounds %struct.cmplx, ptr %1461, i32 0, i32 1
+  %1463 = load double, ptr %1462, align 8
+  %1464 = fsub double %1452, %1463
+  %1465 = getelementptr inbounds %struct.cmplx, ptr %53, i32 0, i32 1
+  store double %1464, ptr %1465, align 8
+  %1466 = load ptr, ptr %9, align 4
+  %1467 = load i32, ptr %44, align 4
+  %1468 = load i32, ptr %7, align 4
+  %1469 = load i32, ptr %24, align 4
+  %1470 = mul i32 11, %1469
+  %1471 = add i32 4, %1470
+  %1472 = mul i32 %1468, %1471
+  %1473 = add i32 %1467, %1472
+  %1474 = getelementptr inbounds %struct.cmplx, ptr %1466, i32 %1473
+  %1475 = getelementptr inbounds %struct.cmplx, ptr %1474, i32 0, i32 0
+  %1476 = load double, ptr %1475, align 8
+  %1477 = load ptr, ptr %9, align 4
+  %1478 = load i32, ptr %44, align 4
+  %1479 = load i32, ptr %7, align 4
+  %1480 = load i32, ptr %24, align 4
+  %1481 = mul i32 11, %1480
+  %1482 = add i32 7, %1481
+  %1483 = mul i32 %1479, %1482
+  %1484 = add i32 %1478, %1483
+  %1485 = getelementptr inbounds %struct.cmplx, ptr %1477, i32 %1484
+  %1486 = getelementptr inbounds %struct.cmplx, ptr %1485, i32 0, i32 0
+  %1487 = load double, ptr %1486, align 8
+  %1488 = fadd double %1476, %1487
+  %1489 = getelementptr inbounds %struct.cmplx, ptr %49, i32 0, i32 0
+  store double %1488, ptr %1489, align 8
+  %1490 = load ptr, ptr %9, align 4
+  %1491 = load i32, ptr %44, align 4
+  %1492 = load i32, ptr %7, align 4
+  %1493 = load i32, ptr %24, align 4
+  %1494 = mul i32 11, %1493
+  %1495 = add i32 4, %1494
+  %1496 = mul i32 %1492, %1495
+  %1497 = add i32 %1491, %1496
+  %1498 = getelementptr inbounds %struct.cmplx, ptr %1490, i32 %1497
+  %1499 = getelementptr inbounds %struct.cmplx, ptr %1498, i32 0, i32 1
+  %1500 = load double, ptr %1499, align 8
+  %1501 = load ptr, ptr %9, align 4
+  %1502 = load i32, ptr %44, align 4
+  %1503 = load i32, ptr %7, align 4
+  %1504 = load i32, ptr %24, align 4
+  %1505 = mul i32 11, %1504
+  %1506 = add i32 7, %1505
+  %1507 = mul i32 %1503, %1506
+  %1508 = add i32 %1502, %1507
+  %1509 = getelementptr inbounds %struct.cmplx, ptr %1501, i32 %1508
+  %1510 = getelementptr inbounds %struct.cmplx, ptr %1509, i32 0, i32 1
+  %1511 = load double, ptr %1510, align 8
+  %1512 = fadd double %1500, %1511
+  %1513 = getelementptr inbounds %struct.cmplx, ptr %49, i32 0, i32 1
+  store double %1512, ptr %1513, align 8
+  %1514 = load ptr, ptr %9, align 4
+  %1515 = load i32, ptr %44, align 4
+  %1516 = load i32, ptr %7, align 4
+  %1517 = load i32, ptr %24, align 4
+  %1518 = mul i32 11, %1517
+  %1519 = add i32 4, %1518
+  %1520 = mul i32 %1516, %1519
+  %1521 = add i32 %1515, %1520
+  %1522 = getelementptr inbounds %struct.cmplx, ptr %1514, i32 %1521
+  %1523 = getelementptr inbounds %struct.cmplx, ptr %1522, i32 0, i32 0
+  %1524 = load double, ptr %1523, align 8
+  %1525 = load ptr, ptr %9, align 4
+  %1526 = load i32, ptr %44, align 4
+  %1527 = load i32, ptr %7, align 4
+  %1528 = load i32, ptr %24, align 4
+  %1529 = mul i32 11, %1528
+  %1530 = add i32 7, %1529
+  %1531 = mul i32 %1527, %1530
+  %1532 = add i32 %1526, %1531
+  %1533 = getelementptr inbounds %struct.cmplx, ptr %1525, i32 %1532
+  %1534 = getelementptr inbounds %struct.cmplx, ptr %1533, i32 0, i32 0
+  %1535 = load double, ptr %1534, align 8
+  %1536 = fsub double %1524, %1535
+  %1537 = getelementptr inbounds %struct.cmplx, ptr %52, i32 0, i32 0
+  store double %1536, ptr %1537, align 8
+  %1538 = load ptr, ptr %9, align 4
+  %1539 = load i32, ptr %44, align 4
+  %1540 = load i32, ptr %7, align 4
+  %1541 = load i32, ptr %24, align 4
+  %1542 = mul i32 11, %1541
+  %1543 = add i32 4, %1542
+  %1544 = mul i32 %1540, %1543
+  %1545 = add i32 %1539, %1544
+  %1546 = getelementptr inbounds %struct.cmplx, ptr %1538, i32 %1545
+  %1547 = getelementptr inbounds %struct.cmplx, ptr %1546, i32 0, i32 1
+  %1548 = load double, ptr %1547, align 8
+  %1549 = load ptr, ptr %9, align 4
+  %1550 = load i32, ptr %44, align 4
+  %1551 = load i32, ptr %7, align 4
+  %1552 = load i32, ptr %24, align 4
+  %1553 = mul i32 11, %1552
+  %1554 = add i32 7, %1553
+  %1555 = mul i32 %1551, %1554
+  %1556 = add i32 %1550, %1555
+  %1557 = getelementptr inbounds %struct.cmplx, ptr %1549, i32 %1556
+  %1558 = getelementptr inbounds %struct.cmplx, ptr %1557, i32 0, i32 1
+  %1559 = load double, ptr %1558, align 8
+  %1560 = fsub double %1548, %1559
+  %1561 = getelementptr inbounds %struct.cmplx, ptr %52, i32 0, i32 1
+  store double %1560, ptr %1561, align 8
+  %1562 = load ptr, ptr %9, align 4
+  %1563 = load i32, ptr %44, align 4
+  %1564 = load i32, ptr %7, align 4
+  %1565 = load i32, ptr %24, align 4
+  %1566 = mul i32 11, %1565
+  %1567 = add i32 5, %1566
+  %1568 = mul i32 %1564, %1567
+  %1569 = add i32 %1563, %1568
+  %1570 = getelementptr inbounds %struct.cmplx, ptr %1562, i32 %1569
+  %1571 = getelementptr inbounds %struct.cmplx, ptr %1570, i32 0, i32 0
+  %1572 = load double, ptr %1571, align 8
+  %1573 = load ptr, ptr %9, align 4
+  %1574 = load i32, ptr %44, align 4
+  %1575 = load i32, ptr %7, align 4
+  %1576 = load i32, ptr %24, align 4
+  %1577 = mul i32 11, %1576
+  %1578 = add i32 6, %1577
+  %1579 = mul i32 %1575, %1578
+  %1580 = add i32 %1574, %1579
+  %1581 = getelementptr inbounds %struct.cmplx, ptr %1573, i32 %1580
+  %1582 = getelementptr inbounds %struct.cmplx, ptr %1581, i32 0, i32 0
+  %1583 = load double, ptr %1582, align 8
+  %1584 = fadd double %1572, %1583
+  %1585 = getelementptr inbounds %struct.cmplx, ptr %50, i32 0, i32 0
+  store double %1584, ptr %1585, align 8
+  %1586 = load ptr, ptr %9, align 4
+  %1587 = load i32, ptr %44, align 4
+  %1588 = load i32, ptr %7, align 4
+  %1589 = load i32, ptr %24, align 4
+  %1590 = mul i32 11, %1589
+  %1591 = add i32 5, %1590
+  %1592 = mul i32 %1588, %1591
+  %1593 = add i32 %1587, %1592
+  %1594 = getelementptr inbounds %struct.cmplx, ptr %1586, i32 %1593
+  %1595 = getelementptr inbounds %struct.cmplx, ptr %1594, i32 0, i32 1
+  %1596 = load double, ptr %1595, align 8
+  %1597 = load ptr, ptr %9, align 4
+  %1598 = load i32, ptr %44, align 4
+  %1599 = load i32, ptr %7, align 4
+  %1600 = load i32, ptr %24, align 4
+  %1601 = mul i32 11, %1600
+  %1602 = add i32 6, %1601
+  %1603 = mul i32 %1599, %1602
+  %1604 = add i32 %1598, %1603
+  %1605 = getelementptr inbounds %struct.cmplx, ptr %1597, i32 %1604
+  %1606 = getelementptr inbounds %struct.cmplx, ptr %1605, i32 0, i32 1
+  %1607 = load double, ptr %1606, align 8
+  %1608 = fadd double %1596, %1607
+  %1609 = getelementptr inbounds %struct.cmplx, ptr %50, i32 0, i32 1
+  store double %1608, ptr %1609, align 8
+  %1610 = load ptr, ptr %9, align 4
+  %1611 = load i32, ptr %44, align 4
+  %1612 = load i32, ptr %7, align 4
+  %1613 = load i32, ptr %24, align 4
+  %1614 = mul i32 11, %1613
+  %1615 = add i32 5, %1614
+  %1616 = mul i32 %1612, %1615
+  %1617 = add i32 %1611, %1616
+  %1618 = getelementptr inbounds %struct.cmplx, ptr %1610, i32 %1617
+  %1619 = getelementptr inbounds %struct.cmplx, ptr %1618, i32 0, i32 0
+  %1620 = load double, ptr %1619, align 8
+  %1621 = load ptr, ptr %9, align 4
+  %1622 = load i32, ptr %44, align 4
+  %1623 = load i32, ptr %7, align 4
+  %1624 = load i32, ptr %24, align 4
+  %1625 = mul i32 11, %1624
+  %1626 = add i32 6, %1625
+  %1627 = mul i32 %1623, %1626
+  %1628 = add i32 %1622, %1627
+  %1629 = getelementptr inbounds %struct.cmplx, ptr %1621, i32 %1628
+  %1630 = getelementptr inbounds %struct.cmplx, ptr %1629, i32 0, i32 0
+  %1631 = load double, ptr %1630, align 8
+  %1632 = fsub double %1620, %1631
+  %1633 = getelementptr inbounds %struct.cmplx, ptr %51, i32 0, i32 0
+  store double %1632, ptr %1633, align 8
+  %1634 = load ptr, ptr %9, align 4
+  %1635 = load i32, ptr %44, align 4
+  %1636 = load i32, ptr %7, align 4
+  %1637 = load i32, ptr %24, align 4
+  %1638 = mul i32 11, %1637
+  %1639 = add i32 5, %1638
+  %1640 = mul i32 %1636, %1639
+  %1641 = add i32 %1635, %1640
+  %1642 = getelementptr inbounds %struct.cmplx, ptr %1634, i32 %1641
+  %1643 = getelementptr inbounds %struct.cmplx, ptr %1642, i32 0, i32 1
+  %1644 = load double, ptr %1643, align 8
+  %1645 = load ptr, ptr %9, align 4
+  %1646 = load i32, ptr %44, align 4
+  %1647 = load i32, ptr %7, align 4
+  %1648 = load i32, ptr %24, align 4
+  %1649 = mul i32 11, %1648
+  %1650 = add i32 6, %1649
+  %1651 = mul i32 %1647, %1650
+  %1652 = add i32 %1646, %1651
+  %1653 = getelementptr inbounds %struct.cmplx, ptr %1645, i32 %1652
+  %1654 = getelementptr inbounds %struct.cmplx, ptr %1653, i32 0, i32 1
+  %1655 = load double, ptr %1654, align 8
+  %1656 = fsub double %1644, %1655
+  %1657 = getelementptr inbounds %struct.cmplx, ptr %51, i32 0, i32 1
+  store double %1656, ptr %1657, align 8
+  %1658 = getelementptr inbounds %struct.cmplx, ptr %45, i32 0, i32 0
+  %1659 = load double, ptr %1658, align 8
+  %1660 = getelementptr inbounds %struct.cmplx, ptr %46, i32 0, i32 0
+  %1661 = load double, ptr %1660, align 8
+  %1662 = fadd double %1659, %1661
