[PowerPC] Add peephole to remove redundant accumulator prime/unprime instructions

[bsaleil]

Summary

In some situations, the compiler may insert an accumulator prime instruction and an accumulator unprime instruction with no use of that accumulator between the two.
That's for example the case when we store an accumulator after assembling it or restoring it. This patch adds a peephole to remove these prime and unprime instructions.

[NeHuang]

llvm/lib/Target/PowerPC/PPCPreEmitPeephole.cpp

373

Question: Besides adding the operand register to the candidate, are we also recording the machine instruction used the register candidate? If I understand it properly, can we also add it in the comment? This would help understand your comment below we mark the two instructions for removal. - line 380

376

nit: do you think we need to add instruction OPC information in the message to differentiate them?

379

nit: operand -> operand register

[bsaleil]

Update comments

[NeHuang]

LGTM.

llvm/lib/Target/PowerPC/PPCPreEmitPeephole.cpp

bool addLinkerOpt(MachineBasicBlock &MBB, const TargetRegisterInfo *TRI) {
             MachineOperand::CreateMCSymbol(Symbol, PPCII::MO_PCREL_OPT_FLAG);
         Pair->DefInst->addOperand(*MF, PCRelLabel);
         Pair->UseInst->addOperand(*MF, PCRelLabel);
         MadeChange |= true;
       }
       return MadeChange;
     }
 
+    // This function removes redundant pairs of accumulator prime/unprime
+    // instructions. In some situations, it's possible the compiler inserts an
+    // accumulator prime instruction followed by an unprime instruction (e.g.
+    // when we store an accumulator after restoring it from a spill). If the
+    // accumulator is not used between the two, they can be removed. This
+    // function removes these redundant pairs from basic blocks.
+    // The algorithm is quite straightforward - every time we encounter a prime
+    // instruction, the primed register is added to a candidate set. Any use
+    // other than a prime removes the candidate from the set and any de-prime
+    // of a current candidate marks both the prime and de-prime for removal.
+    // This way we ensure we only remove prime/de-prime *pairs* with no
+    // intervening uses.
+    bool removeAccPrimeUnprime(MachineBasicBlock &MBB) {
+      DenseSet<MachineInstr *> InstrsToErase;
+      // Initially, none of the acc registers are candidates.
+      SmallVector<MachineInstr *, 8> Candidates(
+          PPC::UACCRCRegClass.getNumRegs(), nullptr);
+
+      for (MachineInstr &BBI : MBB.instrs()) {
+        unsigned Opc = BBI.getOpcode();
+        // If we are visiting a xxmtacc instruction, we add its operand register
+        // to the candidate set.
+        if (Opc == PPC::XXMTACC) {
+          Register Acc = BBI.getOperand(0).getReg();
+          assert(PPC::ACCRCRegClass.contains(Acc) && "Unexpected register");
+          Candidates[Acc - PPC::ACC0] = &BBI;
+        }
+        // If we are visiting a xxmfacc instruction and its operand is in the
+        // candidate set, we mark the two instructions for removal.
+        else if (Opc == PPC::XXMFACC) {
+          Register Acc = BBI.getOperand(0).getReg();
+          assert(PPC::ACCRCRegClass.contains(Acc) && "Unexpected register");
+          if (!Candidates[Acc - PPC::ACC0])
+            continue;
+          InstrsToErase.insert(&BBI);
+          InstrsToErase.insert(Candidates[Acc - PPC::ACC0]);
+        }
+        // If we are visiting an instruction using an accumulator register
+        // as operand, we remove it from the candidate set.
+        else {
+          for (MachineOperand &Operand : BBI.operands()) {
+            if (!Operand.isReg())
+              continue;
+            Register Reg = Operand.getReg();
+            if (PPC::ACCRCRegClass.contains(Reg))
+              Candidates[Reg - PPC::ACC0] = nullptr;
+          }
+        }
+      }
+
+      for (MachineInstr *MI : InstrsToErase)
+        MI->eraseFromParent();
+      NumRemovedInPreEmit += InstrsToErase.size();
+      return !InstrsToErase.empty();
+    }
+
     bool runOnMachineFunction(MachineFunction &MF) override {
       if (skipFunction(MF.getFunction()) || !RunPreEmitPeephole) {
         // Remove UNENCODED_NOP even when this pass is disabled.
         // This needs to be done unconditionally so we don't emit zeros
         // in the instruction stream.
         SmallVector<MachineInstr *, 4> InstrsToErase;
         for (MachineBasicBlock &MBB : MF)
           for (MachineInstr &MI : MBB)
             if (MI.getOpcode() == PPC::UNENCODED_NOP)
               InstrsToErase.push_back(&MI);
         for (MachineInstr *MI : InstrsToErase)
           MI->eraseFromParent();
         return false;
       }
       bool Changed = false;
       const PPCInstrInfo *TII = MF.getSubtarget<PPCSubtarget>().getInstrInfo();
       const TargetRegisterInfo *TRI = MF.getSubtarget().getRegisterInfo();
       SmallVector<MachineInstr *, 4> InstrsToErase;
       for (MachineBasicBlock &MBB : MF) {
         Changed |= removeRedundantLIs(MBB, TRI);
         Changed |= addLinkerOpt(MBB, TRI);
+        Changed |= removeAccPrimeUnprime(MBB);
         for (MachineInstr &MI : MBB) {
           unsigned Opc = MI.getOpcode();
           if (Opc == PPC::UNENCODED_NOP) {
             InstrsToErase.push_back(&MI);
             continue;
           }
           // Detect self copies - these can result from running AADB.
           if (PPCInstrInfo::isSameClassPhysRegCopy(Opc)) {

