[clang] 9f21d6e - [Clang][AArch64] Use generic extract/insert vector for svget/svset/svcreate tuples

[Caroline Concatto via cfe-commits]

Author: Caroline Concatto
Date: 2022-08-19T12:58:59+01:00
New Revision: 9f21d6e953e559d695b1eb372976533eccba2b06

URL: https://github.com/llvm/llvm-project/commit/9f21d6e953e559d695b1eb372976533eccba2b06
DIFF: https://github.com/llvm/llvm-project/commit/9f21d6e953e559d695b1eb372976533eccba2b06.diff

LOG: [Clang][AArch64] Use generic extract/insert vector for svget/svset/svcreate tuples

This patch replaces svget, svset and svcreate aarch64 intrinsics for tuple
types with the generic llvm-ir intrinsics extract/insert vector

Differential Revision: https://reviews.llvm.org/D131547

Added: 
    

Modified: 
    clang/include/clang/Basic/arm_sve.td
    clang/lib/CodeGen/CGBuiltin.cpp
    clang/lib/CodeGen/CodeGenFunction.h
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2-bfloat.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2.c

Removed: 
    


################################################################################
diff  --git a/clang/include/clang/Basic/arm_sve.td b/clang/include/clang/Basic/arm_sve.td
index 1fe26c6fd0ef..175b572ffdab 100644
--- a/clang/include/clang/Basic/arm_sve.td
+++ b/clang/include/clang/Basic/arm_sve.td
@@ -1440,9 +1440,9 @@ def SVUNDEF_2 : SInst<"svundef2_{d}", "2v", "csilUcUsUiUlhfd", MergeNone, "", [I
 def SVUNDEF_3 : SInst<"svundef3_{d}", "3v", "csilUcUsUiUlhfd", MergeNone, "", [IsUndef]>;
 def SVUNDEF_4 : SInst<"svundef4_{d}", "4v", "csilUcUsUiUlhfd", MergeNone, "", [IsUndef]>;
 
-def SVCREATE_2 : SInst<"svcreate2[_{d}]", "2dd",   "csilUcUsUiUlhfd", MergeNone, "aarch64_sve_tuple_create2", [IsTupleCreate]>;
-def SVCREATE_3 : SInst<"svcreate3[_{d}]", "3ddd",  "csilUcUsUiUlhfd", MergeNone, "aarch64_sve_tuple_create3", [IsTupleCreate]>;
-def SVCREATE_4 : SInst<"svcreate4[_{d}]", "4dddd", "csilUcUsUiUlhfd", MergeNone, "aarch64_sve_tuple_create4", [IsTupleCreate]>;
+def SVCREATE_2 : SInst<"svcreate2[_{d}]", "2dd",   "csilUcUsUiUlhfd", MergeNone, "", [IsTupleCreate]>;
+def SVCREATE_3 : SInst<"svcreate3[_{d}]", "3ddd",  "csilUcUsUiUlhfd", MergeNone, "", [IsTupleCreate]>;
+def SVCREATE_4 : SInst<"svcreate4[_{d}]", "4dddd", "csilUcUsUiUlhfd", MergeNone, "", [IsTupleCreate]>;
 
 let ArchGuard = "defined(__ARM_FEATURE_SVE_BF16)" in {
 def SVUNDEF_1_BF16 : SInst<"svundef_{d}",  "dv", "b", MergeNone, "", [IsUndef]>;
@@ -1450,29 +1450,29 @@ def SVUNDEF_2_BF16 : SInst<"svundef2_{d}", "2v", "b", MergeNone, "", [IsUndef]>;
 def SVUNDEF_3_BF16 : SInst<"svundef3_{d}", "3v", "b", MergeNone, "", [IsUndef]>;
 def SVUNDEF_4_BF16 : SInst<"svundef4_{d}", "4v", "b", MergeNone, "", [IsUndef]>;
 
-def SVCREATE_2_BF16 : SInst<"svcreate2[_{d}]", "2dd",   "b", MergeNone, "aarch64_sve_tuple_create2", [IsTupleCreate]>;
-def SVCREATE_3_BF16 : SInst<"svcreate3[_{d}]", "3ddd",  "b", MergeNone, "aarch64_sve_tuple_create3", [IsTupleCreate]>;
-def SVCREATE_4_BF16 : SInst<"svcreate4[_{d}]", "4dddd", "b", MergeNone, "aarch64_sve_tuple_create4", [IsTupleCreate]>;
+def SVCREATE_2_BF16 : SInst<"svcreate2[_{d}]", "2dd",   "b", MergeNone, "", [IsTupleCreate]>;
+def SVCREATE_3_BF16 : SInst<"svcreate3[_{d}]", "3ddd",  "b", MergeNone, "", [IsTupleCreate]>;
+def SVCREATE_4_BF16 : SInst<"svcreate4[_{d}]", "4dddd", "b", MergeNone, "", [IsTupleCreate]>;
 }
 
 ////////////////////////////////////////////////////////////////////////////////
 // Vector insertion and extraction
-def SVGET_2 : SInst<"svget2[_{d}]", "d2i", "csilUcUsUiUlhfd", MergeNone, "aarch64_sve_tuple_get", [IsTupleGet], [ImmCheck<1, ImmCheck0_1>]>;
-def SVGET_3 : SInst<"svget3[_{d}]", "d3i", "csilUcUsUiUlhfd", MergeNone, "aarch64_sve_tuple_get", [IsTupleGet], [ImmCheck<1, ImmCheck0_2>]>;
-def SVGET_4 : SInst<"svget4[_{d}]", "d4i", "csilUcUsUiUlhfd", MergeNone, "aarch64_sve_tuple_get", [IsTupleGet], [ImmCheck<1, ImmCheck0_3>]>;
+def SVGET_2 : SInst<"svget2[_{d}]", "d2i", "csilUcUsUiUlhfd", MergeNone, "", [IsTupleGet], [ImmCheck<1, ImmCheck0_1>]>;
+def SVGET_3 : SInst<"svget3[_{d}]", "d3i", "csilUcUsUiUlhfd", MergeNone, "", [IsTupleGet], [ImmCheck<1, ImmCheck0_2>]>;
+def SVGET_4 : SInst<"svget4[_{d}]", "d4i", "csilUcUsUiUlhfd", MergeNone, "", [IsTupleGet], [ImmCheck<1, ImmCheck0_3>]>;
 
-def SVSET_2 : SInst<"svset2[_{d}]", "22id", "csilUcUsUiUlhfd", MergeNone, "aarch64_sve_tuple_set", [IsTupleSet], [ImmCheck<1, ImmCheck0_1>]>;
-def SVSET_3 : SInst<"svset3[_{d}]", "33id", "csilUcUsUiUlhfd", MergeNone, "aarch64_sve_tuple_set", [IsTupleSet], [ImmCheck<1, ImmCheck0_2>]>;
-def SVSET_4 : SInst<"svset4[_{d}]", "44id", "csilUcUsUiUlhfd", MergeNone, "aarch64_sve_tuple_set", [IsTupleSet], [ImmCheck<1, ImmCheck0_3>]>;
+def SVSET_2 : SInst<"svset2[_{d}]", "22id", "csilUcUsUiUlhfd", MergeNone, "", [IsTupleSet], [ImmCheck<1, ImmCheck0_1>]>;
+def SVSET_3 : SInst<"svset3[_{d}]", "33id", "csilUcUsUiUlhfd", MergeNone, "", [IsTupleSet], [ImmCheck<1, ImmCheck0_2>]>;
+def SVSET_4 : SInst<"svset4[_{d}]", "44id", "csilUcUsUiUlhfd", MergeNone, "", [IsTupleSet], [ImmCheck<1, ImmCheck0_3>]>;
 
 let ArchGuard = "defined(__ARM_FEATURE_SVE_BF16)" in {
-def SVGET_2_BF16 : SInst<"svget2[_{d}]", "d2i", "b", MergeNone, "aarch64_sve_tuple_get", [IsTupleGet], [ImmCheck<1, ImmCheck0_1>]>;
-def SVGET_3_BF16 : SInst<"svget3[_{d}]", "d3i", "b", MergeNone, "aarch64_sve_tuple_get", [IsTupleGet], [ImmCheck<1, ImmCheck0_2>]>;
-def SVGET_4_BF16 : SInst<"svget4[_{d}]", "d4i", "b", MergeNone, "aarch64_sve_tuple_get", [IsTupleGet], [ImmCheck<1, ImmCheck0_3>]>;
+def SVGET_2_BF16 : SInst<"svget2[_{d}]", "d2i", "b", MergeNone, "", [IsTupleGet], [ImmCheck<1, ImmCheck0_1>]>;
+def SVGET_3_BF16 : SInst<"svget3[_{d}]", "d3i", "b", MergeNone, "", [IsTupleGet], [ImmCheck<1, ImmCheck0_2>]>;
+def SVGET_4_BF16 : SInst<"svget4[_{d}]", "d4i", "b", MergeNone, "", [IsTupleGet], [ImmCheck<1, ImmCheck0_3>]>;
 
-def SVSET_2_BF16 : SInst<"svset2[_{d}]", "22id", "b", MergeNone, "aarch64_sve_tuple_set", [IsTupleSet], [ImmCheck<1, ImmCheck0_1>]>;
-def SVSET_3_BF16 : SInst<"svset3[_{d}]", "33id", "b", MergeNone, "aarch64_sve_tuple_set", [IsTupleSet], [ImmCheck<1, ImmCheck0_2>]>;
-def SVSET_4_BF16 : SInst<"svset4[_{d}]", "44id", "b", MergeNone, "aarch64_sve_tuple_set", [IsTupleSet], [ImmCheck<1, ImmCheck0_3>]>;
+def SVSET_2_BF16 : SInst<"svset2[_{d}]", "22id", "b", MergeNone, "", [IsTupleSet], [ImmCheck<1, ImmCheck0_1>]>;
+def SVSET_3_BF16 : SInst<"svset3[_{d}]", "33id", "b", MergeNone, "", [IsTupleSet], [ImmCheck<1, ImmCheck0_2>]>;
+def SVSET_4_BF16 : SInst<"svset4[_{d}]", "44id", "b", MergeNone, "", [IsTupleSet], [ImmCheck<1, ImmCheck0_3>]>;
 }
 
 ////////////////////////////////////////////////////////////////////////////////

diff  --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index db1a016d4061..1910e5301a39 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -8902,8 +8902,6 @@ Value *CodeGenFunction::EmitSVEStructStore(const SVETypeFlags &TypeFlags,
   default:
     llvm_unreachable("unknown intrinsic!");
   }
-  auto TupleTy =
-      llvm::VectorType::get(VTy->getElementType(), VTy->getElementCount() * N);
 
   Value *Predicate = EmitSVEPredicateCast(Ops[0], VTy);
   Value *BasePtr = Builder.CreateBitCast(Ops[1], VecPtrTy);
@@ -8915,10 +8913,11 @@ Value *CodeGenFunction::EmitSVEStructStore(const SVETypeFlags &TypeFlags,
   // The llvm.aarch64.sve.st2/3/4 intrinsics take legal part vectors, so we
   // need to break up the tuple vector.
   SmallVector<llvm::Value*, 5> Operands;
-  Function *FExtr =
-      CGM.getIntrinsic(Intrinsic::aarch64_sve_tuple_get, {VTy, TupleTy});
-  for (unsigned I = 0; I < N; ++I)
-    Operands.push_back(Builder.CreateCall(FExtr, {Val, Builder.getInt32(I)}));
+  unsigned MinElts = VTy->getElementCount().getKnownMinValue();
+  for (unsigned I = 0; I < N; ++I) {
+    Value *Idx = ConstantInt::get(CGM.Int64Ty, I * MinElts);
+    Operands.push_back(Builder.CreateExtractVector(VTy, Val, Idx));
+  }
   Operands.append({Predicate, BasePtr});
 
   Function *F = CGM.getIntrinsic(IntID, { VTy });
@@ -9084,16 +9083,46 @@ CodeGenFunction::getSVEOverloadTypes(const SVETypeFlags &TypeFlags,
   if (TypeFlags.isOverloadWhileRW())
     return {getSVEPredType(TypeFlags), Ops[0]->getType()};
 
-  if (TypeFlags.isOverloadCvt() || TypeFlags.isTupleSet())
+  if (TypeFlags.isOverloadCvt())
     return {Ops[0]->getType(), Ops.back()->getType()};
 
-  if (TypeFlags.isTupleCreate() || TypeFlags.isTupleGet())
-    return {ResultType, Ops[0]->getType()};
-
   assert(TypeFlags.isOverloadDefault() && "Unexpected value for overloads");
   return {DefaultType};
 }
 
+Value *CodeGenFunction::EmitSVETupleSetOrGet(const SVETypeFlags &TypeFlags,
+                                             llvm::Type *Ty,
+                                             ArrayRef<Value *> Ops) {
+  assert((TypeFlags.isTupleSet() || TypeFlags.isTupleGet()) &&
+         "Expects TypleFlag isTupleSet or TypeFlags.isTupleSet()");
+
+  unsigned I = cast<ConstantInt>(Ops[1])->getSExtValue();
+  auto *SingleVecTy = dyn_cast<llvm::ScalableVectorType>(
+                      TypeFlags.isTupleSet() ? Ops[2]->getType() : Ty);
+  Value *Idx = ConstantInt::get(CGM.Int64Ty,
+                                I * SingleVecTy->getMinNumElements());
+
+  if (TypeFlags.isTupleSet())
+    return Builder.CreateInsertVector(Ty, Ops[0], Ops[2], Idx);
+  return Builder.CreateExtractVector(Ty, Ops[0], Idx);
+}
+
+Value *CodeGenFunction::EmitSVETupleCreate(const SVETypeFlags &TypeFlags,
+                                             llvm::Type *Ty,
+                                             ArrayRef<Value *> Ops) {
+  assert(TypeFlags.isTupleCreate() && "Expects TypleFlag isTupleCreate");
+
+  auto *SrcTy = dyn_cast<llvm::ScalableVectorType>(Ops[0]->getType());
+  unsigned MinElts = SrcTy->getMinNumElements();
+  Value *Call = llvm::PoisonValue::get(Ty);
+  for (unsigned I = 0; I < Ops.size(); I++) {
+    Value *Idx = ConstantInt::get(CGM.Int64Ty, I * MinElts);
+    Call = Builder.CreateInsertVector(Ty, Call, Ops[I], Idx);
+  }
+
+  return Call;
+}
+
 Value *CodeGenFunction::EmitAArch64SVEBuiltinExpr(unsigned BuiltinID,
                                                   const CallExpr *E) {
   // Find out if any arguments are required to be integer constant expressions.
@@ -9148,6 +9177,10 @@ Value *CodeGenFunction::EmitAArch64SVEBuiltinExpr(unsigned BuiltinID,
 		return EmitSVEStructLoad(TypeFlags, Ops, Builtin->LLVMIntrinsic);
 	else if (TypeFlags.isStructStore())
 		return EmitSVEStructStore(TypeFlags, Ops, Builtin->LLVMIntrinsic);
+  else if (TypeFlags.isTupleSet() || TypeFlags.isTupleGet())
+        return EmitSVETupleSetOrGet(TypeFlags, Ty, Ops);
+  else if (TypeFlags.isTupleCreate())
+        return EmitSVETupleCreate(TypeFlags, Ty, Ops);
   else if (TypeFlags.isUndef())
     return UndefValue::get(Ty);
   else if (Builtin->LLVMIntrinsic != 0) {
@@ -9361,12 +9394,12 @@ Value *CodeGenFunction::EmitAArch64SVEBuiltinExpr(unsigned BuiltinID,
   case SVE::BI__builtin_sve_svtbl2_f32:
   case SVE::BI__builtin_sve_svtbl2_f64: {
     SVETypeFlags TF(Builtin->TypeModifier);
-    auto VTy = cast<llvm::VectorType>(getSVEType(TF));
-    auto TupleTy = llvm::VectorType::getDoubleElementsVectorType(VTy);
-    Function *FExtr =
-        CGM.getIntrinsic(Intrinsic::aarch64_sve_tuple_get, {VTy, TupleTy});
-    Value *V0 = Builder.CreateCall(FExtr, {Ops[0], Builder.getInt32(0)});
-    Value *V1 = Builder.CreateCall(FExtr, {Ops[0], Builder.getInt32(1)});
+    auto VTy = cast<llvm::ScalableVectorType>(getSVEType(TF));
+    Value *V0 = Builder.CreateExtractVector(VTy, Ops[0],
+                                            ConstantInt::get(CGM.Int64Ty, 0));
+    unsigned MinElts = VTy->getMinNumElements();
+    Value *V1 = Builder.CreateExtractVector(
+        VTy, Ops[0], ConstantInt::get(CGM.Int64Ty, MinElts));
     Function *F = CGM.getIntrinsic(Intrinsic::aarch64_sve_tbl2, VTy);
     return Builder.CreateCall(F, {V0, V1, Ops[1]});
   }

diff  --git a/clang/lib/CodeGen/CodeGenFunction.h b/clang/lib/CodeGen/CodeGenFunction.h
index 0a5b115c625f..efd09d9f8b29 100644
--- a/clang/lib/CodeGen/CodeGenFunction.h
+++ b/clang/lib/CodeGen/CodeGenFunction.h
@@ -4200,6 +4200,12 @@ class CodeGenFunction : public CodeGenTypeCache {
   llvm::Type *getEltType(const SVETypeFlags &TypeFlags);
   llvm::ScalableVectorType *getSVEType(const SVETypeFlags &TypeFlags);
   llvm::ScalableVectorType *getSVEPredType(const SVETypeFlags &TypeFlags);
+  llvm::Value *EmitSVETupleSetOrGet(const SVETypeFlags &TypeFlags,
+                                    llvm::Type *ReturnType,
+                                    ArrayRef<llvm::Value *> Ops);
+  llvm::Value *EmitSVETupleCreate(const SVETypeFlags &TypeFlags,
+                                  llvm::Type *ReturnType,
+                                  ArrayRef<llvm::Value *> Ops);
   llvm::Value *EmitSVEAllTruePred(const SVETypeFlags &TypeFlags);
   llvm::Value *EmitSVEDupX(llvm::Value *Scalar);
   llvm::Value *EmitSVEDupX(llvm::Value *Scalar, llvm::Type *Ty);

