diff --git a/llvm/lib/Analysis/ConstantFolding.cpp b/llvm/lib/Analysis/ConstantFolding.cpp
index 8e447b84d3149..9dbab56348411 100644
--- a/llvm/lib/Analysis/ConstantFolding.cpp
+++ b/llvm/lib/Analysis/ConstantFolding.cpp
@@ -1969,6 +1969,8 @@ bool llvm::canConstantFoldCallTo(const CallBase *Call, const Function *F) {
   case Intrinsic::experimental_constrained_rint:
   case Intrinsic::experimental_constrained_fcmp:
   case Intrinsic::experimental_constrained_fcmps:
+
+  case Intrinsic::experimental_cttz_elts:
     return true;
   default:
     return false;
@@ -3792,6 +3794,27 @@ static Constant *ConstantFoldIntrinsicCall2(Intrinsic::ID IntrinsicID, Type *Ty,
       break;
     }
   }
+
+  if (IntrinsicID == Intrinsic::experimental_cttz_elts) {
+    auto *FVTy = dyn_cast<FixedVectorType>(Operands[0]->getType());
+    bool ZeroIsPoison = cast<ConstantInt>(Operands[1])->isOne();
+    if (!FVTy)
+      return nullptr;
+    unsigned Width = Ty->getIntegerBitWidth();
+    if (APInt::getMaxValue(Width).ult(FVTy->getNumElements()))
+      return PoisonValue::get(Ty);
+    for (unsigned I = 0; I < FVTy->getNumElements(); ++I) {
+      Constant *Elt = Operands[0]->getAggregateElement(I);
+      if (!Elt)
+        return nullptr;
+      if (isa<UndefValue>(Elt) || Elt->isNullValue())
+        continue;
+      return ConstantInt::get(Ty, I);
+    }
+    if (ZeroIsPoison)
+      return PoisonValue::get(Ty);
+    return ConstantInt::get(Ty, FVTy->getNumElements());
+  }
   return nullptr;
 }
 
diff --git a/llvm/test/Transforms/InstSimplify/ConstProp/cttz-elts.ll b/llvm/test/Transforms/InstSimplify/ConstProp/cttz-elts.ll
new file mode 100644
index 0000000000000..db91fd68fcbe6
--- /dev/null
+++ b/llvm/test/Transforms/InstSimplify/ConstProp/cttz-elts.ll
@@ -0,0 +1,74 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt -S -passes=instsimplify < %s | FileCheck %s
+
+define i1 @cttz_elts_i1_v4i32() {
+; CHECK-LABEL: @cttz_elts_i1_v4i32(
+; CHECK-NEXT:    ret i1 poison
+;
+  %res = call i1 @llvm.experimental.cttz.elts.i1.v4i32(<4 x i32> <i32 0, i32 0, i32 0, i32 0>, i1 false)
+  ret i1 %res
+}
+
+define i32 @cttz_elts_v4i1() {
+; CHECK-LABEL: @cttz_elts_v4i1(
+; CHECK-NEXT:    ret i32 3
+;
+  %res = call i32 @llvm.experimental.cttz.elts.i32.v4i1(<4 x i1> <i1 false, i1 false, i1 false, i1 true>, i1 false)
+  ret i32 %res
+}
+
+define i32 @cttz_elts_v4i32() {
+; CHECK-LABEL: @cttz_elts_v4i32(
+; CHECK-NEXT:    ret i32 1
+;
+  %res = call i32 @llvm.experimental.cttz.elts.i32.v4i32(<4 x i32> <i32 0, i32 1, i32 2, i32 3>, i1 false)
+  ret i32 %res
+}
+
+define i32 @cttz_elts_v4i32_unused_lane_undef() {
+; CHECK-LABEL: @cttz_elts_v4i32_unused_lane_undef(
+; CHECK-NEXT:    ret i32 1
+;
+  %res = call i32 @llvm.experimental.cttz.elts.i32.v4i32(<4 x i32> <i32 0, i32 1, i32 undef, i32 3>, i1 false)
+  ret i32 %res
+}
+
+define i32 @cttz_elts_v4i32_used_lane_undef() {
+; CHECK-LABEL: @cttz_elts_v4i32_used_lane_undef(
+; CHECK-NEXT:    ret i32 3
+;
+  %res = call i32 @llvm.experimental.cttz.elts.i32.v4i32(<4 x i32> <i32 0, i32 0, i32 undef, i32 3>, i1 false)
+  ret i32 %res
+}
+
+define i32 @cttz_elts_v4i32_unused_lane_poison() {
+; CHECK-LABEL: @cttz_elts_v4i32_unused_lane_poison(
+; CHECK-NEXT:    ret i32 1
+;
+  %res = call i32 @llvm.experimental.cttz.elts.i32.v4i32(<4 x i32> <i32 0, i32 1, i32 poison, i32 3>, i1 false)
+  ret i32 %res
+}
+
+define i32 @cttz_elts_v4i32_used_lane_poison() {
+; CHECK-LABEL: @cttz_elts_v4i32_used_lane_poison(
+; CHECK-NEXT:    ret i32 3
+;
+  %res = call i32 @llvm.experimental.cttz.elts.i32.v4i32(<4 x i32> <i32 0, i32 0, i32 poison, i32 3>, i1 false)
+  ret i32 %res
+}
+
+define i32 @cttz_elts_v4i32_all_zeros() {
+; CHECK-LABEL: @cttz_elts_v4i32_all_zeros(
+; CHECK-NEXT:    ret i32 4
+;
+  %res = call i32 @llvm.experimental.cttz.elts.i32.v4i32(<4 x i32> <i32 0, i32 0, i32 0, i32 0>, i1 false)
+  ret i32 %res
+}
+
+define i32 @cttz_elts_v4i32_all_zeros_iszeropoison() {
+; CHECK-LABEL: @cttz_elts_v4i32_all_zeros_iszeropoison(
+; CHECK-NEXT:    ret i32 poison
+;
+  %res = call i32 @llvm.experimental.cttz.elts.i32.v4i32(<4 x i32> <i32 0, i32 0, i32 0, i32 0>, i1 true)
+  ret i32 %res
+}