llvm/test/CodeGen/PowerPC/mma-acc-memops.ll

 
 define void @testLdSt(i64 %SrcIdx, i64 %DstIdx) {
 ; LE-PAIRED-LABEL: testLdSt:
 ; LE-PAIRED:       # %bb.0: # %entry
 ; LE-PAIRED-NEXT:    plxv vs1, f@PCREL+96(0), 1
 ; LE-PAIRED-NEXT:    plxv vs0, f@PCREL+112(0), 1
 ; LE-PAIRED-NEXT:    plxv vs3, f@PCREL+64(0), 1
 ; LE-PAIRED-NEXT:    plxv vs2, f@PCREL+80(0), 1
-; LE-PAIRED-NEXT:    xxmtacc acc0
-; LE-PAIRED-NEXT:    xxmfacc acc0
 ; LE-PAIRED-NEXT:    pstxv vs0, f@PCREL+176(0), 1
 ; LE-PAIRED-NEXT:    pstxv vs1, f@PCREL+160(0), 1
 ; LE-PAIRED-NEXT:    pstxv vs2, f@PCREL+144(0), 1
 ; LE-PAIRED-NEXT:    pstxv vs3, f@PCREL+128(0), 1
 ; LE-PAIRED-NEXT:    blr
 ;
 ; BE-PAIRED-LABEL: testLdSt:
 ; BE-PAIRED:       # %bb.0: # %entry
 ; BE-PAIRED-NEXT:    addis r3, r2, .LC0@toc@ha
 ; BE-PAIRED-NEXT:    ld r3, .LC0@toc@l(r3)
 ; BE-PAIRED-NEXT:    lxv vs1, 80(r3)
 ; BE-PAIRED-NEXT:    lxv vs0, 64(r3)
 ; BE-PAIRED-NEXT:    lxv vs3, 112(r3)
 ; BE-PAIRED-NEXT:    lxv vs2, 96(r3)
-; BE-PAIRED-NEXT:    xxmtacc acc0
-; BE-PAIRED-NEXT:    xxmfacc acc0
 ; BE-PAIRED-NEXT:    stxv vs1, 144(r3)
 ; BE-PAIRED-NEXT:    stxv vs0, 128(r3)
 ; BE-PAIRED-NEXT:    stxv vs3, 176(r3)
 ; BE-PAIRED-NEXT:    stxv vs2, 160(r3)
 ; BE-PAIRED-NEXT:    blr
 entry:
   %arrayidx = getelementptr inbounds <512 x i1>, <512 x i1>* @f, i64 1
   %0 = load <512 x i1>, <512 x i1>* %arrayidx, align 64
   %arrayidx1 = getelementptr inbounds <512 x i1>, <512 x i1>* @f, i64 2
   store <512 x i1> %0, <512 x i1>* %arrayidx1, align 64
   ret void
 }
 