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2-bfloat.c
index 3ccc58344e8d..188dc8abdacd 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2-bfloat.c
@@ -17,13 +17,15 @@
 
 // CHECK-LABEL: @test_svcreate2_bf16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x bfloat> @llvm.aarch64.sve.tuple.create2.nxv16bf16.nxv8bf16(<vscale x 8 x bfloat> [[X0:%.*]], <vscale x 8 x bfloat> [[X1:%.*]])
-// CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x bfloat> @llvm.vector.insert.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> poison, <vscale x 8 x bfloat> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x bfloat> @llvm.vector.insert.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[X1:%.*]], i64 8)
+// CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z19test_svcreate2_bf16u14__SVBFloat16_tu14__SVBFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x bfloat> @llvm.aarch64.sve.tuple.create2.nxv16bf16.nxv8bf16(<vscale x 8 x bfloat> [[X0:%.*]], <vscale x 8 x bfloat> [[X1:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x bfloat> @llvm.vector.insert.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> poison, <vscale x 8 x bfloat> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x bfloat> @llvm.vector.insert.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[X1:%.*]], i64 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP1]]
 //
 svbfloat16x2_t test_svcreate2_bf16(svbfloat16_t x0, svbfloat16_t x1)
 {

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2.c
index b79fa3a4df70..78562bec6c20 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2.c
@@ -17,13 +17,15 @@
 
 // CHECK-LABEL: @test_svcreate2_s8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.create2.nxv32i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]])
-// CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i8> @llvm.vector.insert.nxv32i8.nxv16i8(<vscale x 32 x i8> poison, <vscale x 16 x i8> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 32 x i8> @llvm.vector.insert.nxv32i8.nxv16i8(<vscale x 32 x i8> [[TMP0]], <vscale x 16 x i8> [[X1:%.*]], i64 16)
+// CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z17test_svcreate2_s8u10__SVInt8_tu10__SVInt8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.create2.nxv32i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i8> @llvm.vector.insert.nxv32i8.nxv16i8(<vscale x 32 x i8> poison, <vscale x 16 x i8> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 32 x i8> @llvm.vector.insert.nxv32i8.nxv16i8(<vscale x 32 x i8> [[TMP0]], <vscale x 16 x i8> [[X1:%.*]], i64 16)
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP1]]
 //
 svint8x2_t test_svcreate2_s8(svint8_t x0, svint8_t x1)
 {
@@ -32,13 +34,15 @@ svint8x2_t test_svcreate2_s8(svint8_t x0, svint8_t x1)
 
 // CHECK-LABEL: @test_svcreate2_s16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.create2.nxv16i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]])
-// CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i16> @llvm.vector.insert.nxv16i16.nxv8i16(<vscale x 16 x i16> poison, <vscale x 8 x i16> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i16> @llvm.vector.insert.nxv16i16.nxv8i16(<vscale x 16 x i16> [[TMP0]], <vscale x 8 x i16> [[X1:%.*]], i64 8)
+// CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate2_s16u11__SVInt16_tu11__SVInt16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.create2.nxv16i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i16> @llvm.vector.insert.nxv16i16.nxv8i16(<vscale x 16 x i16> poison, <vscale x 8 x i16> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i16> @llvm.vector.insert.nxv16i16.nxv8i16(<vscale x 16 x i16> [[TMP0]], <vscale x 8 x i16> [[X1:%.*]], i64 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP1]]
 //
 svint16x2_t test_svcreate2_s16(svint16_t x0, svint16_t x1)
 {
@@ -47,13 +51,15 @@ svint16x2_t test_svcreate2_s16(svint16_t x0, svint16_t x1)
 
 // CHECK-LABEL: @test_svcreate2_s32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.create2.nxv8i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]])
-// CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i32> @llvm.vector.insert.nxv8i32.nxv4i32(<vscale x 8 x i32> poison, <vscale x 4 x i32> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i32> @llvm.vector.insert.nxv8i32.nxv4i32(<vscale x 8 x i32> [[TMP0]], <vscale x 4 x i32> [[X1:%.*]], i64 4)
+// CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate2_s32u11__SVInt32_tu11__SVInt32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.create2.nxv8i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i32> @llvm.vector.insert.nxv8i32.nxv4i32(<vscale x 8 x i32> poison, <vscale x 4 x i32> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i32> @llvm.vector.insert.nxv8i32.nxv4i32(<vscale x 8 x i32> [[TMP0]], <vscale x 4 x i32> [[X1:%.*]], i64 4)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP1]]
 //
 svint32x2_t test_svcreate2_s32(svint32_t x0, svint32_t x1)
 {
@@ -62,13 +68,15 @@ svint32x2_t test_svcreate2_s32(svint32_t x0, svint32_t x1)
 
 // CHECK-LABEL: @test_svcreate2_s64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.create2.nxv4i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]])
-// CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i64> @llvm.vector.insert.nxv4i64.nxv2i64(<vscale x 4 x i64> poison, <vscale x 2 x i64> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i64> @llvm.vector.insert.nxv4i64.nxv2i64(<vscale x 4 x i64> [[TMP0]], <vscale x 2 x i64> [[X1:%.*]], i64 2)
+// CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate2_s64u11__SVInt64_tu11__SVInt64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.create2.nxv4i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i64> @llvm.vector.insert.nxv4i64.nxv2i64(<vscale x 4 x i64> poison, <vscale x 2 x i64> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i64> @llvm.vector.insert.nxv4i64.nxv2i64(<vscale x 4 x i64> [[TMP0]], <vscale x 2 x i64> [[X1:%.*]], i64 2)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP1]]
 //
 svint64x2_t test_svcreate2_s64(svint64_t x0, svint64_t x1)
 {
@@ -77,13 +85,15 @@ svint64x2_t test_svcreate2_s64(svint64_t x0, svint64_t x1)
 
 // CHECK-LABEL: @test_svcreate2_u8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.create2.nxv32i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]])
-// CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i8> @llvm.vector.insert.nxv32i8.nxv16i8(<vscale x 32 x i8> poison, <vscale x 16 x i8> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 32 x i8> @llvm.vector.insert.nxv32i8.nxv16i8(<vscale x 32 x i8> [[TMP0]], <vscale x 16 x i8> [[X1:%.*]], i64 16)
+// CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z17test_svcreate2_u8u11__SVUint8_tu11__SVUint8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.create2.nxv32i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i8> @llvm.vector.insert.nxv32i8.nxv16i8(<vscale x 32 x i8> poison, <vscale x 16 x i8> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 32 x i8> @llvm.vector.insert.nxv32i8.nxv16i8(<vscale x 32 x i8> [[TMP0]], <vscale x 16 x i8> [[X1:%.*]], i64 16)
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP1]]
 //
 svuint8x2_t test_svcreate2_u8(svuint8_t x0, svuint8_t x1)
 {
@@ -92,13 +102,15 @@ svuint8x2_t test_svcreate2_u8(svuint8_t x0, svuint8_t x1)
 
 // CHECK-LABEL: @test_svcreate2_u16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.create2.nxv16i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]])
-// CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i16> @llvm.vector.insert.nxv16i16.nxv8i16(<vscale x 16 x i16> poison, <vscale x 8 x i16> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i16> @llvm.vector.insert.nxv16i16.nxv8i16(<vscale x 16 x i16> [[TMP0]], <vscale x 8 x i16> [[X1:%.*]], i64 8)
+// CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate2_u16u12__SVUint16_tu12__SVUint16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.create2.nxv16i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i16> @llvm.vector.insert.nxv16i16.nxv8i16(<vscale x 16 x i16> poison, <vscale x 8 x i16> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i16> @llvm.vector.insert.nxv16i16.nxv8i16(<vscale x 16 x i16> [[TMP0]], <vscale x 8 x i16> [[X1:%.*]], i64 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP1]]
 //
 svuint16x2_t test_svcreate2_u16(svuint16_t x0, svuint16_t x1)
 {
@@ -107,13 +119,15 @@ svuint16x2_t test_svcreate2_u16(svuint16_t x0, svuint16_t x1)
 
 // CHECK-LABEL: @test_svcreate2_u32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.create2.nxv8i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]])
-// CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i32> @llvm.vector.insert.nxv8i32.nxv4i32(<vscale x 8 x i32> poison, <vscale x 4 x i32> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i32> @llvm.vector.insert.nxv8i32.nxv4i32(<vscale x 8 x i32> [[TMP0]], <vscale x 4 x i32> [[X1:%.*]], i64 4)
+// CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate2_u32u12__SVUint32_tu12__SVUint32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.create2.nxv8i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i32> @llvm.vector.insert.nxv8i32.nxv4i32(<vscale x 8 x i32> poison, <vscale x 4 x i32> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i32> @llvm.vector.insert.nxv8i32.nxv4i32(<vscale x 8 x i32> [[TMP0]], <vscale x 4 x i32> [[X1:%.*]], i64 4)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP1]]
 //
 svuint32x2_t test_svcreate2_u32(svuint32_t x0, svuint32_t x1)
 {
@@ -122,13 +136,15 @@ svuint32x2_t test_svcreate2_u32(svuint32_t x0, svuint32_t x1)
 
 // CHECK-LABEL: @test_svcreate2_u64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.create2.nxv4i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]])
-// CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i64> @llvm.vector.insert.nxv4i64.nxv2i64(<vscale x 4 x i64> poison, <vscale x 2 x i64> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i64> @llvm.vector.insert.nxv4i64.nxv2i64(<vscale x 4 x i64> [[TMP0]], <vscale x 2 x i64> [[X1:%.*]], i64 2)
+// CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate2_u64u12__SVUint64_tu12__SVUint64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.create2.nxv4i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i64> @llvm.vector.insert.nxv4i64.nxv2i64(<vscale x 4 x i64> poison, <vscale x 2 x i64> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i64> @llvm.vector.insert.nxv4i64.nxv2i64(<vscale x 4 x i64> [[TMP0]], <vscale x 2 x i64> [[X1:%.*]], i64 2)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP1]]
 //
 svuint64x2_t test_svcreate2_u64(svuint64_t x0, svuint64_t x1)
 {
@@ -137,13 +153,15 @@ svuint64x2_t test_svcreate2_u64(svuint64_t x0, svuint64_t x1)
 
 // CHECK-LABEL: @test_svcreate2_f16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x half> @llvm.aarch64.sve.tuple.create2.nxv16f16.nxv8f16(<vscale x 8 x half> [[X0:%.*]], <vscale x 8 x half> [[X1:%.*]])
-// CHECK-NEXT:    ret <vscale x 16 x half> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x half> @llvm.vector.insert.nxv16f16.nxv8f16(<vscale x 16 x half> poison, <vscale x 8 x half> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x half> @llvm.vector.insert.nxv16f16.nxv8f16(<vscale x 16 x half> [[TMP0]], <vscale x 8 x half> [[X1:%.*]], i64 8)
+// CHECK-NEXT:    ret <vscale x 16 x half> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate2_f16u13__SVFloat16_tu13__SVFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x half> @llvm.aarch64.sve.tuple.create2.nxv16f16.nxv8f16(<vscale x 8 x half> [[X0:%.*]], <vscale x 8 x half> [[X1:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 16 x half> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x half> @llvm.vector.insert.nxv16f16.nxv8f16(<vscale x 16 x half> poison, <vscale x 8 x half> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x half> @llvm.vector.insert.nxv16f16.nxv8f16(<vscale x 16 x half> [[TMP0]], <vscale x 8 x half> [[X1:%.*]], i64 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x half> [[TMP1]]
 //
 svfloat16x2_t test_svcreate2_f16(svfloat16_t x0, svfloat16_t x1)
 {
@@ -152,13 +170,15 @@ svfloat16x2_t test_svcreate2_f16(svfloat16_t x0, svfloat16_t x1)
 
 // CHECK-LABEL: @test_svcreate2_f32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x float> @llvm.aarch64.sve.tuple.create2.nxv8f32.nxv4f32(<vscale x 4 x float> [[X0:%.*]], <vscale x 4 x float> [[X1:%.*]])
-// CHECK-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x float> @llvm.vector.insert.nxv8f32.nxv4f32(<vscale x 8 x float> poison, <vscale x 4 x float> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x float> @llvm.vector.insert.nxv8f32.nxv4f32(<vscale x 8 x float> [[TMP0]], <vscale x 4 x float> [[X1:%.*]], i64 4)
+// CHECK-NEXT:    ret <vscale x 8 x float> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate2_f32u13__SVFloat32_tu13__SVFloat32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x float> @llvm.aarch64.sve.tuple.create2.nxv8f32.nxv4f32(<vscale x 4 x float> [[X0:%.*]], <vscale x 4 x float> [[X1:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x float> @llvm.vector.insert.nxv8f32.nxv4f32(<vscale x 8 x float> poison, <vscale x 4 x float> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x float> @llvm.vector.insert.nxv8f32.nxv4f32(<vscale x 8 x float> [[TMP0]], <vscale x 4 x float> [[X1:%.*]], i64 4)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x float> [[TMP1]]
 //
 svfloat32x2_t test_svcreate2_f32(svfloat32_t x0, svfloat32_t x1)
 {
@@ -167,13 +187,15 @@ svfloat32x2_t test_svcreate2_f32(svfloat32_t x0, svfloat32_t x1)
 
 // CHECK-LABEL: @test_svcreate2_f64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x double> @llvm.aarch64.sve.tuple.create2.nxv4f64.nxv2f64(<vscale x 2 x double> [[X0:%.*]], <vscale x 2 x double> [[X1:%.*]])
-// CHECK-NEXT:    ret <vscale x 4 x double> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x double> @llvm.vector.insert.nxv4f64.nxv2f64(<vscale x 4 x double> poison, <vscale x 2 x double> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x double> @llvm.vector.insert.nxv4f64.nxv2f64(<vscale x 4 x double> [[TMP0]], <vscale x 2 x double> [[X1:%.*]], i64 2)
+// CHECK-NEXT:    ret <vscale x 4 x double> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate2_f64u13__SVFloat64_tu13__SVFloat64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x double> @llvm.aarch64.sve.tuple.create2.nxv4f64.nxv2f64(<vscale x 2 x double> [[X0:%.*]], <vscale x 2 x double> [[X1:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 4 x double> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x double> @llvm.vector.insert.nxv4f64.nxv2f64(<vscale x 4 x double> poison, <vscale x 2 x double> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x double> @llvm.vector.insert.nxv4f64.nxv2f64(<vscale x 4 x double> [[TMP0]], <vscale x 2 x double> [[X1:%.*]], i64 2)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x double> [[TMP1]]
 //
 svfloat64x2_t test_svcreate2_f64(svfloat64_t x0, svfloat64_t x1)
 {

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3-bfloat.c
index 14414ab7b325..a52c30cdb3f5 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3-bfloat.c
@@ -17,13 +17,17 @@
 
 // CHECK-LABEL: @test_svcreate3_bf16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.create3.nxv24bf16.nxv8bf16(<vscale x 8 x bfloat> [[X0:%.*]], <vscale x 8 x bfloat> [[X1:%.*]], <vscale x 8 x bfloat> [[X2:%.*]])
-// CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x bfloat> @llvm.vector.insert.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> poison, <vscale x 8 x bfloat> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 24 x bfloat> @llvm.vector.insert.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[X1:%.*]], i64 8)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 24 x bfloat> @llvm.vector.insert.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[X2:%.*]], i64 16)
+// CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z19test_svcreate3_bf16u14__SVBFloat16_tu14__SVBFloat16_tu14__SVBFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.create3.nxv24bf16.nxv8bf16(<vscale x 8 x bfloat> [[X0:%.*]], <vscale x 8 x bfloat> [[X1:%.*]], <vscale x 8 x bfloat> [[X2:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x bfloat> @llvm.vector.insert.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> poison, <vscale x 8 x bfloat> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 24 x bfloat> @llvm.vector.insert.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[X1:%.*]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 24 x bfloat> @llvm.vector.insert.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[X2:%.*]], i64 16)
+// CPP-CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP2]]
 //
 svbfloat16x3_t test_svcreate3_bf16(svbfloat16_t x0, svbfloat16_t x1, svbfloat16_t x2)
 {

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3.c
index 75539180141a..63486eacc6e9 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3.c
@@ -17,13 +17,17 @@
 
