Code Formating

Supporting Arithemtic Operations: `and`, `or`, `xor`

llvm/lib/Target/AMDGPU/SIISelLowering.cpp

@@ -5192,7 +5192,59 @@ static MachineBasicBlock *emitIndirectDst(MachineInstr &MI,
   return LoopBB;
 }
 
-static uint32_t getIdentityValueForWaveReduction(unsigned Opc) {
+static MachineBasicBlock *Expand64BitScalarArithmetic(MachineInstr &MI,
+                                                      MachineBasicBlock *BB) {
+  // For targets older than GFX12, we emit a sequence of 32-bit operations.
+  // For GFX12, we emit s_add_u64 and s_sub_u64.
+  MachineFunction *MF = BB->getParent();
+  const SIInstrInfo *TII = MF->getSubtarget<GCNSubtarget>().getInstrInfo();
+  SIMachineFunctionInfo *MFI = MF->getInfo<SIMachineFunctionInfo>();
+  const GCNSubtarget &ST = MF->getSubtarget<GCNSubtarget>();
+  MachineRegisterInfo &MRI = BB->getParent()->getRegInfo();
+  const DebugLoc &DL = MI.getDebugLoc();
+  MachineOperand &Dest = MI.getOperand(0);
+  MachineOperand &Src0 = MI.getOperand(1);
+  MachineOperand &Src1 = MI.getOperand(2);
+  bool IsAdd = (MI.getOpcode() == AMDGPU::S_ADD_U64_PSEUDO);
+  if (ST.hasScalarAddSub64()) {
+    unsigned Opc = IsAdd ? AMDGPU::S_ADD_U64 : AMDGPU::S_SUB_U64;
+    // clang-format off
+    BuildMI(*BB, MI, DL, TII->get(Opc), Dest.getReg())
+        .add(Src0)
+        .add(Src1);
+    // clang-format on
+  } else {
+    const SIRegisterInfo *TRI = ST.getRegisterInfo();
+    const TargetRegisterClass *BoolRC = TRI->getBoolRC();
+
+    Register DestSub0 = MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
+    Register DestSub1 = MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
+
+    MachineOperand Src0Sub0 = TII->buildExtractSubRegOrImm(
+        MI, MRI, Src0, BoolRC, AMDGPU::sub0, &AMDGPU::SReg_32RegClass);
+    MachineOperand Src0Sub1 = TII->buildExtractSubRegOrImm(
+        MI, MRI, Src0, BoolRC, AMDGPU::sub1, &AMDGPU::SReg_32RegClass);
+
+    MachineOperand Src1Sub0 = TII->buildExtractSubRegOrImm(
+        MI, MRI, Src1, BoolRC, AMDGPU::sub0, &AMDGPU::SReg_32RegClass);
+    MachineOperand Src1Sub1 = TII->buildExtractSubRegOrImm(
+        MI, MRI, Src1, BoolRC, AMDGPU::sub1, &AMDGPU::SReg_32RegClass);
+
+    unsigned LoOpc = IsAdd ? AMDGPU::S_ADD_U32 : AMDGPU::S_SUB_U32;
+    unsigned HiOpc = IsAdd ? AMDGPU::S_ADDC_U32 : AMDGPU::S_SUBB_U32;
+    BuildMI(*BB, MI, DL, TII->get(LoOpc), DestSub0).add(Src0Sub0).add(Src1Sub0);
+    BuildMI(*BB, MI, DL, TII->get(HiOpc), DestSub1).add(Src0Sub1).add(Src1Sub1);
+    BuildMI(*BB, MI, DL, TII->get(TargetOpcode::REG_SEQUENCE), Dest.getReg())
+        .addReg(DestSub0)
+        .addImm(AMDGPU::sub0)
+        .addReg(DestSub1)
+        .addImm(AMDGPU::sub1);
+  }
+  MI.eraseFromParent();
+  return BB;
+}
+
+static uint32_t getIdentityValueFor32BitWaveReduction(unsigned Opc) {
   switch (Opc) {
   case AMDGPU::S_MIN_U32:
     return std::numeric_limits<uint32_t>::max();
@@ -5210,7 +5262,31 @@ static uint32_t getIdentityValueForWaveReduction(unsigned Opc) {
   case AMDGPU::S_AND_B32:
     return std::numeric_limits<uint32_t>::max();
   default:
-    llvm_unreachable("Unexpected opcode in getIdentityValueForWaveReduction");
+    llvm_unreachable(
+        "Unexpected opcode in getIdentityValueFor32BitWaveReduction");
+  }
+}
+
+static uint64_t getIdentityValueFor64BitWaveReduction(unsigned Opc) {
+  switch (Opc) {
+  case AMDGPU::V_CMP_LT_U64_e64: // umin.u64
+    return std::numeric_limits<uint64_t>::max();
+  case AMDGPU::V_CMP_LT_I64_e64: // min.i64
+    return std::numeric_limits<int64_t>::max();
+  case AMDGPU::V_CMP_GT_U64_e64: // umax.u64
+    return std::numeric_limits<uint64_t>::min();
+  case AMDGPU::V_CMP_GT_I64_e64: // max.i64
+    return std::numeric_limits<int64_t>::min();
+  case AMDGPU::S_ADD_U64_PSEUDO:
+  case AMDGPU::S_SUB_U64_PSEUDO:
+  case AMDGPU::S_OR_B64:
+  case AMDGPU::S_XOR_B64:
+    return std::numeric_limits<uint64_t>::min();
+  case AMDGPU::S_AND_B64:
+    return std::numeric_limits<uint64_t>::max();
+  default:
+    llvm_unreachable(
+        "Unexpected opcode in getIdentityValueFor64BitWaveReduction");
   }
 }
 
@@ -5241,53 +5317,99 @@ static MachineBasicBlock *lowerWaveReduce(MachineInstr &MI,
       RetBB = &BB;
       break;
     }
+    case AMDGPU::V_CMP_LT_U64_e64: // umin
+    case AMDGPU::V_CMP_LT_I64_e64: // min
+    case AMDGPU::V_CMP_GT_U64_e64: // umax
+    case AMDGPU::V_CMP_GT_I64_e64: // max
+    case AMDGPU::S_AND_B64:
+    case AMDGPU::S_OR_B64: {
+      // Idempotent operations.
+      BuildMI(BB, MI, DL, TII->get(AMDGPU::S_MOV_B64), DstReg).addReg(SrcReg);
+      RetBB = &BB;
+      break;
+    }
     case AMDGPU::S_XOR_B32:
+    case AMDGPU::S_XOR_B64:
     case AMDGPU::S_ADD_I32:
-    case AMDGPU::S_SUB_I32: {
+    case AMDGPU::S_ADD_U64_PSEUDO:
+    case AMDGPU::S_SUB_I32:
+    case AMDGPU::S_SUB_U64_PSEUDO: {
       const TargetRegisterClass *WaveMaskRegClass = TRI->getWaveMaskRegClass();
       const TargetRegisterClass *DstRegClass = MRI.getRegClass(DstReg);
       Register ExecMask = MRI.createVirtualRegister(WaveMaskRegClass);
-      Register ActiveLanes = MRI.createVirtualRegister(DstRegClass);
+      Register NumActiveLanes =
+          MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
 
       bool IsWave32 = ST.isWave32();
       unsigned MovOpc = IsWave32 ? AMDGPU::S_MOV_B32 : AMDGPU::S_MOV_B64;
       MCRegister ExecReg = IsWave32 ? AMDGPU::EXEC_LO : AMDGPU::EXEC;
-      unsigned CountReg =
+      unsigned BitCountOpc =
           IsWave32 ? AMDGPU::S_BCNT1_I32_B32 : AMDGPU::S_BCNT1_I32_B64;
 
-      auto Exec =
-          BuildMI(BB, MI, DL, TII->get(MovOpc), ExecMask).addReg(ExecReg);
+      BuildMI(BB, MI, DL, TII->get(MovOpc), ExecMask).addReg(ExecReg);
 
-      auto NewAccumulator = BuildMI(BB, MI, DL, TII->get(CountReg), ActiveLanes)
-                                .addReg(Exec->getOperand(0).getReg());
+      auto NewAccumulator =
+          BuildMI(BB, MI, DL, TII->get(BitCountOpc), NumActiveLanes)
+              .addReg(ExecMask);
 
       switch (Opc) {
-      case AMDGPU::S_XOR_B32: {
+      case AMDGPU::S_XOR_B32:
+      case AMDGPU::S_XOR_B64: {
         // Performing an XOR operation on a uniform value
         // depends on the parity of the number of active lanes.
         // For even parity, the result will be 0, for odd
         // parity the result will be the same as the input value.
-        Register ParityRegister = MRI.createVirtualRegister(DstRegClass);
+        Register ParityRegister =
+            MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
 
-        auto ParityReg =
-            BuildMI(BB, MI, DL, TII->get(AMDGPU::S_AND_B32), ParityRegister)
-                .addReg(NewAccumulator->getOperand(0).getReg())
-                .addImm(1);
-        BuildMI(BB, MI, DL, TII->get(AMDGPU::S_MUL_I32), DstReg)
-            .addReg(SrcReg)
-            .addReg(ParityReg->getOperand(0).getReg());
+        BuildMI(BB, MI, DL, TII->get(AMDGPU::S_AND_B32), ParityRegister)
+            .addReg(NewAccumulator->getOperand(0).getReg())
+            .addImm(1)
+            .setOperandDead(3); // Dead scc
+        if (Opc == AMDGPU::S_XOR_B32) {
+          BuildMI(BB, MI, DL, TII->get(AMDGPU::S_MUL_I32), DstReg)
+              .addReg(SrcReg)
+              .addReg(ParityRegister);
+        } else {
+          Register DestSub0 =
+              MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
+          Register DestSub1 =
+              MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
+
+          const TargetRegisterClass *SrcRC = MRI.getRegClass(SrcReg);
+          const TargetRegisterClass *SrcSubRC =
+              TRI->getSubRegisterClass(SrcRC, AMDGPU::sub0);
+
+          MachineOperand Op1L = TII->buildExtractSubRegOrImm(
+              MI, MRI, MI.getOperand(1), SrcRC, AMDGPU::sub0, SrcSubRC);
+          MachineOperand Op1H = TII->buildExtractSubRegOrImm(
+              MI, MRI, MI.getOperand(1), SrcRC, AMDGPU::sub1, SrcSubRC);
+
+          BuildMI(BB, MI, DL, TII->get(AMDGPU::S_MUL_I32), DestSub0)
+              .add(Op1L)
+              .addReg(ParityRegister);
+
+          BuildMI(BB, MI, DL, TII->get(AMDGPU::S_MUL_I32), DestSub1)
+              .add(Op1H)
+              .addReg(ParityRegister);
+
+          BuildMI(BB, MI, DL, TII->get(TargetOpcode::REG_SEQUENCE), DstReg)
+              .addReg(DestSub0)
+              .addImm(AMDGPU::sub0)
+              .addReg(DestSub1)
+              .addImm(AMDGPU::sub1);
+        }
         break;
       }
       case AMDGPU::S_SUB_I32: {
         Register NegatedVal = MRI.createVirtualRegister(DstRegClass);
 
         // Take the negation of the source operand.
-        auto InvertedValReg =
-            BuildMI(BB, MI, DL, TII->get(AMDGPU::S_MUL_I32), NegatedVal)
-                .addImm(-1)
-                .addReg(SrcReg);
+        BuildMI(BB, MI, DL, TII->get(AMDGPU::S_SUB_I32), NegatedVal)
+            .addImm(0)
+            .addReg(SrcReg);
         BuildMI(BB, MI, DL, TII->get(AMDGPU::S_MUL_I32), DstReg)
-            .addReg(InvertedValReg->getOperand(0).getReg())
+            .addReg(NegatedVal)
             .addReg(NewAccumulator->getOperand(0).getReg());
         break;
       }
@@ -5297,6 +5419,75 @@ static MachineBasicBlock *lowerWaveReduce(MachineInstr &MI,
             .addReg(NewAccumulator->getOperand(0).getReg());
         break;
       }
+      case AMDGPU::S_ADD_U64_PSEUDO:
+      case AMDGPU::S_SUB_U64_PSEUDO: {
+        Register DestSub0 = MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
+        Register DestSub1 = MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
+        Register Op1H_Op0L_Reg =
+            MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
+        Register Op1L_Op0H_Reg =
+            MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
+        Register CarryReg = MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
+        Register AddReg = MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
+        Register NegatedValLo =
+            MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
+        Register NegatedValHi =
+            MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
+
+        const TargetRegisterClass *Src1RC = MRI.getRegClass(SrcReg);
+        const TargetRegisterClass *Src1SubRC =
+            TRI->getSubRegisterClass(Src1RC, AMDGPU::sub0);
+
+        MachineOperand Op1L = TII->buildExtractSubRegOrImm(
+            MI, MRI, MI.getOperand(1), Src1RC, AMDGPU::sub0, Src1SubRC);
+        MachineOperand Op1H = TII->buildExtractSubRegOrImm(
+            MI, MRI, MI.getOperand(1), Src1RC, AMDGPU::sub1, Src1SubRC);
+
+        if (Opc == AMDGPU::S_SUB_U64_PSEUDO) {
+          BuildMI(BB, MI, DL, TII->get(AMDGPU::S_SUB_I32), NegatedValLo)
+              .addImm(0)
+              .addReg(NewAccumulator->getOperand(0).getReg())
+              .setOperandDead(3); // Dead scc
+          BuildMI(BB, MI, DL, TII->get(AMDGPU::S_ASHR_I32), NegatedValHi)
+              .addReg(NegatedValLo)
+              .addImm(31)
+              .setOperandDead(3); // Dead scc
+          BuildMI(BB, MI, DL, TII->get(AMDGPU::S_MUL_I32), Op1L_Op0H_Reg)
+              .add(Op1L)
+              .addReg(NegatedValHi);
+        }
+        Register LowOpcode = Opc == AMDGPU::S_SUB_U64_PSEUDO
+                                 ? NegatedValLo
+                                 : NewAccumulator->getOperand(0).getReg();
+        BuildMI(BB, MI, DL, TII->get(AMDGPU::S_MUL_I32), DestSub0)
+            .add(Op1L)
+            .addReg(LowOpcode);
+        BuildMI(BB, MI, DL, TII->get(AMDGPU::S_MUL_HI_U32), CarryReg)
+            .add(Op1L)
+            .addReg(LowOpcode);
+        BuildMI(BB, MI, DL, TII->get(AMDGPU::S_MUL_I32), Op1H_Op0L_Reg)
+            .add(Op1H)
+            .addReg(LowOpcode);
+
+        Register HiVal = Opc == AMDGPU::S_SUB_U64_PSEUDO ? AddReg : DestSub1;
+        BuildMI(BB, MI, DL, TII->get(AMDGPU::S_ADD_U32), HiVal)
+            .addReg(CarryReg)
+            .addReg(Op1H_Op0L_Reg)
+            .setOperandDead(3); // Dead scc
+
+        if (Opc == AMDGPU::S_SUB_U64_PSEUDO) {
+          BuildMI(BB, MI, DL, TII->get(AMDGPU::S_ADD_U32), DestSub1)
+              .addReg(HiVal)
+              .addReg(Op1L_Op0H_Reg)
+              .setOperandDead(3); // Dead scc
+        }
+        BuildMI(BB, MI, DL, TII->get(TargetOpcode::REG_SEQUENCE), DstReg)
+            .addReg(DestSub0)
+            .addImm(AMDGPU::sub0)
+            .addReg(DestSub1)
+            .addImm(AMDGPU::sub1);
+        break;
+      }
       }
       RetBB = &BB;
     }
@@ -5313,6 +5504,11 @@ static MachineBasicBlock *lowerWaveReduce(MachineInstr &MI,
     // so that we will get the next active lane for next iteration.
     MachineBasicBlock::iterator I = BB.end();
     Register SrcReg = MI.getOperand(1).getReg();
+    bool is32BitOpc = (Opc == AMDGPU::S_MIN_U32 || Opc == AMDGPU::S_MIN_I32 ||
+                       Opc == AMDGPU::S_MAX_U32 || Opc == AMDGPU::S_MAX_I32 ||
+                       Opc == AMDGPU::S_ADD_I32 || Opc == AMDGPU::S_SUB_I32 ||
+                       Opc == AMDGPU::S_AND_B32 || Opc == AMDGPU::S_OR_B32 ||
+                       Opc == AMDGPU::S_XOR_B32);
 
     // Create Control flow for loop
     // Split MI's Machine Basic block into For loop
@@ -5322,73 +5518,162 @@ static MachineBasicBlock *lowerWaveReduce(MachineInstr &MI,
     const TargetRegisterClass *WaveMaskRegClass = TRI->getWaveMaskRegClass();
     const TargetRegisterClass *DstRegClass = MRI.getRegClass(DstReg);
     Register LoopIterator = MRI.createVirtualRegister(WaveMaskRegClass);
-    Register InitalValReg = MRI.createVirtualRegister(DstRegClass);
-
+    Register IdentityValReg = MRI.createVirtualRegister(DstRegClass);
     Register AccumulatorReg = MRI.createVirtualRegister(DstRegClass);
     Register ActiveBitsReg = MRI.createVirtualRegister(WaveMaskRegClass);
     Register NewActiveBitsReg = MRI.createVirtualRegister(WaveMaskRegClass);
-
-    Register FF1Reg = MRI.createVirtualRegister(DstRegClass);
-    Register LaneValueReg =
-        MRI.createVirtualRegister(&AMDGPU::SReg_32_XM0RegClass);
+    Register FF1Reg = MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
+    Register LaneValueReg = MRI.createVirtualRegister(DstRegClass);
 
     bool IsWave32 = ST.isWave32();
-    unsigned MovOpc = IsWave32 ? AMDGPU::S_MOV_B32 : AMDGPU::S_MOV_B64;
+    unsigned MovOpcForExec = IsWave32 ? AMDGPU::S_MOV_B32 : AMDGPU::S_MOV_B64;
     unsigned ExecReg = IsWave32 ? AMDGPU::EXEC_LO : AMDGPU::EXEC;
 
     // Create initial values of induction variable from Exec, Accumulator and
     // insert branch instr to newly created ComputeBlock
-    uint32_t InitalValue = getIdentityValueForWaveReduction(Opc);
-    auto TmpSReg =
-        BuildMI(BB, I, DL, TII->get(MovOpc), LoopIterator).addReg(ExecReg);
-    BuildMI(BB, I, DL, TII->get(AMDGPU::S_MOV_B32), InitalValReg)
-        .addImm(InitalValue);
+    BuildMI(BB, I, DL, TII->get(MovOpcForExec), LoopIterator).addReg(ExecReg);
+    if (is32BitOpc) {
+      uint32_t IdentityValue = getIdentityValueFor32BitWaveReduction(Opc);
+      BuildMI(BB, I, DL, TII->get(AMDGPU::S_MOV_B32), IdentityValReg)
+          .addImm(IdentityValue);
+    } else {
+      uint64_t IdentityValue = getIdentityValueFor64BitWaveReduction(Opc);
+      BuildMI(BB, I, DL, TII->get(AMDGPU::S_MOV_B64_IMM_PSEUDO), IdentityValReg)
+          .addImm(IdentityValue);
+    }
     // clang-format off
     BuildMI(BB, I, DL, TII->get(AMDGPU::S_BRANCH))
         .addMBB(ComputeLoop);
     // clang-format on
 
     // Start constructing ComputeLoop
-    I = ComputeLoop->end();
+    I = ComputeLoop->begin();
     auto Accumulator =
         BuildMI(*ComputeLoop, I, DL, TII->get(AMDGPU::PHI), AccumulatorReg)
-            .addReg(InitalValReg)
+            .addReg(IdentityValReg)
             .addMBB(&BB);
     auto ActiveBits =
         BuildMI(*ComputeLoop, I, DL, TII->get(AMDGPU::PHI), ActiveBitsReg)
-            .addReg(TmpSReg->getOperand(0).getReg())
+            .addReg(LoopIterator)
             .addMBB(&BB);
 