 define void @testXLdSt(i64 %SrcIdx, i64 %DstIdx) {
 ; LE-PAIRED-LABEL: testXLdSt:
 ; LE-PAIRED:       # %bb.0: # %entry
 ; LE-PAIRED-NEXT:    sldi r3, r3, 6
 ; LE-PAIRED-NEXT:    paddi r5, 0, f@PCREL, 1
 ; LE-PAIRED-NEXT:    add r6, r5, r3
 ; LE-PAIRED-NEXT:    lxv vs1, 32(r6)
 ; LE-PAIRED-NEXT:    lxv vs0, 48(r6)
 ; LE-PAIRED-NEXT:    lxvx vs3, r5, r3
 ; LE-PAIRED-NEXT:    lxv vs2, 16(r6)
 ; LE-PAIRED-NEXT:    sldi r3, r4, 6
-; LE-PAIRED-NEXT:    xxmtacc acc0
-; LE-PAIRED-NEXT:    xxmfacc acc0
 ; LE-PAIRED-NEXT:    stxvx vs3, r5, r3
 ; LE-PAIRED-NEXT:    add r3, r5, r3
 ; LE-PAIRED-NEXT:    stxv vs0, 48(r3)
 ; LE-PAIRED-NEXT:    stxv vs1, 32(r3)
 ; LE-PAIRED-NEXT:    stxv vs2, 16(r3)
 ; LE-PAIRED-NEXT:    blr
 ;
 ; BE-PAIRED-LABEL: testXLdSt:
 ; BE-PAIRED:       # %bb.0: # %entry
 ; BE-PAIRED-NEXT:    addis r5, r2, .LC0@toc@ha
 ; BE-PAIRED-NEXT:    sldi r3, r3, 6
 ; BE-PAIRED-NEXT:    ld r5, .LC0@toc@l(r5)
 ; BE-PAIRED-NEXT:    add r6, r5, r3
 ; BE-PAIRED-NEXT:    lxvx vs0, r5, r3
 ; BE-PAIRED-NEXT:    sldi r3, r4, 6
 ; BE-PAIRED-NEXT:    lxv vs1, 16(r6)
 ; BE-PAIRED-NEXT:    lxv vs3, 48(r6)
 ; BE-PAIRED-NEXT:    lxv vs2, 32(r6)
-; BE-PAIRED-NEXT:    xxmtacc acc0
-; BE-PAIRED-NEXT:    xxmfacc acc0
 ; BE-PAIRED-NEXT:    stxvx vs0, r5, r3
 ; BE-PAIRED-NEXT:    add r3, r5, r3
 ; BE-PAIRED-NEXT:    stxv vs1, 16(r3)
 ; BE-PAIRED-NEXT:    stxv vs3, 48(r3)
 ; BE-PAIRED-NEXT:    stxv vs2, 32(r3)
 ; BE-PAIRED-NEXT:    blr
 entry:
   %arrayidx = getelementptr inbounds <512 x i1>, <512 x i1>* @f, i64 %SrcIdx
   %0 = load <512 x i1>, <512 x i1>* %arrayidx, align 64
   %arrayidx1 = getelementptr inbounds <512 x i1>, <512 x i1>* @f, i64 %DstIdx
   store <512 x i1> %0, <512 x i1>* %arrayidx1, align 64
   ret void
 }
 