+  %1663 = getelementptr inbounds %struct.cmplx, ptr %47, i32 0, i32 0
+  %1664 = load double, ptr %1663, align 8
+  %1665 = fadd double %1662, %1664
+  %1666 = getelementptr inbounds %struct.cmplx, ptr %48, i32 0, i32 0
+  %1667 = load double, ptr %1666, align 8
+  %1668 = fadd double %1665, %1667
+  %1669 = getelementptr inbounds %struct.cmplx, ptr %49, i32 0, i32 0
+  %1670 = load double, ptr %1669, align 8
+  %1671 = fadd double %1668, %1670
+  %1672 = getelementptr inbounds %struct.cmplx, ptr %50, i32 0, i32 0
+  %1673 = load double, ptr %1672, align 8
+  %1674 = fadd double %1671, %1673
+  %1675 = load ptr, ptr %10, align 4
+  %1676 = load i32, ptr %44, align 4
+  %1677 = load i32, ptr %7, align 4
+  %1678 = load i32, ptr %24, align 4
+  %1679 = load i32, ptr %8, align 4
+  %1680 = mul i32 %1679, 0
+  %1681 = add i32 %1678, %1680
+  %1682 = mul i32 %1677, %1681
+  %1683 = add i32 %1676, %1682
+  %1684 = getelementptr inbounds %struct.cmplx, ptr %1675, i32 %1683
+  %1685 = getelementptr inbounds %struct.cmplx, ptr %1684, i32 0, i32 0
+  store double %1674, ptr %1685, align 8
+  %1686 = getelementptr inbounds %struct.cmplx, ptr %45, i32 0, i32 1
+  %1687 = load double, ptr %1686, align 8
+  %1688 = getelementptr inbounds %struct.cmplx, ptr %46, i32 0, i32 1
+  %1689 = load double, ptr %1688, align 8
+  %1690 = fadd double %1687, %1689
+  %1691 = getelementptr inbounds %struct.cmplx, ptr %47, i32 0, i32 1
+  %1692 = load double, ptr %1691, align 8
+  %1693 = fadd double %1690, %1692
+  %1694 = getelementptr inbounds %struct.cmplx, ptr %48, i32 0, i32 1
+  %1695 = load double, ptr %1694, align 8
+  %1696 = fadd double %1693, %1695
+  %1697 = getelementptr inbounds %struct.cmplx, ptr %49, i32 0, i32 1
+  %1698 = load double, ptr %1697, align 8
+  %1699 = fadd double %1696, %1698
+  %1700 = getelementptr inbounds %struct.cmplx, ptr %50, i32 0, i32 1
+  %1701 = load double, ptr %1700, align 8
+  %1702 = fadd double %1699, %1701
+  %1703 = load ptr, ptr %10, align 4
+  %1704 = load i32, ptr %44, align 4
+  %1705 = load i32, ptr %7, align 4
+  %1706 = load i32, ptr %24, align 4
+  %1707 = load i32, ptr %8, align 4
+  %1708 = mul i32 %1707, 0
+  %1709 = add i32 %1706, %1708
+  %1710 = mul i32 %1705, %1709
+  %1711 = add i32 %1704, %1710
+  %1712 = getelementptr inbounds %struct.cmplx, ptr %1703, i32 %1711
+  %1713 = getelementptr inbounds %struct.cmplx, ptr %1712, i32 0, i32 1
+  store double %1702, ptr %1713, align 8
+  %1714 = getelementptr inbounds %struct.cmplx, ptr %45, i32 0, i32 0
+  %1715 = load double, ptr %1714, align 8
+  %1716 = getelementptr inbounds %struct.cmplx, ptr %46, i32 0, i32 0
+  %1717 = load double, ptr %1716, align 8
+  %1718 = call double @llvm.fmuladd.f64(double 0x3FEAEB8C8764F0BA, double %1717, double %1715)
+  %1719 = getelementptr inbounds %struct.cmplx, ptr %47, i32 0, i32 0
+  %1720 = load double, ptr %1719, align 8
+  %1721 = call double @llvm.fmuladd.f64(double 0x3FDA9628D9C712B6, double %1720, double %1718)
+  %1722 = getelementptr inbounds %struct.cmplx, ptr %48, i32 0, i32 0
+  %1723 = load double, ptr %1722, align 8
+  %1724 = call double @llvm.fmuladd.f64(double 0xBFC2375F640F44DB, double %1723, double %1721)
+  %1725 = getelementptr inbounds %struct.cmplx, ptr %49, i32 0, i32 0
+  %1726 = load double, ptr %1725, align 8
+  %1727 = call double @llvm.fmuladd.f64(double 0xBFE4F49E7F775887, double %1726, double %1724)
+  %1728 = getelementptr inbounds %struct.cmplx, ptr %50, i32 0, i32 0
+  %1729 = load double, ptr %1728, align 8
+  %1730 = call double @llvm.fmuladd.f64(double 0xBFEEB42A9BCD5057, double %1729, double %1727)
+  %1731 = getelementptr inbounds %struct.cmplx, ptr %58, i32 0, i32 0
+  store double %1730, ptr %1731, align 8
+  %1732 = getelementptr inbounds %struct.cmplx, ptr %45, i32 0, i32 1
+  %1733 = load double, ptr %1732, align 8
+  %1734 = getelementptr inbounds %struct.cmplx, ptr %46, i32 0, i32 1
+  %1735 = load double, ptr %1734, align 8
+  %1736 = call double @llvm.fmuladd.f64(double 0x3FEAEB8C8764F0BA, double %1735, double %1733)
+  %1737 = getelementptr inbounds %struct.cmplx, ptr %47, i32 0, i32 1
+  %1738 = load double, ptr %1737, align 8
+  %1739 = call double @llvm.fmuladd.f64(double 0x3FDA9628D9C712B6, double %1738, double %1736)
+  %1740 = getelementptr inbounds %struct.cmplx, ptr %48, i32 0, i32 1
+  %1741 = load double, ptr %1740, align 8
+  %1742 = call double @llvm.fmuladd.f64(double 0xBFC2375F640F44DB, double %1741, double %1739)
+  %1743 = getelementptr inbounds %struct.cmplx, ptr %49, i32 0, i32 1
+  %1744 = load double, ptr %1743, align 8
+  %1745 = call double @llvm.fmuladd.f64(double 0xBFE4F49E7F775887, double %1744, double %1742)
+  %1746 = getelementptr inbounds %struct.cmplx, ptr %50, i32 0, i32 1
+  %1747 = load double, ptr %1746, align 8
+  %1748 = call double @llvm.fmuladd.f64(double 0xBFEEB42A9BCD5057, double %1747, double %1745)
+  %1749 = getelementptr inbounds %struct.cmplx, ptr %58, i32 0, i32 1
+  store double %1748, ptr %1749, align 8
+  %1750 = load double, ptr %15, align 8
+  %1751 = getelementptr inbounds %struct.cmplx, ptr %55, i32 0, i32 0
+  %1752 = load double, ptr %1751, align 8
+  %1753 = load double, ptr %17, align 8
+  %1754 = getelementptr inbounds %struct.cmplx, ptr %54, i32 0, i32 0
+  %1755 = load double, ptr %1754, align 8
+  %1756 = fmul double %1753, %1755
+  %1757 = call double @llvm.fmuladd.f64(double %1750, double %1752, double %1756)
+  %1758 = load double, ptr %19, align 8
+  %1759 = getelementptr inbounds %struct.cmplx, ptr %53, i32 0, i32 0
+  %1760 = load double, ptr %1759, align 8
+  %1761 = call double @llvm.fmuladd.f64(double %1758, double %1760, double %1757)
+  %1762 = load double, ptr %21, align 8
+  %1763 = getelementptr inbounds %struct.cmplx, ptr %52, i32 0, i32 0
+  %1764 = load double, ptr %1763, align 8
+  %1765 = call double @llvm.fmuladd.f64(double %1762, double %1764, double %1761)
+  %1766 = load double, ptr %23, align 8
+  %1767 = getelementptr inbounds %struct.cmplx, ptr %51, i32 0, i32 0
+  %1768 = load double, ptr %1767, align 8
+  %1769 = call double @llvm.fmuladd.f64(double %1766, double %1768, double %1765)
+  %1770 = getelementptr inbounds %struct.cmplx, ptr %59, i32 0, i32 1
+  store double %1769, ptr %1770, align 8
+  %1771 = load double, ptr %15, align 8
+  %1772 = getelementptr inbounds %struct.cmplx, ptr %55, i32 0, i32 1
+  %1773 = load double, ptr %1772, align 8
+  %1774 = load double, ptr %17, align 8
+  %1775 = getelementptr inbounds %struct.cmplx, ptr %54, i32 0, i32 1
+  %1776 = load double, ptr %1775, align 8
+  %1777 = fmul double %1774, %1776
+  %1778 = call double @llvm.fmuladd.f64(double %1771, double %1773, double %1777)
+  %1779 = load double, ptr %19, align 8
+  %1780 = getelementptr inbounds %struct.cmplx, ptr %53, i32 0, i32 1
+  %1781 = load double, ptr %1780, align 8
+  %1782 = call double @llvm.fmuladd.f64(double %1779, double %1781, double %1778)
+  %1783 = load double, ptr %21, align 8
+  %1784 = getelementptr inbounds %struct.cmplx, ptr %52, i32 0, i32 1
+  %1785 = load double, ptr %1784, align 8
+  %1786 = call double @llvm.fmuladd.f64(double %1783, double %1785, double %1782)
+  %1787 = load double, ptr %23, align 8
+  %1788 = getelementptr inbounds %struct.cmplx, ptr %51, i32 0, i32 1
+  %1789 = load double, ptr %1788, align 8
+  %1790 = call double @llvm.fmuladd.f64(double %1787, double %1789, double %1786)
+  %1791 = fneg double %1790
+  %1792 = getelementptr inbounds %struct.cmplx, ptr %59, i32 0, i32 0
+  store double %1791, ptr %1792, align 8
+  %1793 = getelementptr inbounds %struct.cmplx, ptr %58, i32 0, i32 0
+  %1794 = load double, ptr %1793, align 8
+  %1795 = getelementptr inbounds %struct.cmplx, ptr %59, i32 0, i32 0
+  %1796 = load double, ptr %1795, align 8
+  %1797 = fadd double %1794, %1796
+  %1798 = getelementptr inbounds %struct.cmplx, ptr %56, i32 0, i32 0
+  store double %1797, ptr %1798, align 8
+  %1799 = getelementptr inbounds %struct.cmplx, ptr %58, i32 0, i32 1
+  %1800 = load double, ptr %1799, align 8
+  %1801 = getelementptr inbounds %struct.cmplx, ptr %59, i32 0, i32 1
+  %1802 = load double, ptr %1801, align 8
+  %1803 = fadd double %1800, %1802
+  %1804 = getelementptr inbounds %struct.cmplx, ptr %56, i32 0, i32 1
+  store double %1803, ptr %1804, align 8
+  %1805 = getelementptr inbounds %struct.cmplx, ptr %58, i32 0, i32 0
+  %1806 = load double, ptr %1805, align 8
+  %1807 = getelementptr inbounds %struct.cmplx, ptr %59, i32 0, i32 0
+  %1808 = load double, ptr %1807, align 8
+  %1809 = fsub double %1806, %1808
+  %1810 = getelementptr inbounds %struct.cmplx, ptr %57, i32 0, i32 0
+  store double %1809, ptr %1810, align 8
+  %1811 = getelementptr inbounds %struct.cmplx, ptr %58, i32 0, i32 1
+  %1812 = load double, ptr %1811, align 8
+  %1813 = getelementptr inbounds %struct.cmplx, ptr %59, i32 0, i32 1
+  %1814 = load double, ptr %1813, align 8
+  %1815 = fsub double %1812, %1814
+  %1816 = getelementptr inbounds %struct.cmplx, ptr %57, i32 0, i32 1