+    I = ComputeLoop->end();
+    MachineInstr *NewAccumulator;
     // Perform the computations
     unsigned SFFOpc = IsWave32 ? AMDGPU::S_FF1_I32_B32 : AMDGPU::S_FF1_I32_B64;
-    auto FF1 = BuildMI(*ComputeLoop, I, DL, TII->get(SFFOpc), FF1Reg)
-                   .addReg(ActiveBits->getOperand(0).getReg());
-    auto LaneValue = BuildMI(*ComputeLoop, I, DL,
-                             TII->get(AMDGPU::V_READLANE_B32), LaneValueReg)
-                         .addReg(SrcReg)
-                         .addReg(FF1->getOperand(0).getReg());
-    auto NewAccumulator = BuildMI(*ComputeLoop, I, DL, TII->get(Opc), DstReg)
-                              .addReg(Accumulator->getOperand(0).getReg())
-                              .addReg(LaneValue->getOperand(0).getReg());
+    BuildMI(*ComputeLoop, I, DL, TII->get(SFFOpc), FF1Reg)
+        .addReg(ActiveBitsReg);
+    if (is32BitOpc) {
+      BuildMI(*ComputeLoop, I, DL, TII->get(AMDGPU::V_READLANE_B32),
+              LaneValueReg)
+          .addReg(SrcReg)
+          .addReg(FF1Reg);
+      NewAccumulator = BuildMI(*ComputeLoop, I, DL, TII->get(Opc), DstReg)
+                           .addReg(Accumulator->getOperand(0).getReg())
+                           .addReg(LaneValueReg);
+    } else {
+      Register LaneValueLoReg =
+          MRI.createVirtualRegister(&AMDGPU::SReg_32_XM0RegClass);
+      Register LaneValueHiReg =
+          MRI.createVirtualRegister(&AMDGPU::SReg_32_XM0RegClass);
+      Register LaneValReg = MRI.createVirtualRegister(&AMDGPU::SReg_64RegClass);
+      const TargetRegisterClass *SrcRC = MRI.getRegClass(SrcReg);
+      const TargetRegisterClass *SrcSubRC =
+          TRI->getSubRegisterClass(SrcRC, AMDGPU::sub0);
+      MachineOperand Op1L = TII->buildExtractSubRegOrImm(
+          MI, MRI, MI.getOperand(1), SrcRC, AMDGPU::sub0, SrcSubRC);
+      MachineOperand Op1H = TII->buildExtractSubRegOrImm(
+          MI, MRI, MI.getOperand(1), SrcRC, AMDGPU::sub1, SrcSubRC);
+      // lane value input should be in an sgpr
+      MachineInstr *LaneValueLo =
+          BuildMI(*ComputeLoop, I, DL, TII->get(AMDGPU::V_READLANE_B32),
+                  LaneValueLoReg)
+              .add(Op1L)
+              .addReg(FF1Reg);
+      MachineInstr *LaneValueHi =
+          BuildMI(*ComputeLoop, I, DL, TII->get(AMDGPU::V_READLANE_B32),
+                  LaneValueHiReg)
+              .add(Op1H)
+              .addReg(FF1Reg);
+      auto LaneValue = BuildMI(*ComputeLoop, I, DL,
+                               TII->get(TargetOpcode::REG_SEQUENCE), LaneValReg)
+                           .addReg(LaneValueLoReg)
+                           .addImm(AMDGPU::sub0)
+                           .addReg(LaneValueHiReg)
+                           .addImm(AMDGPU::sub1);
+      switch (Opc) {
+      case AMDGPU::S_OR_B64:
+      case AMDGPU::S_AND_B64:
+      case AMDGPU::S_XOR_B64: {
+        NewAccumulator = BuildMI(*ComputeLoop, I, DL, TII->get(Opc), DstReg)
+                             .addReg(Accumulator->getOperand(0).getReg())
+                             .addReg(LaneValue->getOperand(0).getReg())
+                             .setOperandDead(3); // Dead scc
+        break;
+      }
+      case AMDGPU::V_CMP_GT_I64_e64:
+      case AMDGPU::V_CMP_GT_U64_e64:
+      case AMDGPU::V_CMP_LT_I64_e64:
+      case AMDGPU::V_CMP_LT_U64_e64: {
+        Register LaneMaskReg = MRI.createVirtualRegister(WaveMaskRegClass);
+        Register ComparisonResultReg =
+            MRI.createVirtualRegister(WaveMaskRegClass);
+        const TargetRegisterClass *VregClass = TRI->getVGPR64Class();
+        const TargetRegisterClass *VSubRegClass =
+            TRI->getSubRegisterClass(VregClass, AMDGPU::sub0);
+        Register AccumulatorVReg = MRI.createVirtualRegister(VregClass);
+        MachineOperand SrcReg0Sub0 =
+            TII->buildExtractSubRegOrImm(MI, MRI, Accumulator->getOperand(0),
+                                         VregClass, AMDGPU::sub0, VSubRegClass);
+        MachineOperand SrcReg0Sub1 =
+            TII->buildExtractSubRegOrImm(MI, MRI, Accumulator->getOperand(0),
+                                         VregClass, AMDGPU::sub1, VSubRegClass);
+        BuildMI(*ComputeLoop, I, DL, TII->get(TargetOpcode::REG_SEQUENCE),
+                AccumulatorVReg)
+            .add(SrcReg0Sub0)
+            .addImm(AMDGPU::sub0)
+            .add(SrcReg0Sub1)
+            .addImm(AMDGPU::sub1);
+        BuildMI(*ComputeLoop, I, DL, TII->get(Opc), LaneMaskReg)
+            .addReg(LaneValue->getOperand(0).getReg())
+            .addReg(AccumulatorVReg);
 
+        unsigned AndOpc = IsWave32 ? AMDGPU::S_AND_B32 : AMDGPU::S_AND_B64;
+        BuildMI(*ComputeLoop, I, DL, TII->get(AndOpc), ComparisonResultReg)
+            .addReg(LaneMaskReg)
+            .addReg(ActiveBitsReg);
+
+        NewAccumulator = BuildMI(*ComputeLoop, I, DL,
+                                 TII->get(AMDGPU::S_CSELECT_B64), DstReg)
+                             .addReg(LaneValue->getOperand(0).getReg())
+                             .addReg(Accumulator->getOperand(0).getReg());
+        break;
+      }
+      case AMDGPU::S_ADD_U64_PSEUDO:
+      case AMDGPU::S_SUB_U64_PSEUDO: {
+        NewAccumulator = BuildMI(*ComputeLoop, I, DL, TII->get(Opc), DstReg)
+                             .addReg(Accumulator->getOperand(0).getReg())
+                             .addReg(LaneValue->getOperand(0).getReg());
+        ComputeLoop = Expand64BitScalarArithmetic(*NewAccumulator, ComputeLoop);
+        break;
+      }
+      }
+    }
     // Manipulate the iterator to get the next active lane
     unsigned BITSETOpc =
         IsWave32 ? AMDGPU::S_BITSET0_B32 : AMDGPU::S_BITSET0_B64;
-    auto NewActiveBits =
-        BuildMI(*ComputeLoop, I, DL, TII->get(BITSETOpc), NewActiveBitsReg)
-            .addReg(FF1->getOperand(0).getReg())
-            .addReg(ActiveBits->getOperand(0).getReg());
+    BuildMI(*ComputeLoop, I, DL, TII->get(BITSETOpc), NewActiveBitsReg)
+        .addReg(FF1Reg)
+        .addReg(ActiveBitsReg);
 
     // Add phi nodes
-    Accumulator.addReg(NewAccumulator->getOperand(0).getReg())
-        .addMBB(ComputeLoop);
-    ActiveBits.addReg(NewActiveBits->getOperand(0).getReg())
-        .addMBB(ComputeLoop);
+    Accumulator.addReg(DstReg).addMBB(ComputeLoop);
+    ActiveBits.addReg(NewActiveBitsReg).addMBB(ComputeLoop);
 
     // Creating branching
     unsigned CMPOpc = IsWave32 ? AMDGPU::S_CMP_LG_U32 : AMDGPU::S_CMP_LG_U64;
     BuildMI(*ComputeLoop, I, DL, TII->get(CMPOpc))
-        .addReg(NewActiveBits->getOperand(0).getReg())
+        .addReg(NewActiveBitsReg)
         .addImm(0);
     BuildMI(*ComputeLoop, I, DL, TII->get(AMDGPU::S_CBRANCH_SCC1))
         .addMBB(ComputeLoop);
@@ -5410,22 +5695,40 @@ SITargetLowering::EmitInstrWithCustomInserter(MachineInstr &MI,
   switch (MI.getOpcode()) {
   case AMDGPU::WAVE_REDUCE_UMIN_PSEUDO_U32:
     return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::S_MIN_U32);
+  case AMDGPU::WAVE_REDUCE_UMIN_PSEUDO_U64:
+    return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::V_CMP_LT_U64_e64);
   case AMDGPU::WAVE_REDUCE_MIN_PSEUDO_I32:
     return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::S_MIN_I32);
+  case AMDGPU::WAVE_REDUCE_MIN_PSEUDO_I64:
+    return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::V_CMP_LT_I64_e64);
   case AMDGPU::WAVE_REDUCE_UMAX_PSEUDO_U32:
     return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::S_MAX_U32);
+  case AMDGPU::WAVE_REDUCE_UMAX_PSEUDO_U64:
+    return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::V_CMP_GT_U64_e64);
   case AMDGPU::WAVE_REDUCE_MAX_PSEUDO_I32:
     return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::S_MAX_I32);
+  case AMDGPU::WAVE_REDUCE_MAX_PSEUDO_I64:
+    return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::V_CMP_GT_I64_e64);
   case AMDGPU::WAVE_REDUCE_ADD_PSEUDO_I32:
     return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::S_ADD_I32);
+  case AMDGPU::WAVE_REDUCE_ADD_PSEUDO_U64:
+    return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::S_ADD_U64_PSEUDO);
   case AMDGPU::WAVE_REDUCE_SUB_PSEUDO_I32:
     return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::S_SUB_I32);
+  case AMDGPU::WAVE_REDUCE_SUB_PSEUDO_U64:
+    return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::S_SUB_U64_PSEUDO);
   case AMDGPU::WAVE_REDUCE_AND_PSEUDO_B32:
     return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::S_AND_B32);
+  case AMDGPU::WAVE_REDUCE_AND_PSEUDO_B64:
+    return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::S_AND_B64);
   case AMDGPU::WAVE_REDUCE_OR_PSEUDO_B32:
     return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::S_OR_B32);
+  case AMDGPU::WAVE_REDUCE_OR_PSEUDO_B64:
+    return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::S_OR_B64);
   case AMDGPU::WAVE_REDUCE_XOR_PSEUDO_B32:
     return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::S_XOR_B32);
+  case AMDGPU::WAVE_REDUCE_XOR_PSEUDO_B64:
+    return lowerWaveReduce(MI, *BB, *getSubtarget(), AMDGPU::S_XOR_B64);
   case AMDGPU::S_UADDO_PSEUDO:
   case AMDGPU::S_USUBO_PSEUDO: {
     const DebugLoc &DL = MI.getDebugLoc();
@@ -5452,55 +5755,7 @@ SITargetLowering::EmitInstrWithCustomInserter(MachineInstr &MI,
   }
   case AMDGPU::S_ADD_U64_PSEUDO:
   case AMDGPU::S_SUB_U64_PSEUDO: {
-    // For targets older than GFX12, we emit a sequence of 32-bit operations.
-    // For GFX12, we emit s_add_u64 and s_sub_u64.
-    const GCNSubtarget &ST = MF->getSubtarget<GCNSubtarget>();
-    MachineRegisterInfo &MRI = BB->getParent()->getRegInfo();
-    const DebugLoc &DL = MI.getDebugLoc();
-    MachineOperand &Dest = MI.getOperand(0);
-    MachineOperand &Src0 = MI.getOperand(1);
-    MachineOperand &Src1 = MI.getOperand(2);
-    bool IsAdd = (MI.getOpcode() == AMDGPU::S_ADD_U64_PSEUDO);
-    if (Subtarget->hasScalarAddSub64()) {
-      unsigned Opc = IsAdd ? AMDGPU::S_ADD_U64 : AMDGPU::S_SUB_U64;
-      // clang-format off
-      BuildMI(*BB, MI, DL, TII->get(Opc), Dest.getReg())
-          .add(Src0)
-          .add(Src1);
-      // clang-format on
-    } else {
-      const SIRegisterInfo *TRI = ST.getRegisterInfo();
-      const TargetRegisterClass *BoolRC = TRI->getBoolRC();
-
-      Register DestSub0 = MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
-      Register DestSub1 = MRI.createVirtualRegister(&AMDGPU::SReg_32RegClass);
-
-      MachineOperand Src0Sub0 = TII->buildExtractSubRegOrImm(
-          MI, MRI, Src0, BoolRC, AMDGPU::sub0, &AMDGPU::SReg_32RegClass);
-      MachineOperand Src0Sub1 = TII->buildExtractSubRegOrImm(
-          MI, MRI, Src0, BoolRC, AMDGPU::sub1, &AMDGPU::SReg_32RegClass);
-
-      MachineOperand Src1Sub0 = TII->buildExtractSubRegOrImm(
-          MI, MRI, Src1, BoolRC, AMDGPU::sub0, &AMDGPU::SReg_32RegClass);
-      MachineOperand Src1Sub1 = TII->buildExtractSubRegOrImm(
-          MI, MRI, Src1, BoolRC, AMDGPU::sub1, &AMDGPU::SReg_32RegClass);
-
-      unsigned LoOpc = IsAdd ? AMDGPU::S_ADD_U32 : AMDGPU::S_SUB_U32;
-      unsigned HiOpc = IsAdd ? AMDGPU::S_ADDC_U32 : AMDGPU::S_SUBB_U32;
-      BuildMI(*BB, MI, DL, TII->get(LoOpc), DestSub0)
-          .add(Src0Sub0)
-          .add(Src1Sub0);
-      BuildMI(*BB, MI, DL, TII->get(HiOpc), DestSub1)
-          .add(Src0Sub1)
-          .add(Src1Sub1);
-      BuildMI(*BB, MI, DL, TII->get(TargetOpcode::REG_SEQUENCE), Dest.getReg())
-          .addReg(DestSub0)
-          .addImm(AMDGPU::sub0)
-          .addReg(DestSub1)
-          .addImm(AMDGPU::sub1);
-    }
-    MI.eraseFromParent();
-    return BB;
+    return Expand64BitScalarArithmetic(MI, BB);
   }
   case AMDGPU::V_ADD_U64_PSEUDO:
   case AMDGPU::V_SUB_U64_PSEUDO: {

llvm/lib/Target/AMDGPU/SIInstructions.td

@@ -304,28 +304,57 @@ def : GCNPat<(i32 (int_amdgcn_set_inactive_chain_arg i32:$src, i32:$inactive)),
     (V_SET_INACTIVE_B32 0, VGPR_32:$src, 0, VGPR_32:$inactive, (IMPLICIT_DEF))>;
 
 // clang-format off
-defvar int_amdgcn_wave_reduce_ = "int_amdgcn_wave_reduce_";
+
 multiclass
-    AMDGPUWaveReducePseudoGenerator<string Op, string DataType> {
+    AMDGPUWaveReducePseudoGenerator<string Op, string DataType, ValueType ty, RegisterClass RetReg, SrcRegOrImm9 Reg> {
   let usesCustomInserter = 1, hasSideEffects = 0, mayLoad = 0, mayStore = 0, Uses = [EXEC] in {
     def !toupper(Op) #"_PSEUDO_" #DataType
-        : VPseudoInstSI<(outs SGPR_32 : $sdst),
-                        (ins VSrc_b32 : $src, VSrc_b32 : $strategy),
-                        [(set i32 : $sdst, (!cast<AMDGPUWaveReduce>(int_amdgcn_wave_reduce_ #Op) i32 : $src, i32 : $strategy))]> {}
+        : VPseudoInstSI<(outs RetReg : $sdst),
+                        (ins Reg : $src, VSrc_b32 : $strategy),
+                        [(set ty : $sdst, (!cast<AMDGPUWaveReduce>("int_amdgcn_wave_reduce_" #Op) ty : $src, i32 : $strategy))]> {}
   }
 }
 // clang-format on
 
+class WaveReduceOp<string OpName, string TypeStr, ValueType Ty,
+                   RegisterClass ReturnRegisterClass, SrcRegOrImm9 RC> {
+  string Name = OpName;
+  string TypeString = TypeStr;
+  ValueType VT = Ty;
+  RegisterClass RetReg = ReturnRegisterClass;
+  SrcRegOrImm9 Reg = RC;
+}
+
 // Input list : [Operation_name,
-//              type - Signed(I)/Unsigned(U)/Float(F)/Bitwise(B)]
+//              type - Signed(I)/Unsigned(U)/Float(F)/Bitwise(B),
+//              bit-width
+//              output register class,
+//              input register class]
 defvar Operations = [
-  ["umin", "U32"], ["min", "I32"], ["umax", "U32"], ["max", "I32"],
-  ["add", "I32"], ["sub", "I32"], ["and", "B32"], ["or", "B32"],
-  ["xor", "B32"]
+  WaveReduceOp<"umin", "U32", i32, SGPR_32, VSrc_b32>,
+  WaveReduceOp<"min", "I32", i32, SGPR_32, VSrc_b32>,
+  WaveReduceOp<"umax", "U32", i32, SGPR_32, VSrc_b32>,
+  WaveReduceOp<"max", "I32", i32, SGPR_32, VSrc_b32>,
+  WaveReduceOp<"add", "I32", i32, SGPR_32, VSrc_b32>,
+  WaveReduceOp<"sub", "I32", i32, SGPR_32, VSrc_b32>,
+  WaveReduceOp<"and", "B32", i32, SGPR_32, VSrc_b32>,
+  WaveReduceOp<"or", "B32", i32, SGPR_32, VSrc_b32>,
+  WaveReduceOp<"xor", "B32", i32, SGPR_32, VSrc_b32>,
+
+  WaveReduceOp<"umin", "U64", i64, SGPR_64, VSrc_b64>,
+  WaveReduceOp<"min", "I64", i64, SGPR_64, VSrc_b64>,
+  WaveReduceOp<"umax", "U64", i64, SGPR_64, VSrc_b64>,
+  WaveReduceOp<"max", "I64", i64, SGPR_64, VSrc_b64>,
+  WaveReduceOp<"add", "U64", i64, SGPR_64, VSrc_b64>,
+  WaveReduceOp<"sub", "U64", i64, SGPR_64, VSrc_b64>,
+  WaveReduceOp<"and", "B64", i64, SGPR_64, VSrc_b64>,
+  WaveReduceOp<"or", "B64", i64, SGPR_64, VSrc_b64>,
+  WaveReduceOp<"xor", "B64", i64, SGPR_64, VSrc_b64>,
 ];
 
 foreach Op = Operations in {
-  defm WAVE_REDUCE_ : AMDGPUWaveReducePseudoGenerator<Op[0], Op[1]>;
+  defm WAVE_REDUCE_ : AMDGPUWaveReducePseudoGenerator<Op.Name, Op.TypeString,
+                                                      Op.VT, Op.RetReg, Op.Reg>;
 }
 