 define void @testUnalignedLdSt() {
 ; LE-PAIRED-LABEL: testUnalignedLdSt:
 ; LE-PAIRED:       # %bb.0: # %entry
 ; LE-PAIRED-NEXT:    plxv vs1, f@PCREL+43(0), 1
 ; LE-PAIRED-NEXT:    plxv vs0, f@PCREL+59(0), 1
 ; LE-PAIRED-NEXT:    plxv vs3, f@PCREL+11(0), 1
 ; LE-PAIRED-NEXT:    plxv vs2, f@PCREL+27(0), 1
-; LE-PAIRED-NEXT:    xxmtacc acc0
-; LE-PAIRED-NEXT:    xxmfacc acc0
 ; LE-PAIRED-NEXT:    pstxv vs0, f@PCREL+67(0), 1
 ; LE-PAIRED-NEXT:    pstxv vs1, f@PCREL+51(0), 1
 ; LE-PAIRED-NEXT:    pstxv vs2, f@PCREL+35(0), 1
 ; LE-PAIRED-NEXT:    pstxv vs3, f@PCREL+19(0), 1
 ; LE-PAIRED-NEXT:    blr
 ;
 ; BE-PAIRED-LABEL: testUnalignedLdSt:
 ; BE-PAIRED:       # %bb.0: # %entry
 ; BE-PAIRED-NEXT:    addis r3, r2, .LC0@toc@ha
 ; BE-PAIRED-NEXT:    li r4, 11
 ; BE-PAIRED-NEXT:    ld r3, .LC0@toc@l(r3)
 ; BE-PAIRED-NEXT:    lxvx vs0, r3, r4
 ; BE-PAIRED-NEXT:    li r4, 27
 ; BE-PAIRED-NEXT:    lxvx vs1, r3, r4
 ; BE-PAIRED-NEXT:    li r4, 43
 ; BE-PAIRED-NEXT:    lxvx vs2, r3, r4
 ; BE-PAIRED-NEXT:    li r4, 59
 ; BE-PAIRED-NEXT:    lxvx vs3, r3, r4
 ; BE-PAIRED-NEXT:    li r4, 35
-; BE-PAIRED-NEXT:    xxmtacc acc0
-; BE-PAIRED-NEXT:    xxmfacc acc0
 ; BE-PAIRED-NEXT:    stxvx vs1, r3, r4
 ; BE-PAIRED-NEXT:    li r4, 19
 ; BE-PAIRED-NEXT:    stxvx vs0, r3, r4
 ; BE-PAIRED-NEXT:    li r4, 67
 ; BE-PAIRED-NEXT:    stxvx vs3, r3, r4
 ; BE-PAIRED-NEXT:    li r4, 51
 ; BE-PAIRED-NEXT:    stxvx vs2, r3, r4
 ; BE-PAIRED-NEXT:    blr