 // CHECK-LABEL: @test_svcreate3_s8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.create3.nxv48i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]], <vscale x 16 x i8> [[X2:%.*]])
-// CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> poison, <vscale x 16 x i8> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TMP0]], <vscale x 16 x i8> [[X1:%.*]], i64 16)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TMP1]], <vscale x 16 x i8> [[X2:%.*]], i64 32)
+// CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z17test_svcreate3_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.create3.nxv48i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]], <vscale x 16 x i8> [[X2:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> poison, <vscale x 16 x i8> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TMP0]], <vscale x 16 x i8> [[X1:%.*]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TMP1]], <vscale x 16 x i8> [[X2:%.*]], i64 32)
+// CPP-CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP2]]
 //
 svint8x3_t test_svcreate3_s8(svint8_t x0, svint8_t x1, svint8_t x2)
 {
@@ -32,13 +36,17 @@ svint8x3_t test_svcreate3_s8(svint8_t x0, svint8_t x1, svint8_t x2)
 
 // CHECK-LABEL: @test_svcreate3_s16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.create3.nxv24i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]], <vscale x 8 x i16> [[X2:%.*]])
-// CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> poison, <vscale x 8 x i16> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TMP0]], <vscale x 8 x i16> [[X1:%.*]], i64 8)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TMP1]], <vscale x 8 x i16> [[X2:%.*]], i64 16)
+// CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate3_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.create3.nxv24i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]], <vscale x 8 x i16> [[X2:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> poison, <vscale x 8 x i16> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TMP0]], <vscale x 8 x i16> [[X1:%.*]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TMP1]], <vscale x 8 x i16> [[X2:%.*]], i64 16)
+// CPP-CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP2]]
 //
 svint16x3_t test_svcreate3_s16(svint16_t x0, svint16_t x1, svint16_t x2)
 {
@@ -47,13 +55,17 @@ svint16x3_t test_svcreate3_s16(svint16_t x0, svint16_t x1, svint16_t x2)
 
 // CHECK-LABEL: @test_svcreate3_s32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.create3.nxv12i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]], <vscale x 4 x i32> [[X2:%.*]])
-// CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> poison, <vscale x 4 x i32> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TMP0]], <vscale x 4 x i32> [[X1:%.*]], i64 4)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TMP1]], <vscale x 4 x i32> [[X2:%.*]], i64 8)
+// CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate3_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.create3.nxv12i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]], <vscale x 4 x i32> [[X2:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> poison, <vscale x 4 x i32> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TMP0]], <vscale x 4 x i32> [[X1:%.*]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TMP1]], <vscale x 4 x i32> [[X2:%.*]], i64 8)
+// CPP-CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP2]]
 //
 svint32x3_t test_svcreate3_s32(svint32_t x0, svint32_t x1, svint32_t x2)
 {
@@ -62,13 +74,17 @@ svint32x3_t test_svcreate3_s32(svint32_t x0, svint32_t x1, svint32_t x2)
 
 // CHECK-LABEL: @test_svcreate3_s64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.create3.nxv6i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]], <vscale x 2 x i64> [[X2:%.*]])
-// CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> poison, <vscale x 2 x i64> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TMP0]], <vscale x 2 x i64> [[X1:%.*]], i64 2)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TMP1]], <vscale x 2 x i64> [[X2:%.*]], i64 4)
+// CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate3_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.create3.nxv6i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]], <vscale x 2 x i64> [[X2:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> poison, <vscale x 2 x i64> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TMP0]], <vscale x 2 x i64> [[X1:%.*]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TMP1]], <vscale x 2 x i64> [[X2:%.*]], i64 4)
+// CPP-CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP2]]
 //
 svint64x3_t test_svcreate3_s64(svint64_t x0, svint64_t x1, svint64_t x2)
 {
@@ -77,13 +93,17 @@ svint64x3_t test_svcreate3_s64(svint64_t x0, svint64_t x1, svint64_t x2)
 
 // CHECK-LABEL: @test_svcreate3_u8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.create3.nxv48i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]], <vscale x 16 x i8> [[X2:%.*]])
-// CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> poison, <vscale x 16 x i8> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TMP0]], <vscale x 16 x i8> [[X1:%.*]], i64 16)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TMP1]], <vscale x 16 x i8> [[X2:%.*]], i64 32)
+// CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z17test_svcreate3_u8u11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.create3.nxv48i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]], <vscale x 16 x i8> [[X2:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> poison, <vscale x 16 x i8> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TMP0]], <vscale x 16 x i8> [[X1:%.*]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TMP1]], <vscale x 16 x i8> [[X2:%.*]], i64 32)
+// CPP-CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP2]]
 //
 svuint8x3_t test_svcreate3_u8(svuint8_t x0, svuint8_t x1, svuint8_t x2)
 {
@@ -92,13 +112,17 @@ svuint8x3_t test_svcreate3_u8(svuint8_t x0, svuint8_t x1, svuint8_t x2)
 
 // CHECK-LABEL: @test_svcreate3_u16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.create3.nxv24i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]], <vscale x 8 x i16> [[X2:%.*]])
-// CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> poison, <vscale x 8 x i16> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TMP0]], <vscale x 8 x i16> [[X1:%.*]], i64 8)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TMP1]], <vscale x 8 x i16> [[X2:%.*]], i64 16)
+// CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate3_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.create3.nxv24i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]], <vscale x 8 x i16> [[X2:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> poison, <vscale x 8 x i16> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TMP0]], <vscale x 8 x i16> [[X1:%.*]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TMP1]], <vscale x 8 x i16> [[X2:%.*]], i64 16)
+// CPP-CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP2]]
 //
 svuint16x3_t test_svcreate3_u16(svuint16_t x0, svuint16_t x1, svuint16_t x2)
 {
@@ -107,13 +131,17 @@ svuint16x3_t test_svcreate3_u16(svuint16_t x0, svuint16_t x1, svuint16_t x2)
 
 // CHECK-LABEL: @test_svcreate3_u32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.create3.nxv12i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]], <vscale x 4 x i32> [[X2:%.*]])
-// CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> poison, <vscale x 4 x i32> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TMP0]], <vscale x 4 x i32> [[X1:%.*]], i64 4)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TMP1]], <vscale x 4 x i32> [[X2:%.*]], i64 8)
+// CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate3_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.create3.nxv12i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]], <vscale x 4 x i32> [[X2:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> poison, <vscale x 4 x i32> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TMP0]], <vscale x 4 x i32> [[X1:%.*]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TMP1]], <vscale x 4 x i32> [[X2:%.*]], i64 8)
+// CPP-CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP2]]
 //
 svuint32x3_t test_svcreate3_u32(svuint32_t x0, svuint32_t x1, svuint32_t x2)
 {
@@ -122,13 +150,17 @@ svuint32x3_t test_svcreate3_u32(svuint32_t x0, svuint32_t x1, svuint32_t x2)
 
 // CHECK-LABEL: @test_svcreate3_u64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.create3.nxv6i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]], <vscale x 2 x i64> [[X2:%.*]])
-// CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> poison, <vscale x 2 x i64> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TMP0]], <vscale x 2 x i64> [[X1:%.*]], i64 2)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TMP1]], <vscale x 2 x i64> [[X2:%.*]], i64 4)
+// CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate3_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.create3.nxv6i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]], <vscale x 2 x i64> [[X2:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> poison, <vscale x 2 x i64> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TMP0]], <vscale x 2 x i64> [[X1:%.*]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TMP1]], <vscale x 2 x i64> [[X2:%.*]], i64 4)
+// CPP-CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP2]]
 //
 svuint64x3_t test_svcreate3_u64(svuint64_t x0, svuint64_t x1, svuint64_t x2)
 {
@@ -137,13 +169,17 @@ svuint64x3_t test_svcreate3_u64(svuint64_t x0, svuint64_t x1, svuint64_t x2)
 
 // CHECK-LABEL: @test_svcreate3_f16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x half> @llvm.aarch64.sve.tuple.create3.nxv24f16.nxv8f16(<vscale x 8 x half> [[X0:%.*]], <vscale x 8 x half> [[X1:%.*]], <vscale x 8 x half> [[X2:%.*]])
-// CHECK-NEXT:    ret <vscale x 24 x half> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x half> @llvm.vector.insert.nxv24f16.nxv8f16(<vscale x 24 x half> poison, <vscale x 8 x half> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 24 x half> @llvm.vector.insert.nxv24f16.nxv8f16(<vscale x 24 x half> [[TMP0]], <vscale x 8 x half> [[X1:%.*]], i64 8)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 24 x half> @llvm.vector.insert.nxv24f16.nxv8f16(<vscale x 24 x half> [[TMP1]], <vscale x 8 x half> [[X2:%.*]], i64 16)
+// CHECK-NEXT:    ret <vscale x 24 x half> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate3_f16u13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x half> @llvm.aarch64.sve.tuple.create3.nxv24f16.nxv8f16(<vscale x 8 x half> [[X0:%.*]], <vscale x 8 x half> [[X1:%.*]], <vscale x 8 x half> [[X2:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 24 x half> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x half> @llvm.vector.insert.nxv24f16.nxv8f16(<vscale x 24 x half> poison, <vscale x 8 x half> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 24 x half> @llvm.vector.insert.nxv24f16.nxv8f16(<vscale x 24 x half> [[TMP0]], <vscale x 8 x half> [[X1:%.*]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 24 x half> @llvm.vector.insert.nxv24f16.nxv8f16(<vscale x 24 x half> [[TMP1]], <vscale x 8 x half> [[X2:%.*]], i64 16)
+// CPP-CHECK-NEXT:    ret <vscale x 24 x half> [[TMP2]]
 //
 svfloat16x3_t test_svcreate3_f16(svfloat16_t x0, svfloat16_t x1, svfloat16_t x2)
 {
@@ -152,13 +188,17 @@ svfloat16x3_t test_svcreate3_f16(svfloat16_t x0, svfloat16_t x1, svfloat16_t x2)
 
 // CHECK-LABEL: @test_svcreate3_f32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x float> @llvm.aarch64.sve.tuple.create3.nxv12f32.nxv4f32(<vscale x 4 x float> [[X0:%.*]], <vscale x 4 x float> [[X1:%.*]], <vscale x 4 x float> [[X2:%.*]])
-// CHECK-NEXT:    ret <vscale x 12 x float> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x float> @llvm.vector.insert.nxv12f32.nxv4f32(<vscale x 12 x float> poison, <vscale x 4 x float> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 12 x float> @llvm.vector.insert.nxv12f32.nxv4f32(<vscale x 12 x float> [[TMP0]], <vscale x 4 x float> [[X1:%.*]], i64 4)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 12 x float> @llvm.vector.insert.nxv12f32.nxv4f32(<vscale x 12 x float> [[TMP1]], <vscale x 4 x float> [[X2:%.*]], i64 8)
+// CHECK-NEXT:    ret <vscale x 12 x float> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate3_f32u13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x float> @llvm.aarch64.sve.tuple.create3.nxv12f32.nxv4f32(<vscale x 4 x float> [[X0:%.*]], <vscale x 4 x float> [[X1:%.*]], <vscale x 4 x float> [[X2:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 12 x float> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x float> @llvm.vector.insert.nxv12f32.nxv4f32(<vscale x 12 x float> poison, <vscale x 4 x float> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 12 x float> @llvm.vector.insert.nxv12f32.nxv4f32(<vscale x 12 x float> [[TMP0]], <vscale x 4 x float> [[X1:%.*]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 12 x float> @llvm.vector.insert.nxv12f32.nxv4f32(<vscale x 12 x float> [[TMP1]], <vscale x 4 x float> [[X2:%.*]], i64 8)
+// CPP-CHECK-NEXT:    ret <vscale x 12 x float> [[TMP2]]
 //
 svfloat32x3_t test_svcreate3_f32(svfloat32_t x0, svfloat32_t x1, svfloat32_t x2)
 {
@@ -167,13 +207,17 @@ svfloat32x3_t test_svcreate3_f32(svfloat32_t x0, svfloat32_t x1, svfloat32_t x2)
 
 // CHECK-LABEL: @test_svcreate3_f64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x double> @llvm.aarch64.sve.tuple.create3.nxv6f64.nxv2f64(<vscale x 2 x double> [[X0:%.*]], <vscale x 2 x double> [[X1:%.*]], <vscale x 2 x double> [[X2:%.*]])
-// CHECK-NEXT:    ret <vscale x 6 x double> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x double> @llvm.vector.insert.nxv6f64.nxv2f64(<vscale x 6 x double> poison, <vscale x 2 x double> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 6 x double> @llvm.vector.insert.nxv6f64.nxv2f64(<vscale x 6 x double> [[TMP0]], <vscale x 2 x double> [[X1:%.*]], i64 2)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 6 x double> @llvm.vector.insert.nxv6f64.nxv2f64(<vscale x 6 x double> [[TMP1]], <vscale x 2 x double> [[X2:%.*]], i64 4)
+// CHECK-NEXT:    ret <vscale x 6 x double> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate3_f64u13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x double> @llvm.aarch64.sve.tuple.create3.nxv6f64.nxv2f64(<vscale x 2 x double> [[X0:%.*]], <vscale x 2 x double> [[X1:%.*]], <vscale x 2 x double> [[X2:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 6 x double> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x double> @llvm.vector.insert.nxv6f64.nxv2f64(<vscale x 6 x double> poison, <vscale x 2 x double> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 6 x double> @llvm.vector.insert.nxv6f64.nxv2f64(<vscale x 6 x double> [[TMP0]], <vscale x 2 x double> [[X1:%.*]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 6 x double> @llvm.vector.insert.nxv6f64.nxv2f64(<vscale x 6 x double> [[TMP1]], <vscale x 2 x double> [[X2:%.*]], i64 4)
+// CPP-CHECK-NEXT:    ret <vscale x 6 x double> [[TMP2]]
 //
 svfloat64x3_t test_svcreate3_f64(svfloat64_t x0, svfloat64_t x1, svfloat64_t x2)
 {

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4-bfloat.c
index c92362ae1bf8..15c060f90542 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4-bfloat.c
@@ -17,13 +17,19 @@
 
 // CHECK-LABEL: @test_svcreate4_bf16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.create4.nxv32bf16.nxv8bf16(<vscale x 8 x bfloat> [[X0:%.*]], <vscale x 8 x bfloat> [[X1:%.*]], <vscale x 8 x bfloat> [[X2:%.*]], <vscale x 8 x bfloat> [[X4:%.*]])
-// CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.vector.insert.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> poison, <vscale x 8 x bfloat> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 32 x bfloat> @llvm.vector.insert.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[X1:%.*]], i64 8)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 32 x bfloat> @llvm.vector.insert.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[X2:%.*]], i64 16)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 32 x bfloat> @llvm.vector.insert.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[X4:%.*]], i64 24)
+// CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP3]]
 //
 // CPP-CHECK-LABEL: @_Z19test_svcreate4_bf16u14__SVBFloat16_tu14__SVBFloat16_tu14__SVBFloat16_tu14__SVBFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.create4.nxv32bf16.nxv8bf16(<vscale x 8 x bfloat> [[X0:%.*]], <vscale x 8 x bfloat> [[X1:%.*]], <vscale x 8 x bfloat> [[X2:%.*]], <vscale x 8 x bfloat> [[X4:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.vector.insert.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> poison, <vscale x 8 x bfloat> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 32 x bfloat> @llvm.vector.insert.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[X1:%.*]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 32 x bfloat> @llvm.vector.insert.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[X2:%.*]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 32 x bfloat> @llvm.vector.insert.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[X4:%.*]], i64 24)
+// CPP-CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP3]]
 //
 svbfloat16x4_t test_svcreate4_bf16(svbfloat16_t x0, svbfloat16_t x1, svbfloat16_t x2, svbfloat16_t x4)
 {

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4.c
index 6b3f359cb222..e68f6ac91d4d 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4.c
@@ -17,13 +17,19 @@
 
 // CHECK-LABEL: @test_svcreate4_s8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.create4.nxv64i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]], <vscale x 16 x i8> [[X2:%.*]], <vscale x 16 x i8> [[X4:%.*]])
-// CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> poison, <vscale x 16 x i8> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TMP0]], <vscale x 16 x i8> [[X1:%.*]], i64 16)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TMP1]], <vscale x 16 x i8> [[X2:%.*]], i64 32)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TMP2]], <vscale x 16 x i8> [[X4:%.*]], i64 48)
+// CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP3]]
 //
 // CPP-CHECK-LABEL: @_Z17test_svcreate4_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.create4.nxv64i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]], <vscale x 16 x i8> [[X2:%.*]], <vscale x 16 x i8> [[X4:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> poison, <vscale x 16 x i8> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TMP0]], <vscale x 16 x i8> [[X1:%.*]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TMP1]], <vscale x 16 x i8> [[X2:%.*]], i64 32)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TMP2]], <vscale x 16 x i8> [[X4:%.*]], i64 48)
+// CPP-CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP3]]
 //
 svint8x4_t test_svcreate4_s8(svint8_t x0, svint8_t x1, svint8_t x2, svint8_t x4)
 {
@@ -32,13 +38,19 @@ svint8x4_t test_svcreate4_s8(svint8_t x0, svint8_t x1, svint8_t x2, svint8_t x4)
 