 let usesCustomInserter = 1, Defs = [VCC] in {

llvm/test/CodeGen/AMDGPU/llvm.amdgcn.reduce.and.ll

@@ -980,3 +980,845 @@ endif:
   store i32 %combine, ptr addrspace(1) %out
   ret void
 }
+
+define amdgpu_kernel void @uniform_value_i64(ptr addrspace(1) %out, i64 %in) {
+; GFX8DAGISEL-LABEL: uniform_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s2
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s0
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s1
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s3
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: uniform_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: uniform_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: uniform_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: uniform_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX10DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: uniform_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX10GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: uniform_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: uniform_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: uniform_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: uniform_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.and.i64(i64 %in, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @const_value_i64(ptr addrspace(1) %out) {
+; GFX8DAGISEL-LABEL: const_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: const_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: const_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: const_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: const_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: const_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: const_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: const_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: const_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v2, 0
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: const_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v2, 0
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.and.i64(i64 123, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @poison_value_i64(ptr addrspace(1) %out, i64 %in) {
+; GFX8DAGISEL-LABEL: poison_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s0
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s1
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[0:1], v[0:1]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: poison_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, s0
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, s1
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[0:1], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: poison_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v0, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: poison_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    global_store_dwordx2 v0, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: poison_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX10DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10DAGISEL-NEXT:    global_store_dwordx2 v0, v[0:1], s[0:1]
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: poison_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX10GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10GISEL-NEXT:    global_store_dwordx2 v0, v[0:1], s[0:1]
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX11DAGISEL-LABEL: poison_value_i64:
+; GFX11DAGISEL:       ; %bb.0: ; %entry
+; GFX11DAGISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX11DAGISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX11DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX11DAGISEL-NEXT:    global_store_b64 v0, v[0:1], s[0:1]
+; GFX11DAGISEL-NEXT:    s_endpgm
+;
+; GFX11GISEL-LABEL: poison_value_i64:
+; GFX11GISEL:       ; %bb.0: ; %entry
+; GFX11GISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX11GISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX11GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX11GISEL-NEXT:    global_store_b64 v0, v[0:1], s[0:1]
+; GFX11GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.and.i64(i64 poison, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define void @divergent_value_i64(ptr addrspace(1) %out, i64 %id.x) {
+; GFX8DAGISEL-LABEL: divergent_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    s_mov_b64 s[4:5], -1
+; GFX8DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX8DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX8DAGISEL-NEXT:    s_ff1_i32_b64 s10, s[6:7]
+; GFX8DAGISEL-NEXT:    v_readlane_b32 s8, v2, s10
+; GFX8DAGISEL-NEXT:    v_readlane_b32 s9, v3, s10
+; GFX8DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s10
+; GFX8DAGISEL-NEXT:    s_and_b64 s[4:5], s[4:5], s[8:9]
+; GFX8DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX8DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX8DAGISEL-NEXT:  ; %bb.2:
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8DAGISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX8DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX8GISEL-LABEL: divergent_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b64 s[4:5], -1
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX8GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX8GISEL-NEXT:    s_ff1_i32_b64 s10, s[6:7]
+; GFX8GISEL-NEXT:    v_readlane_b32 s8, v2, s10
+; GFX8GISEL-NEXT:    v_readlane_b32 s9, v3, s10
+; GFX8GISEL-NEXT:    s_bitset0_b64 s[6:7], s10
+; GFX8GISEL-NEXT:    s_and_b64 s[4:5], s[4:5], s[8:9]
+; GFX8GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX8GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX8GISEL-NEXT:  ; %bb.2:
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8GISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX8GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX9DAGISEL-LABEL: divergent_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    s_mov_b64 s[4:5], -1
+; GFX9DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX9DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX9DAGISEL-NEXT:    s_ff1_i32_b64 s10, s[6:7]
+; GFX9DAGISEL-NEXT:    v_readlane_b32 s8, v2, s10
+; GFX9DAGISEL-NEXT:    v_readlane_b32 s9, v3, s10
+; GFX9DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s10
+; GFX9DAGISEL-NEXT:    s_and_b64 s[4:5], s[4:5], s[8:9]
+; GFX9DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX9DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX9DAGISEL-NEXT:  ; %bb.2:
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX9DAGISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX9DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX9GISEL-LABEL: divergent_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b64 s[4:5], -1
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX9GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX9GISEL-NEXT:    s_ff1_i32_b64 s10, s[6:7]
+; GFX9GISEL-NEXT:    v_readlane_b32 s8, v2, s10
+; GFX9GISEL-NEXT:    v_readlane_b32 s9, v3, s10
+; GFX9GISEL-NEXT:    s_bitset0_b64 s[6:7], s10
+; GFX9GISEL-NEXT:    s_and_b64 s[4:5], s[4:5], s[8:9]
+; GFX9GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX9GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX9GISEL-NEXT:  ; %bb.2:
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX9GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX9GISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX9GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1064DAGISEL-LABEL: divergent_value_i64:
+; GFX1064DAGISEL:       ; %bb.0: ; %entry
+; GFX1064DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1064DAGISEL-NEXT:    s_mov_b64 s[4:5], -1
+; GFX1064DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1064DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1064DAGISEL-NEXT:    s_ff1_i32_b64 s10, s[6:7]
+; GFX1064DAGISEL-NEXT:    v_readlane_b32 s8, v2, s10
+; GFX1064DAGISEL-NEXT:    v_readlane_b32 s9, v3, s10
+; GFX1064DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s10
+; GFX1064DAGISEL-NEXT:    s_and_b64 s[4:5], s[4:5], s[8:9]
+; GFX1064DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX1064DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1064DAGISEL-NEXT:  ; %bb.2:
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1064DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1064DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1064GISEL-LABEL: divergent_value_i64:
+; GFX1064GISEL:       ; %bb.0: ; %entry
+; GFX1064GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b64 s[4:5], -1
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1064GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1064GISEL-NEXT:    s_ff1_i32_b64 s10, s[6:7]
+; GFX1064GISEL-NEXT:    v_readlane_b32 s8, v2, s10
+; GFX1064GISEL-NEXT:    v_readlane_b32 s9, v3, s10
+; GFX1064GISEL-NEXT:    s_bitset0_b64 s[6:7], s10
+; GFX1064GISEL-NEXT:    s_and_b64 s[4:5], s[4:5], s[8:9]
+; GFX1064GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX1064GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1064GISEL-NEXT:  ; %bb.2:
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1064GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1064GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1032DAGISEL-LABEL: divergent_value_i64:
+; GFX1032DAGISEL:       ; %bb.0: ; %entry
+; GFX1032DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1032DAGISEL-NEXT:    s_mov_b64 s[4:5], -1
+; GFX1032DAGISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1032DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1032DAGISEL-NEXT:    s_ff1_i32_b32 s7, s6
+; GFX1032DAGISEL-NEXT:    v_readlane_b32 s8, v2, s7
+; GFX1032DAGISEL-NEXT:    v_readlane_b32 s9, v3, s7
+; GFX1032DAGISEL-NEXT:    s_bitset0_b32 s6, s7
+; GFX1032DAGISEL-NEXT:    s_and_b64 s[4:5], s[4:5], s[8:9]
+; GFX1032DAGISEL-NEXT:    s_cmp_lg_u32 s6, 0
+; GFX1032DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1032DAGISEL-NEXT:  ; %bb.2:
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1032DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1032DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1032GISEL-LABEL: divergent_value_i64:
+; GFX1032GISEL:       ; %bb.0: ; %entry
+; GFX1032GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b64 s[4:5], -1
+; GFX1032GISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1032GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1032GISEL-NEXT:    s_ff1_i32_b32 s7, s6
+; GFX1032GISEL-NEXT:    v_readlane_b32 s8, v2, s7
+; GFX1032GISEL-NEXT:    v_readlane_b32 s9, v3, s7
+; GFX1032GISEL-NEXT:    s_bitset0_b32 s6, s7
+; GFX1032GISEL-NEXT:    s_and_b64 s[4:5], s[4:5], s[8:9]
+; GFX1032GISEL-NEXT:    s_cmp_lg_u32 s6, 0
+; GFX1032GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1032GISEL-NEXT:  ; %bb.2:
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1032GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1032GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1164DAGISEL-LABEL: divergent_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    s_mov_b64 s[0:1], -1
+; GFX1164DAGISEL-NEXT:    s_mov_b64 s[2:3], exec
+; GFX1164DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1164DAGISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1)
+; GFX1164DAGISEL-NEXT:    s_ctz_i32_b64 s6, s[2:3]
+; GFX1164DAGISEL-NEXT:    v_readlane_b32 s4, v2, s6
+; GFX1164DAGISEL-NEXT:    v_readlane_b32 s5, v3, s6
+; GFX1164DAGISEL-NEXT:    s_bitset0_b64 s[2:3], s6
+; GFX1164DAGISEL-NEXT:    s_and_b64 s[0:1], s[0:1], s[4:5]
+; GFX1164DAGISEL-NEXT:    s_cmp_lg_u64 s[2:3], 0
+; GFX1164DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1164DAGISEL-NEXT:  ; %bb.2:
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX1164DAGISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1164DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1164GISEL-LABEL: divergent_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b64 s[0:1], -1
+; GFX1164GISEL-NEXT:    s_mov_b64 s[2:3], exec
+; GFX1164GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1164GISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1)
+; GFX1164GISEL-NEXT:    s_ctz_i32_b64 s6, s[2:3]
+; GFX1164GISEL-NEXT:    v_readlane_b32 s4, v2, s6
+; GFX1164GISEL-NEXT:    v_readlane_b32 s5, v3, s6
+; GFX1164GISEL-NEXT:    s_bitset0_b64 s[2:3], s6
+; GFX1164GISEL-NEXT:    s_and_b64 s[0:1], s[0:1], s[4:5]
+; GFX1164GISEL-NEXT:    s_cmp_lg_u64 s[2:3], 0
+; GFX1164GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1164GISEL-NEXT:  ; %bb.2:
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX1164GISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1164GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1132DAGISEL-LABEL: divergent_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    s_mov_b64 s[0:1], -1
+; GFX1132DAGISEL-NEXT:    s_mov_b32 s2, exec_lo
+; GFX1132DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1132DAGISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1)
+; GFX1132DAGISEL-NEXT:    s_ctz_i32_b32 s3, s2
+; GFX1132DAGISEL-NEXT:    v_readlane_b32 s4, v2, s3
+; GFX1132DAGISEL-NEXT:    v_readlane_b32 s5, v3, s3
+; GFX1132DAGISEL-NEXT:    s_bitset0_b32 s2, s3
+; GFX1132DAGISEL-NEXT:    s_and_b64 s[0:1], s[0:1], s[4:5]
+; GFX1132DAGISEL-NEXT:    s_cmp_lg_u32 s2, 0
+; GFX1132DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1132DAGISEL-NEXT:  ; %bb.2:
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v3, s1 :: v_dual_mov_b32 v2, s0
+; GFX1132DAGISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1132DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1132GISEL-LABEL: divergent_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b64 s[0:1], -1
+; GFX1132GISEL-NEXT:    s_mov_b32 s2, exec_lo
+; GFX1132GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1132GISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1)
+; GFX1132GISEL-NEXT:    s_ctz_i32_b32 s3, s2
+; GFX1132GISEL-NEXT:    v_readlane_b32 s4, v2, s3
+; GFX1132GISEL-NEXT:    v_readlane_b32 s5, v3, s3
+; GFX1132GISEL-NEXT:    s_bitset0_b32 s2, s3
+; GFX1132GISEL-NEXT:    s_and_b64 s[0:1], s[0:1], s[4:5]
+; GFX1132GISEL-NEXT:    s_cmp_lg_u32 s2, 0
+; GFX1132GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1132GISEL-NEXT:  ; %bb.2:
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v3, s1 :: v_dual_mov_b32 v2, s0
+; GFX1132GISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1132GISEL-NEXT:    s_setpc_b64 s[30:31]
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.and.i64(i64 %id.x, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @divergent_cfg_i64(ptr addrspace(1) %out, i64 %in, i64 %in2) {
+; GFX8DAGISEL-LABEL: divergent_cfg_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x34
+; GFX8DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX8DAGISEL-NEXT:    s_and_saveexec_b64 s[6:7], vcc
+; GFX8DAGISEL-NEXT:    s_xor_b64 s[6:7], exec, s[6:7]
+; GFX8DAGISEL-NEXT:    s_or_saveexec_b64 s[6:7], s[6:7]
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX8DAGISEL-NEXT:    s_xor_b64 exec, exec, s[6:7]
+; GFX8DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s4
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s5
+; GFX8DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX8DAGISEL-NEXT:    s_or_b64 exec, exec, s[6:7]
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: divergent_cfg_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX8GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX8GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX8GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX8GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX8GISEL-NEXT:  ; %bb.1: ; %else
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX8GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX8GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX8GISEL-NEXT:  ; %bb.3: ; %if
+; GFX8GISEL-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x34
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX8GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX8GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: divergent_cfg_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX9DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX9DAGISEL-NEXT:    s_and_saveexec_b64 s[4:5], vcc
+; GFX9DAGISEL-NEXT:    s_xor_b64 s[4:5], exec, s[4:5]
+; GFX9DAGISEL-NEXT:    s_or_saveexec_b64 s[4:5], s[4:5]
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9DAGISEL-NEXT:    s_xor_b64 exec, exec, s[4:5]
+; GFX9DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX9DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX9DAGISEL-NEXT:    s_or_b64 exec, exec, s[4:5]
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: divergent_cfg_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX9GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX9GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX9GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX9GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX9GISEL-NEXT:  ; %bb.1: ; %else
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX9GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX9GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX9GISEL-NEXT:  ; %bb.3: ; %if
+; GFX9GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX9GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX9GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX1064DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1064DAGISEL:       ; %bb.0: ; %entry
+; GFX1064DAGISEL-NEXT:    s_clause 0x1
+; GFX1064DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1064DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1064DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX1064DAGISEL-NEXT:    s_and_saveexec_b64 s[4:5], vcc
+; GFX1064DAGISEL-NEXT:    s_xor_b64 s[4:5], exec, s[4:5]
+; GFX1064DAGISEL-NEXT:    s_or_saveexec_b64 s[4:5], s[4:5]
+; GFX1064DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1064DAGISEL-NEXT:    s_xor_b64 exec, exec, s[4:5]
+; GFX1064DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1064DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1064DAGISEL-NEXT:    s_or_b64 exec, exec, s[4:5]
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1064DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1064DAGISEL-NEXT:    s_endpgm
+;
+; GFX1064GISEL-LABEL: divergent_cfg_i64:
+; GFX1064GISEL:       ; %bb.0: ; %entry
+; GFX1064GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1064GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX1064GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1064GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX1064GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX1064GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1064GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1064GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX1064GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1064GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1064GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX1064GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1064GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1064GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1064GISEL-NEXT:    s_endpgm
+;
+; GFX1032DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1032DAGISEL:       ; %bb.0: ; %entry
+; GFX1032DAGISEL-NEXT:    s_clause 0x1
+; GFX1032DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1032DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1032DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc_lo, 15, v0
+; GFX1032DAGISEL-NEXT:    s_and_saveexec_b32 s4, vcc_lo
+; GFX1032DAGISEL-NEXT:    s_xor_b32 s4, exec_lo, s4
+; GFX1032DAGISEL-NEXT:    s_or_saveexec_b32 s4, s4
+; GFX1032DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1032DAGISEL-NEXT:    s_xor_b32 exec_lo, exec_lo, s4
+; GFX1032DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1032DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1032DAGISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s4
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1032DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1032DAGISEL-NEXT:    s_endpgm
+;
+; GFX1032GISEL-LABEL: divergent_cfg_i64:
+; GFX1032GISEL:       ; %bb.0: ; %entry
+; GFX1032GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1032GISEL-NEXT:    v_cmp_le_u32_e32 vcc_lo, 16, v0
+; GFX1032GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1032GISEL-NEXT:    s_and_saveexec_b32 s8, vcc_lo
+; GFX1032GISEL-NEXT:    s_xor_b32 s8, exec_lo, s8
+; GFX1032GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1032GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1032GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_andn2_saveexec_b32 s2, s8
+; GFX1032GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1032GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1032GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX1032GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1032GISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s2
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1032GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1032GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_clause 0x1
+; GFX1164DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1164DAGISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1164DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1164DAGISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1)
+; GFX1164DAGISEL-NEXT:    v_cmpx_lt_u32_e32 15, v0
+; GFX1164DAGISEL-NEXT:    s_xor_b64 s[6:7], exec, s[6:7]
+; GFX1164DAGISEL-NEXT:    s_or_saveexec_b64 s[6:7], s[6:7]
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164DAGISEL-NEXT:    s_xor_b64 exec, exec, s[6:7]
+; GFX1164DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s4
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s5
+; GFX1164DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1164DAGISEL-NEXT:    s_or_b64 exec, exec, s[6:7]
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: divergent_cfg_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1164GISEL-NEXT:    s_mov_b64 s[8:9], exec
+; GFX1164GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1164GISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1)
+; GFX1164GISEL-NEXT:    v_cmpx_le_u32_e32 16, v0
+; GFX1164GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX1164GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1164GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1164GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_and_not1_saveexec_b64 s[2:3], s[8:9]
+; GFX1164GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1164GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1164GISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX1164GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1164GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_clause 0x1
+; GFX1132DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1132DAGISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1132DAGISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1132DAGISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1)
+; GFX1132DAGISEL-NEXT:    v_cmpx_lt_u32_e32 15, v0
+; GFX1132DAGISEL-NEXT:    s_xor_b32 s6, exec_lo, s6
+; GFX1132DAGISEL-NEXT:    s_or_saveexec_b32 s6, s6
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132DAGISEL-NEXT:    s_xor_b32 exec_lo, exec_lo, s6
+; GFX1132DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s4 :: v_dual_mov_b32 v1, s5
+; GFX1132DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1132DAGISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s6
+; GFX1132DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: divergent_cfg_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1132GISEL-NEXT:    s_mov_b32 s8, exec_lo
+; GFX1132GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1132GISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1)
+; GFX1132GISEL-NEXT:    v_cmpx_le_u32_e32 16, v0
+; GFX1132GISEL-NEXT:    s_xor_b32 s8, exec_lo, s8
+; GFX1132GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1132GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1132GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_and_not1_saveexec_b32 s2, s8
+; GFX1132GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1132GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1132GISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX1132GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1132GISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s2
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v0, s6 :: v_dual_mov_b32 v1, s7
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %tid = call i32 @llvm.amdgcn.workitem.id.x()
+  %d_cmp = icmp ult i32 %tid, 16
+  br i1 %d_cmp, label %if, label %else
+
+if:
+  %reducedValTid = call i64 @llvm.amdgcn.wave.reduce.and.i64(i64 %in2, i32 1)
+  br label %endif
+
+else:
+  %reducedValIn = call i64 @llvm.amdgcn.wave.reduce.and.i64(i64 %in, i32 1)
+  br label %endif
+
+endif:
+  %combine = phi i64 [%reducedValTid, %if], [%reducedValIn, %else]
+  store i64 %combine, ptr addrspace(1) %out
+  ret void
+}