llvm/test/CodeGen/PowerPC/mma-intrinsics.ll

 define void @ass_acc(<512 x i1>* %ptr, <16 x i8> %vc) {
 ; CHECK-LABEL: ass_acc:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    vmr v3, v2
 ; CHECK-NEXT:    xxlor vs0, v2, v2
 ; CHECK-NEXT:    xxlor vs1, v3, v3
 ; CHECK-NEXT:    xxlor vs2, v2, v2
 ; CHECK-NEXT:    xxlor vs3, v3, v3
-; CHECK-NEXT:    xxmtacc acc0
-; CHECK-NEXT:    xxmfacc acc0
 ; CHECK-NEXT:    stxv vs0, 48(r3)
 ; CHECK-NEXT:    stxv vs1, 32(r3)
 ; CHECK-NEXT:    stxv vs2, 16(r3)
 ; CHECK-NEXT:    stxv vs3, 0(r3)
 ; CHECK-NEXT:    blr
 ;
 ; CHECK-BE-LABEL: ass_acc:
 ; CHECK-BE:       # %bb.0: # %entry
 ; CHECK-BE-NEXT:    vmr v3, v2
 ; CHECK-BE-NEXT:    xxlor vs0, v2, v2
 ; CHECK-BE-NEXT:    xxlor vs1, v3, v3
 ; CHECK-BE-NEXT:    xxlor vs2, v2, v2
 ; CHECK-BE-NEXT:    xxlor vs3, v3, v3
-; CHECK-BE-NEXT:    xxmtacc acc0
-; CHECK-BE-NEXT:    xxmfacc acc0
 ; CHECK-BE-NEXT:    stxv vs1, 16(r3)
 ; CHECK-BE-NEXT:    stxv vs0, 0(r3)
 ; CHECK-BE-NEXT:    stxv vs3, 48(r3)
 ; CHECK-BE-NEXT:    stxv vs2, 32(r3)
 ; CHECK-BE-NEXT:    blr
 entry:
   %0 = tail call <512 x i1> @llvm.ppc.mma.assemble.acc(<16 x i8> %vc, <16 x i8> %vc, <16 x i8> %vc, <16 x i8> %vc)
   store <512 x i1> %0, <512 x i1>* %ptr, align 64

 ; CHECK-LABEL: int_xxmtacc:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    vmr v3, v2
 ; CHECK-NEXT:    xxlor vs0, v2, v2
 ; CHECK-NEXT:    xxlor vs1, v3, v3
 ; CHECK-NEXT:    xxlor vs2, v2, v2
 ; CHECK-NEXT:    xxlor vs3, v3, v3
 ; CHECK-NEXT:    xxmtacc acc0
-; CHECK-NEXT:    xxmtacc acc0
-; CHECK-NEXT:    xxmfacc acc0
 ; CHECK-NEXT:    stxv vs0, 48(r3)
 ; CHECK-NEXT:    stxv vs1, 32(r3)
 ; CHECK-NEXT:    stxv vs2, 16(r3)
 ; CHECK-NEXT:    stxv vs3, 0(r3)
 ; CHECK-NEXT:    blr
 ;
 ; CHECK-BE-LABEL: int_xxmtacc:
 ; CHECK-BE:       # %bb.0: # %entry
 ; CHECK-BE-NEXT:    vmr v3, v2
 ; CHECK-BE-NEXT:    xxlor vs0, v2, v2
 ; CHECK-BE-NEXT:    xxlor vs1, v3, v3
 ; CHECK-BE-NEXT:    xxlor vs2, v2, v2
 ; CHECK-BE-NEXT:    xxlor vs3, v3, v3
 ; CHECK-BE-NEXT:    xxmtacc acc0
-; CHECK-BE-NEXT:    xxmtacc acc0
-; CHECK-BE-NEXT:    xxmfacc acc0
 ; CHECK-BE-NEXT:    stxv vs1, 16(r3)
 ; CHECK-BE-NEXT:    stxv vs0, 0(r3)
 ; CHECK-BE-NEXT:    stxv vs3, 48(r3)
 ; CHECK-BE-NEXT:    stxv vs2, 32(r3)
 ; CHECK-BE-NEXT:    blr
 entry:
 ; One xxmtacc is generated from the call to assemble.acc then one xxmtacc is
 ; generated from the call to xxmtacc then one xxmfacc is generated for the store
   %0 = tail call <512 x i1> @llvm.ppc.mma.assemble.acc(<16 x i8> %vc, <16 x i8> %vc, <16 x i8> %vc, <16 x i8> %vc)
   %1 = tail call <512 x i1> @llvm.ppc.mma.xxmtacc(<512 x i1> %0)
   store <512 x i1> %1, <512 x i1>* %ptr, align 64
   ret void
 }
 