 // CHECK-LABEL: @test_svcreate4_s16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.create4.nxv32i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]], <vscale x 8 x i16> [[X2:%.*]], <vscale x 8 x i16> [[X4:%.*]])
-// CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> poison, <vscale x 8 x i16> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TMP0]], <vscale x 8 x i16> [[X1:%.*]], i64 8)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TMP1]], <vscale x 8 x i16> [[X2:%.*]], i64 16)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TMP2]], <vscale x 8 x i16> [[X4:%.*]], i64 24)
+// CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP3]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate4_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.create4.nxv32i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]], <vscale x 8 x i16> [[X2:%.*]], <vscale x 8 x i16> [[X4:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> poison, <vscale x 8 x i16> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TMP0]], <vscale x 8 x i16> [[X1:%.*]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TMP1]], <vscale x 8 x i16> [[X2:%.*]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TMP2]], <vscale x 8 x i16> [[X4:%.*]], i64 24)
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP3]]
 //
 svint16x4_t test_svcreate4_s16(svint16_t x0, svint16_t x1, svint16_t x2, svint16_t x4)
 {
@@ -47,13 +59,19 @@ svint16x4_t test_svcreate4_s16(svint16_t x0, svint16_t x1, svint16_t x2, svint16
 
 // CHECK-LABEL: @test_svcreate4_s32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.create4.nxv16i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]], <vscale x 4 x i32> [[X2:%.*]], <vscale x 4 x i32> [[X4:%.*]])
-// CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> poison, <vscale x 4 x i32> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TMP0]], <vscale x 4 x i32> [[X1:%.*]], i64 4)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TMP1]], <vscale x 4 x i32> [[X2:%.*]], i64 8)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TMP2]], <vscale x 4 x i32> [[X4:%.*]], i64 12)
+// CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP3]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate4_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.create4.nxv16i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]], <vscale x 4 x i32> [[X2:%.*]], <vscale x 4 x i32> [[X4:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> poison, <vscale x 4 x i32> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TMP0]], <vscale x 4 x i32> [[X1:%.*]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TMP1]], <vscale x 4 x i32> [[X2:%.*]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TMP2]], <vscale x 4 x i32> [[X4:%.*]], i64 12)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP3]]
 //
 svint32x4_t test_svcreate4_s32(svint32_t x0, svint32_t x1, svint32_t x2, svint32_t x4)
 {
@@ -62,13 +80,19 @@ svint32x4_t test_svcreate4_s32(svint32_t x0, svint32_t x1, svint32_t x2, svint32
 
 // CHECK-LABEL: @test_svcreate4_s64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.create4.nxv8i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]], <vscale x 2 x i64> [[X2:%.*]], <vscale x 2 x i64> [[X4:%.*]])
-// CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> poison, <vscale x 2 x i64> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TMP0]], <vscale x 2 x i64> [[X1:%.*]], i64 2)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TMP1]], <vscale x 2 x i64> [[X2:%.*]], i64 4)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TMP2]], <vscale x 2 x i64> [[X4:%.*]], i64 6)
+// CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP3]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate4_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.create4.nxv8i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]], <vscale x 2 x i64> [[X2:%.*]], <vscale x 2 x i64> [[X4:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> poison, <vscale x 2 x i64> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TMP0]], <vscale x 2 x i64> [[X1:%.*]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TMP1]], <vscale x 2 x i64> [[X2:%.*]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TMP2]], <vscale x 2 x i64> [[X4:%.*]], i64 6)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP3]]
 //
 svint64x4_t test_svcreate4_s64(svint64_t x0, svint64_t x1, svint64_t x2, svint64_t x4)
 {
@@ -77,13 +101,19 @@ svint64x4_t test_svcreate4_s64(svint64_t x0, svint64_t x1, svint64_t x2, svint64
 
 // CHECK-LABEL: @test_svcreate4_u8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.create4.nxv64i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]], <vscale x 16 x i8> [[X2:%.*]], <vscale x 16 x i8> [[X4:%.*]])
-// CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> poison, <vscale x 16 x i8> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TMP0]], <vscale x 16 x i8> [[X1:%.*]], i64 16)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TMP1]], <vscale x 16 x i8> [[X2:%.*]], i64 32)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TMP2]], <vscale x 16 x i8> [[X4:%.*]], i64 48)
+// CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP3]]
 //
 // CPP-CHECK-LABEL: @_Z17test_svcreate4_u8u11__SVUint8_tu11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.create4.nxv64i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]], <vscale x 16 x i8> [[X2:%.*]], <vscale x 16 x i8> [[X4:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> poison, <vscale x 16 x i8> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TMP0]], <vscale x 16 x i8> [[X1:%.*]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TMP1]], <vscale x 16 x i8> [[X2:%.*]], i64 32)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TMP2]], <vscale x 16 x i8> [[X4:%.*]], i64 48)
+// CPP-CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP3]]
 //
 svuint8x4_t test_svcreate4_u8(svuint8_t x0, svuint8_t x1, svuint8_t x2, svuint8_t x4)
 {
@@ -92,13 +122,19 @@ svuint8x4_t test_svcreate4_u8(svuint8_t x0, svuint8_t x1, svuint8_t x2, svuint8_
 
 // CHECK-LABEL: @test_svcreate4_u16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.create4.nxv32i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]], <vscale x 8 x i16> [[X2:%.*]], <vscale x 8 x i16> [[X4:%.*]])
-// CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> poison, <vscale x 8 x i16> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TMP0]], <vscale x 8 x i16> [[X1:%.*]], i64 8)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TMP1]], <vscale x 8 x i16> [[X2:%.*]], i64 16)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TMP2]], <vscale x 8 x i16> [[X4:%.*]], i64 24)
+// CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP3]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate4_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.create4.nxv32i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]], <vscale x 8 x i16> [[X2:%.*]], <vscale x 8 x i16> [[X4:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> poison, <vscale x 8 x i16> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TMP0]], <vscale x 8 x i16> [[X1:%.*]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TMP1]], <vscale x 8 x i16> [[X2:%.*]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TMP2]], <vscale x 8 x i16> [[X4:%.*]], i64 24)
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP3]]
 //
 svuint16x4_t test_svcreate4_u16(svuint16_t x0, svuint16_t x1, svuint16_t x2, svuint16_t x4)
 {
@@ -107,13 +143,19 @@ svuint16x4_t test_svcreate4_u16(svuint16_t x0, svuint16_t x1, svuint16_t x2, svu
 
 // CHECK-LABEL: @test_svcreate4_u32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.create4.nxv16i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]], <vscale x 4 x i32> [[X2:%.*]], <vscale x 4 x i32> [[X4:%.*]])
-// CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> poison, <vscale x 4 x i32> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TMP0]], <vscale x 4 x i32> [[X1:%.*]], i64 4)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TMP1]], <vscale x 4 x i32> [[X2:%.*]], i64 8)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TMP2]], <vscale x 4 x i32> [[X4:%.*]], i64 12)
+// CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP3]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate4_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.create4.nxv16i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]], <vscale x 4 x i32> [[X2:%.*]], <vscale x 4 x i32> [[X4:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> poison, <vscale x 4 x i32> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TMP0]], <vscale x 4 x i32> [[X1:%.*]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TMP1]], <vscale x 4 x i32> [[X2:%.*]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TMP2]], <vscale x 4 x i32> [[X4:%.*]], i64 12)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP3]]
 //
 svuint32x4_t test_svcreate4_u32(svuint32_t x0, svuint32_t x1, svuint32_t x2, svuint32_t x4)
 {
@@ -122,13 +164,19 @@ svuint32x4_t test_svcreate4_u32(svuint32_t x0, svuint32_t x1, svuint32_t x2, svu
 
 // CHECK-LABEL: @test_svcreate4_u64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.create4.nxv8i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]], <vscale x 2 x i64> [[X2:%.*]], <vscale x 2 x i64> [[X4:%.*]])
-// CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> poison, <vscale x 2 x i64> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TMP0]], <vscale x 2 x i64> [[X1:%.*]], i64 2)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TMP1]], <vscale x 2 x i64> [[X2:%.*]], i64 4)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TMP2]], <vscale x 2 x i64> [[X4:%.*]], i64 6)
+// CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP3]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate4_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.create4.nxv8i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]], <vscale x 2 x i64> [[X2:%.*]], <vscale x 2 x i64> [[X4:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> poison, <vscale x 2 x i64> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TMP0]], <vscale x 2 x i64> [[X1:%.*]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TMP1]], <vscale x 2 x i64> [[X2:%.*]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TMP2]], <vscale x 2 x i64> [[X4:%.*]], i64 6)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP3]]
 //
 svuint64x4_t test_svcreate4_u64(svuint64_t x0, svuint64_t x1, svuint64_t x2, svuint64_t x4)
 {
@@ -137,13 +185,19 @@ svuint64x4_t test_svcreate4_u64(svuint64_t x0, svuint64_t x1, svuint64_t x2, svu
 
 // CHECK-LABEL: @test_svcreate4_f16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x half> @llvm.aarch64.sve.tuple.create4.nxv32f16.nxv8f16(<vscale x 8 x half> [[X0:%.*]], <vscale x 8 x half> [[X1:%.*]], <vscale x 8 x half> [[X2:%.*]], <vscale x 8 x half> [[X4:%.*]])
-// CHECK-NEXT:    ret <vscale x 32 x half> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x half> @llvm.vector.insert.nxv32f16.nxv8f16(<vscale x 32 x half> poison, <vscale x 8 x half> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 32 x half> @llvm.vector.insert.nxv32f16.nxv8f16(<vscale x 32 x half> [[TMP0]], <vscale x 8 x half> [[X1:%.*]], i64 8)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 32 x half> @llvm.vector.insert.nxv32f16.nxv8f16(<vscale x 32 x half> [[TMP1]], <vscale x 8 x half> [[X2:%.*]], i64 16)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 32 x half> @llvm.vector.insert.nxv32f16.nxv8f16(<vscale x 32 x half> [[TMP2]], <vscale x 8 x half> [[X4:%.*]], i64 24)
+// CHECK-NEXT:    ret <vscale x 32 x half> [[TMP3]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate4_f16u13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x half> @llvm.aarch64.sve.tuple.create4.nxv32f16.nxv8f16(<vscale x 8 x half> [[X0:%.*]], <vscale x 8 x half> [[X1:%.*]], <vscale x 8 x half> [[X2:%.*]], <vscale x 8 x half> [[X4:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 32 x half> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x half> @llvm.vector.insert.nxv32f16.nxv8f16(<vscale x 32 x half> poison, <vscale x 8 x half> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 32 x half> @llvm.vector.insert.nxv32f16.nxv8f16(<vscale x 32 x half> [[TMP0]], <vscale x 8 x half> [[X1:%.*]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 32 x half> @llvm.vector.insert.nxv32f16.nxv8f16(<vscale x 32 x half> [[TMP1]], <vscale x 8 x half> [[X2:%.*]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 32 x half> @llvm.vector.insert.nxv32f16.nxv8f16(<vscale x 32 x half> [[TMP2]], <vscale x 8 x half> [[X4:%.*]], i64 24)
+// CPP-CHECK-NEXT:    ret <vscale x 32 x half> [[TMP3]]
 //
 svfloat16x4_t test_svcreate4_f16(svfloat16_t x0, svfloat16_t x1, svfloat16_t x2, svfloat16_t x4)
 {
@@ -152,13 +206,19 @@ svfloat16x4_t test_svcreate4_f16(svfloat16_t x0, svfloat16_t x1, svfloat16_t x2,
 
 // CHECK-LABEL: @test_svcreate4_f32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x float> @llvm.aarch64.sve.tuple.create4.nxv16f32.nxv4f32(<vscale x 4 x float> [[X0:%.*]], <vscale x 4 x float> [[X1:%.*]], <vscale x 4 x float> [[X2:%.*]], <vscale x 4 x float> [[X4:%.*]])
-// CHECK-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x float> @llvm.vector.insert.nxv16f32.nxv4f32(<vscale x 16 x float> poison, <vscale x 4 x float> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x float> @llvm.vector.insert.nxv16f32.nxv4f32(<vscale x 16 x float> [[TMP0]], <vscale x 4 x float> [[X1:%.*]], i64 4)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x float> @llvm.vector.insert.nxv16f32.nxv4f32(<vscale x 16 x float> [[TMP1]], <vscale x 4 x float> [[X2:%.*]], i64 8)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x float> @llvm.vector.insert.nxv16f32.nxv4f32(<vscale x 16 x float> [[TMP2]], <vscale x 4 x float> [[X4:%.*]], i64 12)
+// CHECK-NEXT:    ret <vscale x 16 x float> [[TMP3]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate4_f32u13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x float> @llvm.aarch64.sve.tuple.create4.nxv16f32.nxv4f32(<vscale x 4 x float> [[X0:%.*]], <vscale x 4 x float> [[X1:%.*]], <vscale x 4 x float> [[X2:%.*]], <vscale x 4 x float> [[X4:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x float> @llvm.vector.insert.nxv16f32.nxv4f32(<vscale x 16 x float> poison, <vscale x 4 x float> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x float> @llvm.vector.insert.nxv16f32.nxv4f32(<vscale x 16 x float> [[TMP0]], <vscale x 4 x float> [[X1:%.*]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x float> @llvm.vector.insert.nxv16f32.nxv4f32(<vscale x 16 x float> [[TMP1]], <vscale x 4 x float> [[X2:%.*]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x float> @llvm.vector.insert.nxv16f32.nxv4f32(<vscale x 16 x float> [[TMP2]], <vscale x 4 x float> [[X4:%.*]], i64 12)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x float> [[TMP3]]
 //
 svfloat32x4_t test_svcreate4_f32(svfloat32_t x0, svfloat32_t x1, svfloat32_t x2, svfloat32_t x4)
 {
@@ -167,13 +227,19 @@ svfloat32x4_t test_svcreate4_f32(svfloat32_t x0, svfloat32_t x1, svfloat32_t x2,
 
 // CHECK-LABEL: @test_svcreate4_f64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x double> @llvm.aarch64.sve.tuple.create4.nxv8f64.nxv2f64(<vscale x 2 x double> [[X0:%.*]], <vscale x 2 x double> [[X1:%.*]], <vscale x 2 x double> [[X2:%.*]], <vscale x 2 x double> [[X4:%.*]])
-// CHECK-NEXT:    ret <vscale x 8 x double> [[TMP0]]
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x double> @llvm.vector.insert.nxv8f64.nxv2f64(<vscale x 8 x double> poison, <vscale x 2 x double> [[X0:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x double> @llvm.vector.insert.nxv8f64.nxv2f64(<vscale x 8 x double> [[TMP0]], <vscale x 2 x double> [[X1:%.*]], i64 2)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x double> @llvm.vector.insert.nxv8f64.nxv2f64(<vscale x 8 x double> [[TMP1]], <vscale x 2 x double> [[X2:%.*]], i64 4)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x double> @llvm.vector.insert.nxv8f64.nxv2f64(<vscale x 8 x double> [[TMP2]], <vscale x 2 x double> [[X4:%.*]], i64 6)
+// CHECK-NEXT:    ret <vscale x 8 x double> [[TMP3]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svcreate4_f64u13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x double> @llvm.aarch64.sve.tuple.create4.nxv8f64.nxv2f64(<vscale x 2 x double> [[X0:%.*]], <vscale x 2 x double> [[X1:%.*]], <vscale x 2 x double> [[X2:%.*]], <vscale x 2 x double> [[X4:%.*]])
-// CPP-CHECK-NEXT:    ret <vscale x 8 x double> [[TMP0]]
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x double> @llvm.vector.insert.nxv8f64.nxv2f64(<vscale x 8 x double> poison, <vscale x 2 x double> [[X0:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x double> @llvm.vector.insert.nxv8f64.nxv2f64(<vscale x 8 x double> [[TMP0]], <vscale x 2 x double> [[X1:%.*]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x double> @llvm.vector.insert.nxv8f64.nxv2f64(<vscale x 8 x double> [[TMP1]], <vscale x 2 x double> [[X2:%.*]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x double> @llvm.vector.insert.nxv8f64.nxv2f64(<vscale x 8 x double> [[TMP2]], <vscale x 2 x double> [[X4:%.*]], i64 6)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x double> [[TMP3]]
 //
 svfloat64x4_t test_svcreate4_f64(svfloat64_t x0, svfloat64_t x1, svfloat64_t x2, svfloat64_t x4)
 {

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2-bfloat.c
index ba5f0791eb6a..3fb879f6dc3c 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2-bfloat.c
@@ -17,12 +17,12 @@
 
 // CHECK-LABEL: @test_svget2_bf16_0(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svget2_bf16_014svbfloat16x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 svbfloat16_t test_svget2_bf16_0(svbfloat16x2_t tuple)
@@ -32,12 +32,12 @@ svbfloat16_t test_svget2_bf16_0(svbfloat16x2_t tuple)
 
 // CHECK-LABEL: @test_svget2_bf16_1(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i64 8)
 // CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svget2_bf16_114svbfloat16x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i64 8)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 svbfloat16_t test_svget2_bf16_1(svbfloat16x2_t tuple)

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2.c
index 6e13a8dfc21c..3c5c989ef8e3 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2.c
@@ -16,12 +16,12 @@
 
 // CHECK-LABEL: @test_svget2_s8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[TUPLE:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z14test_svget2_s810svint8x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[TUPLE:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 svint8_t test_svget2_s8(svint8x2_t tuple)
@@ -31,12 +31,12 @@ svint8_t test_svget2_s8(svint8x2_t tuple)
 
 // CHECK-LABEL: @test_svget2_s16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[TUPLE:%.*]], i64 8)
 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget2_s1611svint16x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[TUPLE:%.*]], i64 8)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 svint16_t test_svget2_s16(svint16x2_t tuple)
@@ -46,12 +46,12 @@ svint16_t test_svget2_s16(svint16x2_t tuple)
 