+  store double %1815, ptr %1816, align 8
+  %1817 = load ptr, ptr %11, align 4
+  %1818 = load i32, ptr %44, align 4
+  %1819 = sub i32 %1818, 1
+  %1820 = load i32, ptr %7, align 4
+  %1821 = sub i32 %1820, 1
+  %1822 = mul i32 0, %1821
+  %1823 = add i32 %1819, %1822
+  %1824 = getelementptr inbounds %struct.cmplx, ptr %1817, i32 %1823
+  %1825 = getelementptr inbounds %struct.cmplx, ptr %1824, i32 0, i32 0
+  %1826 = load double, ptr %1825, align 8
+  %1827 = getelementptr inbounds %struct.cmplx, ptr %56, i32 0, i32 0
+  %1828 = load double, ptr %1827, align 8
+  %1829 = load i32, ptr %12, align 4
+  %1830 = sitofp i32 %1829 to double
+  %1831 = load ptr, ptr %11, align 4
+  %1832 = load i32, ptr %44, align 4
+  %1833 = sub i32 %1832, 1
+  %1834 = load i32, ptr %7, align 4
+  %1835 = sub i32 %1834, 1
+  %1836 = mul i32 0, %1835
+  %1837 = add i32 %1833, %1836
+  %1838 = getelementptr inbounds %struct.cmplx, ptr %1831, i32 %1837
+  %1839 = getelementptr inbounds %struct.cmplx, ptr %1838, i32 0, i32 1
+  %1840 = load double, ptr %1839, align 8
+  %1841 = fmul double %1830, %1840
+  %1842 = getelementptr inbounds %struct.cmplx, ptr %56, i32 0, i32 1
+  %1843 = load double, ptr %1842, align 8
+  %1844 = fmul double %1841, %1843
+  %1845 = fneg double %1844
+  %1846 = call double @llvm.fmuladd.f64(double %1826, double %1828, double %1845)
+  %1847 = load ptr, ptr %10, align 4
+  %1848 = load i32, ptr %44, align 4
+  %1849 = load i32, ptr %7, align 4
+  %1850 = load i32, ptr %24, align 4
+  %1851 = load i32, ptr %8, align 4
+  %1852 = mul i32 %1851, 1
+  %1853 = add i32 %1850, %1852
+  %1854 = mul i32 %1849, %1853
+  %1855 = add i32 %1848, %1854
+  %1856 = getelementptr inbounds %struct.cmplx, ptr %1847, i32 %1855
+  %1857 = getelementptr inbounds %struct.cmplx, ptr %1856, i32 0, i32 0
+  store double %1846, ptr %1857, align 8
+  %1858 = load ptr, ptr %11, align 4
+  %1859 = load i32, ptr %44, align 4
+  %1860 = sub i32 %1859, 1
+  %1861 = load i32, ptr %7, align 4
+  %1862 = sub i32 %1861, 1
+  %1863 = mul i32 0, %1862
+  %1864 = add i32 %1860, %1863
+  %1865 = getelementptr inbounds %struct.cmplx, ptr %1858, i32 %1864
+  %1866 = getelementptr inbounds %struct.cmplx, ptr %1865, i32 0, i32 0
+  %1867 = load double, ptr %1866, align 8
+  %1868 = getelementptr inbounds %struct.cmplx, ptr %56, i32 0, i32 1
+  %1869 = load double, ptr %1868, align 8
+  %1870 = load i32, ptr %12, align 4
+  %1871 = sitofp i32 %1870 to double
+  %1872 = load ptr, ptr %11, align 4
+  %1873 = load i32, ptr %44, align 4
+  %1874 = sub i32 %1873, 1
+  %1875 = load i32, ptr %7, align 4
+  %1876 = sub i32 %1875, 1
+  %1877 = mul i32 0, %1876
+  %1878 = add i32 %1874, %1877
+  %1879 = getelementptr inbounds %struct.cmplx, ptr %1872, i32 %1878
+  %1880 = getelementptr inbounds %struct.cmplx, ptr %1879, i32 0, i32 1
+  %1881 = load double, ptr %1880, align 8
+  %1882 = fmul double %1871, %1881
+  %1883 = getelementptr inbounds %struct.cmplx, ptr %56, i32 0, i32 0
+  %1884 = load double, ptr %1883, align 8
+  %1885 = fmul double %1882, %1884
+  %1886 = call double @llvm.fmuladd.f64(double %1867, double %1869, double %1885)
+  %1887 = load ptr, ptr %10, align 4
+  %1888 = load i32, ptr %44, align 4
+  %1889 = load i32, ptr %7, align 4
+  %1890 = load i32, ptr %24, align 4
+  %1891 = load i32, ptr %8, align 4
+  %1892 = mul i32 %1891, 1
+  %1893 = add i32 %1890, %1892
+  %1894 = mul i32 %1889, %1893
+  %1895 = add i32 %1888, %1894
+  %1896 = getelementptr inbounds %struct.cmplx, ptr %1887, i32 %1895
+  %1897 = getelementptr inbounds %struct.cmplx, ptr %1896, i32 0, i32 1
+  store double %1886, ptr %1897, align 8
+  %1898 = load ptr, ptr %11, align 4
+  %1899 = load i32, ptr %44, align 4
+  %1900 = sub i32 %1899, 1
+  %1901 = load i32, ptr %7, align 4
+  %1902 = sub i32 %1901, 1
+  %1903 = mul i32 9, %1902
+  %1904 = add i32 %1900, %1903
+  %1905 = getelementptr inbounds %struct.cmplx, ptr %1898, i32 %1904
+  %1906 = getelementptr inbounds %struct.cmplx, ptr %1905, i32 0, i32 0
+  %1907 = load double, ptr %1906, align 8
+  %1908 = getelementptr inbounds %struct.cmplx, ptr %57, i32 0, i32 0
+  %1909 = load double, ptr %1908, align 8
+  %1910 = load i32, ptr %12, align 4
+  %1911 = sitofp i32 %1910 to double
+  %1912 = load ptr, ptr %11, align 4
+  %1913 = load i32, ptr %44, align 4
+  %1914 = sub i32 %1913, 1
+  %1915 = load i32, ptr %7, align 4
+  %1916 = sub i32 %1915, 1
+  %1917 = mul i32 9, %1916
+  %1918 = add i32 %1914, %1917
+  %1919 = getelementptr inbounds %struct.cmplx, ptr %1912, i32 %1918
+  %1920 = getelementptr inbounds %struct.cmplx, ptr %1919, i32 0, i32 1
+  %1921 = load double, ptr %1920, align 8
+  %1922 = fmul double %1911, %1921
+  %1923 = getelementptr inbounds %struct.cmplx, ptr %57, i32 0, i32 1
+  %1924 = load double, ptr %1923, align 8
+  %1925 = fmul double %1922, %1924
+  %1926 = fneg double %1925
+  %1927 = call double @llvm.fmuladd.f64(double %1907, double %1909, double %1926)
+  %1928 = load ptr, ptr %10, align 4
+  %1929 = load i32, ptr %44, align 4
+  %1930 = load i32, ptr %7, align 4
+  %1931 = load i32, ptr %24, align 4
+  %1932 = load i32, ptr %8, align 4
+  %1933 = mul i32 %1932, 10
+  %1934 = add i32 %1931, %1933
+  %1935 = mul i32 %1930, %1934
+  %1936 = add i32 %1929, %1935
+  %1937 = getelementptr inbounds %struct.cmplx, ptr %1928, i32 %1936
+  %1938 = getelementptr inbounds %struct.cmplx, ptr %1937, i32 0, i32 0
+  store double %1927, ptr %1938, align 8
+  %1939 = load ptr, ptr %11, align 4
+  %1940 = load i32, ptr %44, align 4
+  %1941 = sub i32 %1940, 1
+  %1942 = load i32, ptr %7, align 4
+  %1943 = sub i32 %1942, 1
+  %1944 = mul i32 9, %1943
+  %1945 = add i32 %1941, %1944
+  %1946 = getelementptr inbounds %struct.cmplx, ptr %1939, i32 %1945
+  %1947 = getelementptr inbounds %struct.cmplx, ptr %1946, i32 0, i32 0
+  %1948 = load double, ptr %1947, align 8
+  %1949 = getelementptr inbounds %struct.cmplx, ptr %57, i32 0, i32 1
+  %1950 = load double, ptr %1949, align 8
+  %1951 = load i32, ptr %12, align 4
+  %1952 = sitofp i32 %1951 to double
+  %1953 = load ptr, ptr %11, align 4
+  %1954 = load i32, ptr %44, align 4
+  %1955 = sub i32 %1954, 1
+  %1956 = load i32, ptr %7, align 4
+  %1957 = sub i32 %1956, 1
+  %1958 = mul i32 9, %1957
+  %1959 = add i32 %1955, %1958
+  %1960 = getelementptr inbounds %struct.cmplx, ptr %1953, i32 %1959
+  %1961 = getelementptr inbounds %struct.cmplx, ptr %1960, i32 0, i32 1
+  %1962 = load double, ptr %1961, align 8
+  %1963 = fmul double %1952, %1962
+  %1964 = getelementptr inbounds %struct.cmplx, ptr %57, i32 0, i32 0
+  %1965 = load double, ptr %1964, align 8
+  %1966 = fmul double %1963, %1965
+  %1967 = call double @llvm.fmuladd.f64(double %1948, double %1950, double %1966)
+  %1968 = load ptr, ptr %10, align 4
+  %1969 = load i32, ptr %44, align 4
+  %1970 = load i32, ptr %7, align 4
+  %1971 = load i32, ptr %24, align 4
+  %1972 = load i32, ptr %8, align 4
+  %1973 = mul i32 %1972, 10
+  %1974 = add i32 %1971, %1973
+  %1975 = mul i32 %1970, %1974
+  %1976 = add i32 %1969, %1975
+  %1977 = getelementptr inbounds %struct.cmplx, ptr %1968, i32 %1976
+  %1978 = getelementptr inbounds %struct.cmplx, ptr %1977, i32 0, i32 1
+  store double %1967, ptr %1978, align 8
+  %1979 = getelementptr inbounds %struct.cmplx, ptr %45, i32 0, i32 0
+  %1980 = load double, ptr %1979, align 8
+  %1981 = getelementptr inbounds %struct.cmplx, ptr %46, i32 0, i32 0
+  %1982 = load double, ptr %1981, align 8
+  %1983 = call double @llvm.fmuladd.f64(double 0x3FDA9628D9C712B6, double %1982, double %1980)
+  %1984 = getelementptr inbounds %struct.cmplx, ptr %47, i32 0, i32 0
+  %1985 = load double, ptr %1984, align 8
+  %1986 = call double @llvm.fmuladd.f64(double 0xBFE4F49E7F775887, double %1985, double %1983)
+  %1987 = getelementptr inbounds %struct.cmplx, ptr %48, i32 0, i32 0
+  %1988 = load double, ptr %1987, align 8
+  %1989 = call double @llvm.fmuladd.f64(double 0xBFEEB42A9BCD5057, double %1988, double %1986)
+  %1990 = getelementptr inbounds %struct.cmplx, ptr %49, i32 0, i32 0
+  %1991 = load double, ptr %1990, align 8
+  %1992 = call double @llvm.fmuladd.f64(double 0xBFC2375F640F44DB, double %1991, double %1989)
+  %1993 = getelementptr inbounds %struct.cmplx, ptr %50, i32 0, i32 0
+  %1994 = load double, ptr %1993, align 8
+  %1995 = call double @llvm.fmuladd.f64(double 0x3FEAEB8C8764F0BA, double %1994, double %1992)
+  %1996 = getelementptr inbounds %struct.cmplx, ptr %62, i32 0, i32 0
+  store double %1995, ptr %1996, align 8
+  %1997 = getelementptr inbounds %struct.cmplx, ptr %45, i32 0, i32 1
+  %1998 = load double, ptr %1997, align 8
+  %1999 = getelementptr inbounds %struct.cmplx, ptr %46, i32 0, i32 1
+  %2000 = load double, ptr %1999, align 8
+  %2001 = call double @llvm.fmuladd.f64(double 0x3FDA9628D9C712B6, double %2000, double %1998)
+  %2002 = getelementptr inbounds %struct.cmplx, ptr %47, i32 0, i32 1
+  %2003 = load double, ptr %2002, align 8
+  %2004 = call double @llvm.fmuladd.f64(double 0xBFE4F49E7F775887, double %2003, double %2001)