llvm/test/CodeGen/AMDGPU/llvm.amdgcn.reduce.max.ll

@@ -980,3 +980,903 @@ endif:
   store i32 %combine, ptr addrspace(1) %out
   ret void
 }
+
+define amdgpu_kernel void @uniform_value_i64(ptr addrspace(1) %out, i64 %in) {
+; GFX8DAGISEL-LABEL: uniform_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s2
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s0
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s1
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s3
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: uniform_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: uniform_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: uniform_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: uniform_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX10DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: uniform_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX10GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: uniform_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: uniform_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: uniform_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: uniform_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.max.i64(i64 %in, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @const_value_i64(ptr addrspace(1) %out) {
+; GFX8DAGISEL-LABEL: const_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: const_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: const_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: const_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: const_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: const_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: const_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: const_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: const_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v2, 0
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: const_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v2, 0
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.max.i64(i64 123, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @poison_value_i64(ptr addrspace(1) %out, i64 %in) {
+; GFX8DAGISEL-LABEL: poison_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s0
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s1
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[0:1], v[0:1]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: poison_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, s0
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, s1
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[0:1], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: poison_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v0, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: poison_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    global_store_dwordx2 v0, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: poison_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX10DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10DAGISEL-NEXT:    global_store_dwordx2 v0, v[0:1], s[0:1]
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: poison_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX10GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10GISEL-NEXT:    global_store_dwordx2 v0, v[0:1], s[0:1]
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX11DAGISEL-LABEL: poison_value_i64:
+; GFX11DAGISEL:       ; %bb.0: ; %entry
+; GFX11DAGISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX11DAGISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX11DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX11DAGISEL-NEXT:    global_store_b64 v0, v[0:1], s[0:1]
+; GFX11DAGISEL-NEXT:    s_endpgm
+;
+; GFX11GISEL-LABEL: poison_value_i64:
+; GFX11GISEL:       ; %bb.0: ; %entry
+; GFX11GISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX11GISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX11GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX11GISEL-NEXT:    global_store_b64 v0, v[0:1], s[0:1]
+; GFX11GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.max.i64(i64 poison, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define void @divergent_value_i64(ptr addrspace(1) %out, i64 %id.x) {
+; GFX8DAGISEL-LABEL: divergent_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    s_mov_b32 s4, 0
+; GFX8DAGISEL-NEXT:    s_brev_b32 s5, 1
+; GFX8DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX8DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX8DAGISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX8DAGISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX8DAGISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX8DAGISEL-NEXT:    v_cmp_gt_i64_e32 vcc, s[8:9], v[4:5]
+; GFX8DAGISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX8DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX8DAGISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX8DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX8DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX8DAGISEL-NEXT:  ; %bb.2:
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8DAGISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX8DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX8GISEL-LABEL: divergent_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b32 s4, 0
+; GFX8GISEL-NEXT:    s_brev_b32 s5, 1
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX8GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX8GISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX8GISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX8GISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX8GISEL-NEXT:    v_cmp_gt_i64_e32 vcc, s[8:9], v[4:5]
+; GFX8GISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX8GISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX8GISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX8GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX8GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX8GISEL-NEXT:  ; %bb.2:
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8GISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX8GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX9DAGISEL-LABEL: divergent_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    s_mov_b32 s4, 0
+; GFX9DAGISEL-NEXT:    s_brev_b32 s5, 1
+; GFX9DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX9DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX9DAGISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX9DAGISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX9DAGISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX9DAGISEL-NEXT:    v_cmp_gt_i64_e32 vcc, s[8:9], v[4:5]
+; GFX9DAGISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX9DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX9DAGISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX9DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX9DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX9DAGISEL-NEXT:  ; %bb.2:
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX9DAGISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX9DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX9GISEL-LABEL: divergent_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b32 s4, 0
+; GFX9GISEL-NEXT:    s_brev_b32 s5, 1
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX9GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX9GISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX9GISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX9GISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX9GISEL-NEXT:    v_cmp_gt_i64_e32 vcc, s[8:9], v[4:5]
+; GFX9GISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX9GISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX9GISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX9GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX9GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX9GISEL-NEXT:  ; %bb.2:
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX9GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX9GISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX9GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1064DAGISEL-LABEL: divergent_value_i64:
+; GFX1064DAGISEL:       ; %bb.0: ; %entry
+; GFX1064DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1064DAGISEL-NEXT:    s_mov_b32 s4, 0
+; GFX1064DAGISEL-NEXT:    s_brev_b32 s5, 1
+; GFX1064DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1064DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1064DAGISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX1064DAGISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX1064DAGISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX1064DAGISEL-NEXT:    v_cmp_gt_i64_e32 vcc, s[8:9], v[4:5]
+; GFX1064DAGISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX1064DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX1064DAGISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX1064DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX1064DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1064DAGISEL-NEXT:  ; %bb.2:
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1064DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1064DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1064GISEL-LABEL: divergent_value_i64:
+; GFX1064GISEL:       ; %bb.0: ; %entry
+; GFX1064GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b32 s4, 0
+; GFX1064GISEL-NEXT:    s_brev_b32 s5, 1
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1064GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1064GISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX1064GISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX1064GISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX1064GISEL-NEXT:    v_cmp_gt_i64_e32 vcc, s[8:9], v[4:5]
+; GFX1064GISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX1064GISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX1064GISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX1064GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX1064GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1064GISEL-NEXT:  ; %bb.2:
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1064GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1064GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1032DAGISEL-LABEL: divergent_value_i64:
+; GFX1032DAGISEL:       ; %bb.0: ; %entry
+; GFX1032DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1032DAGISEL-NEXT:    s_mov_b32 s4, 0
+; GFX1032DAGISEL-NEXT:    s_brev_b32 s5, 1
+; GFX1032DAGISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1032DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1032DAGISEL-NEXT:    s_ff1_i32_b32 s7, s6
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX1032DAGISEL-NEXT:    v_readlane_b32 s8, v2, s7
+; GFX1032DAGISEL-NEXT:    v_readlane_b32 s9, v3, s7
+; GFX1032DAGISEL-NEXT:    v_cmp_gt_i64_e32 vcc_lo, s[8:9], v[4:5]
+; GFX1032DAGISEL-NEXT:    s_and_b32 s10, vcc_lo, s6
+; GFX1032DAGISEL-NEXT:    s_bitset0_b32 s6, s7
+; GFX1032DAGISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX1032DAGISEL-NEXT:    s_cmp_lg_u32 s6, 0
+; GFX1032DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1032DAGISEL-NEXT:  ; %bb.2:
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1032DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1032DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1032GISEL-LABEL: divergent_value_i64:
+; GFX1032GISEL:       ; %bb.0: ; %entry
+; GFX1032GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b32 s4, 0
+; GFX1032GISEL-NEXT:    s_brev_b32 s5, 1
+; GFX1032GISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1032GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1032GISEL-NEXT:    s_ff1_i32_b32 s7, s6
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX1032GISEL-NEXT:    v_readlane_b32 s8, v2, s7
+; GFX1032GISEL-NEXT:    v_readlane_b32 s9, v3, s7
+; GFX1032GISEL-NEXT:    v_cmp_gt_i64_e32 vcc_lo, s[8:9], v[4:5]
+; GFX1032GISEL-NEXT:    s_and_b32 s10, vcc_lo, s6
+; GFX1032GISEL-NEXT:    s_bitset0_b32 s6, s7
+; GFX1032GISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX1032GISEL-NEXT:    s_cmp_lg_u32 s6, 0
+; GFX1032GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1032GISEL-NEXT:  ; %bb.2:
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1032GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1032GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1164DAGISEL-LABEL: divergent_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    s_mov_b32 s0, 0
+; GFX1164DAGISEL-NEXT:    s_brev_b32 s1, 1
+; GFX1164DAGISEL-NEXT:    s_mov_b64 s[2:3], exec
+; GFX1164DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1164DAGISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_4) | instid1(VALU_DEP_1)
+; GFX1164DAGISEL-NEXT:    s_ctz_i32_b64 s8, s[2:3]
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v4, s0
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v5, s1
+; GFX1164DAGISEL-NEXT:    v_readlane_b32 s4, v2, s8
+; GFX1164DAGISEL-NEXT:    v_readlane_b32 s5, v3, s8
+; GFX1164DAGISEL-NEXT:    v_cmp_gt_i64_e32 vcc, s[4:5], v[4:5]
+; GFX1164DAGISEL-NEXT:    s_and_b64 s[6:7], vcc, s[2:3]
+; GFX1164DAGISEL-NEXT:    s_bitset0_b64 s[2:3], s8
+; GFX1164DAGISEL-NEXT:    s_cselect_b64 s[0:1], s[4:5], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_cmp_lg_u64 s[2:3], 0
+; GFX1164DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1164DAGISEL-NEXT:  ; %bb.2:
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX1164DAGISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1164DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1164GISEL-LABEL: divergent_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b32 s0, 0
+; GFX1164GISEL-NEXT:    s_brev_b32 s1, 1
+; GFX1164GISEL-NEXT:    s_mov_b64 s[2:3], exec
+; GFX1164GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1164GISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_4) | instid1(VALU_DEP_1)
+; GFX1164GISEL-NEXT:    s_ctz_i32_b64 s8, s[2:3]
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v4, s0
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v5, s1
+; GFX1164GISEL-NEXT:    v_readlane_b32 s4, v2, s8
+; GFX1164GISEL-NEXT:    v_readlane_b32 s5, v3, s8
+; GFX1164GISEL-NEXT:    v_cmp_gt_i64_e32 vcc, s[4:5], v[4:5]
+; GFX1164GISEL-NEXT:    s_and_b64 s[6:7], vcc, s[2:3]
+; GFX1164GISEL-NEXT:    s_bitset0_b64 s[2:3], s8
+; GFX1164GISEL-NEXT:    s_cselect_b64 s[0:1], s[4:5], s[0:1]
+; GFX1164GISEL-NEXT:    s_cmp_lg_u64 s[2:3], 0
+; GFX1164GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1164GISEL-NEXT:  ; %bb.2:
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX1164GISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1164GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1132DAGISEL-LABEL: divergent_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    s_mov_b32 s0, 0
+; GFX1132DAGISEL-NEXT:    s_brev_b32 s1, 1
+; GFX1132DAGISEL-NEXT:    s_mov_b32 s2, exec_lo
+; GFX1132DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1132DAGISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_3) | instid1(VALU_DEP_1)
+; GFX1132DAGISEL-NEXT:    s_ctz_i32_b32 s3, s2
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v4, s0 :: v_dual_mov_b32 v5, s1
+; GFX1132DAGISEL-NEXT:    v_readlane_b32 s4, v2, s3
+; GFX1132DAGISEL-NEXT:    v_readlane_b32 s5, v3, s3
+; GFX1132DAGISEL-NEXT:    v_cmp_gt_i64_e32 vcc_lo, s[4:5], v[4:5]
+; GFX1132DAGISEL-NEXT:    s_and_b32 s6, vcc_lo, s2
+; GFX1132DAGISEL-NEXT:    s_bitset0_b32 s2, s3
+; GFX1132DAGISEL-NEXT:    s_cselect_b64 s[0:1], s[4:5], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_cmp_lg_u32 s2, 0
+; GFX1132DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1132DAGISEL-NEXT:  ; %bb.2:
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v3, s1 :: v_dual_mov_b32 v2, s0
+; GFX1132DAGISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1132DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1132GISEL-LABEL: divergent_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b32 s0, 0
+; GFX1132GISEL-NEXT:    s_brev_b32 s1, 1
+; GFX1132GISEL-NEXT:    s_mov_b32 s2, exec_lo
+; GFX1132GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1132GISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_3) | instid1(VALU_DEP_1)
+; GFX1132GISEL-NEXT:    s_ctz_i32_b32 s3, s2
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v4, s0 :: v_dual_mov_b32 v5, s1
+; GFX1132GISEL-NEXT:    v_readlane_b32 s4, v2, s3
+; GFX1132GISEL-NEXT:    v_readlane_b32 s5, v3, s3
+; GFX1132GISEL-NEXT:    v_cmp_gt_i64_e32 vcc_lo, s[4:5], v[4:5]
+; GFX1132GISEL-NEXT:    s_and_b32 s6, vcc_lo, s2
+; GFX1132GISEL-NEXT:    s_bitset0_b32 s2, s3
+; GFX1132GISEL-NEXT:    s_cselect_b64 s[0:1], s[4:5], s[0:1]
+; GFX1132GISEL-NEXT:    s_cmp_lg_u32 s2, 0
+; GFX1132GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1132GISEL-NEXT:  ; %bb.2:
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v3, s1 :: v_dual_mov_b32 v2, s0
+; GFX1132GISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1132GISEL-NEXT:    s_setpc_b64 s[30:31]
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.max.i64(i64 %id.x, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @divergent_cfg_i64(ptr addrspace(1) %out, i64 %in, i64 %in2) {
+; GFX8DAGISEL-LABEL: divergent_cfg_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x34
+; GFX8DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX8DAGISEL-NEXT:    s_and_saveexec_b64 s[6:7], vcc
+; GFX8DAGISEL-NEXT:    s_xor_b64 s[6:7], exec, s[6:7]
+; GFX8DAGISEL-NEXT:    s_or_saveexec_b64 s[6:7], s[6:7]
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX8DAGISEL-NEXT:    s_xor_b64 exec, exec, s[6:7]
+; GFX8DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s4
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s5
+; GFX8DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX8DAGISEL-NEXT:    s_or_b64 exec, exec, s[6:7]
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: divergent_cfg_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX8GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX8GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX8GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX8GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX8GISEL-NEXT:  ; %bb.1: ; %else
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX8GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX8GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX8GISEL-NEXT:  ; %bb.3: ; %if
+; GFX8GISEL-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x34
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX8GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX8GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: divergent_cfg_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX9DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX9DAGISEL-NEXT:    s_and_saveexec_b64 s[4:5], vcc
+; GFX9DAGISEL-NEXT:    s_xor_b64 s[4:5], exec, s[4:5]
+; GFX9DAGISEL-NEXT:    s_or_saveexec_b64 s[4:5], s[4:5]
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9DAGISEL-NEXT:    s_xor_b64 exec, exec, s[4:5]
+; GFX9DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX9DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX9DAGISEL-NEXT:    s_or_b64 exec, exec, s[4:5]
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: divergent_cfg_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX9GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX9GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX9GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX9GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX9GISEL-NEXT:  ; %bb.1: ; %else
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX9GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX9GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX9GISEL-NEXT:  ; %bb.3: ; %if
+; GFX9GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX9GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX9GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX1064DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1064DAGISEL:       ; %bb.0: ; %entry
+; GFX1064DAGISEL-NEXT:    s_clause 0x1
+; GFX1064DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1064DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1064DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX1064DAGISEL-NEXT:    s_and_saveexec_b64 s[4:5], vcc
+; GFX1064DAGISEL-NEXT:    s_xor_b64 s[4:5], exec, s[4:5]
+; GFX1064DAGISEL-NEXT:    s_or_saveexec_b64 s[4:5], s[4:5]
+; GFX1064DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1064DAGISEL-NEXT:    s_xor_b64 exec, exec, s[4:5]
+; GFX1064DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1064DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1064DAGISEL-NEXT:    s_or_b64 exec, exec, s[4:5]
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1064DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1064DAGISEL-NEXT:    s_endpgm
+;
+; GFX1064GISEL-LABEL: divergent_cfg_i64:
+; GFX1064GISEL:       ; %bb.0: ; %entry
+; GFX1064GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1064GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX1064GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1064GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX1064GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX1064GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1064GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1064GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX1064GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1064GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1064GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX1064GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1064GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1064GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1064GISEL-NEXT:    s_endpgm
+;
+; GFX1032DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1032DAGISEL:       ; %bb.0: ; %entry
+; GFX1032DAGISEL-NEXT:    s_clause 0x1
+; GFX1032DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1032DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1032DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc_lo, 15, v0
+; GFX1032DAGISEL-NEXT:    s_and_saveexec_b32 s4, vcc_lo
+; GFX1032DAGISEL-NEXT:    s_xor_b32 s4, exec_lo, s4
+; GFX1032DAGISEL-NEXT:    s_or_saveexec_b32 s4, s4
+; GFX1032DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1032DAGISEL-NEXT:    s_xor_b32 exec_lo, exec_lo, s4
+; GFX1032DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1032DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1032DAGISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s4
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1032DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1032DAGISEL-NEXT:    s_endpgm
+;
+; GFX1032GISEL-LABEL: divergent_cfg_i64:
+; GFX1032GISEL:       ; %bb.0: ; %entry
+; GFX1032GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1032GISEL-NEXT:    v_cmp_le_u32_e32 vcc_lo, 16, v0
+; GFX1032GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1032GISEL-NEXT:    s_and_saveexec_b32 s8, vcc_lo
+; GFX1032GISEL-NEXT:    s_xor_b32 s8, exec_lo, s8
+; GFX1032GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1032GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1032GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_andn2_saveexec_b32 s2, s8
+; GFX1032GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1032GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1032GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX1032GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1032GISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s2
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1032GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1032GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_clause 0x1
+; GFX1164DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1164DAGISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1164DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1164DAGISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1)
+; GFX1164DAGISEL-NEXT:    v_cmpx_lt_u32_e32 15, v0
+; GFX1164DAGISEL-NEXT:    s_xor_b64 s[6:7], exec, s[6:7]
+; GFX1164DAGISEL-NEXT:    s_or_saveexec_b64 s[6:7], s[6:7]
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164DAGISEL-NEXT:    s_xor_b64 exec, exec, s[6:7]
+; GFX1164DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s4
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s5
+; GFX1164DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1164DAGISEL-NEXT:    s_or_b64 exec, exec, s[6:7]
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: divergent_cfg_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1164GISEL-NEXT:    s_mov_b64 s[8:9], exec
+; GFX1164GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1164GISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1)
+; GFX1164GISEL-NEXT:    v_cmpx_le_u32_e32 16, v0
+; GFX1164GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX1164GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1164GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1164GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_and_not1_saveexec_b64 s[2:3], s[8:9]
+; GFX1164GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1164GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1164GISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX1164GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1164GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_clause 0x1
+; GFX1132DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1132DAGISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1132DAGISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1132DAGISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1)
+; GFX1132DAGISEL-NEXT:    v_cmpx_lt_u32_e32 15, v0
+; GFX1132DAGISEL-NEXT:    s_xor_b32 s6, exec_lo, s6
+; GFX1132DAGISEL-NEXT:    s_or_saveexec_b32 s6, s6
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132DAGISEL-NEXT:    s_xor_b32 exec_lo, exec_lo, s6
+; GFX1132DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s4 :: v_dual_mov_b32 v1, s5
+; GFX1132DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1132DAGISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s6
+; GFX1132DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: divergent_cfg_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1132GISEL-NEXT:    s_mov_b32 s8, exec_lo
+; GFX1132GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1132GISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1)
+; GFX1132GISEL-NEXT:    v_cmpx_le_u32_e32 16, v0
+; GFX1132GISEL-NEXT:    s_xor_b32 s8, exec_lo, s8
+; GFX1132GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1132GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1132GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_and_not1_saveexec_b32 s2, s8
+; GFX1132GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1132GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1132GISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX1132GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1132GISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s2
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v0, s6 :: v_dual_mov_b32 v1, s7
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %tid = call i32 @llvm.amdgcn.workitem.id.x()
+  %d_cmp = icmp ult i32 %tid, 16
+  br i1 %d_cmp, label %if, label %else
+
+if:
+  %reducedValTid = call i64 @llvm.amdgcn.wave.reduce.max.i64(i64 %in2, i32 1)
+  br label %endif
+
+else:
+  %reducedValIn = call i64 @llvm.amdgcn.wave.reduce.max.i64(i64 %in, i32 1)
+  br label %endif
+
+endif:
+  %combine = phi i64 [%reducedValTid, %if], [%reducedValIn, %else]
+  store i64 %combine, ptr addrspace(1) %out
+  ret void
+}