 // CHECK-LABEL: @test_svget2_s32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[TUPLE:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget2_s3211svint32x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[TUPLE:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 svint32_t test_svget2_s32(svint32x2_t tuple)
@@ -61,12 +61,12 @@ svint32_t test_svget2_s32(svint32x2_t tuple)
 
 // CHECK-LABEL: @test_svget2_s64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[TUPLE:%.*]], i64 2)
 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget2_s6411svint64x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[TUPLE:%.*]], i64 2)
 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 svint64_t test_svget2_s64(svint64x2_t tuple)
@@ -76,12 +76,12 @@ svint64_t test_svget2_s64(svint64x2_t tuple)
 
 // CHECK-LABEL: @test_svget2_u8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[TUPLE:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z14test_svget2_u811svuint8x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[TUPLE:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 svuint8_t test_svget2_u8(svuint8x2_t tuple)
@@ -91,12 +91,12 @@ svuint8_t test_svget2_u8(svuint8x2_t tuple)
 
 // CHECK-LABEL: @test_svget2_u16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[TUPLE:%.*]], i64 8)
 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget2_u1612svuint16x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[TUPLE:%.*]], i64 8)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 svuint16_t test_svget2_u16(svuint16x2_t tuple)
@@ -106,12 +106,12 @@ svuint16_t test_svget2_u16(svuint16x2_t tuple)
 
 // CHECK-LABEL: @test_svget2_u32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[TUPLE:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget2_u3212svuint32x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[TUPLE:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 svuint32_t test_svget2_u32(svuint32x2_t tuple)
@@ -121,12 +121,12 @@ svuint32_t test_svget2_u32(svuint32x2_t tuple)
 
 // CHECK-LABEL: @test_svget2_u64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[TUPLE:%.*]], i64 2)
 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget2_u6412svuint64x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[TUPLE:%.*]], i64 2)
 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 svuint64_t test_svget2_u64(svuint64x2_t tuple)
@@ -136,12 +136,12 @@ svuint64_t test_svget2_u64(svuint64x2_t tuple)
 
 // CHECK-LABEL: @test_svget2_f16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[TUPLE:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget2_f1613svfloat16x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[TUPLE:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 svfloat16_t test_svget2_f16(svfloat16x2_t tuple)
@@ -151,12 +151,12 @@ svfloat16_t test_svget2_f16(svfloat16x2_t tuple)
 
 // CHECK-LABEL: @test_svget2_f32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[TUPLE:%.*]], i64 4)
 // CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget2_f3213svfloat32x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[TUPLE:%.*]], i64 4)
 // CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 svfloat32_t test_svget2_f32(svfloat32x2_t tuple)
@@ -166,12 +166,12 @@ svfloat32_t test_svget2_f32(svfloat32x2_t tuple)
 
 // CHECK-LABEL: @test_svget2_f64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[TUPLE:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget2_f6413svfloat64x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[TUPLE:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 svfloat64_t test_svget2_f64(svfloat64x2_t tuple)

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3-bfloat.c
index 757ef4c56d24..c613c0498130 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3-bfloat.c
@@ -17,12 +17,12 @@
 
 // CHECK-LABEL: @test_svget3_bf16_0(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svget3_bf16_014svbfloat16x3_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 svbfloat16_t test_svget3_bf16_0(svbfloat16x3_t tuple)
@@ -32,12 +32,12 @@ svbfloat16_t test_svget3_bf16_0(svbfloat16x3_t tuple)
 
 // CHECK-LABEL: @test_svget3_bf16_1(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i64 8)
 // CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svget3_bf16_114svbfloat16x3_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i64 8)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 svbfloat16_t test_svget3_bf16_1(svbfloat16x3_t tuple)
@@ -47,12 +47,12 @@ svbfloat16_t test_svget3_bf16_1(svbfloat16x3_t tuple)
 
 // CHECK-LABEL: @test_svget3_bf16_2(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i64 16)
 // CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svget3_bf16_214svbfloat16x3_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i64 16)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 svbfloat16_t test_svget3_bf16_2(svbfloat16x3_t tuple)

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3.c
index 55541c260146..6059eb294887 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3.c
@@ -16,12 +16,12 @@
 
 // CHECK-LABEL: @test_svget3_s8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[TUPLE:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z14test_svget3_s810svint8x3_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[TUPLE:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 svint8_t test_svget3_s8(svint8x3_t tuple)
@@ -31,12 +31,12 @@ svint8_t test_svget3_s8(svint8x3_t tuple)
 
 // CHECK-LABEL: @test_svget3_s16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[TUPLE:%.*]], i64 16)
 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget3_s1611svint16x3_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[TUPLE:%.*]], i64 16)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 svint16_t test_svget3_s16(svint16x3_t tuple)
@@ -46,12 +46,12 @@ svint16_t test_svget3_s16(svint16x3_t tuple)
 
 // CHECK-LABEL: @test_svget3_s32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[TUPLE:%.*]], i64 4)
 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget3_s3211svint32x3_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[TUPLE:%.*]], i64 4)
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 svint32_t test_svget3_s32(svint32x3_t tuple)
@@ -61,12 +61,12 @@ svint32_t test_svget3_s32(svint32x3_t tuple)
 
 // CHECK-LABEL: @test_svget3_s64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[TUPLE:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget3_s6411svint64x3_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[TUPLE:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 svint64_t test_svget3_s64(svint64x3_t tuple)
@@ -76,12 +76,12 @@ svint64_t test_svget3_s64(svint64x3_t tuple)
 
 // CHECK-LABEL: @test_svget3_u8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[TUPLE:%.*]], i64 32)
 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z14test_svget3_u811svuint8x3_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[TUPLE:%.*]], i64 32)
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 svuint8_t test_svget3_u8(svuint8x3_t tuple)
@@ -91,12 +91,12 @@ svuint8_t test_svget3_u8(svuint8x3_t tuple)
 
 // CHECK-LABEL: @test_svget3_u16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[TUPLE:%.*]], i64 8)
 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget3_u1612svuint16x3_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[TUPLE:%.*]], i64 8)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 svuint16_t test_svget3_u16(svuint16x3_t tuple)
@@ -106,12 +106,12 @@ svuint16_t test_svget3_u16(svuint16x3_t tuple)
 
 // CHECK-LABEL: @test_svget3_u32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[TUPLE:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget3_u3212svuint32x3_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[TUPLE:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 svuint32_t test_svget3_u32(svuint32x3_t tuple)
@@ -121,12 +121,12 @@ svuint32_t test_svget3_u32(svuint32x3_t tuple)
 
 // CHECK-LABEL: @test_svget3_u64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[TUPLE:%.*]], i64 4)
 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget3_u6412svuint64x3_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[TUPLE:%.*]], i64 4)
 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 svuint64_t test_svget3_u64(svuint64x3_t tuple)
@@ -136,12 +136,12 @@ svuint64_t test_svget3_u64(svuint64x3_t tuple)
 
 // CHECK-LABEL: @test_svget3_f16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv24f16(<vscale x 24 x half> [[TUPLE:%.*]], i64 8)
 // CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget3_f1613svfloat16x3_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv24f16(<vscale x 24 x half> [[TUPLE:%.*]], i64 8)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 svfloat16_t test_svget3_f16(svfloat16x3_t tuple)
@@ -151,12 +151,12 @@ svfloat16_t test_svget3_f16(svfloat16x3_t tuple)
 
 // CHECK-LABEL: @test_svget3_f32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv12f32(<vscale x 12 x float> [[TUPLE:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget3_f3213svfloat32x3_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv12f32(<vscale x 12 x float> [[TUPLE:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 svfloat32_t test_svget3_f32(svfloat32x3_t tuple)
@@ -166,12 +166,12 @@ svfloat32_t test_svget3_f32(svfloat32x3_t tuple)
 
 // CHECK-LABEL: @test_svget3_f64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv6f64(<vscale x 6 x double> [[TUPLE:%.*]], i64 4)
 // CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget3_f6413svfloat64x3_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv6f64(<vscale x 6 x double> [[TUPLE:%.*]], i64 4)
 // CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 svfloat64_t test_svget3_f64(svfloat64x3_t tuple)

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4-bfloat.c
index ec8d2795cc20..bc9d04720c4e 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4-bfloat.c
@@ -17,12 +17,12 @@
 
 // CHECK-LABEL: @test_svget4_bf16_0(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svget4_bf16_014svbfloat16x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 svbfloat16_t test_svget4_bf16_0(svbfloat16x4_t tuple)
@@ -32,12 +32,12 @@ svbfloat16_t test_svget4_bf16_0(svbfloat16x4_t tuple)
 
 // CHECK-LABEL: @test_svget4_bf16_1(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i64 8)
 // CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svget4_bf16_114svbfloat16x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i64 8)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 svbfloat16_t test_svget4_bf16_1(svbfloat16x4_t tuple)
@@ -47,12 +47,12 @@ svbfloat16_t test_svget4_bf16_1(svbfloat16x4_t tuple)
 
 // CHECK-LABEL: @test_svget4_bf16_2(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i64 16)
 // CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svget4_bf16_214svbfloat16x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i64 16)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 svbfloat16_t test_svget4_bf16_2(svbfloat16x4_t tuple)
@@ -62,12 +62,12 @@ svbfloat16_t test_svget4_bf16_2(svbfloat16x4_t tuple)
 
 // CHECK-LABEL: @test_svget4_bf16_3(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 3)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i64 24)
 // CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svget4_bf16_314svbfloat16x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i64 24)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
 //
 svbfloat16_t test_svget4_bf16_3(svbfloat16x4_t tuple)

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4.c
index 5fbd4e36e8d1..862327ae4120 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4.c
@@ -18,12 +18,12 @@
 // several parameters, one for each member of the original struct.
 // CHECK-LABEL: @test_svget4_s8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[TUPLE:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z14test_svget4_s810svint8x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[TUPLE:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 svint8_t test_svget4_s8(svint8x4_t tuple)
@@ -33,12 +33,12 @@ svint8_t test_svget4_s8(svint8x4_t tuple)
 
 // CHECK-LABEL: @test_svget4_s16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[TUPLE:%.*]], i64 16)
 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget4_s1611svint16x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[TUPLE:%.*]], i64 16)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 svint16_t test_svget4_s16(svint16x4_t tuple)
@@ -48,12 +48,12 @@ svint16_t test_svget4_s16(svint16x4_t tuple)
 
 // CHECK-LABEL: @test_svget4_s32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[TUPLE:%.*]], i64 8)
 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget4_s3211svint32x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[TUPLE:%.*]], i64 8)
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 svint32_t test_svget4_s32(svint32x4_t tuple)
@@ -63,12 +63,12 @@ svint32_t test_svget4_s32(svint32x4_t tuple)
 
 // CHECK-LABEL: @test_svget4_s64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[TUPLE:%.*]], i32 3)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[TUPLE:%.*]], i64 6)
 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget4_s6411svint64x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[TUPLE:%.*]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[TUPLE:%.*]], i64 6)
 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 svint64_t test_svget4_s64(svint64x4_t tuple)
@@ -78,12 +78,12 @@ svint64_t test_svget4_s64(svint64x4_t tuple)
 
 // CHECK-LABEL: @test_svget4_u8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[TUPLE:%.*]], i64 32)
 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z14test_svget4_u811svuint8x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[TUPLE:%.*]], i64 32)
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 svuint8_t test_svget4_u8(svuint8x4_t tuple)
@@ -93,12 +93,12 @@ svuint8_t test_svget4_u8(svuint8x4_t tuple)
 
 // CHECK-LABEL: @test_svget4_u16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[TUPLE:%.*]], i32 3)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[TUPLE:%.*]], i64 24)
 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget4_u1612svuint16x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[TUPLE:%.*]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[TUPLE:%.*]], i64 24)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 svuint16_t test_svget4_u16(svuint16x4_t tuple)
@@ -108,12 +108,12 @@ svuint16_t test_svget4_u16(svuint16x4_t tuple)
 
 // CHECK-LABEL: @test_svget4_u32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[TUPLE:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget4_u3212svuint32x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[TUPLE:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 svuint32_t test_svget4_u32(svuint32x4_t tuple)
@@ -123,12 +123,12 @@ svuint32_t test_svget4_u32(svuint32x4_t tuple)
 
 // CHECK-LABEL: @test_svget4_u64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[TUPLE:%.*]], i32 3)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[TUPLE:%.*]], i64 6)
 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget4_u6412svuint64x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[TUPLE:%.*]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[TUPLE:%.*]], i64 6)
 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 svuint64_t test_svget4_u64(svuint64x4_t tuple)
@@ -138,12 +138,12 @@ svuint64_t test_svget4_u64(svuint64x4_t tuple)
 
 // CHECK-LABEL: @test_svget4_f16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[TUPLE:%.*]], i64 16)
 // CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget4_f1613svfloat16x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[TUPLE:%.*]], i64 16)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 svfloat16_t test_svget4_f16(svfloat16x4_t tuple)
@@ -153,12 +153,12 @@ svfloat16_t test_svget4_f16(svfloat16x4_t tuple)
 
 // CHECK-LABEL: @test_svget4_f32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[TUPLE:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget4_f3213svfloat32x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[TUPLE:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 svfloat32_t test_svget4_f32(svfloat32x4_t tuple)
@@ -168,12 +168,12 @@ svfloat32_t test_svget4_f32(svfloat32x4_t tuple)
 
 // CHECK-LABEL: @test_svget4_f64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[TUPLE:%.*]], i64 4)
 // CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svget4_f6413svfloat64x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[TUPLE:%.*]], i64 4)
 // CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 svfloat64_t test_svget4_f64(svfloat64x4_t tuple)

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2-bfloat.c
index 6ee67e7c9ade..8b9220abe615 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2-bfloat.c
@@ -17,12 +17,12 @@
 
 // CHECK-LABEL: @test_svset2_bf16_0(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x bfloat> @llvm.aarch64.sve.tuple.set.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i32 0, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x bfloat> @llvm.vector.insert.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svset2_bf16_014svbfloat16x2_tu14__SVBFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x bfloat> @llvm.aarch64.sve.tuple.set.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i32 0, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x bfloat> @llvm.vector.insert.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
 //
 svbfloat16x2_t test_svset2_bf16_0(svbfloat16x2_t tuple, svbfloat16_t x)
@@ -32,12 +32,12 @@ svbfloat16x2_t test_svset2_bf16_0(svbfloat16x2_t tuple, svbfloat16_t x)
 
 // CHECK-LABEL: @test_svset2_bf16_1(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x bfloat> @llvm.aarch64.sve.tuple.set.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i32 1, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x bfloat> @llvm.vector.insert.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 8)
 // CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svset2_bf16_114svbfloat16x2_tu14__SVBFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x bfloat> @llvm.aarch64.sve.tuple.set.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i32 1, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x bfloat> @llvm.vector.insert.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 8)
 // CPP-CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
 //
 svbfloat16x2_t test_svset2_bf16_1(svbfloat16x2_t tuple, svbfloat16_t x)

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2.c
index c67f9f6cb626..6812c5ebc7aa 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2.c
@@ -17,12 +17,12 @@
 
 // CHECK-LABEL: @test_svset2_s8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.set.nxv32i8.nxv16i8(<vscale x 32 x i8> [[TUPLE:%.*]], i32 1, <vscale x 16 x i8> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i8> @llvm.vector.insert.nxv32i8.nxv16i8(<vscale x 32 x i8> [[TUPLE:%.*]], <vscale x 16 x i8> [[X:%.*]], i64 16)
 // CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z14test_svset2_s810svint8x2_tu10__SVInt8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.set.nxv32i8.nxv16i8(<vscale x 32 x i8> [[TUPLE:%.*]], i32 1, <vscale x 16 x i8> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i8> @llvm.vector.insert.nxv32i8.nxv16i8(<vscale x 32 x i8> [[TUPLE:%.*]], <vscale x 16 x i8> [[X:%.*]], i64 16)
 // CPP-CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 svint8x2_t test_svset2_s8(svint8x2_t tuple, svint8_t x)
@@ -32,12 +32,12 @@ svint8x2_t test_svset2_s8(svint8x2_t tuple, svint8_t x)
 
 // CHECK-LABEL: @test_svset2_s16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.set.nxv16i16.nxv8i16(<vscale x 16 x i16> [[TUPLE:%.*]], i32 0, <vscale x 8 x i16> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i16> @llvm.vector.insert.nxv16i16.nxv8i16(<vscale x 16 x i16> [[TUPLE:%.*]], <vscale x 8 x i16> [[X:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset2_s1611svint16x2_tu11__SVInt16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.set.nxv16i16.nxv8i16(<vscale x 16 x i16> [[TUPLE:%.*]], i32 0, <vscale x 8 x i16> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i16> @llvm.vector.insert.nxv16i16.nxv8i16(<vscale x 16 x i16> [[TUPLE:%.*]], <vscale x 8 x i16> [[X:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 svint16x2_t test_svset2_s16(svint16x2_t tuple, svint16_t x)
@@ -47,12 +47,12 @@ svint16x2_t test_svset2_s16(svint16x2_t tuple, svint16_t x)
 