+  %2005 = getelementptr inbounds %struct.cmplx, ptr %48, i32 0, i32 1
+  %2006 = load double, ptr %2005, align 8
+  %2007 = call double @llvm.fmuladd.f64(double 0xBFEEB42A9BCD5057, double %2006, double %2004)
+  %2008 = getelementptr inbounds %struct.cmplx, ptr %49, i32 0, i32 1
+  %2009 = load double, ptr %2008, align 8
+  %2010 = call double @llvm.fmuladd.f64(double 0xBFC2375F640F44DB, double %2009, double %2007)
+  %2011 = getelementptr inbounds %struct.cmplx, ptr %50, i32 0, i32 1
+  %2012 = load double, ptr %2011, align 8
+  %2013 = call double @llvm.fmuladd.f64(double 0x3FEAEB8C8764F0BA, double %2012, double %2010)
+  %2014 = getelementptr inbounds %struct.cmplx, ptr %62, i32 0, i32 1
+  store double %2013, ptr %2014, align 8
+  %2015 = load double, ptr %17, align 8
+  %2016 = getelementptr inbounds %struct.cmplx, ptr %55, i32 0, i32 0
+  %2017 = load double, ptr %2016, align 8
+  %2018 = load double, ptr %21, align 8
+  %2019 = getelementptr inbounds %struct.cmplx, ptr %54, i32 0, i32 0
+  %2020 = load double, ptr %2019, align 8
+  %2021 = fmul double %2018, %2020
+  %2022 = call double @llvm.fmuladd.f64(double %2015, double %2017, double %2021)
+  %2023 = load double, ptr %23, align 8
+  %2024 = getelementptr inbounds %struct.cmplx, ptr %53, i32 0, i32 0
+  %2025 = load double, ptr %2024, align 8
+  %2026 = fneg double %2023
+  %2027 = call double @llvm.fmuladd.f64(double %2026, double %2025, double %2022)
+  %2028 = load double, ptr %19, align 8
+  %2029 = getelementptr inbounds %struct.cmplx, ptr %52, i32 0, i32 0
+  %2030 = load double, ptr %2029, align 8
+  %2031 = fneg double %2028
+  %2032 = call double @llvm.fmuladd.f64(double %2031, double %2030, double %2027)
+  %2033 = load double, ptr %15, align 8
+  %2034 = getelementptr inbounds %struct.cmplx, ptr %51, i32 0, i32 0
+  %2035 = load double, ptr %2034, align 8
+  %2036 = fneg double %2033
+  %2037 = call double @llvm.fmuladd.f64(double %2036, double %2035, double %2032)
+  %2038 = getelementptr inbounds %struct.cmplx, ptr %63, i32 0, i32 1
+  store double %2037, ptr %2038, align 8
+  %2039 = load double, ptr %17, align 8
+  %2040 = getelementptr inbounds %struct.cmplx, ptr %55, i32 0, i32 1
+  %2041 = load double, ptr %2040, align 8
+  %2042 = load double, ptr %21, align 8
+  %2043 = getelementptr inbounds %struct.cmplx, ptr %54, i32 0, i32 1
+  %2044 = load double, ptr %2043, align 8
+  %2045 = fmul double %2042, %2044
+  %2046 = call double @llvm.fmuladd.f64(double %2039, double %2041, double %2045)
+  %2047 = load double, ptr %23, align 8
+  %2048 = getelementptr inbounds %struct.cmplx, ptr %53, i32 0, i32 1
+  %2049 = load double, ptr %2048, align 8
+  %2050 = fneg double %2047
+  %2051 = call double @llvm.fmuladd.f64(double %2050, double %2049, double %2046)
+  %2052 = load double, ptr %19, align 8
+  %2053 = getelementptr inbounds %struct.cmplx, ptr %52, i32 0, i32 1
+  %2054 = load double, ptr %2053, align 8
+  %2055 = fneg double %2052
+  %2056 = call double @llvm.fmuladd.f64(double %2055, double %2054, double %2051)
+  %2057 = load double, ptr %15, align 8
+  %2058 = getelementptr inbounds %struct.cmplx, ptr %51, i32 0, i32 1
+  %2059 = load double, ptr %2058, align 8
+  %2060 = fneg double %2057
+  %2061 = call double @llvm.fmuladd.f64(double %2060, double %2059, double %2056)
+  %2062 = fneg double %2061
+  %2063 = getelementptr inbounds %struct.cmplx, ptr %63, i32 0, i32 0
+  store double %2062, ptr %2063, align 8
+  %2064 = getelementptr inbounds %struct.cmplx, ptr %62, i32 0, i32 0
+  %2065 = load double, ptr %2064, align 8
+  %2066 = getelementptr inbounds %struct.cmplx, ptr %63, i32 0, i32 0
+  %2067 = load double, ptr %2066, align 8
+  %2068 = fadd double %2065, %2067
+  %2069 = getelementptr inbounds %struct.cmplx, ptr %60, i32 0, i32 0
+  store double %2068, ptr %2069, align 8
+  %2070 = getelementptr inbounds %struct.cmplx, ptr %62, i32 0, i32 1
+  %2071 = load double, ptr %2070, align 8
+  %2072 = getelementptr inbounds %struct.cmplx, ptr %63, i32 0, i32 1
+  %2073 = load double, ptr %2072, align 8
+  %2074 = fadd double %2071, %2073
+  %2075 = getelementptr inbounds %struct.cmplx, ptr %60, i32 0, i32 1
+  store double %2074, ptr %2075, align 8
+  %2076 = getelementptr inbounds %struct.cmplx, ptr %62, i32 0, i32 0
+  %2077 = load double, ptr %2076, align 8
+  %2078 = getelementptr inbounds %struct.cmplx, ptr %63, i32 0, i32 0
+  %2079 = load double, ptr %2078, align 8
+  %2080 = fsub double %2077, %2079
+  %2081 = getelementptr inbounds %struct.cmplx, ptr %61, i32 0, i32 0
+  store double %2080, ptr %2081, align 8
+  %2082 = getelementptr inbounds %struct.cmplx, ptr %62, i32 0, i32 1
+  %2083 = load double, ptr %2082, align 8
+  %2084 = getelementptr inbounds %struct.cmplx, ptr %63, i32 0, i32 1
+  %2085 = load double, ptr %2084, align 8
+  %2086 = fsub double %2083, %2085
+  %2087 = getelementptr inbounds %struct.cmplx, ptr %61, i32 0, i32 1
+  store double %2086, ptr %2087, align 8
+  %2088 = load ptr, ptr %11, align 4
+  %2089 = load i32, ptr %44, align 4
+  %2090 = sub i32 %2089, 1
+  %2091 = load i32, ptr %7, align 4
+  %2092 = sub i32 %2091, 1
+  %2093 = mul i32 1, %2092
+  %2094 = add i32 %2090, %2093
+  %2095 = getelementptr inbounds %struct.cmplx, ptr %2088, i32 %2094
+  %2096 = getelementptr inbounds %struct.cmplx, ptr %2095, i32 0, i32 0
+  %2097 = load double, ptr %2096, align 8
+  %2098 = getelementptr inbounds %struct.cmplx, ptr %60, i32 0, i32 0
+  %2099 = load double, ptr %2098, align 8
+  %2100 = load i32, ptr %12, align 4
+  %2101 = sitofp i32 %2100 to double
+  %2102 = load ptr, ptr %11, align 4
+  %2103 = load i32, ptr %44, align 4
+  %2104 = sub i32 %2103, 1
+  %2105 = load i32, ptr %7, align 4
+  %2106 = sub i32 %2105, 1
+  %2107 = mul i32 1, %2106
+  %2108 = add i32 %2104, %2107
+  %2109 = getelementptr inbounds %struct.cmplx, ptr %2102, i32 %2108
+  %2110 = getelementptr inbounds %struct.cmplx, ptr %2109, i32 0, i32 1
+  %2111 = load double, ptr %2110, align 8
+  %2112 = fmul double %2101, %2111
+  %2113 = getelementptr inbounds %struct.cmplx, ptr %60, i32 0, i32 1
+  %2114 = load double, ptr %2113, align 8
+  %2115 = fmul double %2112, %2114
+  %2116 = fneg double %2115
+  %2117 = call double @llvm.fmuladd.f64(double %2097, double %2099, double %2116)
+  %2118 = load ptr, ptr %10, align 4
+  %2119 = load i32, ptr %44, align 4
+  %2120 = load i32, ptr %7, align 4
+  %2121 = load i32, ptr %24, align 4
+  %2122 = load i32, ptr %8, align 4
+  %2123 = mul i32 %2122, 2
+  %2124 = add i32 %2121, %2123
+  %2125 = mul i32 %2120, %2124
+  %2126 = add i32 %2119, %2125
+  %2127 = getelementptr inbounds %struct.cmplx, ptr %2118, i32 %2126
+  %2128 = getelementptr inbounds %struct.cmplx, ptr %2127, i32 0, i32 0
+  store double %2117, ptr %2128, align 8
+  %2129 = load ptr, ptr %11, align 4
+  %2130 = load i32, ptr %44, align 4
+  %2131 = sub i32 %2130, 1
+  %2132 = load i32, ptr %7, align 4
+  %2133 = sub i32 %2132, 1
+  %2134 = mul i32 1, %2133
+  %2135 = add i32 %2131, %2134
+  %2136 = getelementptr inbounds %struct.cmplx, ptr %2129, i32 %2135
+  %2137 = getelementptr inbounds %struct.cmplx, ptr %2136, i32 0, i32 0
+  %2138 = load double, ptr %2137, align 8
+  %2139 = getelementptr inbounds %struct.cmplx, ptr %60, i32 0, i32 1
+  %2140 = load double, ptr %2139, align 8
+  %2141 = load i32, ptr %12, align 4
+  %2142 = sitofp i32 %2141 to double
+  %2143 = load ptr, ptr %11, align 4
+  %2144 = load i32, ptr %44, align 4
+  %2145 = sub i32 %2144, 1
+  %2146 = load i32, ptr %7, align 4
+  %2147 = sub i32 %2146, 1
+  %2148 = mul i32 1, %2147
+  %2149 = add i32 %2145, %2148
+  %2150 = getelementptr inbounds %struct.cmplx, ptr %2143, i32 %2149
+  %2151 = getelementptr inbounds %struct.cmplx, ptr %2150, i32 0, i32 1
+  %2152 = load double, ptr %2151, align 8
+  %2153 = fmul double %2142, %2152
+  %2154 = getelementptr inbounds %struct.cmplx, ptr %60, i32 0, i32 0
+  %2155 = load double, ptr %2154, align 8
+  %2156 = fmul double %2153, %2155
+  %2157 = call double @llvm.fmuladd.f64(double %2138, double %2140, double %2156)
+  %2158 = load ptr, ptr %10, align 4
+  %2159 = load i32, ptr %44, align 4
+  %2160 = load i32, ptr %7, align 4
+  %2161 = load i32, ptr %24, align 4
+  %2162 = load i32, ptr %8, align 4
+  %2163 = mul i32 %2162, 2
+  %2164 = add i32 %2161, %2163
+  %2165 = mul i32 %2160, %2164
+  %2166 = add i32 %2159, %2165
+  %2167 = getelementptr inbounds %struct.cmplx, ptr %2158, i32 %2166
+  %2168 = getelementptr inbounds %struct.cmplx, ptr %2167, i32 0, i32 1
+  store double %2157, ptr %2168, align 8
+  %2169 = load ptr, ptr %11, align 4
+  %2170 = load i32, ptr %44, align 4
+  %2171 = sub i32 %2170, 1
+  %2172 = load i32, ptr %7, align 4
+  %2173 = sub i32 %2172, 1
+  %2174 = mul i32 8, %2173
+  %2175 = add i32 %2171, %2174
+  %2176 = getelementptr inbounds %struct.cmplx, ptr %2169, i32 %2175
+  %2177 = getelementptr inbounds %struct.cmplx, ptr %2176, i32 0, i32 0
+  %2178 = load double, ptr %2177, align 8