llvm/test/CodeGen/AMDGPU/llvm.amdgcn.reduce.min.ll

@@ -980,3 +980,903 @@ endif:
   store i32 %combine, ptr addrspace(1) %out
   ret void
 }
+
+define amdgpu_kernel void @uniform_value_i64(ptr addrspace(1) %out, i64 %in) {
+; GFX8DAGISEL-LABEL: uniform_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s2
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s0
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s1
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s3
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: uniform_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: uniform_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: uniform_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: uniform_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX10DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: uniform_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX10GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: uniform_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: uniform_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: uniform_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: uniform_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.min.i64(i64 %in, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @const_value_i64(ptr addrspace(1) %out) {
+; GFX8DAGISEL-LABEL: const_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: const_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: const_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: const_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: const_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: const_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: const_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: const_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: const_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v2, 0
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: const_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v2, 0
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.min.i64(i64 123, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @poison_value_i64(ptr addrspace(1) %out, i64 %in) {
+; GFX8DAGISEL-LABEL: poison_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s0
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s1
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[0:1], v[0:1]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: poison_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, s0
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, s1
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[0:1], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: poison_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v0, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: poison_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    global_store_dwordx2 v0, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: poison_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX10DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10DAGISEL-NEXT:    global_store_dwordx2 v0, v[0:1], s[0:1]
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: poison_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX10GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10GISEL-NEXT:    global_store_dwordx2 v0, v[0:1], s[0:1]
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX11DAGISEL-LABEL: poison_value_i64:
+; GFX11DAGISEL:       ; %bb.0: ; %entry
+; GFX11DAGISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX11DAGISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX11DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX11DAGISEL-NEXT:    global_store_b64 v0, v[0:1], s[0:1]
+; GFX11DAGISEL-NEXT:    s_endpgm
+;
+; GFX11GISEL-LABEL: poison_value_i64:
+; GFX11GISEL:       ; %bb.0: ; %entry
+; GFX11GISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX11GISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX11GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX11GISEL-NEXT:    global_store_b64 v0, v[0:1], s[0:1]
+; GFX11GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.min.i64(i64 poison, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define void @divergent_value_i64(ptr addrspace(1) %out, i64 %id.x) {
+; GFX8DAGISEL-LABEL: divergent_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    s_mov_b32 s4, -1
+; GFX8DAGISEL-NEXT:    s_brev_b32 s5, -2
+; GFX8DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX8DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX8DAGISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX8DAGISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX8DAGISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX8DAGISEL-NEXT:    v_cmp_lt_i64_e32 vcc, s[8:9], v[4:5]
+; GFX8DAGISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX8DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX8DAGISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX8DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX8DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX8DAGISEL-NEXT:  ; %bb.2:
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8DAGISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX8DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX8GISEL-LABEL: divergent_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b32 s4, -1
+; GFX8GISEL-NEXT:    s_brev_b32 s5, -2
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX8GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX8GISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX8GISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX8GISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX8GISEL-NEXT:    v_cmp_lt_i64_e32 vcc, s[8:9], v[4:5]
+; GFX8GISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX8GISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX8GISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX8GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX8GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX8GISEL-NEXT:  ; %bb.2:
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8GISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX8GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX9DAGISEL-LABEL: divergent_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    s_mov_b32 s4, -1
+; GFX9DAGISEL-NEXT:    s_brev_b32 s5, -2
+; GFX9DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX9DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX9DAGISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX9DAGISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX9DAGISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX9DAGISEL-NEXT:    v_cmp_lt_i64_e32 vcc, s[8:9], v[4:5]
+; GFX9DAGISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX9DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX9DAGISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX9DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX9DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX9DAGISEL-NEXT:  ; %bb.2:
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX9DAGISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX9DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX9GISEL-LABEL: divergent_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b32 s4, -1
+; GFX9GISEL-NEXT:    s_brev_b32 s5, -2
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX9GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX9GISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX9GISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX9GISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX9GISEL-NEXT:    v_cmp_lt_i64_e32 vcc, s[8:9], v[4:5]
+; GFX9GISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX9GISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX9GISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX9GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX9GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX9GISEL-NEXT:  ; %bb.2:
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX9GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX9GISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX9GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1064DAGISEL-LABEL: divergent_value_i64:
+; GFX1064DAGISEL:       ; %bb.0: ; %entry
+; GFX1064DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1064DAGISEL-NEXT:    s_mov_b32 s4, -1
+; GFX1064DAGISEL-NEXT:    s_brev_b32 s5, -2
+; GFX1064DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1064DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1064DAGISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX1064DAGISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX1064DAGISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX1064DAGISEL-NEXT:    v_cmp_lt_i64_e32 vcc, s[8:9], v[4:5]
+; GFX1064DAGISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX1064DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX1064DAGISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX1064DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX1064DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1064DAGISEL-NEXT:  ; %bb.2:
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1064DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1064DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1064GISEL-LABEL: divergent_value_i64:
+; GFX1064GISEL:       ; %bb.0: ; %entry
+; GFX1064GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b32 s4, -1
+; GFX1064GISEL-NEXT:    s_brev_b32 s5, -2
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1064GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1064GISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX1064GISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX1064GISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX1064GISEL-NEXT:    v_cmp_lt_i64_e32 vcc, s[8:9], v[4:5]
+; GFX1064GISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX1064GISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX1064GISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX1064GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX1064GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1064GISEL-NEXT:  ; %bb.2:
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1064GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1064GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1032DAGISEL-LABEL: divergent_value_i64:
+; GFX1032DAGISEL:       ; %bb.0: ; %entry
+; GFX1032DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1032DAGISEL-NEXT:    s_mov_b32 s4, -1
+; GFX1032DAGISEL-NEXT:    s_brev_b32 s5, -2
+; GFX1032DAGISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1032DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1032DAGISEL-NEXT:    s_ff1_i32_b32 s7, s6
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX1032DAGISEL-NEXT:    v_readlane_b32 s8, v2, s7
+; GFX1032DAGISEL-NEXT:    v_readlane_b32 s9, v3, s7
+; GFX1032DAGISEL-NEXT:    v_cmp_lt_i64_e32 vcc_lo, s[8:9], v[4:5]
+; GFX1032DAGISEL-NEXT:    s_and_b32 s10, vcc_lo, s6
+; GFX1032DAGISEL-NEXT:    s_bitset0_b32 s6, s7
+; GFX1032DAGISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX1032DAGISEL-NEXT:    s_cmp_lg_u32 s6, 0
+; GFX1032DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1032DAGISEL-NEXT:  ; %bb.2:
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1032DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1032DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1032GISEL-LABEL: divergent_value_i64:
+; GFX1032GISEL:       ; %bb.0: ; %entry
+; GFX1032GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b32 s4, -1
+; GFX1032GISEL-NEXT:    s_brev_b32 s5, -2
+; GFX1032GISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1032GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1032GISEL-NEXT:    s_ff1_i32_b32 s7, s6
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX1032GISEL-NEXT:    v_readlane_b32 s8, v2, s7
+; GFX1032GISEL-NEXT:    v_readlane_b32 s9, v3, s7
+; GFX1032GISEL-NEXT:    v_cmp_lt_i64_e32 vcc_lo, s[8:9], v[4:5]
+; GFX1032GISEL-NEXT:    s_and_b32 s10, vcc_lo, s6
+; GFX1032GISEL-NEXT:    s_bitset0_b32 s6, s7
+; GFX1032GISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX1032GISEL-NEXT:    s_cmp_lg_u32 s6, 0
+; GFX1032GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1032GISEL-NEXT:  ; %bb.2:
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1032GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1032GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1164DAGISEL-LABEL: divergent_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    s_mov_b32 s0, -1
+; GFX1164DAGISEL-NEXT:    s_brev_b32 s1, -2
+; GFX1164DAGISEL-NEXT:    s_mov_b64 s[2:3], exec
+; GFX1164DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1164DAGISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_4) | instid1(VALU_DEP_1)
+; GFX1164DAGISEL-NEXT:    s_ctz_i32_b64 s8, s[2:3]
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v4, s0
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v5, s1
+; GFX1164DAGISEL-NEXT:    v_readlane_b32 s4, v2, s8
+; GFX1164DAGISEL-NEXT:    v_readlane_b32 s5, v3, s8
+; GFX1164DAGISEL-NEXT:    v_cmp_lt_i64_e32 vcc, s[4:5], v[4:5]
+; GFX1164DAGISEL-NEXT:    s_and_b64 s[6:7], vcc, s[2:3]
+; GFX1164DAGISEL-NEXT:    s_bitset0_b64 s[2:3], s8
+; GFX1164DAGISEL-NEXT:    s_cselect_b64 s[0:1], s[4:5], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_cmp_lg_u64 s[2:3], 0
+; GFX1164DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1164DAGISEL-NEXT:  ; %bb.2:
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX1164DAGISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1164DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1164GISEL-LABEL: divergent_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b32 s0, -1
+; GFX1164GISEL-NEXT:    s_brev_b32 s1, -2
+; GFX1164GISEL-NEXT:    s_mov_b64 s[2:3], exec
+; GFX1164GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1164GISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_4) | instid1(VALU_DEP_1)
+; GFX1164GISEL-NEXT:    s_ctz_i32_b64 s8, s[2:3]
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v4, s0
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v5, s1
+; GFX1164GISEL-NEXT:    v_readlane_b32 s4, v2, s8
+; GFX1164GISEL-NEXT:    v_readlane_b32 s5, v3, s8
+; GFX1164GISEL-NEXT:    v_cmp_lt_i64_e32 vcc, s[4:5], v[4:5]
+; GFX1164GISEL-NEXT:    s_and_b64 s[6:7], vcc, s[2:3]
+; GFX1164GISEL-NEXT:    s_bitset0_b64 s[2:3], s8
+; GFX1164GISEL-NEXT:    s_cselect_b64 s[0:1], s[4:5], s[0:1]
+; GFX1164GISEL-NEXT:    s_cmp_lg_u64 s[2:3], 0
+; GFX1164GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1164GISEL-NEXT:  ; %bb.2:
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX1164GISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1164GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1132DAGISEL-LABEL: divergent_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    s_mov_b32 s0, -1
+; GFX1132DAGISEL-NEXT:    s_brev_b32 s1, -2
+; GFX1132DAGISEL-NEXT:    s_mov_b32 s2, exec_lo
+; GFX1132DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1132DAGISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_3) | instid1(VALU_DEP_1)
+; GFX1132DAGISEL-NEXT:    s_ctz_i32_b32 s3, s2
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v4, s0 :: v_dual_mov_b32 v5, s1
+; GFX1132DAGISEL-NEXT:    v_readlane_b32 s4, v2, s3
+; GFX1132DAGISEL-NEXT:    v_readlane_b32 s5, v3, s3
+; GFX1132DAGISEL-NEXT:    v_cmp_lt_i64_e32 vcc_lo, s[4:5], v[4:5]
+; GFX1132DAGISEL-NEXT:    s_and_b32 s6, vcc_lo, s2
+; GFX1132DAGISEL-NEXT:    s_bitset0_b32 s2, s3
+; GFX1132DAGISEL-NEXT:    s_cselect_b64 s[0:1], s[4:5], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_cmp_lg_u32 s2, 0
+; GFX1132DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1132DAGISEL-NEXT:  ; %bb.2:
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v3, s1 :: v_dual_mov_b32 v2, s0
+; GFX1132DAGISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1132DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1132GISEL-LABEL: divergent_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b32 s0, -1
+; GFX1132GISEL-NEXT:    s_brev_b32 s1, -2
+; GFX1132GISEL-NEXT:    s_mov_b32 s2, exec_lo
+; GFX1132GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1132GISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_3) | instid1(VALU_DEP_1)
+; GFX1132GISEL-NEXT:    s_ctz_i32_b32 s3, s2
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v4, s0 :: v_dual_mov_b32 v5, s1
+; GFX1132GISEL-NEXT:    v_readlane_b32 s4, v2, s3
+; GFX1132GISEL-NEXT:    v_readlane_b32 s5, v3, s3
+; GFX1132GISEL-NEXT:    v_cmp_lt_i64_e32 vcc_lo, s[4:5], v[4:5]
+; GFX1132GISEL-NEXT:    s_and_b32 s6, vcc_lo, s2
+; GFX1132GISEL-NEXT:    s_bitset0_b32 s2, s3
+; GFX1132GISEL-NEXT:    s_cselect_b64 s[0:1], s[4:5], s[0:1]
+; GFX1132GISEL-NEXT:    s_cmp_lg_u32 s2, 0
+; GFX1132GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1132GISEL-NEXT:  ; %bb.2:
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v3, s1 :: v_dual_mov_b32 v2, s0
+; GFX1132GISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1132GISEL-NEXT:    s_setpc_b64 s[30:31]
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.min.i64(i64 %id.x, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @divergent_cfg_i64(ptr addrspace(1) %out, i64 %in, i64 %in2) {
+; GFX8DAGISEL-LABEL: divergent_cfg_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x34
+; GFX8DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX8DAGISEL-NEXT:    s_and_saveexec_b64 s[6:7], vcc
+; GFX8DAGISEL-NEXT:    s_xor_b64 s[6:7], exec, s[6:7]
+; GFX8DAGISEL-NEXT:    s_or_saveexec_b64 s[6:7], s[6:7]
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX8DAGISEL-NEXT:    s_xor_b64 exec, exec, s[6:7]
+; GFX8DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s4
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s5
+; GFX8DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX8DAGISEL-NEXT:    s_or_b64 exec, exec, s[6:7]
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: divergent_cfg_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX8GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX8GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX8GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX8GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX8GISEL-NEXT:  ; %bb.1: ; %else
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX8GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX8GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX8GISEL-NEXT:  ; %bb.3: ; %if
+; GFX8GISEL-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x34
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX8GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX8GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: divergent_cfg_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX9DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX9DAGISEL-NEXT:    s_and_saveexec_b64 s[4:5], vcc
+; GFX9DAGISEL-NEXT:    s_xor_b64 s[4:5], exec, s[4:5]
+; GFX9DAGISEL-NEXT:    s_or_saveexec_b64 s[4:5], s[4:5]
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9DAGISEL-NEXT:    s_xor_b64 exec, exec, s[4:5]
+; GFX9DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX9DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX9DAGISEL-NEXT:    s_or_b64 exec, exec, s[4:5]
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: divergent_cfg_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX9GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX9GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX9GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX9GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX9GISEL-NEXT:  ; %bb.1: ; %else
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX9GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX9GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX9GISEL-NEXT:  ; %bb.3: ; %if
+; GFX9GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX9GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX9GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX1064DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1064DAGISEL:       ; %bb.0: ; %entry
+; GFX1064DAGISEL-NEXT:    s_clause 0x1
+; GFX1064DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1064DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1064DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX1064DAGISEL-NEXT:    s_and_saveexec_b64 s[4:5], vcc
+; GFX1064DAGISEL-NEXT:    s_xor_b64 s[4:5], exec, s[4:5]
+; GFX1064DAGISEL-NEXT:    s_or_saveexec_b64 s[4:5], s[4:5]
+; GFX1064DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1064DAGISEL-NEXT:    s_xor_b64 exec, exec, s[4:5]
+; GFX1064DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1064DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1064DAGISEL-NEXT:    s_or_b64 exec, exec, s[4:5]
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1064DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1064DAGISEL-NEXT:    s_endpgm
+;
+; GFX1064GISEL-LABEL: divergent_cfg_i64:
+; GFX1064GISEL:       ; %bb.0: ; %entry
+; GFX1064GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1064GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX1064GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1064GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX1064GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX1064GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1064GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1064GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX1064GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1064GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1064GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX1064GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1064GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1064GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1064GISEL-NEXT:    s_endpgm
+;
+; GFX1032DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1032DAGISEL:       ; %bb.0: ; %entry
+; GFX1032DAGISEL-NEXT:    s_clause 0x1
+; GFX1032DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1032DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1032DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc_lo, 15, v0
+; GFX1032DAGISEL-NEXT:    s_and_saveexec_b32 s4, vcc_lo
+; GFX1032DAGISEL-NEXT:    s_xor_b32 s4, exec_lo, s4
+; GFX1032DAGISEL-NEXT:    s_or_saveexec_b32 s4, s4
+; GFX1032DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1032DAGISEL-NEXT:    s_xor_b32 exec_lo, exec_lo, s4
+; GFX1032DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1032DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1032DAGISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s4
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1032DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1032DAGISEL-NEXT:    s_endpgm
+;
+; GFX1032GISEL-LABEL: divergent_cfg_i64:
+; GFX1032GISEL:       ; %bb.0: ; %entry
+; GFX1032GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1032GISEL-NEXT:    v_cmp_le_u32_e32 vcc_lo, 16, v0
+; GFX1032GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1032GISEL-NEXT:    s_and_saveexec_b32 s8, vcc_lo
+; GFX1032GISEL-NEXT:    s_xor_b32 s8, exec_lo, s8
+; GFX1032GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1032GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1032GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_andn2_saveexec_b32 s2, s8
+; GFX1032GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1032GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1032GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX1032GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1032GISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s2
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1032GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1032GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_clause 0x1
+; GFX1164DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1164DAGISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1164DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1164DAGISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1)
+; GFX1164DAGISEL-NEXT:    v_cmpx_lt_u32_e32 15, v0
+; GFX1164DAGISEL-NEXT:    s_xor_b64 s[6:7], exec, s[6:7]
+; GFX1164DAGISEL-NEXT:    s_or_saveexec_b64 s[6:7], s[6:7]
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164DAGISEL-NEXT:    s_xor_b64 exec, exec, s[6:7]
+; GFX1164DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s4
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s5
+; GFX1164DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1164DAGISEL-NEXT:    s_or_b64 exec, exec, s[6:7]
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: divergent_cfg_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1164GISEL-NEXT:    s_mov_b64 s[8:9], exec
+; GFX1164GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1164GISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1)
+; GFX1164GISEL-NEXT:    v_cmpx_le_u32_e32 16, v0
+; GFX1164GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX1164GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1164GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1164GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_and_not1_saveexec_b64 s[2:3], s[8:9]
+; GFX1164GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1164GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1164GISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX1164GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1164GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_clause 0x1
+; GFX1132DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1132DAGISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1132DAGISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1132DAGISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1)
+; GFX1132DAGISEL-NEXT:    v_cmpx_lt_u32_e32 15, v0
+; GFX1132DAGISEL-NEXT:    s_xor_b32 s6, exec_lo, s6
+; GFX1132DAGISEL-NEXT:    s_or_saveexec_b32 s6, s6
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132DAGISEL-NEXT:    s_xor_b32 exec_lo, exec_lo, s6
+; GFX1132DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s4 :: v_dual_mov_b32 v1, s5
+; GFX1132DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1132DAGISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s6
+; GFX1132DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: divergent_cfg_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1132GISEL-NEXT:    s_mov_b32 s8, exec_lo
+; GFX1132GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1132GISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1)
+; GFX1132GISEL-NEXT:    v_cmpx_le_u32_e32 16, v0
+; GFX1132GISEL-NEXT:    s_xor_b32 s8, exec_lo, s8
+; GFX1132GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1132GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1132GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_and_not1_saveexec_b32 s2, s8
+; GFX1132GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1132GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1132GISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX1132GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1132GISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s2
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v0, s6 :: v_dual_mov_b32 v1, s7
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %tid = call i32 @llvm.amdgcn.workitem.id.x()
+  %d_cmp = icmp ult i32 %tid, 16
+  br i1 %d_cmp, label %if, label %else
+
+if:
+  %reducedValTid = call i64 @llvm.amdgcn.wave.reduce.min.i64(i64 %in2, i32 1)
+  br label %endif
+
+else:
+  %reducedValIn = call i64 @llvm.amdgcn.wave.reduce.min.i64(i64 %in, i32 1)
+  br label %endif
+
+endif:
+  %combine = phi i64 [%reducedValTid, %if], [%reducedValIn, %else]
+  store i64 %combine, ptr addrspace(1) %out
+  ret void
+}