 // CHECK-LABEL: @test_svset2_s32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.set.nxv8i32.nxv4i32(<vscale x 8 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i32> @llvm.vector.insert.nxv8i32.nxv4i32(<vscale x 8 x i32> [[TUPLE:%.*]], <vscale x 4 x i32> [[X:%.*]], i64 4)
 // CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset2_s3211svint32x2_tu11__SVInt32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.set.nxv8i32.nxv4i32(<vscale x 8 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i32> @llvm.vector.insert.nxv8i32.nxv4i32(<vscale x 8 x i32> [[TUPLE:%.*]], <vscale x 4 x i32> [[X:%.*]], i64 4)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 svint32x2_t test_svset2_s32(svint32x2_t tuple, svint32_t x)
@@ -62,12 +62,12 @@ svint32x2_t test_svset2_s32(svint32x2_t tuple, svint32_t x)
 
 // CHECK-LABEL: @test_svset2_s64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.set.nxv4i64.nxv2i64(<vscale x 4 x i64> [[TUPLE:%.*]], i32 0, <vscale x 2 x i64> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i64> @llvm.vector.insert.nxv4i64.nxv2i64(<vscale x 4 x i64> [[TUPLE:%.*]], <vscale x 2 x i64> [[X:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset2_s6411svint64x2_tu11__SVInt64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.set.nxv4i64.nxv2i64(<vscale x 4 x i64> [[TUPLE:%.*]], i32 0, <vscale x 2 x i64> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i64> @llvm.vector.insert.nxv4i64.nxv2i64(<vscale x 4 x i64> [[TUPLE:%.*]], <vscale x 2 x i64> [[X:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 svint64x2_t test_svset2_s64(svint64x2_t tuple, svint64_t x)
@@ -77,12 +77,12 @@ svint64x2_t test_svset2_s64(svint64x2_t tuple, svint64_t x)
 
 // CHECK-LABEL: @test_svset2_u8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.set.nxv32i8.nxv16i8(<vscale x 32 x i8> [[TUPLE:%.*]], i32 1, <vscale x 16 x i8> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i8> @llvm.vector.insert.nxv32i8.nxv16i8(<vscale x 32 x i8> [[TUPLE:%.*]], <vscale x 16 x i8> [[X:%.*]], i64 16)
 // CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z14test_svset2_u811svuint8x2_tu11__SVUint8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.set.nxv32i8.nxv16i8(<vscale x 32 x i8> [[TUPLE:%.*]], i32 1, <vscale x 16 x i8> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i8> @llvm.vector.insert.nxv32i8.nxv16i8(<vscale x 32 x i8> [[TUPLE:%.*]], <vscale x 16 x i8> [[X:%.*]], i64 16)
 // CPP-CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 svuint8x2_t test_svset2_u8(svuint8x2_t tuple, svuint8_t x)
@@ -92,12 +92,12 @@ svuint8x2_t test_svset2_u8(svuint8x2_t tuple, svuint8_t x)
 
 // CHECK-LABEL: @test_svset2_u16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.set.nxv16i16.nxv8i16(<vscale x 16 x i16> [[TUPLE:%.*]], i32 0, <vscale x 8 x i16> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i16> @llvm.vector.insert.nxv16i16.nxv8i16(<vscale x 16 x i16> [[TUPLE:%.*]], <vscale x 8 x i16> [[X:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset2_u1612svuint16x2_tu12__SVUint16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.set.nxv16i16.nxv8i16(<vscale x 16 x i16> [[TUPLE:%.*]], i32 0, <vscale x 8 x i16> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i16> @llvm.vector.insert.nxv16i16.nxv8i16(<vscale x 16 x i16> [[TUPLE:%.*]], <vscale x 8 x i16> [[X:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 svuint16x2_t test_svset2_u16(svuint16x2_t tuple, svuint16_t x)
@@ -107,12 +107,12 @@ svuint16x2_t test_svset2_u16(svuint16x2_t tuple, svuint16_t x)
 
 // CHECK-LABEL: @test_svset2_u32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.set.nxv8i32.nxv4i32(<vscale x 8 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i32> @llvm.vector.insert.nxv8i32.nxv4i32(<vscale x 8 x i32> [[TUPLE:%.*]], <vscale x 4 x i32> [[X:%.*]], i64 4)
 // CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset2_u3212svuint32x2_tu12__SVUint32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.set.nxv8i32.nxv4i32(<vscale x 8 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i32> @llvm.vector.insert.nxv8i32.nxv4i32(<vscale x 8 x i32> [[TUPLE:%.*]], <vscale x 4 x i32> [[X:%.*]], i64 4)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 svuint32x2_t test_svset2_u32(svuint32x2_t tuple, svuint32_t x)
@@ -122,12 +122,12 @@ svuint32x2_t test_svset2_u32(svuint32x2_t tuple, svuint32_t x)
 
 // CHECK-LABEL: @test_svset2_u64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.set.nxv4i64.nxv2i64(<vscale x 4 x i64> [[TUPLE:%.*]], i32 0, <vscale x 2 x i64> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i64> @llvm.vector.insert.nxv4i64.nxv2i64(<vscale x 4 x i64> [[TUPLE:%.*]], <vscale x 2 x i64> [[X:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset2_u6412svuint64x2_tu12__SVUint64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.set.nxv4i64.nxv2i64(<vscale x 4 x i64> [[TUPLE:%.*]], i32 0, <vscale x 2 x i64> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i64> @llvm.vector.insert.nxv4i64.nxv2i64(<vscale x 4 x i64> [[TUPLE:%.*]], <vscale x 2 x i64> [[X:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 svuint64x2_t test_svset2_u64(svuint64x2_t tuple, svuint64_t x)
@@ -137,12 +137,12 @@ svuint64x2_t test_svset2_u64(svuint64x2_t tuple, svuint64_t x)
 
 // CHECK-LABEL: @test_svset2_f16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x half> @llvm.aarch64.sve.tuple.set.nxv16f16.nxv8f16(<vscale x 16 x half> [[TUPLE:%.*]], i32 1, <vscale x 8 x half> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x half> @llvm.vector.insert.nxv16f16.nxv8f16(<vscale x 16 x half> [[TUPLE:%.*]], <vscale x 8 x half> [[X:%.*]], i64 8)
 // CHECK-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset2_f1613svfloat16x2_tu13__SVFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x half> @llvm.aarch64.sve.tuple.set.nxv16f16.nxv8f16(<vscale x 16 x half> [[TUPLE:%.*]], i32 1, <vscale x 8 x half> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x half> @llvm.vector.insert.nxv16f16.nxv8f16(<vscale x 16 x half> [[TUPLE:%.*]], <vscale x 8 x half> [[X:%.*]], i64 8)
 // CPP-CHECK-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 svfloat16x2_t test_svset2_f16(svfloat16x2_t tuple, svfloat16_t x)
@@ -152,12 +152,12 @@ svfloat16x2_t test_svset2_f16(svfloat16x2_t tuple, svfloat16_t x)
 
 // CHECK-LABEL: @test_svset2_f32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x float> @llvm.aarch64.sve.tuple.set.nxv8f32.nxv4f32(<vscale x 8 x float> [[TUPLE:%.*]], i32 0, <vscale x 4 x float> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x float> @llvm.vector.insert.nxv8f32.nxv4f32(<vscale x 8 x float> [[TUPLE:%.*]], <vscale x 4 x float> [[X:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset2_f3213svfloat32x2_tu13__SVFloat32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x float> @llvm.aarch64.sve.tuple.set.nxv8f32.nxv4f32(<vscale x 8 x float> [[TUPLE:%.*]], i32 0, <vscale x 4 x float> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x float> @llvm.vector.insert.nxv8f32.nxv4f32(<vscale x 8 x float> [[TUPLE:%.*]], <vscale x 4 x float> [[X:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 svfloat32x2_t test_svset2_f32(svfloat32x2_t tuple, svfloat32_t x)
@@ -167,12 +167,12 @@ svfloat32x2_t test_svset2_f32(svfloat32x2_t tuple, svfloat32_t x)
 
 // CHECK-LABEL: @test_svset2_f64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x double> @llvm.aarch64.sve.tuple.set.nxv4f64.nxv2f64(<vscale x 4 x double> [[TUPLE:%.*]], i32 1, <vscale x 2 x double> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x double> @llvm.vector.insert.nxv4f64.nxv2f64(<vscale x 4 x double> [[TUPLE:%.*]], <vscale x 2 x double> [[X:%.*]], i64 2)
 // CHECK-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset2_f6413svfloat64x2_tu13__SVFloat64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x double> @llvm.aarch64.sve.tuple.set.nxv4f64.nxv2f64(<vscale x 4 x double> [[TUPLE:%.*]], i32 1, <vscale x 2 x double> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x double> @llvm.vector.insert.nxv4f64.nxv2f64(<vscale x 4 x double> [[TUPLE:%.*]], <vscale x 2 x double> [[X:%.*]], i64 2)
 // CPP-CHECK-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 svfloat64x2_t test_svset2_f64(svfloat64x2_t tuple, svfloat64_t x)

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3-bfloat.c
index 91527228ebee..7749ba7a07e5 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3-bfloat.c
@@ -18,12 +18,12 @@
 
 // CHECK-LABEL: @test_svset3_bf16_0(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.set.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 0, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x bfloat> @llvm.vector.insert.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svset3_bf16_014svbfloat16x3_tu14__SVBFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.set.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 0, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x bfloat> @llvm.vector.insert.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP0]]
 //
 svbfloat16x3_t test_svset3_bf16_0(svbfloat16x3_t tuple, svbfloat16_t x)
@@ -33,12 +33,12 @@ svbfloat16x3_t test_svset3_bf16_0(svbfloat16x3_t tuple, svbfloat16_t x)
 
 // CHECK-LABEL: @test_svset3_bf16_1(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.set.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 1, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x bfloat> @llvm.vector.insert.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 8)
 // CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svset3_bf16_114svbfloat16x3_tu14__SVBFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.set.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 1, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x bfloat> @llvm.vector.insert.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 8)
 // CPP-CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP0]]
 //
 svbfloat16x3_t test_svset3_bf16_1(svbfloat16x3_t tuple, svbfloat16_t x)
@@ -48,12 +48,12 @@ svbfloat16x3_t test_svset3_bf16_1(svbfloat16x3_t tuple, svbfloat16_t x)
 
 // CHECK-LABEL: @test_svset3_bf16_2(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.set.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 2, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x bfloat> @llvm.vector.insert.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 16)
 // CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svset3_bf16_214svbfloat16x3_tu14__SVBFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.set.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 2, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x bfloat> @llvm.vector.insert.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 16)
 // CPP-CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP0]]
 //
 svbfloat16x3_t test_svset3_bf16_2(svbfloat16x3_t tuple, svbfloat16_t x)

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3.c
index 91eec7d76257..1c0f1580d938 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3.c
@@ -20,12 +20,12 @@
 // several parameters, one for each member of the original struct.
 // CHECK-LABEL: @test_svset3_s8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.set.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TUPLE:%.*]], i32 1, <vscale x 16 x i8> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TUPLE:%.*]], <vscale x 16 x i8> [[X:%.*]], i64 16)
 // CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z14test_svset3_s810svint8x3_tu10__SVInt8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.set.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TUPLE:%.*]], i32 1, <vscale x 16 x i8> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TUPLE:%.*]], <vscale x 16 x i8> [[X:%.*]], i64 16)
 // CPP-CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
 //
 svint8x3_t test_svset3_s8(svint8x3_t tuple, svint8_t x)
@@ -35,12 +35,12 @@ svint8x3_t test_svset3_s8(svint8x3_t tuple, svint8_t x)
 
 // CHECK-LABEL: @test_svset3_s16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.set.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TUPLE:%.*]], i32 2, <vscale x 8 x i16> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TUPLE:%.*]], <vscale x 8 x i16> [[X:%.*]], i64 16)
 // CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset3_s1611svint16x3_tu11__SVInt16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.set.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TUPLE:%.*]], i32 2, <vscale x 8 x i16> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TUPLE:%.*]], <vscale x 8 x i16> [[X:%.*]], i64 16)
 // CPP-CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
 //
 svint16x3_t test_svset3_s16(svint16x3_t tuple, svint16_t x)
@@ -50,12 +50,12 @@ svint16x3_t test_svset3_s16(svint16x3_t tuple, svint16_t x)
 
 // CHECK-LABEL: @test_svset3_s32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.set.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TUPLE:%.*]], i32 0, <vscale x 4 x i32> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TUPLE:%.*]], <vscale x 4 x i32> [[X:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset3_s3211svint32x3_tu11__SVInt32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.set.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TUPLE:%.*]], i32 0, <vscale x 4 x i32> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TUPLE:%.*]], <vscale x 4 x i32> [[X:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
 //
 svint32x3_t test_svset3_s32(svint32x3_t tuple, svint32_t x)
@@ -65,12 +65,12 @@ svint32x3_t test_svset3_s32(svint32x3_t tuple, svint32_t x)
 
 // CHECK-LABEL: @test_svset3_s64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.set.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TUPLE:%.*]], i32 1, <vscale x 2 x i64> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TUPLE:%.*]], <vscale x 2 x i64> [[X:%.*]], i64 2)
 // CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset3_s6411svint64x3_tu11__SVInt64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.set.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TUPLE:%.*]], i32 1, <vscale x 2 x i64> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TUPLE:%.*]], <vscale x 2 x i64> [[X:%.*]], i64 2)
 // CPP-CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
 //
 svint64x3_t test_svset3_s64(svint64x3_t tuple, svint64_t x)
@@ -80,12 +80,12 @@ svint64x3_t test_svset3_s64(svint64x3_t tuple, svint64_t x)
 
 // CHECK-LABEL: @test_svset3_u8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.set.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TUPLE:%.*]], i32 2, <vscale x 16 x i8> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TUPLE:%.*]], <vscale x 16 x i8> [[X:%.*]], i64 32)
 // CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z14test_svset3_u811svuint8x3_tu11__SVUint8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.set.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TUPLE:%.*]], i32 2, <vscale x 16 x i8> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TUPLE:%.*]], <vscale x 16 x i8> [[X:%.*]], i64 32)
 // CPP-CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
 //
 svuint8x3_t test_svset3_u8(svuint8x3_t tuple, svuint8_t x)
@@ -95,12 +95,12 @@ svuint8x3_t test_svset3_u8(svuint8x3_t tuple, svuint8_t x)
 
 // CHECK-LABEL: @test_svset3_u16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.set.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TUPLE:%.*]], i32 0, <vscale x 8 x i16> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TUPLE:%.*]], <vscale x 8 x i16> [[X:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset3_u1612svuint16x3_tu12__SVUint16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.set.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TUPLE:%.*]], i32 0, <vscale x 8 x i16> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TUPLE:%.*]], <vscale x 8 x i16> [[X:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
 //
 svuint16x3_t test_svset3_u16(svuint16x3_t tuple, svuint16_t x)
@@ -110,12 +110,12 @@ svuint16x3_t test_svset3_u16(svuint16x3_t tuple, svuint16_t x)
 
 // CHECK-LABEL: @test_svset3_u32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.set.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TUPLE:%.*]], <vscale x 4 x i32> [[X:%.*]], i64 4)
 // CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset3_u3212svuint32x3_tu12__SVUint32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.set.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TUPLE:%.*]], <vscale x 4 x i32> [[X:%.*]], i64 4)
 // CPP-CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
 //
 svuint32x3_t test_svset3_u32(svuint32x3_t tuple, svuint32_t x)
@@ -125,12 +125,12 @@ svuint32x3_t test_svset3_u32(svuint32x3_t tuple, svuint32_t x)
 
 // CHECK-LABEL: @test_svset3_u64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.set.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TUPLE:%.*]], i32 2, <vscale x 2 x i64> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TUPLE:%.*]], <vscale x 2 x i64> [[X:%.*]], i64 4)
 // CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset3_u6412svuint64x3_tu12__SVUint64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.set.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TUPLE:%.*]], i32 2, <vscale x 2 x i64> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TUPLE:%.*]], <vscale x 2 x i64> [[X:%.*]], i64 4)
 // CPP-CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
 //
 svuint64x3_t test_svset3_u64(svuint64x3_t tuple, svuint64_t x)
@@ -140,12 +140,12 @@ svuint64x3_t test_svset3_u64(svuint64x3_t tuple, svuint64_t x)
 
 // CHECK-LABEL: @test_svset3_f16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x half> @llvm.aarch64.sve.tuple.set.nxv24f16.nxv8f16(<vscale x 24 x half> [[TUPLE:%.*]], i32 0, <vscale x 8 x half> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x half> @llvm.vector.insert.nxv24f16.nxv8f16(<vscale x 24 x half> [[TUPLE:%.*]], <vscale x 8 x half> [[X:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 24 x half> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset3_f1613svfloat16x3_tu13__SVFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x half> @llvm.aarch64.sve.tuple.set.nxv24f16.nxv8f16(<vscale x 24 x half> [[TUPLE:%.*]], i32 0, <vscale x 8 x half> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x half> @llvm.vector.insert.nxv24f16.nxv8f16(<vscale x 24 x half> [[TUPLE:%.*]], <vscale x 8 x half> [[X:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 24 x half> [[TMP0]]
 //
 svfloat16x3_t test_svset3_f16(svfloat16x3_t tuple, svfloat16_t x)
@@ -155,12 +155,12 @@ svfloat16x3_t test_svset3_f16(svfloat16x3_t tuple, svfloat16_t x)
 