+  %2179 = getelementptr inbounds %struct.cmplx, ptr %61, i32 0, i32 0
+  %2180 = load double, ptr %2179, align 8
+  %2181 = load i32, ptr %12, align 4
+  %2182 = sitofp i32 %2181 to double
+  %2183 = load ptr, ptr %11, align 4
+  %2184 = load i32, ptr %44, align 4
+  %2185 = sub i32 %2184, 1
+  %2186 = load i32, ptr %7, align 4
+  %2187 = sub i32 %2186, 1
+  %2188 = mul i32 8, %2187
+  %2189 = add i32 %2185, %2188
+  %2190 = getelementptr inbounds %struct.cmplx, ptr %2183, i32 %2189
+  %2191 = getelementptr inbounds %struct.cmplx, ptr %2190, i32 0, i32 1
+  %2192 = load double, ptr %2191, align 8
+  %2193 = fmul double %2182, %2192
+  %2194 = getelementptr inbounds %struct.cmplx, ptr %61, i32 0, i32 1
+  %2195 = load double, ptr %2194, align 8
+  %2196 = fmul double %2193, %2195
+  %2197 = fneg double %2196
+  %2198 = call double @llvm.fmuladd.f64(double %2178, double %2180, double %2197)
+  %2199 = load ptr, ptr %10, align 4
+  %2200 = load i32, ptr %44, align 4
+  %2201 = load i32, ptr %7, align 4
+  %2202 = load i32, ptr %24, align 4
+  %2203 = load i32, ptr %8, align 4
+  %2204 = mul i32 %2203, 9
+  %2205 = add i32 %2202, %2204
+  %2206 = mul i32 %2201, %2205
+  %2207 = add i32 %2200, %2206
+  %2208 = getelementptr inbounds %struct.cmplx, ptr %2199, i32 %2207
+  %2209 = getelementptr inbounds %struct.cmplx, ptr %2208, i32 0, i32 0
+  store double %2198, ptr %2209, align 8
+  %2210 = load ptr, ptr %11, align 4
+  %2211 = load i32, ptr %44, align 4
+  %2212 = sub i32 %2211, 1
+  %2213 = load i32, ptr %7, align 4
+  %2214 = sub i32 %2213, 1
+  %2215 = mul i32 8, %2214
+  %2216 = add i32 %2212, %2215
+  %2217 = getelementptr inbounds %struct.cmplx, ptr %2210, i32 %2216
+  %2218 = getelementptr inbounds %struct.cmplx, ptr %2217, i32 0, i32 0
+  %2219 = load double, ptr %2218, align 8
+  %2220 = getelementptr inbounds %struct.cmplx, ptr %61, i32 0, i32 1
+  %2221 = load double, ptr %2220, align 8
+  %2222 = load i32, ptr %12, align 4
+  %2223 = sitofp i32 %2222 to double
+  %2224 = load ptr, ptr %11, align 4
+  %2225 = load i32, ptr %44, align 4
+  %2226 = sub i32 %2225, 1
+  %2227 = load i32, ptr %7, align 4
+  %2228 = sub i32 %2227, 1
+  %2229 = mul i32 8, %2228
+  %2230 = add i32 %2226, %2229
+  %2231 = getelementptr inbounds %struct.cmplx, ptr %2224, i32 %2230
+  %2232 = getelementptr inbounds %struct.cmplx, ptr %2231, i32 0, i32 1
+  %2233 = load double, ptr %2232, align 8
+  %2234 = fmul double %2223, %2233
+  %2235 = getelementptr inbounds %struct.cmplx, ptr %61, i32 0, i32 0
+  %2236 = load double, ptr %2235, align 8
+  %2237 = fmul double %2234, %2236
+  %2238 = call double @llvm.fmuladd.f64(double %2219, double %2221, double %2237)
+  %2239 = load ptr, ptr %10, align 4
+  %2240 = load i32, ptr %44, align 4
+  %2241 = load i32, ptr %7, align 4
+  %2242 = load i32, ptr %24, align 4
+  %2243 = load i32, ptr %8, align 4
+  %2244 = mul i32 %2243, 9
+  %2245 = add i32 %2242, %2244
+  %2246 = mul i32 %2241, %2245
+  %2247 = add i32 %2240, %2246
+  %2248 = getelementptr inbounds %struct.cmplx, ptr %2239, i32 %2247
+  %2249 = getelementptr inbounds %struct.cmplx, ptr %2248, i32 0, i32 1
+  store double %2238, ptr %2249, align 8
+  %2250 = getelementptr inbounds %struct.cmplx, ptr %45, i32 0, i32 0
+  %2251 = load double, ptr %2250, align 8
+  %2252 = getelementptr inbounds %struct.cmplx, ptr %46, i32 0, i32 0
+  %2253 = load double, ptr %2252, align 8
+  %2254 = call double @llvm.fmuladd.f64(double 0xBFC2375F640F44DB, double %2253, double %2251)
+  %2255 = getelementptr inbounds %struct.cmplx, ptr %47, i32 0, i32 0
+  %2256 = load double, ptr %2255, align 8
+  %2257 = call double @llvm.fmuladd.f64(double 0xBFEEB42A9BCD5057, double %2256, double %2254)
+  %2258 = getelementptr inbounds %struct.cmplx, ptr %48, i32 0, i32 0
+  %2259 = load double, ptr %2258, align 8
+  %2260 = call double @llvm.fmuladd.f64(double 0x3FDA9628D9C712B6, double %2259, double %2257)
+  %2261 = getelementptr inbounds %struct.cmplx, ptr %49, i32 0, i32 0
+  %2262 = load double, ptr %2261, align 8
+  %2263 = call double @llvm.fmuladd.f64(double 0x3FEAEB8C8764F0BA, double %2262, double %2260)
+  %2264 = getelementptr inbounds %struct.cmplx, ptr %50, i32 0, i32 0
+  %2265 = load double, ptr %2264, align 8
+  %2266 = call double @llvm.fmuladd.f64(double 0xBFE4F49E7F775887, double %2265, double %2263)
+  %2267 = getelementptr inbounds %struct.cmplx, ptr %66, i32 0, i32 0
+  store double %2266, ptr %2267, align 8
+  %2268 = getelementptr inbounds %struct.cmplx, ptr %45, i32 0, i32 1
+  %2269 = load double, ptr %2268, align 8
+  %2270 = getelementptr inbounds %struct.cmplx, ptr %46, i32 0, i32 1
+  %2271 = load double, ptr %2270, align 8
+  %2272 = call double @llvm.fmuladd.f64(double 0xBFC2375F640F44DB, double %2271, double %2269)
+  %2273 = getelementptr inbounds %struct.cmplx, ptr %47, i32 0, i32 1
+  %2274 = load double, ptr %2273, align 8
+  %2275 = call double @llvm.fmuladd.f64(double 0xBFEEB42A9BCD5057, double %2274, double %2272)
+  %2276 = getelementptr inbounds %struct.cmplx, ptr %48, i32 0, i32 1
+  %2277 = load double, ptr %2276, align 8
+  %2278 = call double @llvm.fmuladd.f64(double 0x3FDA9628D9C712B6, double %2277, double %2275)
+  %2279 = getelementptr inbounds %struct.cmplx, ptr %49, i32 0, i32 1
+  %2280 = load double, ptr %2279, align 8
+  %2281 = call double @llvm.fmuladd.f64(double 0x3FEAEB8C8764F0BA, double %2280, double %2278)
+  %2282 = getelementptr inbounds %struct.cmplx, ptr %50, i32 0, i32 1
+  %2283 = load double, ptr %2282, align 8
+  %2284 = call double @llvm.fmuladd.f64(double 0xBFE4F49E7F775887, double %2283, double %2281)
+  %2285 = getelementptr inbounds %struct.cmplx, ptr %66, i32 0, i32 1
+  store double %2284, ptr %2285, align 8
+  %2286 = load double, ptr %19, align 8
+  %2287 = getelementptr inbounds %struct.cmplx, ptr %55, i32 0, i32 0
+  %2288 = load double, ptr %2287, align 8
+  %2289 = load double, ptr %23, align 8
+  %2290 = getelementptr inbounds %struct.cmplx, ptr %54, i32 0, i32 0
+  %2291 = load double, ptr %2290, align 8
+  %2292 = fmul double %2289, %2291
+  %2293 = fneg double %2292
+  %2294 = call double @llvm.fmuladd.f64(double %2286, double %2288, double %2293)
+  %2295 = load double, ptr %17, align 8
+  %2296 = getelementptr inbounds %struct.cmplx, ptr %53, i32 0, i32 0
+  %2297 = load double, ptr %2296, align 8
+  %2298 = fneg double %2295
+  %2299 = call double @llvm.fmuladd.f64(double %2298, double %2297, double %2294)
+  %2300 = load double, ptr %15, align 8
+  %2301 = getelementptr inbounds %struct.cmplx, ptr %52, i32 0, i32 0
+  %2302 = load double, ptr %2301, align 8
+  %2303 = call double @llvm.fmuladd.f64(double %2300, double %2302, double %2299)
+  %2304 = load double, ptr %21, align 8
+  %2305 = getelementptr inbounds %struct.cmplx, ptr %51, i32 0, i32 0
+  %2306 = load double, ptr %2305, align 8
+  %2307 = call double @llvm.fmuladd.f64(double %2304, double %2306, double %2303)
+  %2308 = getelementptr inbounds %struct.cmplx, ptr %67, i32 0, i32 1
+  store double %2307, ptr %2308, align 8
+  %2309 = load double, ptr %19, align 8
+  %2310 = getelementptr inbounds %struct.cmplx, ptr %55, i32 0, i32 1
+  %2311 = load double, ptr %2310, align 8
+  %2312 = load double, ptr %23, align 8
+  %2313 = getelementptr inbounds %struct.cmplx, ptr %54, i32 0, i32 1
+  %2314 = load double, ptr %2313, align 8
+  %2315 = fmul double %2312, %2314
+  %2316 = fneg double %2315
+  %2317 = call double @llvm.fmuladd.f64(double %2309, double %2311, double %2316)
+  %2318 = load double, ptr %17, align 8
+  %2319 = getelementptr inbounds %struct.cmplx, ptr %53, i32 0, i32 1
+  %2320 = load double, ptr %2319, align 8
+  %2321 = fneg double %2318
+  %2322 = call double @llvm.fmuladd.f64(double %2321, double %2320, double %2317)
+  %2323 = load double, ptr %15, align 8
+  %2324 = getelementptr inbounds %struct.cmplx, ptr %52, i32 0, i32 1
+  %2325 = load double, ptr %2324, align 8
+  %2326 = call double @llvm.fmuladd.f64(double %2323, double %2325, double %2322)
+  %2327 = load double, ptr %21, align 8
+  %2328 = getelementptr inbounds %struct.cmplx, ptr %51, i32 0, i32 1
+  %2329 = load double, ptr %2328, align 8
+  %2330 = call double @llvm.fmuladd.f64(double %2327, double %2329, double %2326)
+  %2331 = fneg double %2330
+  %2332 = getelementptr inbounds %struct.cmplx, ptr %67, i32 0, i32 0
+  store double %2331, ptr %2332, align 8
+  %2333 = getelementptr inbounds %struct.cmplx, ptr %66, i32 0, i32 0
+  %2334 = load double, ptr %2333, align 8
+  %2335 = getelementptr inbounds %struct.cmplx, ptr %67, i32 0, i32 0
+  %2336 = load double, ptr %2335, align 8
+  %2337 = fadd double %2334, %2336
+  %2338 = getelementptr inbounds %struct.cmplx, ptr %64, i32 0, i32 0
+  store double %2337, ptr %2338, align 8
+  %2339 = getelementptr inbounds %struct.cmplx, ptr %66, i32 0, i32 1
+  %2340 = load double, ptr %2339, align 8
+  %2341 = getelementptr inbounds %struct.cmplx, ptr %67, i32 0, i32 1
+  %2342 = load double, ptr %2341, align 8