llvm/test/CodeGen/AMDGPU/llvm.amdgcn.reduce.or.ll

@@ -980,3 +980,846 @@ endif:
   store i32 %combine, ptr addrspace(1) %out
   ret void
 }
+
+define amdgpu_kernel void @uniform_value_i64(ptr addrspace(1) %out, i64 %in) {
+; GFX8DAGISEL-LABEL: uniform_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s2
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s0
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s1
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s3
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: uniform_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: uniform_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: uniform_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: uniform_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX10DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: uniform_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX10GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: uniform_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: uniform_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: uniform_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: uniform_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.or.i64(i64 %in, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @const_value_i64(ptr addrspace(1) %out) {
+; GFX8DAGISEL-LABEL: const_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: const_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: const_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: const_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: const_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: const_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: const_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: const_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: const_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v2, 0
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: const_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v2, 0
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.or.i64(i64 123, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @poison_value_i64(ptr addrspace(1) %out, i64 %in) {
+; GFX8DAGISEL-LABEL: poison_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s0
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s1
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[0:1], v[0:1]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: poison_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, s0
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, s1
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[0:1], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: poison_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v0, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: poison_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    global_store_dwordx2 v0, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: poison_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX10DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10DAGISEL-NEXT:    global_store_dwordx2 v0, v[0:1], s[0:1]
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: poison_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX10GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10GISEL-NEXT:    global_store_dwordx2 v0, v[0:1], s[0:1]
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX11DAGISEL-LABEL: poison_value_i64:
+; GFX11DAGISEL:       ; %bb.0: ; %entry
+; GFX11DAGISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX11DAGISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX11DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX11DAGISEL-NEXT:    global_store_b64 v0, v[0:1], s[0:1]
+; GFX11DAGISEL-NEXT:    s_endpgm
+;
+; GFX11GISEL-LABEL: poison_value_i64:
+; GFX11GISEL:       ; %bb.0: ; %entry
+; GFX11GISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX11GISEL-NEXT:    v_mov_b32_e32 v0, 0
+; GFX11GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX11GISEL-NEXT:    global_store_b64 v0, v[0:1], s[0:1]
+; GFX11GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.or.i64(i64 poison, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define void @divergent_value_i64(ptr addrspace(1) %out, i64 %id.x) {
+; GFX8DAGISEL-LABEL: divergent_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    s_mov_b64 s[4:5], 0
+; GFX8DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX8DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX8DAGISEL-NEXT:    s_ff1_i32_b64 s10, s[6:7]
+; GFX8DAGISEL-NEXT:    v_readlane_b32 s8, v2, s10
+; GFX8DAGISEL-NEXT:    v_readlane_b32 s9, v3, s10
+; GFX8DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s10
+; GFX8DAGISEL-NEXT:    s_or_b64 s[4:5], s[4:5], s[8:9]
+; GFX8DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX8DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX8DAGISEL-NEXT:  ; %bb.2:
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8DAGISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX8DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX8GISEL-LABEL: divergent_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b64 s[4:5], 0
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX8GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX8GISEL-NEXT:    s_ff1_i32_b64 s10, s[6:7]
+; GFX8GISEL-NEXT:    v_readlane_b32 s8, v2, s10
+; GFX8GISEL-NEXT:    v_readlane_b32 s9, v3, s10
+; GFX8GISEL-NEXT:    s_bitset0_b64 s[6:7], s10
+; GFX8GISEL-NEXT:    s_or_b64 s[4:5], s[4:5], s[8:9]
+; GFX8GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX8GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX8GISEL-NEXT:  ; %bb.2:
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8GISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX8GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX9DAGISEL-LABEL: divergent_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    s_mov_b64 s[4:5], 0
+; GFX9DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX9DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX9DAGISEL-NEXT:    s_ff1_i32_b64 s10, s[6:7]
+; GFX9DAGISEL-NEXT:    v_readlane_b32 s8, v2, s10
+; GFX9DAGISEL-NEXT:    v_readlane_b32 s9, v3, s10
+; GFX9DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s10
+; GFX9DAGISEL-NEXT:    s_or_b64 s[4:5], s[4:5], s[8:9]
+; GFX9DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX9DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX9DAGISEL-NEXT:  ; %bb.2:
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX9DAGISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX9DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX9GISEL-LABEL: divergent_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b64 s[4:5], 0
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX9GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX9GISEL-NEXT:    s_ff1_i32_b64 s10, s[6:7]
+; GFX9GISEL-NEXT:    v_readlane_b32 s8, v2, s10
+; GFX9GISEL-NEXT:    v_readlane_b32 s9, v3, s10
+; GFX9GISEL-NEXT:    s_bitset0_b64 s[6:7], s10
+; GFX9GISEL-NEXT:    s_or_b64 s[4:5], s[4:5], s[8:9]
+; GFX9GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX9GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX9GISEL-NEXT:  ; %bb.2:
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX9GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX9GISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX9GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1064DAGISEL-LABEL: divergent_value_i64:
+; GFX1064DAGISEL:       ; %bb.0: ; %entry
+; GFX1064DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1064DAGISEL-NEXT:    s_mov_b64 s[4:5], 0
+; GFX1064DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1064DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1064DAGISEL-NEXT:    s_ff1_i32_b64 s10, s[6:7]
+; GFX1064DAGISEL-NEXT:    v_readlane_b32 s8, v2, s10
+; GFX1064DAGISEL-NEXT:    v_readlane_b32 s9, v3, s10
+; GFX1064DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s10
+; GFX1064DAGISEL-NEXT:    s_or_b64 s[4:5], s[4:5], s[8:9]
+; GFX1064DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX1064DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1064DAGISEL-NEXT:  ; %bb.2:
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1064DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1064DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1064GISEL-LABEL: divergent_value_i64:
+; GFX1064GISEL:       ; %bb.0: ; %entry
+; GFX1064GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b64 s[4:5], 0
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1064GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1064GISEL-NEXT:    s_ff1_i32_b64 s10, s[6:7]
+; GFX1064GISEL-NEXT:    v_readlane_b32 s8, v2, s10
+; GFX1064GISEL-NEXT:    v_readlane_b32 s9, v3, s10
+; GFX1064GISEL-NEXT:    s_bitset0_b64 s[6:7], s10
+; GFX1064GISEL-NEXT:    s_or_b64 s[4:5], s[4:5], s[8:9]
+; GFX1064GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX1064GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1064GISEL-NEXT:  ; %bb.2:
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1064GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1064GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1032DAGISEL-LABEL: divergent_value_i64:
+; GFX1032DAGISEL:       ; %bb.0: ; %entry
+; GFX1032DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1032DAGISEL-NEXT:    s_mov_b64 s[4:5], 0
+; GFX1032DAGISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1032DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1032DAGISEL-NEXT:    s_ff1_i32_b32 s7, s6
+; GFX1032DAGISEL-NEXT:    v_readlane_b32 s8, v2, s7
+; GFX1032DAGISEL-NEXT:    v_readlane_b32 s9, v3, s7
+; GFX1032DAGISEL-NEXT:    s_bitset0_b32 s6, s7
+; GFX1032DAGISEL-NEXT:    s_or_b64 s[4:5], s[4:5], s[8:9]
+; GFX1032DAGISEL-NEXT:    s_cmp_lg_u32 s6, 0
+; GFX1032DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1032DAGISEL-NEXT:  ; %bb.2:
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1032DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1032DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1032GISEL-LABEL: divergent_value_i64:
+; GFX1032GISEL:       ; %bb.0: ; %entry
+; GFX1032GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b64 s[4:5], 0
+; GFX1032GISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1032GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1032GISEL-NEXT:    s_ff1_i32_b32 s7, s6
+; GFX1032GISEL-NEXT:    v_readlane_b32 s8, v2, s7
+; GFX1032GISEL-NEXT:    v_readlane_b32 s9, v3, s7
+; GFX1032GISEL-NEXT:    s_bitset0_b32 s6, s7
+; GFX1032GISEL-NEXT:    s_or_b64 s[4:5], s[4:5], s[8:9]
+; GFX1032GISEL-NEXT:    s_cmp_lg_u32 s6, 0
+; GFX1032GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1032GISEL-NEXT:  ; %bb.2:
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1032GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1032GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1164DAGISEL-LABEL: divergent_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    s_mov_b64 s[0:1], 0
+; GFX1164DAGISEL-NEXT:    s_mov_b64 s[2:3], exec
+; GFX1164DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1164DAGISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1)
+; GFX1164DAGISEL-NEXT:    s_ctz_i32_b64 s6, s[2:3]
+; GFX1164DAGISEL-NEXT:    v_readlane_b32 s4, v2, s6
+; GFX1164DAGISEL-NEXT:    v_readlane_b32 s5, v3, s6
+; GFX1164DAGISEL-NEXT:    s_bitset0_b64 s[2:3], s6
+; GFX1164DAGISEL-NEXT:    s_or_b64 s[0:1], s[0:1], s[4:5]
+; GFX1164DAGISEL-NEXT:    s_cmp_lg_u64 s[2:3], 0
+; GFX1164DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1164DAGISEL-NEXT:  ; %bb.2:
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX1164DAGISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1164DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1164GISEL-LABEL: divergent_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b64 s[0:1], 0
+; GFX1164GISEL-NEXT:    s_mov_b64 s[2:3], exec
+; GFX1164GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1164GISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1)
+; GFX1164GISEL-NEXT:    s_ctz_i32_b64 s6, s[2:3]
+; GFX1164GISEL-NEXT:    v_readlane_b32 s4, v2, s6
+; GFX1164GISEL-NEXT:    v_readlane_b32 s5, v3, s6
+; GFX1164GISEL-NEXT:    s_bitset0_b64 s[2:3], s6
+; GFX1164GISEL-NEXT:    s_or_b64 s[0:1], s[0:1], s[4:5]
+; GFX1164GISEL-NEXT:    s_cmp_lg_u64 s[2:3], 0
+; GFX1164GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1164GISEL-NEXT:  ; %bb.2:
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX1164GISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1164GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1132DAGISEL-LABEL: divergent_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    s_mov_b64 s[0:1], 0
+; GFX1132DAGISEL-NEXT:    s_mov_b32 s2, exec_lo
+; GFX1132DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1132DAGISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1)
+; GFX1132DAGISEL-NEXT:    s_ctz_i32_b32 s3, s2
+; GFX1132DAGISEL-NEXT:    v_readlane_b32 s4, v2, s3
+; GFX1132DAGISEL-NEXT:    v_readlane_b32 s5, v3, s3
+; GFX1132DAGISEL-NEXT:    s_bitset0_b32 s2, s3
+; GFX1132DAGISEL-NEXT:    s_or_b64 s[0:1], s[0:1], s[4:5]
+; GFX1132DAGISEL-NEXT:    s_cmp_lg_u32 s2, 0
+; GFX1132DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1132DAGISEL-NEXT:  ; %bb.2:
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v3, s1 :: v_dual_mov_b32 v2, s0
+; GFX1132DAGISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1132DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1132GISEL-LABEL: divergent_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b64 s[0:1], 0
+; GFX1132GISEL-NEXT:    s_mov_b32 s2, exec_lo
+; GFX1132GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1132GISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(NEXT) | instid1(SALU_CYCLE_1)
+; GFX1132GISEL-NEXT:    s_ctz_i32_b32 s3, s2
+; GFX1132GISEL-NEXT:    v_readlane_b32 s4, v2, s3
+; GFX1132GISEL-NEXT:    v_readlane_b32 s5, v3, s3
+; GFX1132GISEL-NEXT:    s_bitset0_b32 s2, s3
+; GFX1132GISEL-NEXT:    s_or_b64 s[0:1], s[0:1], s[4:5]
+; GFX1132GISEL-NEXT:    s_cmp_lg_u32 s2, 0
+; GFX1132GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1132GISEL-NEXT:  ; %bb.2:
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v3, s1 :: v_dual_mov_b32 v2, s0
+; GFX1132GISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1132GISEL-NEXT:    s_setpc_b64 s[30:31]
+entry:
+  ; %id.x = call i32 @llvm.amdgcn.workitem.id.x()
+  %result = call i64 @llvm.amdgcn.wave.reduce.or.i64(i64 %id.x, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @divergent_cfg_i64(ptr addrspace(1) %out, i64 %in, i64 %in2) {
+; GFX8DAGISEL-LABEL: divergent_cfg_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x34
+; GFX8DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX8DAGISEL-NEXT:    s_and_saveexec_b64 s[6:7], vcc
+; GFX8DAGISEL-NEXT:    s_xor_b64 s[6:7], exec, s[6:7]
+; GFX8DAGISEL-NEXT:    s_or_saveexec_b64 s[6:7], s[6:7]
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX8DAGISEL-NEXT:    s_xor_b64 exec, exec, s[6:7]
+; GFX8DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s4
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s5
+; GFX8DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX8DAGISEL-NEXT:    s_or_b64 exec, exec, s[6:7]
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: divergent_cfg_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX8GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX8GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX8GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX8GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX8GISEL-NEXT:  ; %bb.1: ; %else
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX8GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX8GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX8GISEL-NEXT:  ; %bb.3: ; %if
+; GFX8GISEL-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x34
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX8GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX8GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: divergent_cfg_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX9DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX9DAGISEL-NEXT:    s_and_saveexec_b64 s[4:5], vcc
+; GFX9DAGISEL-NEXT:    s_xor_b64 s[4:5], exec, s[4:5]
+; GFX9DAGISEL-NEXT:    s_or_saveexec_b64 s[4:5], s[4:5]
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9DAGISEL-NEXT:    s_xor_b64 exec, exec, s[4:5]
+; GFX9DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX9DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX9DAGISEL-NEXT:    s_or_b64 exec, exec, s[4:5]
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: divergent_cfg_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX9GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX9GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX9GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX9GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX9GISEL-NEXT:  ; %bb.1: ; %else
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX9GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX9GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX9GISEL-NEXT:  ; %bb.3: ; %if
+; GFX9GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX9GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX9GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX1064DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1064DAGISEL:       ; %bb.0: ; %entry
+; GFX1064DAGISEL-NEXT:    s_clause 0x1
+; GFX1064DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1064DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1064DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX1064DAGISEL-NEXT:    s_and_saveexec_b64 s[4:5], vcc
+; GFX1064DAGISEL-NEXT:    s_xor_b64 s[4:5], exec, s[4:5]
+; GFX1064DAGISEL-NEXT:    s_or_saveexec_b64 s[4:5], s[4:5]
+; GFX1064DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1064DAGISEL-NEXT:    s_xor_b64 exec, exec, s[4:5]
+; GFX1064DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1064DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1064DAGISEL-NEXT:    s_or_b64 exec, exec, s[4:5]
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1064DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1064DAGISEL-NEXT:    s_endpgm
+;
+; GFX1064GISEL-LABEL: divergent_cfg_i64:
+; GFX1064GISEL:       ; %bb.0: ; %entry
+; GFX1064GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1064GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX1064GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1064GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX1064GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX1064GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1064GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1064GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX1064GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1064GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1064GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX1064GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1064GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1064GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1064GISEL-NEXT:    s_endpgm
+;
+; GFX1032DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1032DAGISEL:       ; %bb.0: ; %entry
+; GFX1032DAGISEL-NEXT:    s_clause 0x1
+; GFX1032DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1032DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1032DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc_lo, 15, v0
+; GFX1032DAGISEL-NEXT:    s_and_saveexec_b32 s4, vcc_lo
+; GFX1032DAGISEL-NEXT:    s_xor_b32 s4, exec_lo, s4
+; GFX1032DAGISEL-NEXT:    s_or_saveexec_b32 s4, s4
+; GFX1032DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1032DAGISEL-NEXT:    s_xor_b32 exec_lo, exec_lo, s4
+; GFX1032DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1032DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1032DAGISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s4
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1032DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1032DAGISEL-NEXT:    s_endpgm
+;
+; GFX1032GISEL-LABEL: divergent_cfg_i64:
+; GFX1032GISEL:       ; %bb.0: ; %entry
+; GFX1032GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1032GISEL-NEXT:    v_cmp_le_u32_e32 vcc_lo, 16, v0
+; GFX1032GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1032GISEL-NEXT:    s_and_saveexec_b32 s8, vcc_lo
+; GFX1032GISEL-NEXT:    s_xor_b32 s8, exec_lo, s8
+; GFX1032GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1032GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1032GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_andn2_saveexec_b32 s2, s8
+; GFX1032GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1032GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1032GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX1032GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1032GISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s2
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1032GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1032GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_clause 0x1
+; GFX1164DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1164DAGISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1164DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1164DAGISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1)
+; GFX1164DAGISEL-NEXT:    v_cmpx_lt_u32_e32 15, v0
+; GFX1164DAGISEL-NEXT:    s_xor_b64 s[6:7], exec, s[6:7]
+; GFX1164DAGISEL-NEXT:    s_or_saveexec_b64 s[6:7], s[6:7]
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164DAGISEL-NEXT:    s_xor_b64 exec, exec, s[6:7]
+; GFX1164DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s4
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s5
+; GFX1164DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1164DAGISEL-NEXT:    s_or_b64 exec, exec, s[6:7]
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: divergent_cfg_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1164GISEL-NEXT:    s_mov_b64 s[8:9], exec
+; GFX1164GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1164GISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1)
+; GFX1164GISEL-NEXT:    v_cmpx_le_u32_e32 16, v0
+; GFX1164GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX1164GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1164GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1164GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_and_not1_saveexec_b64 s[2:3], s[8:9]
+; GFX1164GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1164GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1164GISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX1164GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1164GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_clause 0x1
+; GFX1132DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1132DAGISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1132DAGISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1132DAGISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1)
+; GFX1132DAGISEL-NEXT:    v_cmpx_lt_u32_e32 15, v0
+; GFX1132DAGISEL-NEXT:    s_xor_b32 s6, exec_lo, s6
+; GFX1132DAGISEL-NEXT:    s_or_saveexec_b32 s6, s6
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132DAGISEL-NEXT:    s_xor_b32 exec_lo, exec_lo, s6
+; GFX1132DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s4 :: v_dual_mov_b32 v1, s5
+; GFX1132DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1132DAGISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s6
+; GFX1132DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: divergent_cfg_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1132GISEL-NEXT:    s_mov_b32 s8, exec_lo
+; GFX1132GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1132GISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1)
+; GFX1132GISEL-NEXT:    v_cmpx_le_u32_e32 16, v0
+; GFX1132GISEL-NEXT:    s_xor_b32 s8, exec_lo, s8
+; GFX1132GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1132GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1132GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_and_not1_saveexec_b32 s2, s8
+; GFX1132GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1132GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1132GISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX1132GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1132GISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s2
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v0, s6 :: v_dual_mov_b32 v1, s7
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %tid = call i32 @llvm.amdgcn.workitem.id.x()
+  %d_cmp = icmp ult i32 %tid, 16
+  br i1 %d_cmp, label %if, label %else
+
+if:
+  %reducedValTid = call i64 @llvm.amdgcn.wave.reduce.or.i64(i64 %in2, i32 1)
+  br label %endif
+
+else:
+  %reducedValIn = call i64 @llvm.amdgcn.wave.reduce.or.i64(i64 %in, i32 1)
+  br label %endif
+
+endif:
+  %combine = phi i64 [%reducedValTid, %if], [%reducedValIn, %else]
+  store i64 %combine, ptr addrspace(1) %out
+  ret void
+}