 // CHECK-LABEL: @test_svset3_f32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x float> @llvm.aarch64.sve.tuple.set.nxv12f32.nxv4f32(<vscale x 12 x float> [[TUPLE:%.*]], i32 1, <vscale x 4 x float> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x float> @llvm.vector.insert.nxv12f32.nxv4f32(<vscale x 12 x float> [[TUPLE:%.*]], <vscale x 4 x float> [[X:%.*]], i64 4)
 // CHECK-NEXT:    ret <vscale x 12 x float> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset3_f3213svfloat32x3_tu13__SVFloat32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x float> @llvm.aarch64.sve.tuple.set.nxv12f32.nxv4f32(<vscale x 12 x float> [[TUPLE:%.*]], i32 1, <vscale x 4 x float> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x float> @llvm.vector.insert.nxv12f32.nxv4f32(<vscale x 12 x float> [[TUPLE:%.*]], <vscale x 4 x float> [[X:%.*]], i64 4)
 // CPP-CHECK-NEXT:    ret <vscale x 12 x float> [[TMP0]]
 //
 svfloat32x3_t test_svset3_f32(svfloat32x3_t tuple, svfloat32_t x)
@@ -170,12 +170,12 @@ svfloat32x3_t test_svset3_f32(svfloat32x3_t tuple, svfloat32_t x)
 
 // CHECK-LABEL: @test_svset3_f64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x double> @llvm.aarch64.sve.tuple.set.nxv6f64.nxv2f64(<vscale x 6 x double> [[TUPLE:%.*]], i32 2, <vscale x 2 x double> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x double> @llvm.vector.insert.nxv6f64.nxv2f64(<vscale x 6 x double> [[TUPLE:%.*]], <vscale x 2 x double> [[X:%.*]], i64 4)
 // CHECK-NEXT:    ret <vscale x 6 x double> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset3_f6413svfloat64x3_tu13__SVFloat64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x double> @llvm.aarch64.sve.tuple.set.nxv6f64.nxv2f64(<vscale x 6 x double> [[TUPLE:%.*]], i32 2, <vscale x 2 x double> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x double> @llvm.vector.insert.nxv6f64.nxv2f64(<vscale x 6 x double> [[TUPLE:%.*]], <vscale x 2 x double> [[X:%.*]], i64 4)
 // CPP-CHECK-NEXT:    ret <vscale x 6 x double> [[TMP0]]
 //
 svfloat64x3_t test_svset3_f64(svfloat64x3_t tuple, svfloat64_t x)

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4-bfloat.c
index e17055528037..bd9666366df8 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4-bfloat.c
@@ -18,12 +18,12 @@
 
 // CHECK-LABEL: @test_svset4_bf16_0(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 0, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.vector.insert.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 0)
 // CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svset4_bf16_014svbfloat16x4_tu14__SVBFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 0, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.vector.insert.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 0)
 // CPP-CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
 //
 svbfloat16x4_t test_svset4_bf16_0(svbfloat16x4_t tuple, svbfloat16_t x)
@@ -33,12 +33,12 @@ svbfloat16x4_t test_svset4_bf16_0(svbfloat16x4_t tuple, svbfloat16_t x)
 
 // CHECK-LABEL: @test_svset4_bf16_1(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 1, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.vector.insert.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 8)
 // CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svset4_bf16_114svbfloat16x4_tu14__SVBFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 1, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.vector.insert.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 8)
 // CPP-CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
 //
 svbfloat16x4_t test_svset4_bf16_1(svbfloat16x4_t tuple, svbfloat16_t x)
@@ -48,12 +48,12 @@ svbfloat16x4_t test_svset4_bf16_1(svbfloat16x4_t tuple, svbfloat16_t x)
 
 // CHECK-LABEL: @test_svset4_bf16_2(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 2, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.vector.insert.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 16)
 // CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svset4_bf16_214svbfloat16x4_tu14__SVBFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 2, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.vector.insert.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 16)
 // CPP-CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
 //
 svbfloat16x4_t test_svset4_bf16_2(svbfloat16x4_t tuple, svbfloat16_t x)
@@ -63,12 +63,12 @@ svbfloat16x4_t test_svset4_bf16_2(svbfloat16x4_t tuple, svbfloat16_t x)
 
 // CHECK-LABEL: @test_svset4_bf16_3(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 3, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.vector.insert.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 24)
 // CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svset4_bf16_314svbfloat16x4_tu14__SVBFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 3, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x bfloat> @llvm.vector.insert.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], <vscale x 8 x bfloat> [[X:%.*]], i64 24)
 // CPP-CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
 //
 svbfloat16x4_t test_svset4_bf16_3(svbfloat16x4_t tuple, svbfloat16_t x)

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4.c
index 93a85a164f48..1f6ee26451d7 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4.c
@@ -18,12 +18,12 @@
 
 // CHECK-LABEL: @test_svset4_s8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.set.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TUPLE:%.*]], i32 1, <vscale x 16 x i8> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TUPLE:%.*]], <vscale x 16 x i8> [[X:%.*]], i64 16)
 // CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z14test_svset4_s810svint8x4_tu10__SVInt8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.set.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TUPLE:%.*]], i32 1, <vscale x 16 x i8> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TUPLE:%.*]], <vscale x 16 x i8> [[X:%.*]], i64 16)
 // CPP-CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 svint8x4_t test_svset4_s8(svint8x4_t tuple, svint8_t x)
@@ -33,12 +33,12 @@ svint8x4_t test_svset4_s8(svint8x4_t tuple, svint8_t x)
 
 // CHECK-LABEL: @test_svset4_s16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.set.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TUPLE:%.*]], i32 3, <vscale x 8 x i16> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TUPLE:%.*]], <vscale x 8 x i16> [[X:%.*]], i64 24)
 // CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset4_s1611svint16x4_tu11__SVInt16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.set.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TUPLE:%.*]], i32 3, <vscale x 8 x i16> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TUPLE:%.*]], <vscale x 8 x i16> [[X:%.*]], i64 24)
 // CPP-CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 svint16x4_t test_svset4_s16(svint16x4_t tuple, svint16_t x)
@@ -48,12 +48,12 @@ svint16x4_t test_svset4_s16(svint16x4_t tuple, svint16_t x)
 
 // CHECK-LABEL: @test_svset4_s32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.set.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TUPLE:%.*]], <vscale x 4 x i32> [[X:%.*]], i64 4)
 // CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset4_s3211svint32x4_tu11__SVInt32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.set.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TUPLE:%.*]], <vscale x 4 x i32> [[X:%.*]], i64 4)
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 svint32x4_t test_svset4_s32(svint32x4_t tuple, svint32_t x)
@@ -63,12 +63,12 @@ svint32x4_t test_svset4_s32(svint32x4_t tuple, svint32_t x)
 
 // CHECK-LABEL: @test_svset4_s64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.set.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TUPLE:%.*]], i32 1, <vscale x 2 x i64> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TUPLE:%.*]], <vscale x 2 x i64> [[X:%.*]], i64 2)
 // CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset4_s6411svint64x4_tu11__SVInt64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.set.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TUPLE:%.*]], i32 1, <vscale x 2 x i64> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TUPLE:%.*]], <vscale x 2 x i64> [[X:%.*]], i64 2)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 svint64x4_t test_svset4_s64(svint64x4_t tuple, svint64_t x)
@@ -78,12 +78,12 @@ svint64x4_t test_svset4_s64(svint64x4_t tuple, svint64_t x)
 
 // CHECK-LABEL: @test_svset4_u8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.set.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TUPLE:%.*]], i32 3, <vscale x 16 x i8> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TUPLE:%.*]], <vscale x 16 x i8> [[X:%.*]], i64 48)
 // CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z14test_svset4_u811svuint8x4_tu11__SVUint8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.set.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TUPLE:%.*]], i32 3, <vscale x 16 x i8> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 64 x i8> @llvm.vector.insert.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TUPLE:%.*]], <vscale x 16 x i8> [[X:%.*]], i64 48)
 // CPP-CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 svuint8x4_t test_svset4_u8(svuint8x4_t tuple, svuint8_t x)
@@ -93,12 +93,12 @@ svuint8x4_t test_svset4_u8(svuint8x4_t tuple, svuint8_t x)
 
 // CHECK-LABEL: @test_svset4_u16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.set.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TUPLE:%.*]], i32 1, <vscale x 8 x i16> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TUPLE:%.*]], <vscale x 8 x i16> [[X:%.*]], i64 8)
 // CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset4_u1612svuint16x4_tu12__SVUint16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.set.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TUPLE:%.*]], i32 1, <vscale x 8 x i16> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x i16> @llvm.vector.insert.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TUPLE:%.*]], <vscale x 8 x i16> [[X:%.*]], i64 8)
 // CPP-CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 svuint16x4_t test_svset4_u16(svuint16x4_t tuple, svuint16_t x)
@@ -108,12 +108,12 @@ svuint16x4_t test_svset4_u16(svuint16x4_t tuple, svuint16_t x)
 
 // CHECK-LABEL: @test_svset4_u32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.set.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TUPLE:%.*]], <vscale x 4 x i32> [[X:%.*]], i64 4)
 // CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset4_u3212svuint32x4_tu12__SVUint32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.set.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TUPLE:%.*]], <vscale x 4 x i32> [[X:%.*]], i64 4)
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 svuint32x4_t test_svset4_u32(svuint32x4_t tuple, svuint32_t x)
@@ -123,12 +123,12 @@ svuint32x4_t test_svset4_u32(svuint32x4_t tuple, svuint32_t x)
 
 // CHECK-LABEL: @test_svset4_u64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.set.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TUPLE:%.*]], i32 3, <vscale x 2 x i64> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TUPLE:%.*]], <vscale x 2 x i64> [[X:%.*]], i64 6)
 // CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset4_u6412svuint64x4_tu12__SVUint64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.set.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TUPLE:%.*]], i32 3, <vscale x 2 x i64> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i64> @llvm.vector.insert.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TUPLE:%.*]], <vscale x 2 x i64> [[X:%.*]], i64 6)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 svuint64x4_t test_svset4_u64(svuint64x4_t tuple, svuint64_t x)
@@ -138,12 +138,12 @@ svuint64x4_t test_svset4_u64(svuint64x4_t tuple, svuint64_t x)
 
 // CHECK-LABEL: @test_svset4_f16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x half> @llvm.aarch64.sve.tuple.set.nxv32f16.nxv8f16(<vscale x 32 x half> [[TUPLE:%.*]], i32 1, <vscale x 8 x half> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x half> @llvm.vector.insert.nxv32f16.nxv8f16(<vscale x 32 x half> [[TUPLE:%.*]], <vscale x 8 x half> [[X:%.*]], i64 8)
 // CHECK-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset4_f1613svfloat16x4_tu13__SVFloat16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x half> @llvm.aarch64.sve.tuple.set.nxv32f16.nxv8f16(<vscale x 32 x half> [[TUPLE:%.*]], i32 1, <vscale x 8 x half> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 32 x half> @llvm.vector.insert.nxv32f16.nxv8f16(<vscale x 32 x half> [[TUPLE:%.*]], <vscale x 8 x half> [[X:%.*]], i64 8)
 // CPP-CHECK-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 svfloat16x4_t test_svset4_f16(svfloat16x4_t tuple, svfloat16_t x)
@@ -153,12 +153,12 @@ svfloat16x4_t test_svset4_f16(svfloat16x4_t tuple, svfloat16_t x)
 
 // CHECK-LABEL: @test_svset4_f32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x float> @llvm.aarch64.sve.tuple.set.nxv16f32.nxv4f32(<vscale x 16 x float> [[TUPLE:%.*]], i32 1, <vscale x 4 x float> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x float> @llvm.vector.insert.nxv16f32.nxv4f32(<vscale x 16 x float> [[TUPLE:%.*]], <vscale x 4 x float> [[X:%.*]], i64 4)
 // CHECK-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset4_f3213svfloat32x4_tu13__SVFloat32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x float> @llvm.aarch64.sve.tuple.set.nxv16f32.nxv4f32(<vscale x 16 x float> [[TUPLE:%.*]], i32 1, <vscale x 4 x float> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x float> @llvm.vector.insert.nxv16f32.nxv4f32(<vscale x 16 x float> [[TUPLE:%.*]], <vscale x 4 x float> [[X:%.*]], i64 4)
 // CPP-CHECK-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 svfloat32x4_t test_svset4_f32(svfloat32x4_t tuple, svfloat32_t x)
@@ -168,12 +168,12 @@ svfloat32x4_t test_svset4_f32(svfloat32x4_t tuple, svfloat32_t x)
 
 // CHECK-LABEL: @test_svset4_f64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x double> @llvm.aarch64.sve.tuple.set.nxv8f64.nxv2f64(<vscale x 8 x double> [[TUPLE:%.*]], i32 3, <vscale x 2 x double> [[X:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x double> @llvm.vector.insert.nxv8f64.nxv2f64(<vscale x 8 x double> [[TUPLE:%.*]], <vscale x 2 x double> [[X:%.*]], i64 6)
 // CHECK-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svset4_f6413svfloat64x4_tu13__SVFloat64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x double> @llvm.aarch64.sve.tuple.set.nxv8f64.nxv2f64(<vscale x 8 x double> [[TUPLE:%.*]], i32 3, <vscale x 2 x double> [[X:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x double> @llvm.vector.insert.nxv8f64.nxv2f64(<vscale x 8 x double> [[TUPLE:%.*]], <vscale x 2 x double> [[X:%.*]], i64 6)
 // CPP-CHECK-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 svfloat64x4_t test_svset4_f64(svfloat64x4_t tuple, svfloat64_t x)

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2-bfloat.c
index 0a82576afbd9..5b99f4d69b14 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2-bfloat.c
@@ -17,16 +17,16 @@
 // CHECK-LABEL: @test_svst2_bf16(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i64 8)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv8bf16(<vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z15test_svst2_bf16u10__SVBool_tPu6__bf1614svbfloat16x2_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv8bf16(<vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -40,8 +40,8 @@ void test_svst2_bf16(svbool_t pg, bfloat16_t *base, svbfloat16x2_t data)
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i64 8)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv8bf16(<vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -50,8 +50,8 @@ void test_svst2_bf16(svbool_t pg, bfloat16_t *base, svbfloat16x2_t data)
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv8bf16(<vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2.c
index d54d5467c79f..5fc572362859 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2.c
@@ -16,15 +16,15 @@
 
 // CHECK-LABEL: @test_svst2_s8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i64 16)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z13test_svst2_s8u10__SVBool_tPa10svint8x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i64 16)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -36,16 +36,16 @@ void test_svst2_s8(svbool_t pg, int8_t *base, svint8x2_t data)
 // CHECK-LABEL: @test_svst2_s16(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i64 8)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst2_s16u10__SVBool_tPs11svint16x2_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -57,16 +57,16 @@ void test_svst2_s16(svbool_t pg, int16_t *base, svint16x2_t data)
 // CHECK-LABEL: @test_svst2_s32(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i64 4)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst2_s32u10__SVBool_tPi11svint32x2_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i64 4)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -78,16 +78,16 @@ void test_svst2_s32(svbool_t pg, int32_t *base, svint32x2_t data)
 // CHECK-LABEL: @test_svst2_s64(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i64 2)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst2_s64u10__SVBool_tPl11svint64x2_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i64 2)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -98,15 +98,15 @@ void test_svst2_s64(svbool_t pg, int64_t *base, svint64x2_t data)
 