+  %2343 = fadd double %2340, %2342
+  %2344 = getelementptr inbounds %struct.cmplx, ptr %64, i32 0, i32 1
+  store double %2343, ptr %2344, align 8
+  %2345 = getelementptr inbounds %struct.cmplx, ptr %66, i32 0, i32 0
+  %2346 = load double, ptr %2345, align 8
+  %2347 = getelementptr inbounds %struct.cmplx, ptr %67, i32 0, i32 0
+  %2348 = load double, ptr %2347, align 8
+  %2349 = fsub double %2346, %2348
+  %2350 = getelementptr inbounds %struct.cmplx, ptr %65, i32 0, i32 0
+  store double %2349, ptr %2350, align 8
+  %2351 = getelementptr inbounds %struct.cmplx, ptr %66, i32 0, i32 1
+  %2352 = load double, ptr %2351, align 8
+  %2353 = getelementptr inbounds %struct.cmplx, ptr %67, i32 0, i32 1
+  %2354 = load double, ptr %2353, align 8
+  %2355 = fsub double %2352, %2354
+  %2356 = getelementptr inbounds %struct.cmplx, ptr %65, i32 0, i32 1
+  store double %2355, ptr %2356, align 8
+  %2357 = load ptr, ptr %11, align 4
+  %2358 = load i32, ptr %44, align 4
+  %2359 = sub i32 %2358, 1
+  %2360 = load i32, ptr %7, align 4
+  %2361 = sub i32 %2360, 1
+  %2362 = mul i32 2, %2361
+  %2363 = add i32 %2359, %2362
+  %2364 = getelementptr inbounds %struct.cmplx, ptr %2357, i32 %2363
+  %2365 = getelementptr inbounds %struct.cmplx, ptr %2364, i32 0, i32 0
+  %2366 = load double, ptr %2365, align 8
+  %2367 = getelementptr inbounds %struct.cmplx, ptr %64, i32 0, i32 0
+  %2368 = load double, ptr %2367, align 8
+  %2369 = load i32, ptr %12, align 4
+  %2370 = sitofp i32 %2369 to double
+  %2371 = load ptr, ptr %11, align 4
+  %2372 = load i32, ptr %44, align 4
+  %2373 = sub i32 %2372, 1
+  %2374 = load i32, ptr %7, align 4
+  %2375 = sub i32 %2374, 1
+  %2376 = mul i32 2, %2375
+  %2377 = add i32 %2373, %2376
+  %2378 = getelementptr inbounds %struct.cmplx, ptr %2371, i32 %2377
+  %2379 = getelementptr inbounds %struct.cmplx, ptr %2378, i32 0, i32 1
+  %2380 = load double, ptr %2379, align 8
+  %2381 = fmul double %2370, %2380
+  %2382 = getelementptr inbounds %struct.cmplx, ptr %64, i32 0, i32 1
+  %2383 = load double, ptr %2382, align 8
+  %2384 = fmul double %2381, %2383
+  %2385 = fneg double %2384
+  %2386 = call double @llvm.fmuladd.f64(double %2366, double %2368, double %2385)
+  %2387 = load ptr, ptr %10, align 4
+  %2388 = load i32, ptr %44, align 4
+  %2389 = load i32, ptr %7, align 4
+  %2390 = load i32, ptr %24, align 4
+  %2391 = load i32, ptr %8, align 4
+  %2392 = mul i32 %2391, 3
+  %2393 = add i32 %2390, %2392
+  %2394 = mul i32 %2389, %2393
+  %2395 = add i32 %2388, %2394
+  %2396 = getelementptr inbounds %struct.cmplx, ptr %2387, i32 %2395
+  %2397 = getelementptr inbounds %struct.cmplx, ptr %2396, i32 0, i32 0
+  store double %2386, ptr %2397, align 8
+  %2398 = load ptr, ptr %11, align 4
+  %2399 = load i32, ptr %44, align 4
+  %2400 = sub i32 %2399, 1
+  %2401 = load i32, ptr %7, align 4
+  %2402 = sub i32 %2401, 1
+  %2403 = mul i32 2, %2402
+  %2404 = add i32 %2400, %2403
+  %2405 = getelementptr inbounds %struct.cmplx, ptr %2398, i32 %2404
+  %2406 = getelementptr inbounds %struct.cmplx, ptr %2405, i32 0, i32 0
+  %2407 = load double, ptr %2406, align 8
+  %2408 = getelementptr inbounds %struct.cmplx, ptr %64, i32 0, i32 1
+  %2409 = load double, ptr %2408, align 8
+  %2410 = load i32, ptr %12, align 4
+  %2411 = sitofp i32 %2410 to double
+  %2412 = load ptr, ptr %11, align 4
+  %2413 = load i32, ptr %44, align 4
+  %2414 = sub i32 %2413, 1
+  %2415 = load i32, ptr %7, align 4
+  %2416 = sub i32 %2415, 1
+  %2417 = mul i32 2, %2416
+  %2418 = add i32 %2414, %2417
+  %2419 = getelementptr inbounds %struct.cmplx, ptr %2412, i32 %2418
+  %2420 = getelementptr inbounds %struct.cmplx, ptr %2419, i32 0, i32 1
+  %2421 = load double, ptr %2420, align 8
+  %2422 = fmul double %2411, %2421
+  %2423 = getelementptr inbounds %struct.cmplx, ptr %64, i32 0, i32 0
+  %2424 = load double, ptr %2423, align 8
+  %2425 = fmul double %2422, %2424
+  %2426 = call double @llvm.fmuladd.f64(double %2407, double %2409, double %2425)
+  %2427 = load ptr, ptr %10, align 4
+  %2428 = load i32, ptr %44, align 4
+  %2429 = load i32, ptr %7, align 4
+  %2430 = load i32, ptr %24, align 4
+  %2431 = load i32, ptr %8, align 4
+  %2432 = mul i32 %2431, 3
+  %2433 = add i32 %2430, %2432
+  %2434 = mul i32 %2429, %2433
+  %2435 = add i32 %2428, %2434
+  %2436 = getelementptr inbounds %struct.cmplx, ptr %2427, i32 %2435
+  %2437 = getelementptr inbounds %struct.cmplx, ptr %2436, i32 0, i32 1
+  store double %2426, ptr %2437, align 8
+  %2438 = load ptr, ptr %11, align 4
+  %2439 = load i32, ptr %44, align 4
+  %2440 = sub i32 %2439, 1
+  %2441 = load i32, ptr %7, align 4
+  %2442 = sub i32 %2441, 1
+  %2443 = mul i32 7, %2442
+  %2444 = add i32 %2440, %2443
+  %2445 = getelementptr inbounds %struct.cmplx, ptr %2438, i32 %2444
+  %2446 = getelementptr inbounds %struct.cmplx, ptr %2445, i32 0, i32 0
+  %2447 = load double, ptr %2446, align 8
+  %2448 = getelementptr inbounds %struct.cmplx, ptr %65, i32 0, i32 0
+  %2449 = load double, ptr %2448, align 8
+  %2450 = load i32, ptr %12, align 4
+  %2451 = sitofp i32 %2450 to double
+  %2452 = load ptr, ptr %11, align 4
+  %2453 = load i32, ptr %44, align 4
+  %2454 = sub i32 %2453, 1
+  %2455 = load i32, ptr %7, align 4
+  %2456 = sub i32 %2455, 1
+  %2457 = mul i32 7, %2456
+  %2458 = add i32 %2454, %2457
+  %2459 = getelementptr inbounds %struct.cmplx, ptr %2452, i32 %2458
+  %2460 = getelementptr inbounds %struct.cmplx, ptr %2459, i32 0, i32 1
+  %2461 = load double, ptr %2460, align 8
+  %2462 = fmul double %2451, %2461
+  %2463 = getelementptr inbounds %struct.cmplx, ptr %65, i32 0, i32 1
+  %2464 = load double, ptr %2463, align 8
+  %2465 = fmul double %2462, %2464
+  %2466 = fneg double %2465
+  %2467 = call double @llvm.fmuladd.f64(double %2447, double %2449, double %2466)
+  %2468 = load ptr, ptr %10, align 4
+  %2469 = load i32, ptr %44, align 4
+  %2470 = load i32, ptr %7, align 4
+  %2471 = load i32, ptr %24, align 4
+  %2472 = load i32, ptr %8, align 4
+  %2473 = mul i32 %2472, 8
+  %2474 = add i32 %2471, %2473
+  %2475 = mul i32 %2470, %2474
+  %2476 = add i32 %2469, %2475
+  %2477 = getelementptr inbounds %struct.cmplx, ptr %2468, i32 %2476
+  %2478 = getelementptr inbounds %struct.cmplx, ptr %2477, i32 0, i32 0
+  store double %2467, ptr %2478, align 8
+  %2479 = load ptr, ptr %11, align 4
+  %2480 = load i32, ptr %44, align 4
+  %2481 = sub i32 %2480, 1
+  %2482 = load i32, ptr %7, align 4
+  %2483 = sub i32 %2482, 1
+  %2484 = mul i32 7, %2483
+  %2485 = add i32 %2481, %2484
+  %2486 = getelementptr inbounds %struct.cmplx, ptr %2479, i32 %2485
+  %2487 = getelementptr inbounds %struct.cmplx, ptr %2486, i32 0, i32 0
+  %2488 = load double, ptr %2487, align 8
+  %2489 = getelementptr inbounds %struct.cmplx, ptr %65, i32 0, i32 1
+  %2490 = load double, ptr %2489, align 8
+  %2491 = load i32, ptr %12, align 4
+  %2492 = sitofp i32 %2491 to double
+  %2493 = load ptr, ptr %11, align 4
+  %2494 = load i32, ptr %44, align 4
+  %2495 = sub i32 %2494, 1
+  %2496 = load i32, ptr %7, align 4
+  %2497 = sub i32 %2496, 1
+  %2498 = mul i32 7, %2497
+  %2499 = add i32 %2495, %2498
+  %2500 = getelementptr inbounds %struct.cmplx, ptr %2493, i32 %2499
+  %2501 = getelementptr inbounds %struct.cmplx, ptr %2500, i32 0, i32 1
+  %2502 = load double, ptr %2501, align 8
+  %2503 = fmul double %2492, %2502
+  %2504 = getelementptr inbounds %struct.cmplx, ptr %65, i32 0, i32 0
+  %2505 = load double, ptr %2504, align 8
+  %2506 = fmul double %2503, %2505
+  %2507 = call double @llvm.fmuladd.f64(double %2488, double %2490, double %2506)
+  %2508 = load ptr, ptr %10, align 4
+  %2509 = load i32, ptr %44, align 4
+  %2510 = load i32, ptr %7, align 4
+  %2511 = load i32, ptr %24, align 4
+  %2512 = load i32, ptr %8, align 4
+  %2513 = mul i32 %2512, 8
+  %2514 = add i32 %2511, %2513
+  %2515 = mul i32 %2510, %2514
+  %2516 = add i32 %2509, %2515
+  %2517 = getelementptr inbounds %struct.cmplx, ptr %2508, i32 %2516
+  %2518 = getelementptr inbounds %struct.cmplx, ptr %2517, i32 0, i32 1
+  store double %2507, ptr %2518, align 8
+  %2519 = getelementptr inbounds %struct.cmplx, ptr %45, i32 0, i32 0
+  %2520 = load double, ptr %2519, align 8
+  %2521 = getelementptr inbounds %struct.cmplx, ptr %46, i32 0, i32 0
+  %2522 = load double, ptr %2521, align 8
+  %2523 = call double @llvm.fmuladd.f64(double 0xBFE4F49E7F775887, double %2522, double %2520)
+  %2524 = getelementptr inbounds %struct.cmplx, ptr %47, i32 0, i32 0
+  %2525 = load double, ptr %2524, align 8
+  %2526 = call double @llvm.fmuladd.f64(double 0xBFC2375F640F44DB, double %2525, double %2523)
+  %2527 = getelementptr inbounds %struct.cmplx, ptr %48, i32 0, i32 0
+  %2528 = load double, ptr %2527, align 8
+  %2529 = call double @llvm.fmuladd.f64(double 0x3FEAEB8C8764F0BA, double %2528, double %2526)
+  %2530 = getelementptr inbounds %struct.cmplx, ptr %49, i32 0, i32 0
+  %2531 = load double, ptr %2530, align 8
+  %2532 = call double @llvm.fmuladd.f64(double 0xBFEEB42A9BCD5057, double %2531, double %2529)