llvm/test/CodeGen/AMDGPU/llvm.amdgcn.reduce.umax.ll

@@ -947,3 +947,853 @@ endif:
   store i32 %combine, ptr addrspace(1) %out
   ret void
 }
+
+define amdgpu_kernel void @uniform_value_i64(ptr addrspace(1) %out, i64 %in) {
+; GFX8DAGISEL-LABEL: uniform_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s2
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s0
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s1
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s3
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: uniform_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: uniform_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: uniform_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: uniform_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX10DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: uniform_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX10GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: uniform_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: uniform_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: uniform_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: uniform_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.umax.i64(i64 %in, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @const_value_i64(ptr addrspace(1) %out) {
+; GFX8DAGISEL-LABEL: const_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: const_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: const_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v1, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: const_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: const_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10DAGISEL-NEXT:    global_store_dwordx2 v1, v[0:1], s[0:1]
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: const_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: const_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    global_store_b64 v1, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: const_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: const_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v0, 0x7b
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    global_store_b64 v1, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: const_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v2, 0
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.umax.i64(i64 123, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @poison_value_i64(ptr addrspace(1) %out, i64 %in) {
+; GFX8DAGISEL-LABEL: poison_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: poison_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: poison_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: poison_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: poison_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: poison_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX11DAGISEL-LABEL: poison_value_i64:
+; GFX11DAGISEL:       ; %bb.0: ; %entry
+; GFX11DAGISEL-NEXT:    s_endpgm
+;
+; GFX11GISEL-LABEL: poison_value_i64:
+; GFX11GISEL:       ; %bb.0: ; %entry
+; GFX11GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.umax.i64(i64 poison, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define void @divergent_value_i64(ptr addrspace(1) %out, i64 %id.x) {
+; GFX8DAGISEL-LABEL: divergent_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    s_mov_b64 s[4:5], 0
+; GFX8DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX8DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX8DAGISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX8DAGISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX8DAGISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX8DAGISEL-NEXT:    v_cmp_gt_u64_e32 vcc, s[8:9], v[4:5]
+; GFX8DAGISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX8DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX8DAGISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX8DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX8DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX8DAGISEL-NEXT:  ; %bb.2:
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8DAGISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX8DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX8GISEL-LABEL: divergent_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b64 s[4:5], 0
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX8GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX8GISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX8GISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX8GISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX8GISEL-NEXT:    v_cmp_gt_u64_e32 vcc, s[8:9], v[4:5]
+; GFX8GISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX8GISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX8GISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX8GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX8GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX8GISEL-NEXT:  ; %bb.2:
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8GISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX8GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX9DAGISEL-LABEL: divergent_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    s_mov_b64 s[4:5], 0
+; GFX9DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX9DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX9DAGISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX9DAGISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX9DAGISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX9DAGISEL-NEXT:    v_cmp_gt_u64_e32 vcc, s[8:9], v[4:5]
+; GFX9DAGISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX9DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX9DAGISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX9DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX9DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX9DAGISEL-NEXT:  ; %bb.2:
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX9DAGISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX9DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX9GISEL-LABEL: divergent_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b64 s[4:5], 0
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX9GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX9GISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX9GISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX9GISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX9GISEL-NEXT:    v_cmp_gt_u64_e32 vcc, s[8:9], v[4:5]
+; GFX9GISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX9GISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX9GISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX9GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX9GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX9GISEL-NEXT:  ; %bb.2:
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX9GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX9GISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX9GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1064DAGISEL-LABEL: divergent_value_i64:
+; GFX1064DAGISEL:       ; %bb.0: ; %entry
+; GFX1064DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1064DAGISEL-NEXT:    s_mov_b64 s[4:5], 0
+; GFX1064DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1064DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1064DAGISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX1064DAGISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX1064DAGISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX1064DAGISEL-NEXT:    v_cmp_gt_u64_e32 vcc, s[8:9], v[4:5]
+; GFX1064DAGISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX1064DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX1064DAGISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX1064DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX1064DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1064DAGISEL-NEXT:  ; %bb.2:
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1064DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1064DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1064GISEL-LABEL: divergent_value_i64:
+; GFX1064GISEL:       ; %bb.0: ; %entry
+; GFX1064GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b64 s[4:5], 0
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1064GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1064GISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX1064GISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX1064GISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX1064GISEL-NEXT:    v_cmp_gt_u64_e32 vcc, s[8:9], v[4:5]
+; GFX1064GISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX1064GISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX1064GISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX1064GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX1064GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1064GISEL-NEXT:  ; %bb.2:
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1064GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1064GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1032DAGISEL-LABEL: divergent_value_i64:
+; GFX1032DAGISEL:       ; %bb.0: ; %entry
+; GFX1032DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1032DAGISEL-NEXT:    s_mov_b64 s[4:5], 0
+; GFX1032DAGISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1032DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1032DAGISEL-NEXT:    s_ff1_i32_b32 s7, s6
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX1032DAGISEL-NEXT:    v_readlane_b32 s8, v2, s7
+; GFX1032DAGISEL-NEXT:    v_readlane_b32 s9, v3, s7
+; GFX1032DAGISEL-NEXT:    v_cmp_gt_u64_e32 vcc_lo, s[8:9], v[4:5]
+; GFX1032DAGISEL-NEXT:    s_and_b32 s10, vcc_lo, s6
+; GFX1032DAGISEL-NEXT:    s_bitset0_b32 s6, s7
+; GFX1032DAGISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX1032DAGISEL-NEXT:    s_cmp_lg_u32 s6, 0
+; GFX1032DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1032DAGISEL-NEXT:  ; %bb.2:
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1032DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1032DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1032GISEL-LABEL: divergent_value_i64:
+; GFX1032GISEL:       ; %bb.0: ; %entry
+; GFX1032GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b64 s[4:5], 0
+; GFX1032GISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1032GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1032GISEL-NEXT:    s_ff1_i32_b32 s7, s6
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX1032GISEL-NEXT:    v_readlane_b32 s8, v2, s7
+; GFX1032GISEL-NEXT:    v_readlane_b32 s9, v3, s7
+; GFX1032GISEL-NEXT:    v_cmp_gt_u64_e32 vcc_lo, s[8:9], v[4:5]
+; GFX1032GISEL-NEXT:    s_and_b32 s10, vcc_lo, s6
+; GFX1032GISEL-NEXT:    s_bitset0_b32 s6, s7
+; GFX1032GISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX1032GISEL-NEXT:    s_cmp_lg_u32 s6, 0
+; GFX1032GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1032GISEL-NEXT:  ; %bb.2:
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1032GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1032GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1164DAGISEL-LABEL: divergent_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    s_mov_b64 s[0:1], 0
+; GFX1164DAGISEL-NEXT:    s_mov_b64 s[2:3], exec
+; GFX1164DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1164DAGISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_4) | instid1(VALU_DEP_1)
+; GFX1164DAGISEL-NEXT:    s_ctz_i32_b64 s8, s[2:3]
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v4, s0
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v5, s1
+; GFX1164DAGISEL-NEXT:    v_readlane_b32 s4, v2, s8
+; GFX1164DAGISEL-NEXT:    v_readlane_b32 s5, v3, s8
+; GFX1164DAGISEL-NEXT:    v_cmp_gt_u64_e32 vcc, s[4:5], v[4:5]
+; GFX1164DAGISEL-NEXT:    s_and_b64 s[6:7], vcc, s[2:3]
+; GFX1164DAGISEL-NEXT:    s_bitset0_b64 s[2:3], s8
+; GFX1164DAGISEL-NEXT:    s_cselect_b64 s[0:1], s[4:5], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_cmp_lg_u64 s[2:3], 0
+; GFX1164DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1164DAGISEL-NEXT:  ; %bb.2:
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX1164DAGISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1164DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1164GISEL-LABEL: divergent_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b64 s[0:1], 0
+; GFX1164GISEL-NEXT:    s_mov_b64 s[2:3], exec
+; GFX1164GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1164GISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_4) | instid1(VALU_DEP_1)
+; GFX1164GISEL-NEXT:    s_ctz_i32_b64 s8, s[2:3]
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v4, s0
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v5, s1
+; GFX1164GISEL-NEXT:    v_readlane_b32 s4, v2, s8
+; GFX1164GISEL-NEXT:    v_readlane_b32 s5, v3, s8
+; GFX1164GISEL-NEXT:    v_cmp_gt_u64_e32 vcc, s[4:5], v[4:5]
+; GFX1164GISEL-NEXT:    s_and_b64 s[6:7], vcc, s[2:3]
+; GFX1164GISEL-NEXT:    s_bitset0_b64 s[2:3], s8
+; GFX1164GISEL-NEXT:    s_cselect_b64 s[0:1], s[4:5], s[0:1]
+; GFX1164GISEL-NEXT:    s_cmp_lg_u64 s[2:3], 0
+; GFX1164GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1164GISEL-NEXT:  ; %bb.2:
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX1164GISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1164GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1132DAGISEL-LABEL: divergent_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    s_mov_b64 s[0:1], 0
+; GFX1132DAGISEL-NEXT:    s_mov_b32 s2, exec_lo
+; GFX1132DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1132DAGISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_3) | instid1(VALU_DEP_1)
+; GFX1132DAGISEL-NEXT:    s_ctz_i32_b32 s3, s2
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v4, s0 :: v_dual_mov_b32 v5, s1
+; GFX1132DAGISEL-NEXT:    v_readlane_b32 s4, v2, s3
+; GFX1132DAGISEL-NEXT:    v_readlane_b32 s5, v3, s3
+; GFX1132DAGISEL-NEXT:    v_cmp_gt_u64_e32 vcc_lo, s[4:5], v[4:5]
+; GFX1132DAGISEL-NEXT:    s_and_b32 s6, vcc_lo, s2
+; GFX1132DAGISEL-NEXT:    s_bitset0_b32 s2, s3
+; GFX1132DAGISEL-NEXT:    s_cselect_b64 s[0:1], s[4:5], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_cmp_lg_u32 s2, 0
+; GFX1132DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1132DAGISEL-NEXT:  ; %bb.2:
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v3, s1 :: v_dual_mov_b32 v2, s0
+; GFX1132DAGISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1132DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1132GISEL-LABEL: divergent_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b64 s[0:1], 0
+; GFX1132GISEL-NEXT:    s_mov_b32 s2, exec_lo
+; GFX1132GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1132GISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_3) | instid1(VALU_DEP_1)
+; GFX1132GISEL-NEXT:    s_ctz_i32_b32 s3, s2
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v4, s0 :: v_dual_mov_b32 v5, s1
+; GFX1132GISEL-NEXT:    v_readlane_b32 s4, v2, s3
+; GFX1132GISEL-NEXT:    v_readlane_b32 s5, v3, s3
+; GFX1132GISEL-NEXT:    v_cmp_gt_u64_e32 vcc_lo, s[4:5], v[4:5]
+; GFX1132GISEL-NEXT:    s_and_b32 s6, vcc_lo, s2
+; GFX1132GISEL-NEXT:    s_bitset0_b32 s2, s3
+; GFX1132GISEL-NEXT:    s_cselect_b64 s[0:1], s[4:5], s[0:1]
+; GFX1132GISEL-NEXT:    s_cmp_lg_u32 s2, 0
+; GFX1132GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1132GISEL-NEXT:  ; %bb.2:
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v3, s1 :: v_dual_mov_b32 v2, s0
+; GFX1132GISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1132GISEL-NEXT:    s_setpc_b64 s[30:31]
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.umax.i64(i64 %id.x, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @divergent_cfg_i64(ptr addrspace(1) %out, i64 %in, i64 %in2) {
+; GFX8DAGISEL-LABEL: divergent_cfg_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x34
+; GFX8DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX8DAGISEL-NEXT:    s_and_saveexec_b64 s[6:7], vcc
+; GFX8DAGISEL-NEXT:    s_xor_b64 s[6:7], exec, s[6:7]
+; GFX8DAGISEL-NEXT:    s_or_saveexec_b64 s[6:7], s[6:7]
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX8DAGISEL-NEXT:    s_xor_b64 exec, exec, s[6:7]
+; GFX8DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s4
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s5
+; GFX8DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX8DAGISEL-NEXT:    s_or_b64 exec, exec, s[6:7]
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: divergent_cfg_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX8GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX8GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX8GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX8GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX8GISEL-NEXT:  ; %bb.1: ; %else
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX8GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX8GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX8GISEL-NEXT:  ; %bb.3: ; %if
+; GFX8GISEL-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x34
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX8GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX8GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: divergent_cfg_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX9DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX9DAGISEL-NEXT:    s_and_saveexec_b64 s[4:5], vcc
+; GFX9DAGISEL-NEXT:    s_xor_b64 s[4:5], exec, s[4:5]
+; GFX9DAGISEL-NEXT:    s_or_saveexec_b64 s[4:5], s[4:5]
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9DAGISEL-NEXT:    s_xor_b64 exec, exec, s[4:5]
+; GFX9DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX9DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX9DAGISEL-NEXT:    s_or_b64 exec, exec, s[4:5]
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: divergent_cfg_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX9GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX9GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX9GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX9GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX9GISEL-NEXT:  ; %bb.1: ; %else
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX9GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX9GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX9GISEL-NEXT:  ; %bb.3: ; %if
+; GFX9GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX9GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX9GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX1064DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1064DAGISEL:       ; %bb.0: ; %entry
+; GFX1064DAGISEL-NEXT:    s_clause 0x1
+; GFX1064DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1064DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1064DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX1064DAGISEL-NEXT:    s_and_saveexec_b64 s[4:5], vcc
+; GFX1064DAGISEL-NEXT:    s_xor_b64 s[4:5], exec, s[4:5]
+; GFX1064DAGISEL-NEXT:    s_or_saveexec_b64 s[4:5], s[4:5]
+; GFX1064DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1064DAGISEL-NEXT:    s_xor_b64 exec, exec, s[4:5]
+; GFX1064DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1064DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1064DAGISEL-NEXT:    s_or_b64 exec, exec, s[4:5]
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1064DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1064DAGISEL-NEXT:    s_endpgm
+;
+; GFX1064GISEL-LABEL: divergent_cfg_i64:
+; GFX1064GISEL:       ; %bb.0: ; %entry
+; GFX1064GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1064GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX1064GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1064GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX1064GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX1064GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1064GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1064GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX1064GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1064GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1064GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX1064GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1064GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1064GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1064GISEL-NEXT:    s_endpgm
+;
+; GFX1032DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1032DAGISEL:       ; %bb.0: ; %entry
+; GFX1032DAGISEL-NEXT:    s_clause 0x1
+; GFX1032DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1032DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1032DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc_lo, 15, v0
+; GFX1032DAGISEL-NEXT:    s_and_saveexec_b32 s4, vcc_lo
+; GFX1032DAGISEL-NEXT:    s_xor_b32 s4, exec_lo, s4
+; GFX1032DAGISEL-NEXT:    s_or_saveexec_b32 s4, s4
+; GFX1032DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1032DAGISEL-NEXT:    s_xor_b32 exec_lo, exec_lo, s4
+; GFX1032DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1032DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1032DAGISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s4
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1032DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1032DAGISEL-NEXT:    s_endpgm
+;
+; GFX1032GISEL-LABEL: divergent_cfg_i64:
+; GFX1032GISEL:       ; %bb.0: ; %entry
+; GFX1032GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1032GISEL-NEXT:    v_cmp_le_u32_e32 vcc_lo, 16, v0
+; GFX1032GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1032GISEL-NEXT:    s_and_saveexec_b32 s8, vcc_lo
+; GFX1032GISEL-NEXT:    s_xor_b32 s8, exec_lo, s8
+; GFX1032GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1032GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1032GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_andn2_saveexec_b32 s2, s8
+; GFX1032GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1032GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1032GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX1032GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1032GISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s2
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1032GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1032GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_clause 0x1
+; GFX1164DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1164DAGISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1164DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1164DAGISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1)
+; GFX1164DAGISEL-NEXT:    v_cmpx_lt_u32_e32 15, v0
+; GFX1164DAGISEL-NEXT:    s_xor_b64 s[6:7], exec, s[6:7]
+; GFX1164DAGISEL-NEXT:    s_or_saveexec_b64 s[6:7], s[6:7]
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164DAGISEL-NEXT:    s_xor_b64 exec, exec, s[6:7]
+; GFX1164DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s4
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s5
+; GFX1164DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1164DAGISEL-NEXT:    s_or_b64 exec, exec, s[6:7]
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: divergent_cfg_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1164GISEL-NEXT:    s_mov_b64 s[8:9], exec
+; GFX1164GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1164GISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1)
+; GFX1164GISEL-NEXT:    v_cmpx_le_u32_e32 16, v0
+; GFX1164GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX1164GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1164GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1164GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_and_not1_saveexec_b64 s[2:3], s[8:9]
+; GFX1164GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1164GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1164GISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX1164GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1164GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_clause 0x1
+; GFX1132DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1132DAGISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1132DAGISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1132DAGISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1)
+; GFX1132DAGISEL-NEXT:    v_cmpx_lt_u32_e32 15, v0
+; GFX1132DAGISEL-NEXT:    s_xor_b32 s6, exec_lo, s6
+; GFX1132DAGISEL-NEXT:    s_or_saveexec_b32 s6, s6
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132DAGISEL-NEXT:    s_xor_b32 exec_lo, exec_lo, s6
+; GFX1132DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s4 :: v_dual_mov_b32 v1, s5
+; GFX1132DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1132DAGISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s6
+; GFX1132DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: divergent_cfg_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1132GISEL-NEXT:    s_mov_b32 s8, exec_lo
+; GFX1132GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1132GISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1)
+; GFX1132GISEL-NEXT:    v_cmpx_le_u32_e32 16, v0
+; GFX1132GISEL-NEXT:    s_xor_b32 s8, exec_lo, s8
+; GFX1132GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1132GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1132GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_and_not1_saveexec_b32 s2, s8
+; GFX1132GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1132GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1132GISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX1132GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1132GISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s2
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v0, s6 :: v_dual_mov_b32 v1, s7
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %tid = call i32 @llvm.amdgcn.workitem.id.x()
+  %d_cmp = icmp ult i32 %tid, 16
+  br i1 %d_cmp, label %if, label %else
+
+if:
+  %reducedValTid = call i64 @llvm.amdgcn.wave.reduce.umax.i64(i64 %in2, i32 1)
+  br label %endif
+
+else:
+  %reducedValIn = call i64 @llvm.amdgcn.wave.reduce.umax.i64(i64 %in, i32 1)
+  br label %endif
+
+endif:
+  %combine = phi i64 [%reducedValTid, %if], [%reducedValIn, %else]
+  store i64 %combine, ptr addrspace(1) %out
+  ret void
+}

