Index: lib/Target/NVPTX/NVPTXPeephole.cpp
===================================================================
--- lib/Target/NVPTX/NVPTXPeephole.cpp
+++ lib/Target/NVPTX/NVPTXPeephole.cpp
@@ -36,7 +36,6 @@
 #include "llvm/CodeGen/MachineFunctionPass.h"
 #include "llvm/CodeGen/MachineInstrBuilder.h"
 #include "llvm/CodeGen/MachineRegisterInfo.h"
-#include "llvm/CodeGen/MachineFrameInfo.h"
 #include "llvm/Target/TargetRegisterInfo.h"
 #include "llvm/Target/TargetInstrInfo.h"
 
@@ -96,7 +95,7 @@
 
   // Check the LEA_ADDRi operand is Frame index
   auto &BaseAddrOp = GenericAddrDef->getOperand(1);
-  if (BaseAddrOp.getType() == MachineOperand::MO_FrameIndex) {
+  if (BaseAddrOp.isReg() && BaseAddrOp.getReg() == NVPTX::VRFrame) {
     return true;
   }
 
@@ -111,9 +110,7 @@
   auto &Prev = *MRI.getUniqueVRegDef(Root.getOperand(1).getReg());
 
   // Get the correct offset
-  int FrameIndex = Prev.getOperand(1).getIndex();
-  int Offset = MF.getFrameInfo()->getObjectOffset(FrameIndex) +
-               Prev.getOperand(2).getImm();
+  auto Offset = Prev.getOperand(2).getImm();
 
   MachineInstrBuilder MIB =
       BuildMI(MF, Root.getDebugLoc(), TII->get(Prev.getOpcode()),
@@ -145,6 +142,14 @@
       }
     }  // Instruction
   }    // Basic Block
+
+  const auto &MRI = MF.getRegInfo();
+  if (MRI.use_empty(NVPTX::VRFrame)) {
+    if (auto MI = MRI.getUniqueVRegDef(NVPTX::VRFrame)) {
+      MI->eraseFromParentAndMarkDBGValuesForRemoval();
+    }
+  }
+
   return Changed;
 }
 
Index: lib/Target/NVPTX/NVPTXTargetMachine.cpp
===================================================================
--- lib/Target/NVPTX/NVPTXTargetMachine.cpp
+++ lib/Target/NVPTX/NVPTXTargetMachine.cpp
@@ -205,13 +205,12 @@
   if (!ST.hasImageHandles())
     addPass(createNVPTXReplaceImageHandlesPass());
 
-  addPass(createNVPTXPeephole());
-
   return false;
 }
 
 void NVPTXPassConfig::addPostRegAlloc() {
   addPass(createNVPTXPrologEpilogPass(), false);
+  addPass(createNVPTXPeephole());
 }
 
 FunctionPass *NVPTXPassConfig::createTargetRegisterAllocator(bool) {
Index: test/CodeGen/NVPTX/local-stack-frame.ll
===================================================================
--- test/CodeGen/NVPTX/local-stack-frame.ll
+++ test/CodeGen/NVPTX/local-stack-frame.ll
@@ -59,10 +59,16 @@
 
 ; PTX32:        cvta.local.u32   %SP, %SPL;
 ; PTX32:        add.u32          {{%r[0-9]+}}, %SP, 0;
+; PTX32:        add.u32          {{%r[0-9]+}}, %SPL, 0;
+; PTX32:        add.u32          {{%r[0-9]+}}, %SP, 4;
+; PTX32:        add.u32          {{%r[0-9]+}}, %SPL, 4;
 ; PTX32:        st.local.u32     [{{%r[0-9]+}}], {{%r[0-9]+}}
 ; PTX32:        st.local.u32     [{{%r[0-9]+}}], {{%r[0-9]+}}
 ; PTX64:        cvta.local.u64   %SP, %SPL;
 ; PTX64:        add.u64          {{%rd[0-9]+}}, %SP, 0;
+; PTX64:        add.u64          {{%rd[0-9]+}}, %SPL, 0;
+; PTX64:        add.u64          {{%rd[0-9]+}}, %SP, 4;
+; PTX64:        add.u64          {{%rd[0-9]+}}, %SPL, 4;
 ; PTX64:        st.local.u32     [{{%rd[0-9]+}}], {{%r[0-9]+}}
 ; PTX64:        st.local.u32     [{{%rd[0-9]+}}], {{%r[0-9]+}}
 define void @foo4() {