+  %2533 = getelementptr inbounds %struct.cmplx, ptr %50, i32 0, i32 0
+  %2534 = load double, ptr %2533, align 8
+  %2535 = call double @llvm.fmuladd.f64(double 0x3FDA9628D9C712B6, double %2534, double %2532)
+  %2536 = getelementptr inbounds %struct.cmplx, ptr %70, i32 0, i32 0
+  store double %2535, ptr %2536, align 8
+  %2537 = getelementptr inbounds %struct.cmplx, ptr %45, i32 0, i32 1
+  %2538 = load double, ptr %2537, align 8
+  %2539 = getelementptr inbounds %struct.cmplx, ptr %46, i32 0, i32 1
+  %2540 = load double, ptr %2539, align 8
+  %2541 = call double @llvm.fmuladd.f64(double 0xBFE4F49E7F775887, double %2540, double %2538)
+  %2542 = getelementptr inbounds %struct.cmplx, ptr %47, i32 0, i32 1
+  %2543 = load double, ptr %2542, align 8
+  %2544 = call double @llvm.fmuladd.f64(double 0xBFC2375F640F44DB, double %2543, double %2541)
+  %2545 = getelementptr inbounds %struct.cmplx, ptr %48, i32 0, i32 1
+  %2546 = load double, ptr %2545, align 8
+  %2547 = call double @llvm.fmuladd.f64(double 0x3FEAEB8C8764F0BA, double %2546, double %2544)
+  %2548 = getelementptr inbounds %struct.cmplx, ptr %49, i32 0, i32 1
+  %2549 = load double, ptr %2548, align 8
+  %2550 = call double @llvm.fmuladd.f64(double 0xBFEEB42A9BCD5057, double %2549, double %2547)
+  %2551 = getelementptr inbounds %struct.cmplx, ptr %50, i32 0, i32 1
+  %2552 = load double, ptr %2551, align 8
+  %2553 = call double @llvm.fmuladd.f64(double 0x3FDA9628D9C712B6, double %2552, double %2550)
+  %2554 = getelementptr inbounds %struct.cmplx, ptr %70, i32 0, i32 1
+  store double %2553, ptr %2554, align 8
+  %2555 = load double, ptr %21, align 8
+  %2556 = getelementptr inbounds %struct.cmplx, ptr %55, i32 0, i32 0
+  %2557 = load double, ptr %2556, align 8
+  %2558 = load double, ptr %19, align 8
+  %2559 = getelementptr inbounds %struct.cmplx, ptr %54, i32 0, i32 0
+  %2560 = load double, ptr %2559, align 8
+  %2561 = fmul double %2558, %2560
+  %2562 = fneg double %2561
+  %2563 = call double @llvm.fmuladd.f64(double %2555, double %2557, double %2562)
+  %2564 = load double, ptr %15, align 8
+  %2565 = getelementptr inbounds %struct.cmplx, ptr %53, i32 0, i32 0
+  %2566 = load double, ptr %2565, align 8
+  %2567 = call double @llvm.fmuladd.f64(double %2564, double %2566, double %2563)
+  %2568 = load double, ptr %23, align 8
+  %2569 = getelementptr inbounds %struct.cmplx, ptr %52, i32 0, i32 0
+  %2570 = load double, ptr %2569, align 8
+  %2571 = call double @llvm.fmuladd.f64(double %2568, double %2570, double %2567)
+  %2572 = load double, ptr %17, align 8
+  %2573 = getelementptr inbounds %struct.cmplx, ptr %51, i32 0, i32 0
+  %2574 = load double, ptr %2573, align 8
+  %2575 = fneg double %2572
+  %2576 = call double @llvm.fmuladd.f64(double %2575, double %2574, double %2571)
+  %2577 = getelementptr inbounds %struct.cmplx, ptr %71, i32 0, i32 1
+  store double %2576, ptr %2577, align 8
+  %2578 = load double, ptr %21, align 8
+  %2579 = getelementptr inbounds %struct.cmplx, ptr %55, i32 0, i32 1
+  %2580 = load double, ptr %2579, align 8
+  %2581 = load double, ptr %19, align 8
+  %2582 = getelementptr inbounds %struct.cmplx, ptr %54, i32 0, i32 1
+  %2583 = load double, ptr %2582, align 8
+  %2584 = fmul double %2581, %2583
+  %2585 = fneg double %2584
+  %2586 = call double @llvm.fmuladd.f64(double %2578, double %2580, double %2585)
+  %2587 = load double, ptr %15, align 8
+  %2588 = getelementptr inbounds %struct.cmplx, ptr %53, i32 0, i32 1
+  %2589 = load double, ptr %2588, align 8
+  %2590 = call double @llvm.fmuladd.f64(double %2587, double %2589, double %2586)
+  %2591 = load double, ptr %23, align 8
+  %2592 = getelementptr inbounds %struct.cmplx, ptr %52, i32 0, i32 1
+  %2593 = load double, ptr %2592, align 8
+  %2594 = call double @llvm.fmuladd.f64(double %2591, double %2593, double %2590)
+  %2595 = load double, ptr %17, align 8
+  %2596 = getelementptr inbounds %struct.cmplx, ptr %51, i32 0, i32 1
+  %2597 = load double, ptr %2596, align 8
+  %2598 = fneg double %2595
+  %2599 = call double @llvm.fmuladd.f64(double %2598, double %2597, double %2594)
+  %2600 = fneg double %2599
+  %2601 = getelementptr inbounds %struct.cmplx, ptr %71, i32 0, i32 0
+  store double %2600, ptr %2601, align 8
+  %2602 = getelementptr inbounds %struct.cmplx, ptr %70, i32 0, i32 0
+  %2603 = load double, ptr %2602, align 8
+  %2604 = getelementptr inbounds %struct.cmplx, ptr %71, i32 0, i32 0
+  %2605 = load double, ptr %2604, align 8
+  %2606 = fadd double %2603, %2605
+  %2607 = getelementptr inbounds %struct.cmplx, ptr %68, i32 0, i32 0
+  store double %2606, ptr %2607, align 8
+  %2608 = getelementptr inbounds %struct.cmplx, ptr %70, i32 0, i32 1
+  %2609 = load double, ptr %2608, align 8
+  %2610 = getelementptr inbounds %struct.cmplx, ptr %71, i32 0, i32 1
+  %2611 = load double, ptr %2610, align 8
+  %2612 = fadd double %2609, %2611
+  %2613 = getelementptr inbounds %struct.cmplx, ptr %68, i32 0, i32 1
+  store double %2612, ptr %2613, align 8
+  %2614 = getelementptr inbounds %struct.cmplx, ptr %70, i32 0, i32 0
+  %2615 = load double, ptr %2614, align 8
+  %2616 = getelementptr inbounds %struct.cmplx, ptr %71, i32 0, i32 0
+  %2617 = load double, ptr %2616, align 8
+  %2618 = fsub double %2615, %2617
+  %2619 = getelementptr inbounds %struct.cmplx, ptr %69, i32 0, i32 0
+  store double %2618, ptr %2619, align 8
+  %2620 = getelementptr inbounds %struct.cmplx, ptr %70, i32 0, i32 1
+  %2621 = load double, ptr %2620, align 8
+  %2622 = getelementptr inbounds %struct.cmplx, ptr %71, i32 0, i32 1
+  %2623 = load double, ptr %2622, align 8
+  %2624 = fsub double %2621, %2623
+  %2625 = getelementptr inbounds %struct.cmplx, ptr %69, i32 0, i32 1
+  store double %2624, ptr %2625, align 8
+  %2626 = load ptr, ptr %11, align 4
+  %2627 = load i32, ptr %44, align 4
+  %2628 = sub i32 %2627, 1
+  %2629 = load i32, ptr %7, align 4
+  %2630 = sub i32 %2629, 1
+  %2631 = mul i32 3, %2630
+  %2632 = add i32 %2628, %2631
+  %2633 = getelementptr inbounds %struct.cmplx, ptr %2626, i32 %2632
+  %2634 = getelementptr inbounds %struct.cmplx, ptr %2633, i32 0, i32 0
+  %2635 = load double, ptr %2634, align 8
+  %2636 = getelementptr inbounds %struct.cmplx, ptr %68, i32 0, i32 0
+  %2637 = load double, ptr %2636, align 8
+  %2638 = load i32, ptr %12, align 4
+  %2639 = sitofp i32 %2638 to double
+  %2640 = load ptr, ptr %11, align 4
+  %2641 = load i32, ptr %44, align 4
+  %2642 = sub i32 %2641, 1
+  %2643 = load i32, ptr %7, align 4
+  %2644 = sub i32 %2643, 1
+  %2645 = mul i32 3, %2644
+  %2646 = add i32 %2642, %2645
+  %2647 = getelementptr inbounds %struct.cmplx, ptr %2640, i32 %2646
+  %2648 = getelementptr inbounds %struct.cmplx, ptr %2647, i32 0, i32 1
+  %2649 = load double, ptr %2648, align 8
+  %2650 = fmul double %2639, %2649
+  %2651 = getelementptr inbounds %struct.cmplx, ptr %68, i32 0, i32 1
+  %2652 = load double, ptr %2651, align 8
+  %2653 = fmul double %2650, %2652
+  %2654 = fneg double %2653
+  %2655 = call double @llvm.fmuladd.f64(double %2635, double %2637, double %2654)
+  %2656 = load ptr, ptr %10, align 4
+  %2657 = load i32, ptr %44, align 4
+  %2658 = load i32, ptr %7, align 4
+  %2659 = load i32, ptr %24, align 4
+  %2660 = load i32, ptr %8, align 4
+  %2661 = mul i32 %2660, 4
+  %2662 = add i32 %2659, %2661
+  %2663 = mul i32 %2658, %2662
+  %2664 = add i32 %2657, %2663
+  %2665 = getelementptr inbounds %struct.cmplx, ptr %2656, i32 %2664
+  %2666 = getelementptr inbounds %struct.cmplx, ptr %2665, i32 0, i32 0
+  store double %2655, ptr %2666, align 8
+  %2667 = load ptr, ptr %11, align 4
+  %2668 = load i32, ptr %44, align 4
+  %2669 = sub i32 %2668, 1
+  %2670 = load i32, ptr %7, align 4
+  %2671 = sub i32 %2670, 1
+  %2672 = mul i32 3, %2671
+  %2673 = add i32 %2669, %2672
+  %2674 = getelementptr inbounds %struct.cmplx, ptr %2667, i32 %2673
+  %2675 = getelementptr inbounds %struct.cmplx, ptr %2674, i32 0, i32 0
+  %2676 = load double, ptr %2675, align 8
+  %2677 = getelementptr inbounds %struct.cmplx, ptr %68, i32 0, i32 1
+  %2678 = load double, ptr %2677, align 8
+  %2679 = load i32, ptr %12, align 4
+  %2680 = sitofp i32 %2679 to double
+  %2681 = load ptr, ptr %11, align 4
+  %2682 = load i32, ptr %44, align 4
+  %2683 = sub i32 %2682, 1
+  %2684 = load i32, ptr %7, align 4
+  %2685 = sub i32 %2684, 1
+  %2686 = mul i32 3, %2685
+  %2687 = add i32 %2683, %2686
+  %2688 = getelementptr inbounds %struct.cmplx, ptr %2681, i32 %2687
+  %2689 = getelementptr inbounds %struct.cmplx, ptr %2688, i32 0, i32 1
+  %2690 = load double, ptr %2689, align 8
+  %2691 = fmul double %2680, %2690
+  %2692 = getelementptr inbounds %struct.cmplx, ptr %68, i32 0, i32 0
+  %2693 = load double, ptr %2692, align 8
+  %2694 = fmul double %2691, %2693
+  %2695 = call double @llvm.fmuladd.f64(double %2676, double %2678, double %2694)
+  %2696 = load ptr, ptr %10, align 4
+  %2697 = load i32, ptr %44, align 4
+  %2698 = load i32, ptr %7, align 4