 ; xxmfacc
 declare <512 x i1> @llvm.ppc.mma.xxmfacc(<512 x i1>)
 define void @int_xxmfacc(<512 x i1>* %ptr, <16 x i8> %vc) {
 ; CHECK-LABEL: int_xxmfacc:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    vmr v3, v2
 ; CHECK-NEXT:    xxlor vs0, v2, v2
 ; CHECK-NEXT:    xxlor vs1, v3, v3
 ; CHECK-NEXT:    xxlor vs2, v2, v2
 ; CHECK-NEXT:    xxlor vs3, v3, v3
-; CHECK-NEXT:    xxmtacc acc0
-; CHECK-NEXT:    xxmfacc acc0
-; CHECK-NEXT:    xxmfacc acc0
 ; CHECK-NEXT:    stxv vs0, 48(r3)
 ; CHECK-NEXT:    stxv vs1, 32(r3)
 ; CHECK-NEXT:    stxv vs2, 16(r3)
 ; CHECK-NEXT:    stxv vs3, 0(r3)
 ; CHECK-NEXT:    blr
 ;
 ; CHECK-BE-LABEL: int_xxmfacc:
 ; CHECK-BE:       # %bb.0: # %entry
 ; CHECK-BE-NEXT:    vmr v3, v2
 ; CHECK-BE-NEXT:    xxlor vs0, v2, v2
 ; CHECK-BE-NEXT:    xxlor vs1, v3, v3
 ; CHECK-BE-NEXT:    xxlor vs2, v2, v2
 ; CHECK-BE-NEXT:    xxlor vs3, v3, v3
-; CHECK-BE-NEXT:    xxmtacc acc0
-; CHECK-BE-NEXT:    xxmfacc acc0
-; CHECK-BE-NEXT:    xxmfacc acc0
 ; CHECK-BE-NEXT:    stxv vs1, 16(r3)
 ; CHECK-BE-NEXT:    stxv vs0, 0(r3)
 ; CHECK-BE-NEXT:    stxv vs3, 48(r3)
 ; CHECK-BE-NEXT:    stxv vs2, 32(r3)
 ; CHECK-BE-NEXT:    blr
 entry:
 ; One xxmtacc is generated from the call to assemble.acc then one xxmfacc is
 ; generated from the call to xxmfacc then one xxmfacc is generated for the store

 ; CHECK-NEXT:    lxv vs1, 32(r3)
 ; CHECK-NEXT:    lxv vs0, 48(r3)
 ; CHECK-NEXT:    lxv vs3, 0(r3)
 ; CHECK-NEXT:    lxv vs2, 16(r3)
 ; CHECK-NEXT:    xxmtacc acc0
 ; CHECK-NEXT:    xvi4ger8pp acc0, v2, v2
 ; CHECK-NEXT:  .LBB7_3: # %if.end
 ; CHECK-NEXT:    xxmfacc acc0
-; CHECK-NEXT:    xxmtacc acc0
-; CHECK-NEXT:    xxmfacc acc0
 ; CHECK-NEXT:    stxv vs0, 48(r3)
 ; CHECK-NEXT:    stxv vs1, 32(r3)
 ; CHECK-NEXT:    stxv vs2, 16(r3)
 ; CHECK-NEXT:    stxv vs3, 0(r3)
 ; CHECK-NEXT:    blr
 ;
 ; CHECK-BE-LABEL: testBranch:
 ; CHECK-BE:       # %bb.0: # %entry
 ; CHECK-BE-NEXT:    cmplwi r7, 0
 ; CHECK-BE-NEXT:    beq cr0, .LBB7_2
 ; CHECK-BE-NEXT:  # %bb.1: # %if.then
 ; CHECK-BE-NEXT:    xxsetaccz acc0
 ; CHECK-BE-NEXT:    b .LBB7_3
 ; CHECK-BE-NEXT:  .LBB7_2: # %if.else
 ; CHECK-BE-NEXT:    lxv vs1, 16(r3)
 ; CHECK-BE-NEXT:    lxv vs0, 0(r3)
 ; CHECK-BE-NEXT:    lxv vs3, 48(r3)
 ; CHECK-BE-NEXT:    lxv vs2, 32(r3)
 ; CHECK-BE-NEXT:    xxmtacc acc0
 ; CHECK-BE-NEXT:    xvi4ger8pp acc0, v2, v2
 ; CHECK-BE-NEXT:  .LBB7_3: # %if.end
 ; CHECK-BE-NEXT:    xxmfacc acc0
-; CHECK-BE-NEXT:    xxmtacc acc0
-; CHECK-BE-NEXT:    xxmfacc acc0
 ; CHECK-BE-NEXT:    stxv vs1, 16(r3)
 ; CHECK-BE-NEXT:    stxv vs0, 0(r3)
 ; CHECK-BE-NEXT:    stxv vs3, 48(r3)
 ; CHECK-BE-NEXT:    stxv vs2, 32(r3)
 ; CHECK-BE-NEXT:    blr
 entry:
   %tobool = icmp eq i32 %val, 0
   br i1 %tobool, label %if.else, label %if.then