llvm/test/CodeGen/AMDGPU/llvm.amdgcn.reduce.umin.ll

@@ -947,3 +947,853 @@ endif:
   store i32 %combine, ptr addrspace(1) %out
   ret void
 }
+
+define amdgpu_kernel void @uniform_value_i64(ptr addrspace(1) %out, i64 %in) {
+; GFX8DAGISEL-LABEL: uniform_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s2
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s0
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s1
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s3
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: uniform_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: uniform_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: uniform_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: uniform_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX10DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: uniform_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX10GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: uniform_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: uniform_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: uniform_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: uniform_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.umin.i64(i64 %in, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @const_value_i64(ptr addrspace(1) %out) {
+; GFX8DAGISEL-LABEL: const_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: const_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: const_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v1, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: const_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: const_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX10DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10DAGISEL-NEXT:    global_store_dwordx2 v1, v[0:1], s[0:1]
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: const_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX10GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX10GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX10GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: const_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    global_store_b64 v1, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: const_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, 0
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: const_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v0, 0x7b
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    global_store_b64 v1, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: const_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b64 s[0:1], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v0, 0x7b
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v2, 0
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.umin.i64(i64 123, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @poison_value_i64(ptr addrspace(1) %out) {
+; GFX8DAGISEL-LABEL: poison_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: poison_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: poison_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: poison_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX10DAGISEL-LABEL: poison_value_i64:
+; GFX10DAGISEL:       ; %bb.0: ; %entry
+; GFX10DAGISEL-NEXT:    s_endpgm
+;
+; GFX10GISEL-LABEL: poison_value_i64:
+; GFX10GISEL:       ; %bb.0: ; %entry
+; GFX10GISEL-NEXT:    s_endpgm
+;
+; GFX11DAGISEL-LABEL: poison_value_i64:
+; GFX11DAGISEL:       ; %bb.0: ; %entry
+; GFX11DAGISEL-NEXT:    s_endpgm
+;
+; GFX11GISEL-LABEL: poison_value_i64:
+; GFX11GISEL:       ; %bb.0: ; %entry
+; GFX11GISEL-NEXT:    s_endpgm
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.umin.i64(i64 poison, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define void @divergent_value_i64(ptr addrspace(1) %out, i64 %id.x) {
+; GFX8DAGISEL-LABEL: divergent_value_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    s_mov_b64 s[4:5], -1
+; GFX8DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX8DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX8DAGISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX8DAGISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX8DAGISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX8DAGISEL-NEXT:    v_cmp_lt_u64_e32 vcc, s[8:9], v[4:5]
+; GFX8DAGISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX8DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX8DAGISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX8DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX8DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX8DAGISEL-NEXT:  ; %bb.2:
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8DAGISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX8DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX8GISEL-LABEL: divergent_value_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b64 s[4:5], -1
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX8GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX8GISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX8GISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX8GISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX8GISEL-NEXT:    v_cmp_lt_u64_e32 vcc, s[8:9], v[4:5]
+; GFX8GISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX8GISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX8GISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX8GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX8GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX8GISEL-NEXT:  ; %bb.2:
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
+; GFX8GISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX8GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX9DAGISEL-LABEL: divergent_value_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    s_mov_b64 s[4:5], -1
+; GFX9DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX9DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX9DAGISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX9DAGISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX9DAGISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX9DAGISEL-NEXT:    v_cmp_lt_u64_e32 vcc, s[8:9], v[4:5]
+; GFX9DAGISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX9DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX9DAGISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX9DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX9DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX9DAGISEL-NEXT:  ; %bb.2:
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX9DAGISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX9DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX9GISEL-LABEL: divergent_value_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b64 s[4:5], -1
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX9GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX9GISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX9GISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX9GISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX9GISEL-NEXT:    v_cmp_lt_u64_e32 vcc, s[8:9], v[4:5]
+; GFX9GISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX9GISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX9GISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX9GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX9GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX9GISEL-NEXT:  ; %bb.2:
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX9GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX9GISEL-NEXT:    s_waitcnt vmcnt(0)
+; GFX9GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1064DAGISEL-LABEL: divergent_value_i64:
+; GFX1064DAGISEL:       ; %bb.0: ; %entry
+; GFX1064DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1064DAGISEL-NEXT:    s_mov_b64 s[4:5], -1
+; GFX1064DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1064DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1064DAGISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX1064DAGISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX1064DAGISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX1064DAGISEL-NEXT:    v_cmp_lt_u64_e32 vcc, s[8:9], v[4:5]
+; GFX1064DAGISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX1064DAGISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX1064DAGISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX1064DAGISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX1064DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1064DAGISEL-NEXT:  ; %bb.2:
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1064DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1064DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1064GISEL-LABEL: divergent_value_i64:
+; GFX1064GISEL:       ; %bb.0: ; %entry
+; GFX1064GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b64 s[4:5], -1
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1064GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1064GISEL-NEXT:    s_ff1_i32_b64 s12, s[6:7]
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX1064GISEL-NEXT:    v_readlane_b32 s8, v2, s12
+; GFX1064GISEL-NEXT:    v_readlane_b32 s9, v3, s12
+; GFX1064GISEL-NEXT:    v_cmp_lt_u64_e32 vcc, s[8:9], v[4:5]
+; GFX1064GISEL-NEXT:    s_and_b64 s[10:11], vcc, s[6:7]
+; GFX1064GISEL-NEXT:    s_bitset0_b64 s[6:7], s12
+; GFX1064GISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX1064GISEL-NEXT:    s_cmp_lg_u64 s[6:7], 0
+; GFX1064GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1064GISEL-NEXT:  ; %bb.2:
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1064GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1064GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1032DAGISEL-LABEL: divergent_value_i64:
+; GFX1032DAGISEL:       ; %bb.0: ; %entry
+; GFX1032DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1032DAGISEL-NEXT:    s_mov_b64 s[4:5], -1
+; GFX1032DAGISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1032DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1032DAGISEL-NEXT:    s_ff1_i32_b32 s7, s6
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX1032DAGISEL-NEXT:    v_readlane_b32 s8, v2, s7
+; GFX1032DAGISEL-NEXT:    v_readlane_b32 s9, v3, s7
+; GFX1032DAGISEL-NEXT:    v_cmp_lt_u64_e32 vcc_lo, s[8:9], v[4:5]
+; GFX1032DAGISEL-NEXT:    s_and_b32 s10, vcc_lo, s6
+; GFX1032DAGISEL-NEXT:    s_bitset0_b32 s6, s7
+; GFX1032DAGISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX1032DAGISEL-NEXT:    s_cmp_lg_u32 s6, 0
+; GFX1032DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1032DAGISEL-NEXT:  ; %bb.2:
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1032DAGISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1032DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1032GISEL-LABEL: divergent_value_i64:
+; GFX1032GISEL:       ; %bb.0: ; %entry
+; GFX1032GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b64 s[4:5], -1
+; GFX1032GISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1032GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1032GISEL-NEXT:    s_ff1_i32_b32 s7, s6
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v4, s4
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v5, s5
+; GFX1032GISEL-NEXT:    v_readlane_b32 s8, v2, s7
+; GFX1032GISEL-NEXT:    v_readlane_b32 s9, v3, s7
+; GFX1032GISEL-NEXT:    v_cmp_lt_u64_e32 vcc_lo, s[8:9], v[4:5]
+; GFX1032GISEL-NEXT:    s_and_b32 s10, vcc_lo, s6
+; GFX1032GISEL-NEXT:    s_bitset0_b32 s6, s7
+; GFX1032GISEL-NEXT:    s_cselect_b64 s[4:5], s[8:9], s[4:5]
+; GFX1032GISEL-NEXT:    s_cmp_lg_u32 s6, 0
+; GFX1032GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1032GISEL-NEXT:  ; %bb.2:
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v2, s4
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v3, s5
+; GFX1032GISEL-NEXT:    global_store_dwordx2 v[0:1], v[2:3], off
+; GFX1032GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1164DAGISEL-LABEL: divergent_value_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    s_mov_b64 s[0:1], -1
+; GFX1164DAGISEL-NEXT:    s_mov_b64 s[2:3], exec
+; GFX1164DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1164DAGISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_4) | instid1(VALU_DEP_1)
+; GFX1164DAGISEL-NEXT:    s_ctz_i32_b64 s8, s[2:3]
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v4, s0
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v5, s1
+; GFX1164DAGISEL-NEXT:    v_readlane_b32 s4, v2, s8
+; GFX1164DAGISEL-NEXT:    v_readlane_b32 s5, v3, s8
+; GFX1164DAGISEL-NEXT:    v_cmp_lt_u64_e32 vcc, s[4:5], v[4:5]
+; GFX1164DAGISEL-NEXT:    s_and_b64 s[6:7], vcc, s[2:3]
+; GFX1164DAGISEL-NEXT:    s_bitset0_b64 s[2:3], s8
+; GFX1164DAGISEL-NEXT:    s_cselect_b64 s[0:1], s[4:5], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_cmp_lg_u64 s[2:3], 0
+; GFX1164DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1164DAGISEL-NEXT:  ; %bb.2:
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX1164DAGISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1164DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1164GISEL-LABEL: divergent_value_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b64 s[0:1], -1
+; GFX1164GISEL-NEXT:    s_mov_b64 s[2:3], exec
+; GFX1164GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1164GISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_4) | instid1(VALU_DEP_1)
+; GFX1164GISEL-NEXT:    s_ctz_i32_b64 s8, s[2:3]
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v4, s0
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v5, s1
+; GFX1164GISEL-NEXT:    v_readlane_b32 s4, v2, s8
+; GFX1164GISEL-NEXT:    v_readlane_b32 s5, v3, s8
+; GFX1164GISEL-NEXT:    v_cmp_lt_u64_e32 vcc, s[4:5], v[4:5]
+; GFX1164GISEL-NEXT:    s_and_b64 s[6:7], vcc, s[2:3]
+; GFX1164GISEL-NEXT:    s_bitset0_b64 s[2:3], s8
+; GFX1164GISEL-NEXT:    s_cselect_b64 s[0:1], s[4:5], s[0:1]
+; GFX1164GISEL-NEXT:    s_cmp_lg_u64 s[2:3], 0
+; GFX1164GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1164GISEL-NEXT:  ; %bb.2:
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX1164GISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1164GISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1132DAGISEL-LABEL: divergent_value_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    s_mov_b64 s[0:1], -1
+; GFX1132DAGISEL-NEXT:    s_mov_b32 s2, exec_lo
+; GFX1132DAGISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1132DAGISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_3) | instid1(VALU_DEP_1)
+; GFX1132DAGISEL-NEXT:    s_ctz_i32_b32 s3, s2
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v4, s0 :: v_dual_mov_b32 v5, s1
+; GFX1132DAGISEL-NEXT:    v_readlane_b32 s4, v2, s3
+; GFX1132DAGISEL-NEXT:    v_readlane_b32 s5, v3, s3
+; GFX1132DAGISEL-NEXT:    v_cmp_lt_u64_e32 vcc_lo, s[4:5], v[4:5]
+; GFX1132DAGISEL-NEXT:    s_and_b32 s6, vcc_lo, s2
+; GFX1132DAGISEL-NEXT:    s_bitset0_b32 s2, s3
+; GFX1132DAGISEL-NEXT:    s_cselect_b64 s[0:1], s[4:5], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_cmp_lg_u32 s2, 0
+; GFX1132DAGISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1132DAGISEL-NEXT:  ; %bb.2:
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v3, s1 :: v_dual_mov_b32 v2, s0
+; GFX1132DAGISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1132DAGISEL-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX1132GISEL-LABEL: divergent_value_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b64 s[0:1], -1
+; GFX1132GISEL-NEXT:    s_mov_b32 s2, exec_lo
+; GFX1132GISEL-NEXT:  .LBB8_1: ; =>This Inner Loop Header: Depth=1
+; GFX1132GISEL-NEXT:    s_delay_alu instid0(SALU_CYCLE_1) | instskip(SKIP_3) | instid1(VALU_DEP_1)
+; GFX1132GISEL-NEXT:    s_ctz_i32_b32 s3, s2
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v4, s0 :: v_dual_mov_b32 v5, s1
+; GFX1132GISEL-NEXT:    v_readlane_b32 s4, v2, s3
+; GFX1132GISEL-NEXT:    v_readlane_b32 s5, v3, s3
+; GFX1132GISEL-NEXT:    v_cmp_lt_u64_e32 vcc_lo, s[4:5], v[4:5]
+; GFX1132GISEL-NEXT:    s_and_b32 s6, vcc_lo, s2
+; GFX1132GISEL-NEXT:    s_bitset0_b32 s2, s3
+; GFX1132GISEL-NEXT:    s_cselect_b64 s[0:1], s[4:5], s[0:1]
+; GFX1132GISEL-NEXT:    s_cmp_lg_u32 s2, 0
+; GFX1132GISEL-NEXT:    s_cbranch_scc1 .LBB8_1
+; GFX1132GISEL-NEXT:  ; %bb.2:
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v3, s1 :: v_dual_mov_b32 v2, s0
+; GFX1132GISEL-NEXT:    global_store_b64 v[0:1], v[2:3], off
+; GFX1132GISEL-NEXT:    s_setpc_b64 s[30:31]
+entry:
+  %result = call i64 @llvm.amdgcn.wave.reduce.umin.i64(i64 %id.x, i32 1)
+  store i64 %result, ptr addrspace(1) %out
+  ret void
+}
+
+define amdgpu_kernel void @divergent_cfg_i64(ptr addrspace(1) %out, i64 %in, i64 %in2) {
+; GFX8DAGISEL-LABEL: divergent_cfg_i64:
+; GFX8DAGISEL:       ; %bb.0: ; %entry
+; GFX8DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8DAGISEL-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x34
+; GFX8DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX8DAGISEL-NEXT:    s_and_saveexec_b64 s[6:7], vcc
+; GFX8DAGISEL-NEXT:    s_xor_b64 s[6:7], exec, s[6:7]
+; GFX8DAGISEL-NEXT:    s_or_saveexec_b64 s[6:7], s[6:7]
+; GFX8DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX8DAGISEL-NEXT:    s_xor_b64 exec, exec, s[6:7]
+; GFX8DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v0, s4
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v1, s5
+; GFX8DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX8DAGISEL-NEXT:    s_or_b64 exec, exec, s[6:7]
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8DAGISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8DAGISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8DAGISEL-NEXT:    s_endpgm
+;
+; GFX8GISEL-LABEL: divergent_cfg_i64:
+; GFX8GISEL:       ; %bb.0: ; %entry
+; GFX8GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX8GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX8GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX8GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX8GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX8GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX8GISEL-NEXT:  ; %bb.1: ; %else
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX8GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX8GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX8GISEL-NEXT:  ; %bb.3: ; %if
+; GFX8GISEL-NEXT:    s_load_dwordx2 s[4:5], s[4:5], 0x34
+; GFX8GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX8GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX8GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX8GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v3, s1
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX8GISEL-NEXT:    v_mov_b32_e32 v2, s0
+; GFX8GISEL-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GFX8GISEL-NEXT:    s_endpgm
+;
+; GFX9DAGISEL-LABEL: divergent_cfg_i64:
+; GFX9DAGISEL:       ; %bb.0: ; %entry
+; GFX9DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX9DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX9DAGISEL-NEXT:    s_and_saveexec_b64 s[4:5], vcc
+; GFX9DAGISEL-NEXT:    s_xor_b64 s[4:5], exec, s[4:5]
+; GFX9DAGISEL-NEXT:    s_or_saveexec_b64 s[4:5], s[4:5]
+; GFX9DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX9DAGISEL-NEXT:    s_xor_b64 exec, exec, s[4:5]
+; GFX9DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX9DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX9DAGISEL-NEXT:    s_or_b64 exec, exec, s[4:5]
+; GFX9DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9DAGISEL-NEXT:    s_endpgm
+;
+; GFX9GISEL-LABEL: divergent_cfg_i64:
+; GFX9GISEL:       ; %bb.0: ; %entry
+; GFX9GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX9GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX9GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX9GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX9GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX9GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX9GISEL-NEXT:  ; %bb.1: ; %else
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX9GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX9GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX9GISEL-NEXT:  ; %bb.3: ; %if
+; GFX9GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX9GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX9GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX9GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX9GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX9GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX9GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX9GISEL-NEXT:    s_endpgm
+;
+; GFX1064DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1064DAGISEL:       ; %bb.0: ; %entry
+; GFX1064DAGISEL-NEXT:    s_clause 0x1
+; GFX1064DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1064DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1064DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc, 15, v0
+; GFX1064DAGISEL-NEXT:    s_and_saveexec_b64 s[4:5], vcc
+; GFX1064DAGISEL-NEXT:    s_xor_b64 s[4:5], exec, s[4:5]
+; GFX1064DAGISEL-NEXT:    s_or_saveexec_b64 s[4:5], s[4:5]
+; GFX1064DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1064DAGISEL-NEXT:    s_xor_b64 exec, exec, s[4:5]
+; GFX1064DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1064DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1064DAGISEL-NEXT:    s_or_b64 exec, exec, s[4:5]
+; GFX1064DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1064DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1064DAGISEL-NEXT:    s_endpgm
+;
+; GFX1064GISEL-LABEL: divergent_cfg_i64:
+; GFX1064GISEL:       ; %bb.0: ; %entry
+; GFX1064GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1064GISEL-NEXT:    v_cmp_le_u32_e32 vcc, 16, v0
+; GFX1064GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1064GISEL-NEXT:    s_and_saveexec_b64 s[8:9], vcc
+; GFX1064GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX1064GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1064GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1064GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_andn2_saveexec_b64 s[2:3], s[8:9]
+; GFX1064GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1064GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1064GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1064GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1064GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX1064GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1064GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1064GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1064GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1064GISEL-NEXT:    s_endpgm
+;
+; GFX1032DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1032DAGISEL:       ; %bb.0: ; %entry
+; GFX1032DAGISEL-NEXT:    s_clause 0x1
+; GFX1032DAGISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1032DAGISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1032DAGISEL-NEXT:    v_cmp_lt_u32_e32 vcc_lo, 15, v0
+; GFX1032DAGISEL-NEXT:    s_and_saveexec_b32 s4, vcc_lo
+; GFX1032DAGISEL-NEXT:    s_xor_b32 s4, exec_lo, s4
+; GFX1032DAGISEL-NEXT:    s_or_saveexec_b32 s4, s4
+; GFX1032DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1032DAGISEL-NEXT:    s_xor_b32 exec_lo, exec_lo, s4
+; GFX1032DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1032DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1032DAGISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s4
+; GFX1032DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1032DAGISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1032DAGISEL-NEXT:    s_endpgm
+;
+; GFX1032GISEL-LABEL: divergent_cfg_i64:
+; GFX1032GISEL:       ; %bb.0: ; %entry
+; GFX1032GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
+; GFX1032GISEL-NEXT:    v_cmp_le_u32_e32 vcc_lo, 16, v0
+; GFX1032GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1032GISEL-NEXT:    s_and_saveexec_b32 s8, vcc_lo
+; GFX1032GISEL-NEXT:    s_xor_b32 s8, exec_lo, s8
+; GFX1032GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1032GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1032GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_andn2_saveexec_b32 s2, s8
+; GFX1032GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1032GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1032GISEL-NEXT:    s_load_dwordx2 s[6:7], s[4:5], 0x34
+; GFX1032GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1032GISEL-NEXT:    s_mov_b64 s[6:7], s[6:7]
+; GFX1032GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1032GISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s2
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1032GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1032GISEL-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
+; GFX1032GISEL-NEXT:    s_endpgm
+;
+; GFX1164DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1164DAGISEL:       ; %bb.0: ; %entry
+; GFX1164DAGISEL-NEXT:    s_clause 0x1
+; GFX1164DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164DAGISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1164DAGISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1164DAGISEL-NEXT:    s_mov_b64 s[6:7], exec
+; GFX1164DAGISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1)
+; GFX1164DAGISEL-NEXT:    v_cmpx_lt_u32_e32 15, v0
+; GFX1164DAGISEL-NEXT:    s_xor_b64 s[6:7], exec, s[6:7]
+; GFX1164DAGISEL-NEXT:    s_or_saveexec_b64 s[6:7], s[6:7]
+; GFX1164DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s2
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s3
+; GFX1164DAGISEL-NEXT:    s_xor_b64 exec, exec, s[6:7]
+; GFX1164DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v0, s4
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v1, s5
+; GFX1164DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1164DAGISEL-NEXT:    s_or_b64 exec, exec, s[6:7]
+; GFX1164DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164DAGISEL-NEXT:    s_endpgm
+;
+; GFX1164GISEL-LABEL: divergent_cfg_i64:
+; GFX1164GISEL:       ; %bb.0: ; %entry
+; GFX1164GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1164GISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1164GISEL-NEXT:    s_mov_b64 s[8:9], exec
+; GFX1164GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1164GISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1)
+; GFX1164GISEL-NEXT:    v_cmpx_le_u32_e32 16, v0
+; GFX1164GISEL-NEXT:    s_xor_b64 s[8:9], exec, s[8:9]
+; GFX1164GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1164GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1164GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_and_not1_saveexec_b64 s[2:3], s[8:9]
+; GFX1164GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1164GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1164GISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1164GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1164GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX1164GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1164GISEL-NEXT:    s_or_b64 exec, exec, s[2:3]
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v0, s6
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v1, s7
+; GFX1164GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1164GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1164GISEL-NEXT:    s_endpgm
+;
+; GFX1132DAGISEL-LABEL: divergent_cfg_i64:
+; GFX1132DAGISEL:       ; %bb.0: ; %entry
+; GFX1132DAGISEL-NEXT:    s_clause 0x1
+; GFX1132DAGISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132DAGISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1132DAGISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1132DAGISEL-NEXT:    s_mov_b32 s6, exec_lo
+; GFX1132DAGISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1) | instskip(SKIP_1) | instid1(SALU_CYCLE_1)
+; GFX1132DAGISEL-NEXT:    v_cmpx_lt_u32_e32 15, v0
+; GFX1132DAGISEL-NEXT:    s_xor_b32 s6, exec_lo, s6
+; GFX1132DAGISEL-NEXT:    s_or_saveexec_b32 s6, s6
+; GFX1132DAGISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s2 :: v_dual_mov_b32 v1, s3
+; GFX1132DAGISEL-NEXT:    s_xor_b32 exec_lo, exec_lo, s6
+; GFX1132DAGISEL-NEXT:  ; %bb.1: ; %if
+; GFX1132DAGISEL-NEXT:    v_dual_mov_b32 v0, s4 :: v_dual_mov_b32 v1, s5
+; GFX1132DAGISEL-NEXT:  ; %bb.2: ; %endif
+; GFX1132DAGISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s6
+; GFX1132DAGISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132DAGISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132DAGISEL-NEXT:    s_endpgm
+;
+; GFX1132GISEL-LABEL: divergent_cfg_i64:
+; GFX1132GISEL:       ; %bb.0: ; %entry
+; GFX1132GISEL-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
+; GFX1132GISEL-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
+; GFX1132GISEL-NEXT:    s_mov_b32 s8, exec_lo
+; GFX1132GISEL-NEXT:    ; implicit-def: $sgpr6_sgpr7
+; GFX1132GISEL-NEXT:    s_delay_alu instid0(VALU_DEP_1)
+; GFX1132GISEL-NEXT:    v_cmpx_le_u32_e32 16, v0
+; GFX1132GISEL-NEXT:    s_xor_b32 s8, exec_lo, s8
+; GFX1132GISEL-NEXT:    s_cbranch_execz .LBB9_2
+; GFX1132GISEL-NEXT:  ; %bb.1: ; %else
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GFX1132GISEL-NEXT:  .LBB9_2: ; %Flow
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_and_not1_saveexec_b32 s2, s8
+; GFX1132GISEL-NEXT:    s_cbranch_execz .LBB9_4
+; GFX1132GISEL-NEXT:  ; %bb.3: ; %if
+; GFX1132GISEL-NEXT:    s_load_b64 s[4:5], s[4:5], 0x34
+; GFX1132GISEL-NEXT:    s_waitcnt lgkmcnt(0)
+; GFX1132GISEL-NEXT:    s_mov_b64 s[6:7], s[4:5]
+; GFX1132GISEL-NEXT:  .LBB9_4: ; %endif
+; GFX1132GISEL-NEXT:    s_or_b32 exec_lo, exec_lo, s2
+; GFX1132GISEL-NEXT:    v_dual_mov_b32 v0, s6 :: v_dual_mov_b32 v1, s7
+; GFX1132GISEL-NEXT:    v_mov_b32_e32 v2, 0
+; GFX1132GISEL-NEXT:    global_store_b64 v2, v[0:1], s[0:1]
+; GFX1132GISEL-NEXT:    s_endpgm
+entry:
+  %tid = call i32 @llvm.amdgcn.workitem.id.x()
+  %d_cmp = icmp ult i32 %tid, 16
+  br i1 %d_cmp, label %if, label %else
+
+if:
+  %reducedValTid = call i64 @llvm.amdgcn.wave.reduce.umin.i64(i64 %in2, i32 1)
+  br label %endif
+
+else:
+  %reducedValIn = call i64 @llvm.amdgcn.wave.reduce.umin.i64(i64 %in, i32 1)
+  br label %endif
+
+endif:
+  %combine = phi i64 [%reducedValTid, %if], [%reducedValIn, %else]
+  store i64 %combine, ptr addrspace(1) %out
+  ret void
+}