diff --git a/llvm/lib/Target/NVPTX/NVPTXISelLowering.h b/llvm/lib/Target/NVPTX/NVPTXISelLowering.h
--- a/llvm/lib/Target/NVPTX/NVPTXISelLowering.h
+++ b/llvm/lib/Target/NVPTX/NVPTXISelLowering.h
@@ -491,7 +491,8 @@
 
   std::string getPrototype(const DataLayout &DL, Type *, const ArgListTy &,
                            const SmallVectorImpl<ISD::OutputArg> &,
-                           MaybeAlign retAlignment, const CallBase &CB) const;
+                           MaybeAlign retAlignment, const CallBase &CB,
+                           const unsigned &uniqueCallSite) const;
 
   SDValue LowerReturn(SDValue Chain, CallingConv::ID CallConv, bool isVarArg,
                       const SmallVectorImpl<ISD::OutputArg> &Outs,
diff --git a/llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp b/llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp
--- a/llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp
+++ b/llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp
@@ -65,7 +65,7 @@
 
 using namespace llvm;
 
-static unsigned int uniqueCallSite = 0;
+static std::atomic<unsigned> globalUniqueCallSite = ATOMIC_VAR_INIT(0);
 
 static cl::opt<bool> sched4reg(
     "nvptx-sched4reg",
@@ -1243,7 +1243,7 @@
 std::string NVPTXTargetLowering::getPrototype(
     const DataLayout &DL, Type *retTy, const ArgListTy &Args,
     const SmallVectorImpl<ISD::OutputArg> &Outs, MaybeAlign retAlignment,
-    const CallBase &CB) const {
+    const CallBase &CB, const unsigned &uniqueCallSite) const {
   auto PtrVT = getPointerTy(DL);
 
   bool isABI = (STI.getSmVersion() >= 20);
@@ -1422,6 +1422,7 @@
   if (!isABI)
     return Chain;
 
+  unsigned uniqueCallSite = globalUniqueCallSite.fetch_add(1);
   SDValue tempChain = Chain;
   Chain = DAG.getCALLSEQ_START(Chain, uniqueCallSite, 0, dl);
   SDValue InFlag = Chain.getValue(1);
@@ -1678,7 +1679,8 @@
     // The prototype is embedded in a string and put as the operand for a
     // CallPrototype SDNode which will print out to the value of the string.
     SDVTList ProtoVTs = DAG.getVTList(MVT::Other, MVT::Glue);
-    std::string Proto = getPrototype(DL, RetTy, Args, Outs, retAlignment, *CB);
+    std::string Proto =
+        getPrototype(DL, RetTy, Args, Outs, retAlignment, *CB, uniqueCallSite);
     const char *ProtoStr =
       nvTM->getManagedStrPool()->getManagedString(Proto.c_str())->c_str();
     SDValue ProtoOps[] = {
@@ -1838,7 +1840,6 @@
                                                    true),
                              InFlag, dl);
   InFlag = Chain.getValue(1);
-  uniqueCallSite++;
 
   // Append ProxyReg instructions to the chain to make sure that `callseq_end`
   // will not get lost. Otherwise, during libcalls expansion, the nodes can become