 ; CHECK-NEXT:    stxvp vsp4, r1(r3)
 ; CHECK-NEXT:    li r3, 32
 ; CHECK-NEXT:    stxvp vsp6, r1(r3)
 ; CHECK-NEXT:    bl testRedundantPrimeUnprimeF@notoc
 ; CHECK-NEXT:    li r3, 64
 ; CHECK-NEXT:    lxvp vsp0, r1(r3)
 ; CHECK-NEXT:    li r3, 32
 ; CHECK-NEXT:    lxvp vsp2, r1(r3)
-; CHECK-NEXT:    xxmtacc acc0
-; CHECK-NEXT:    xxmfacc acc0
 ; CHECK-NEXT:    stxv vs0, 112(r30)
 ; CHECK-NEXT:    stxv vs1, 96(r30)
 ; CHECK-NEXT:    stxv vs2, 80(r30)
 ; CHECK-NEXT:    stxv vs3, 64(r30)
 ; CHECK-NEXT:    addi r1, r1, 112
 ; CHECK-NEXT:    ld r0, 16(r1)
 ; CHECK-NEXT:    ld r30, -16(r1) # 8-byte Folded Reload
 ; CHECK-NEXT:    mtlr r0

 ; CHECK-BE-NEXT:    li r3, 144
 ; CHECK-BE-NEXT:    stxvp vsp6, r1(r3)
 ; CHECK-BE-NEXT:    bl testRedundantPrimeUnprimeF
 ; CHECK-BE-NEXT:    nop
 ; CHECK-BE-NEXT:    li r3, 112
 ; CHECK-BE-NEXT:    lxvp vsp0, r1(r3)
 ; CHECK-BE-NEXT:    li r3, 144
 ; CHECK-BE-NEXT:    lxvp vsp2, r1(r3)
-; CHECK-BE-NEXT:    xxmtacc acc0
-; CHECK-BE-NEXT:    xxmfacc acc0
 ; CHECK-BE-NEXT:    stxv vs3, 112(r30)
 ; CHECK-BE-NEXT:    stxv vs2, 96(r30)
 ; CHECK-BE-NEXT:    stxv vs1, 80(r30)
 ; CHECK-BE-NEXT:    stxv vs0, 64(r30)
 ; CHECK-BE-NEXT:    ld r30, 176(r1) # 8-byte Folded Reload
 ; CHECK-BE-NEXT:    addi r1, r1, 192
 ; CHECK-BE-NEXT:    ld r0, 16(r1)
 ; CHECK-BE-NEXT:    mtlr r0