+  %2699 = load i32, ptr %24, align 4
+  %2700 = load i32, ptr %8, align 4
+  %2701 = mul i32 %2700, 4
+  %2702 = add i32 %2699, %2701
+  %2703 = mul i32 %2698, %2702
+  %2704 = add i32 %2697, %2703
+  %2705 = getelementptr inbounds %struct.cmplx, ptr %2696, i32 %2704
+  %2706 = getelementptr inbounds %struct.cmplx, ptr %2705, i32 0, i32 1
+  store double %2695, ptr %2706, align 8
+  %2707 = load ptr, ptr %11, align 4
+  %2708 = load i32, ptr %44, align 4
+  %2709 = sub i32 %2708, 1
+  %2710 = load i32, ptr %7, align 4
+  %2711 = sub i32 %2710, 1
+  %2712 = mul i32 6, %2711
+  %2713 = add i32 %2709, %2712
+  %2714 = getelementptr inbounds %struct.cmplx, ptr %2707, i32 %2713
+  %2715 = getelementptr inbounds %struct.cmplx, ptr %2714, i32 0, i32 0
+  %2716 = load double, ptr %2715, align 8
+  %2717 = getelementptr inbounds %struct.cmplx, ptr %69, i32 0, i32 0
+  %2718 = load double, ptr %2717, align 8
+  %2719 = load i32, ptr %12, align 4
+  %2720 = sitofp i32 %2719 to double
+  %2721 = load ptr, ptr %11, align 4
+  %2722 = load i32, ptr %44, align 4
+  %2723 = sub i32 %2722, 1
+  %2724 = load i32, ptr %7, align 4
+  %2725 = sub i32 %2724, 1
+  %2726 = mul i32 6, %2725
+  %2727 = add i32 %2723, %2726
+  %2728 = getelementptr inbounds %struct.cmplx, ptr %2721, i32 %2727
+  %2729 = getelementptr inbounds %struct.cmplx, ptr %2728, i32 0, i32 1
+  %2730 = load double, ptr %2729, align 8
+  %2731 = fmul double %2720, %2730
+  %2732 = getelementptr inbounds %struct.cmplx, ptr %69, i32 0, i32 1
+  %2733 = load double, ptr %2732, align 8
+  %2734 = fmul double %2731, %2733
+  %2735 = fneg double %2734
+  %2736 = call double @llvm.fmuladd.f64(double %2716, double %2718, double %2735)
+  %2737 = load ptr, ptr %10, align 4
+  %2738 = load i32, ptr %44, align 4
+  %2739 = load i32, ptr %7, align 4
+  %2740 = load i32, ptr %24, align 4
+  %2741 = load i32, ptr %8, align 4
+  %2742 = mul i32 %2741, 7
+  %2743 = add i32 %2740, %2742
+  %2744 = mul i32 %2739, %2743
+  %2745 = add i32 %2738, %2744
+  %2746 = getelementptr inbounds %struct.cmplx, ptr %2737, i32 %2745
+  %2747 = getelementptr inbounds %struct.cmplx, ptr %2746, i32 0, i32 0
+  store double %2736, ptr %2747, align 8
+  %2748 = load ptr, ptr %11, align 4
+  %2749 = load i32, ptr %44, align 4
+  %2750 = sub i32 %2749, 1
+  %2751 = load i32, ptr %7, align 4
+  %2752 = sub i32 %2751, 1
+  %2753 = mul i32 6, %2752
+  %2754 = add i32 %2750, %2753
+  %2755 = getelementptr inbounds %struct.cmplx, ptr %2748, i32 %2754
+  %2756 = getelementptr inbounds %struct.cmplx, ptr %2755, i32 0, i32 0
+  %2757 = load double, ptr %2756, align 8
+  %2758 = getelementptr inbounds %struct.cmplx, ptr %69, i32 0, i32 1
+  %2759 = load double, ptr %2758, align 8
+  %2760 = load i32, ptr %12, align 4
+  %2761 = sitofp i32 %2760 to double
+  %2762 = load ptr, ptr %11, align 4
+  %2763 = load i32, ptr %44, align 4
+  %2764 = sub i32 %2763, 1
+  %2765 = load i32, ptr %7, align 4
+  %2766 = sub i32 %2765, 1
+  %2767 = mul i32 6, %2766
+  %2768 = add i32 %2764, %2767
+  %2769 = getelementptr inbounds %struct.cmplx, ptr %2762, i32 %2768
+  %2770 = getelementptr inbounds %struct.cmplx, ptr %2769, i32 0, i32 1
+  %2771 = load double, ptr %2770, align 8
+  %2772 = fmul double %2761, %2771
+  %2773 = getelementptr inbounds %struct.cmplx, ptr %69, i32 0, i32 0
+  %2774 = load double, ptr %2773, align 8
+  %2775 = fmul double %2772, %2774
+  %2776 = call double @llvm.fmuladd.f64(double %2757, double %2759, double %2775)
+  %2777 = load ptr, ptr %10, align 4
+  %2778 = load i32, ptr %44, align 4
+  %2779 = load i32, ptr %7, align 4
+  %2780 = load i32, ptr %24, align 4
+  %2781 = load i32, ptr %8, align 4
+  %2782 = mul i32 %2781, 7
+  %2783 = add i32 %2780, %2782
+  %2784 = mul i32 %2779, %2783
+  %2785 = add i32 %2778, %2784
+  %2786 = getelementptr inbounds %struct.cmplx, ptr %2777, i32 %2785
+  %2787 = getelementptr inbounds %struct.cmplx, ptr %2786, i32 0, i32 1
+  store double %2776, ptr %2787, align 8
+  br label %2788
+
+2788:                                             ; preds = %1168
+  %2789 = load i32, ptr %44, align 4
+  %2790 = add i32 %2789, 1
+  store i32 %2790, ptr %44, align 4
+  br label %1164, !llvm.loop !4
+
+2791:                                             ; preds = %1164
+  br label %2792
+
+2792:                                             ; preds = %2791
+  %2793 = load i32, ptr %24, align 4
+  %2794 = add i32 %2793, 1
+  store i32 %2794, ptr %24, align 4
+  br label %87, !llvm.loop !6
+
+2795:                                             ; preds = %87
+  ret void
+}
+
+; Function Attrs: nocallback nofree nounwind willreturn memory(argmem: readwrite)
+declare void @llvm.memcpy.p0.p0.i32(ptr noalias nocapture writeonly, ptr noalias nocapture readonly, i32, i1 immarg) #2
+
+; Function Attrs: nocallback nofree nosync nounwind speculatable willreturn memory(none)
+declare double @llvm.fmuladd.f64(double, double, double) #3
+
+attributes #0 = { noinline nounwind optnone uwtable "frame-pointer"="all" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="e500" "target-features"="+spe,-altivec,-bpermd,-crbits,-crypto,-direct-move,-extdiv,-htm,-isa-v206-instructions,-isa-v207-instructions,-isa-v30-instructions,-power8-vector,-power9-vector,-privileged,-quadword-atomics,-rop-protect,-vsx" }
+attributes #1 = { "frame-pointer"="all" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="e500" "target-features"="+spe,-altivec,-bpermd,-crbits,-crypto,-direct-move,-extdiv,-htm,-isa-v206-instructions,-isa-v207-instructions,-isa-v30-instructions,-power8-vector,-power9-vector,-privileged,-quadword-atomics,-rop-protect,-vsx" }
+attributes #2 = { nocallback nofree nounwind willreturn memory(argmem: readwrite) }
+attributes #3 = { nocallback nofree nosync nounwind speculatable willreturn memory(none) }
+
+!llvm.module.flags = !{!0, !1, !2}
+!llvm.ident = !{!3}
+
+!0 = !{i32 1, !"wchar_size", i32 4}
+!1 = !{i32 7, !"uwtable", i32 2}
+!2 = !{i32 7, !"frame-pointer", i32 2}
+!3 = !{!"clang version 17.0.0 (https://github.com/llvm/llvm-project.git 69db592f762ade86508826a7b3c9d5434c4837e2)"}
+!4 = distinct !{!4, !5}
+!5 = !{!"llvm.loop.mustprogress"}
+!6 = distinct !{!6, !5}
+;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
+; SPE: {{.*}}

llvm/test/CodeGen/PowerPC/spe.ll

 ; SPE-NEXT:    lwz 0, 292(1)
 ; SPE-NEXT:    addi 1, 1, 288
 ; SPE-NEXT:    mtlr 0
 ; SPE-NEXT:    blr
 ;
 ; EFPU2-LABEL: test_spill:
 ; EFPU2:       # %bb.0: # %entry
 ; EFPU2-NEXT:    mflr 0
-; EFPU2-NEXT:    stwu 1, -144(1)
+; EFPU2-NEXT:    stwu 1, -160(1)
 ; EFPU2-NEXT:    mr 5, 3
 ; EFPU2-NEXT:    mr 6, 4
-; EFPU2-NEXT:    stw 0, 148(1)
-; EFPU2-NEXT:    evstdd 27, 104(1) # 8-byte Folded Spill
-; EFPU2-NEXT:    evstdd 28, 112(1) # 8-byte Folded Spill
-; EFPU2-NEXT:    evstdd 29, 120(1) # 8-byte Folded Spill
-; EFPU2-NEXT:    evstdd 30, 128(1) # 8-byte Folded Spill
-; EFPU2-NEXT:    lwz 28, 152(1)
+; EFPU2-NEXT:    stw 0, 164(1)
+; EFPU2-NEXT:    evstdd 27, 120(1) # 8-byte Folded Spill
+; EFPU2-NEXT:    evstdd 28, 128(1) # 8-byte Folded Spill
+; EFPU2-NEXT:    evstdd 29, 136(1) # 8-byte Folded Spill
+; EFPU2-NEXT:    evstdd 30, 144(1) # 8-byte Folded Spill
+; EFPU2-NEXT:    lwz 28, 168(1)
 ; EFPU2-NEXT:    bl __adddf3
 ; EFPU2-NEXT:    lis 5, 16393
 ; EFPU2-NEXT:    lis 6, -4069
 ; EFPU2-NEXT:    ori 5, 5, 8697
 ; EFPU2-NEXT:    ori 6, 6, 34414
 ; EFPU2-NEXT:    #APP
 ; EFPU2-NEXT:    #NO_APP
 ; EFPU2-NEXT:    bl __adddf3
 ; EFPU2-NEXT:    mr 30, 3
 ; EFPU2-NEXT:    mr 29, 4
-; EFPU2-NEXT:    addi 3, 1, 52
+; EFPU2-NEXT:    addi 3, 1, 68
 ; EFPU2-NEXT:    li 4, 0
 ; EFPU2-NEXT:    li 5, 24
 ; EFPU2-NEXT:    li 6, 1
 ; EFPU2-NEXT:    li 27, 0
 ; EFPU2-NEXT:    bl test_memset
 ; EFPU2-NEXT:    stw 27, 0(28)
 ; EFPU2-NEXT:    bl test_func2
-; EFPU2-NEXT:    addi 3, 1, 8
+; EFPU2-NEXT:    addi 3, 1, 24
 ; EFPU2-NEXT:    li 4, 0
 ; EFPU2-NEXT:    li 5, 20
 ; EFPU2-NEXT:    li 6, 1
 ; EFPU2-NEXT:    bl test_memset
 ; EFPU2-NEXT:    mr 3, 30
 ; EFPU2-NEXT:    mr 4, 29
-; EFPU2-NEXT:    evldd 30, 128(1) # 8-byte Folded Reload
-; EFPU2-NEXT:    evldd 29, 120(1) # 8-byte Folded Reload
-; EFPU2-NEXT:    evldd 28, 112(1) # 8-byte Folded Reload
-; EFPU2-NEXT:    evldd 27, 104(1) # 8-byte Folded Reload
-; EFPU2-NEXT:    lwz 0, 148(1)
-; EFPU2-NEXT:    addi 1, 1, 144
+; EFPU2-NEXT:    evldd 30, 144(1) # 8-byte Folded Reload
+; EFPU2-NEXT:    evldd 29, 136(1) # 8-byte Folded Reload
+; EFPU2-NEXT:    evldd 28, 128(1) # 8-byte Folded Reload
+; EFPU2-NEXT:    evldd 27, 120(1) # 8-byte Folded Reload
+; EFPU2-NEXT:    lwz 0, 164(1)
+; EFPU2-NEXT:    addi 1, 1, 160
 ; EFPU2-NEXT:    mtlr 0
 ; EFPU2-NEXT:    blr
 entry:
   %v1 = alloca [13 x i32], align 4
   %v2 = alloca [11 x i32], align 4
   %0 = fadd double %a, %a
   call void asm sideeffect "","~{s0},~{s3},~{s4},~{s5},~{s6},~{s7},~{s8},~{s9},~{s10},~{s11},~{s12},~{s13},~{s14},~{s15},~{s16},~{s17},~{s18},~{s19},~{s20},~{s21},~{s22},~{s23},~{s24},~{s25},~{s26},~{s27},~{s28},~{s29},~{s30},~{s31}"() nounwind
   %1 = fadd double %0, 3.14159