 // CHECK-LABEL: @test_svst2_u8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i64 16)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z13test_svst2_u8u10__SVBool_tPh11svuint8x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i64 16)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -118,16 +118,16 @@ void test_svst2_u8(svbool_t pg, uint8_t *base, svuint8x2_t data)
 // CHECK-LABEL: @test_svst2_u16(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i64 8)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst2_u16u10__SVBool_tPt12svuint16x2_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -139,16 +139,16 @@ void test_svst2_u16(svbool_t pg, uint16_t *base, svuint16x2_t data)
 // CHECK-LABEL: @test_svst2_u32(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i64 4)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst2_u32u10__SVBool_tPj12svuint32x2_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i64 4)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -160,16 +160,16 @@ void test_svst2_u32(svbool_t pg, uint32_t *base, svuint32x2_t data)
 // CHECK-LABEL: @test_svst2_u64(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i64 2)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst2_u64u10__SVBool_tPm12svuint64x2_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i64 2)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -181,16 +181,16 @@ void test_svst2_u64(svbool_t pg, uint64_t *base, svuint64x2_t data)
 // CHECK-LABEL: @test_svst2_f16(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i64 8)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv8f16(<vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst2_f16u10__SVBool_tPDh13svfloat16x2_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv8f16(<vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -202,16 +202,16 @@ void test_svst2_f16(svbool_t pg, float16_t *base, svfloat16x2_t data)
 // CHECK-LABEL: @test_svst2_f32(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i64 4)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv4f32(<vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst2_f32u10__SVBool_tPf13svfloat32x2_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i64 4)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv4f32(<vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -223,16 +223,16 @@ void test_svst2_f32(svbool_t pg, float32_t *base, svfloat32x2_t data)
 // CHECK-LABEL: @test_svst2_f64(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i64 2)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv2f64(<vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst2_f64u10__SVBool_tPd13svfloat64x2_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i64 2)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv2f64(<vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -245,8 +245,8 @@ void test_svst2_f64(svbool_t pg, float64_t *base, svfloat64x2_t data)
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
 // CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i64 16)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
 // CHECK-NEXT:    ret void
 //
@@ -254,8 +254,8 @@ void test_svst2_f64(svbool_t pg, float64_t *base, svfloat64x2_t data)
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i64 16)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -269,8 +269,8 @@ void test_svst2_vnum_s8(svbool_t pg, int8_t *base, int64_t vnum, svint8x2_t data
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i64 8)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -279,8 +279,8 @@ void test_svst2_vnum_s8(svbool_t pg, int8_t *base, int64_t vnum, svint8x2_t data
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -294,8 +294,8 @@ void test_svst2_vnum_s16(svbool_t pg, int16_t *base, int64_t vnum, svint16x2_t d
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i64 4)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -304,8 +304,8 @@ void test_svst2_vnum_s16(svbool_t pg, int16_t *base, int64_t vnum, svint16x2_t d
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i64 4)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -319,8 +319,8 @@ void test_svst2_vnum_s32(svbool_t pg, int32_t *base, int64_t vnum, svint32x2_t d
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i64 2)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -329,8 +329,8 @@ void test_svst2_vnum_s32(svbool_t pg, int32_t *base, int64_t vnum, svint32x2_t d
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i64 2)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -343,8 +343,8 @@ void test_svst2_vnum_s64(svbool_t pg, int64_t *base, int64_t vnum, svint64x2_t d
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
 // CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i64 16)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
 // CHECK-NEXT:    ret void
 //
@@ -352,8 +352,8 @@ void test_svst2_vnum_s64(svbool_t pg, int64_t *base, int64_t vnum, svint64x2_t d
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i64 16)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -367,8 +367,8 @@ void test_svst2_vnum_u8(svbool_t pg, uint8_t *base, int64_t vnum, svuint8x2_t da
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i64 8)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -377,8 +377,8 @@ void test_svst2_vnum_u8(svbool_t pg, uint8_t *base, int64_t vnum, svuint8x2_t da
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -392,8 +392,8 @@ void test_svst2_vnum_u16(svbool_t pg, uint16_t *base, int64_t vnum, svuint16x2_t
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i64 4)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -402,8 +402,8 @@ void test_svst2_vnum_u16(svbool_t pg, uint16_t *base, int64_t vnum, svuint16x2_t
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i64 4)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -417,8 +417,8 @@ void test_svst2_vnum_u32(svbool_t pg, uint32_t *base, int64_t vnum, svuint32x2_t
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i64 2)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -427,8 +427,8 @@ void test_svst2_vnum_u32(svbool_t pg, uint32_t *base, int64_t vnum, svuint32x2_t
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i64 2)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -442,8 +442,8 @@ void test_svst2_vnum_u64(svbool_t pg, uint64_t *base, int64_t vnum, svuint64x2_t
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i64 8)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv8f16(<vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -452,8 +452,8 @@ void test_svst2_vnum_u64(svbool_t pg, uint64_t *base, int64_t vnum, svuint64x2_t
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv8f16(<vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -467,8 +467,8 @@ void test_svst2_vnum_f16(svbool_t pg, float16_t *base, int64_t vnum, svfloat16x2
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i64 4)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv4f32(<vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], <vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -477,8 +477,8 @@ void test_svst2_vnum_f16(svbool_t pg, float16_t *base, int64_t vnum, svfloat16x2
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i64 4)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv4f32(<vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], <vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -492,8 +492,8 @@ void test_svst2_vnum_f32(svbool_t pg, float32_t *base, int64_t vnum, svfloat32x2
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i64 2)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv2f64(<vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], <vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -502,8 +502,8 @@ void test_svst2_vnum_f32(svbool_t pg, float32_t *base, int64_t vnum, svfloat32x2
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i64 2)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st2.nxv2f64(<vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], <vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3-bfloat.c
index 4db1f6ec6660..80065a5661f8 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3-bfloat.c
@@ -18,18 +18,18 @@
 // CHECK-LABEL: @test_svst3_bf16(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i64 16)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv8bf16(<vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z15test_svst3_bf16u10__SVBool_tPu6__bf1614svbfloat16x3_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i64 16)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv8bf16(<vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -43,9 +43,9 @@ void test_svst3_bf16(svbool_t pg, bfloat16_t *base, svbfloat16x3_t data)
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i64 16)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv8bf16(<vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x bfloat> [[TMP5]], <vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -54,9 +54,9 @@ void test_svst3_bf16(svbool_t pg, bfloat16_t *base, svbfloat16x3_t data)
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i64 16)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv8bf16(<vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x bfloat> [[TMP5]], <vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3.c
index 00c754eba7ab..60a54859dfba 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3.c
@@ -16,17 +16,17 @@
 
 // CHECK-LABEL: @test_svst3_s8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i64 16)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i64 32)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z13test_svst3_s8u10__SVBool_tPa10svint8x3_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i64 32)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -38,18 +38,18 @@ void test_svst3_s8(svbool_t pg, int8_t *base, svint8x3_t data)
 // CHECK-LABEL: @test_svst3_s16(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i64 16)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst3_s16u10__SVBool_tPs11svint16x3_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i64 16)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -61,18 +61,18 @@ void test_svst3_s16(svbool_t pg, int16_t *base, svint16x3_t data)
 // CHECK-LABEL: @test_svst3_s32(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i64 8)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst3_s32u10__SVBool_tPi11svint32x3_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -84,18 +84,18 @@ void test_svst3_s32(svbool_t pg, int32_t *base, svint32x3_t data)
 // CHECK-LABEL: @test_svst3_s64(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i64 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i64 4)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst3_s64u10__SVBool_tPl11svint64x3_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i64 4)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -106,17 +106,17 @@ void test_svst3_s64(svbool_t pg, int64_t *base, svint64x3_t data)
 
 // CHECK-LABEL: @test_svst3_u8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i64 16)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i64 32)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z13test_svst3_u8u10__SVBool_tPh11svuint8x3_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i64 32)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -128,18 +128,18 @@ void test_svst3_u8(svbool_t pg, uint8_t *base, svuint8x3_t data)
 // CHECK-LABEL: @test_svst3_u16(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i64 16)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst3_u16u10__SVBool_tPt12svuint16x3_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i64 16)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -151,18 +151,18 @@ void test_svst3_u16(svbool_t pg, uint16_t *base, svuint16x3_t data)
 // CHECK-LABEL: @test_svst3_u32(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i64 8)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst3_u32u10__SVBool_tPj12svuint32x3_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -174,18 +174,18 @@ void test_svst3_u32(svbool_t pg, uint32_t *base, svuint32x3_t data)
 // CHECK-LABEL: @test_svst3_u64(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i64 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i64 4)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst3_u64u10__SVBool_tPm12svuint64x3_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i64 4)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -197,18 +197,18 @@ void test_svst3_u64(svbool_t pg, uint64_t *base, svuint64x3_t data)
 // CHECK-LABEL: @test_svst3_f16(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i64 16)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv8f16(<vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], <vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst3_f16u10__SVBool_tPDh13svfloat16x3_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i64 16)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv8f16(<vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], <vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -220,18 +220,18 @@ void test_svst3_f16(svbool_t pg, float16_t *base, svfloat16x3_t data)
 // CHECK-LABEL: @test_svst3_f32(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i64 8)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv4f32(<vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP3]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst3_f32u10__SVBool_tPf13svfloat32x3_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv4f32(<vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP3]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -243,18 +243,18 @@ void test_svst3_f32(svbool_t pg, float32_t *base, svfloat32x3_t data)
 // CHECK-LABEL: @test_svst3_f64(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i64 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i64 4)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv2f64(<vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP3]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst3_f64u10__SVBool_tPd13svfloat64x3_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i64 4)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv2f64(<vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP3]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -267,9 +267,9 @@ void test_svst3_f64(svbool_t pg, float64_t *base, svfloat64x3_t data)
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
 // CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i64 16)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i64 32)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
 // CHECK-NEXT:    ret void
 //
@@ -277,9 +277,9 @@ void test_svst3_f64(svbool_t pg, float64_t *base, svfloat64x3_t data)
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i64 32)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -293,9 +293,9 @@ void test_svst3_vnum_s8(svbool_t pg, int8_t *base, int64_t vnum, svint8x3_t data
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i64 16)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -304,9 +304,9 @@ void test_svst3_vnum_s8(svbool_t pg, int8_t *base, int64_t vnum, svint8x3_t data
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i64 16)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -320,9 +320,9 @@ void test_svst3_vnum_s16(svbool_t pg, int16_t *base, int64_t vnum, svint16x3_t d
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i64 8)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i32> [[TMP5]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -331,9 +331,9 @@ void test_svst3_vnum_s16(svbool_t pg, int16_t *base, int64_t vnum, svint16x3_t d
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i32> [[TMP5]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -347,9 +347,9 @@ void test_svst3_vnum_s32(svbool_t pg, int32_t *base, int64_t vnum, svint32x3_t d
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i64 2)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i64 4)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -358,9 +358,9 @@ void test_svst3_vnum_s32(svbool_t pg, int32_t *base, int64_t vnum, svint32x3_t d
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i64 4)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -373,9 +373,9 @@ void test_svst3_vnum_s64(svbool_t pg, int64_t *base, int64_t vnum, svint64x3_t d
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
 // CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i64 16)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i64 32)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
 // CHECK-NEXT:    ret void
 //
@@ -383,9 +383,9 @@ void test_svst3_vnum_s64(svbool_t pg, int64_t *base, int64_t vnum, svint64x3_t d
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i64 32)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -399,9 +399,9 @@ void test_svst3_vnum_u8(svbool_t pg, uint8_t *base, int64_t vnum, svuint8x3_t da
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i64 16)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -410,9 +410,9 @@ void test_svst3_vnum_u8(svbool_t pg, uint8_t *base, int64_t vnum, svuint8x3_t da
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i64 16)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -426,9 +426,9 @@ void test_svst3_vnum_u16(svbool_t pg, uint16_t *base, int64_t vnum, svuint16x3_t
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i64 8)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i32> [[TMP5]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -437,9 +437,9 @@ void test_svst3_vnum_u16(svbool_t pg, uint16_t *base, int64_t vnum, svuint16x3_t
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i32> [[TMP5]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -453,9 +453,9 @@ void test_svst3_vnum_u32(svbool_t pg, uint32_t *base, int64_t vnum, svuint32x3_t
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i64 2)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i64 4)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -464,9 +464,9 @@ void test_svst3_vnum_u32(svbool_t pg, uint32_t *base, int64_t vnum, svuint32x3_t
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i64 4)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -480,9 +480,9 @@ void test_svst3_vnum_u64(svbool_t pg, uint64_t *base, int64_t vnum, svuint64x3_t
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i64 16)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv8f16(<vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x half> [[TMP5]], <vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -491,9 +491,9 @@ void test_svst3_vnum_u64(svbool_t pg, uint64_t *base, int64_t vnum, svuint64x3_t
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i64 16)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv8f16(<vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x half> [[TMP5]], <vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -507,9 +507,9 @@ void test_svst3_vnum_f16(svbool_t pg, float16_t *base, int64_t vnum, svfloat16x3
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i64 8)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv4f32(<vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], <vscale x 4 x float> [[TMP5]], <vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -518,9 +518,9 @@ void test_svst3_vnum_f16(svbool_t pg, float16_t *base, int64_t vnum, svfloat16x3
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv4f32(<vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], <vscale x 4 x float> [[TMP5]], <vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -534,9 +534,9 @@ void test_svst3_vnum_f32(svbool_t pg, float32_t *base, int64_t vnum, svfloat32x3
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i64 2)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i64 4)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv2f64(<vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], <vscale x 2 x double> [[TMP5]], <vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -545,9 +545,9 @@ void test_svst3_vnum_f32(svbool_t pg, float32_t *base, int64_t vnum, svfloat32x3
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i64 4)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st3.nxv2f64(<vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], <vscale x 2 x double> [[TMP5]], <vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4-bfloat.c
index b0bf73b1451c..d4d7c682eba0 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4-bfloat.c
@@ -18,20 +18,20 @@
 // CHECK-LABEL: @test_svst4_bf16(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i64 16)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i64 24)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv8bf16(<vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z15test_svst4_bf16u10__SVBool_tPu6__bf1614svbfloat16x4_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i64 24)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv8bf16(<vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -45,10 +45,10 @@ void test_svst4_bf16(svbool_t pg, bfloat16_t *base, svbfloat16x4_t data)
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i64 16)
+// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i64 24)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv8bf16(<vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x bfloat> [[TMP5]], <vscale x 8 x bfloat> [[TMP6]], <vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -57,10 +57,10 @@ void test_svst4_bf16(svbool_t pg, bfloat16_t *base, svbfloat16x4_t data)
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i64 24)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv8bf16(<vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x bfloat> [[TMP5]], <vscale x 8 x bfloat> [[TMP6]], <vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4.c
index cecb91f0adbf..0027341f57b2 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4.c
@@ -16,19 +16,19 @@
 
 // CHECK-LABEL: @test_svst4_s8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 16)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 32)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 48)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z13test_svst4_s8u10__SVBool_tPa10svint8x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 32)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 48)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -40,20 +40,20 @@ void test_svst4_s8(svbool_t pg, int8_t *base, svint8x4_t data)
 // CHECK-LABEL: @test_svst4_s16(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 16)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 24)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst4_s16u10__SVBool_tPs11svint16x4_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 24)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -65,20 +65,20 @@ void test_svst4_s16(svbool_t pg, int16_t *base, svint16x4_t data)
 // CHECK-LABEL: @test_svst4_s32(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 12)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst4_s32u10__SVBool_tPi11svint32x4_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 12)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -90,20 +90,20 @@ void test_svst4_s32(svbool_t pg, int32_t *base, svint32x4_t data)
 // CHECK-LABEL: @test_svst4_s64(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 6)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst4_s64u10__SVBool_tPl11svint64x4_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 6)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -114,19 +114,19 @@ void test_svst4_s64(svbool_t pg, int64_t *base, svint64x4_t data)
 
 // CHECK-LABEL: @test_svst4_u8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 16)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 32)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 48)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z13test_svst4_u8u10__SVBool_tPh11svuint8x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 32)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 48)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -138,20 +138,20 @@ void test_svst4_u8(svbool_t pg, uint8_t *base, svuint8x4_t data)
 // CHECK-LABEL: @test_svst4_u16(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 16)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 24)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst4_u16u10__SVBool_tPt12svuint16x4_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 24)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -163,20 +163,20 @@ void test_svst4_u16(svbool_t pg, uint16_t *base, svuint16x4_t data)
 // CHECK-LABEL: @test_svst4_u32(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 12)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst4_u32u10__SVBool_tPj12svuint32x4_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 12)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -188,20 +188,20 @@ void test_svst4_u32(svbool_t pg, uint32_t *base, svuint32x4_t data)
 // CHECK-LABEL: @test_svst4_u64(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 6)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst4_u64u10__SVBool_tPm12svuint64x4_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 6)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -213,20 +213,20 @@ void test_svst4_u64(svbool_t pg, uint64_t *base, svuint64x4_t data)
 // CHECK-LABEL: @test_svst4_f16(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i64 16)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i64 24)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv8f16(<vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst4_f16u10__SVBool_tPDh13svfloat16x4_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i64 24)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv8f16(<vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -238,20 +238,20 @@ void test_svst4_f16(svbool_t pg, float16_t *base, svfloat16x4_t data)
 // CHECK-LABEL: @test_svst4_f32(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i64 12)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv4f32(<vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst4_f32u10__SVBool_tPf13svfloat32x4_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i64 12)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv4f32(<vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -263,20 +263,20 @@ void test_svst4_f32(svbool_t pg, float32_t *base, svfloat32x4_t data)
 // CHECK-LABEL: @test_svst4_f64(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i64 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i64 6)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv2f64(<vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
 // CHECK-NEXT:    ret void
 //
 // CPP-CHECK-LABEL: @_Z14test_svst4_f64u10__SVBool_tPd13svfloat64x4_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i64 6)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv2f64(<vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -289,10 +289,10 @@ void test_svst4_f64(svbool_t pg, float64_t *base, svfloat64x4_t data)
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
 // CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 16)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 32)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 48)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], <vscale x 16 x i8> [[TMP5]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
 // CHECK-NEXT:    ret void
 //
@@ -300,10 +300,10 @@ void test_svst4_f64(svbool_t pg, float64_t *base, svfloat64x4_t data)
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 32)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 48)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], <vscale x 16 x i8> [[TMP5]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -317,10 +317,10 @@ void test_svst4_vnum_s8(svbool_t pg, int8_t *base, int64_t vnum, svint8x4_t data
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 16)
+// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 24)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i16> [[TMP6]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -329,10 +329,10 @@ void test_svst4_vnum_s8(svbool_t pg, int8_t *base, int64_t vnum, svint8x4_t data
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 24)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i16> [[TMP6]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -346,10 +346,10 @@ void test_svst4_vnum_s16(svbool_t pg, int16_t *base, int64_t vnum, svint16x4_t d
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 12)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i32> [[TMP5]], <vscale x 4 x i32> [[TMP6]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -358,10 +358,10 @@ void test_svst4_vnum_s16(svbool_t pg, int16_t *base, int64_t vnum, svint16x4_t d
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 12)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i32> [[TMP5]], <vscale x 4 x i32> [[TMP6]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -375,10 +375,10 @@ void test_svst4_vnum_s32(svbool_t pg, int32_t *base, int64_t vnum, svint32x4_t d
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 2)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 6)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i64> [[TMP6]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -387,10 +387,10 @@ void test_svst4_vnum_s32(svbool_t pg, int32_t *base, int64_t vnum, svint32x4_t d
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 6)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i64> [[TMP6]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -403,10 +403,10 @@ void test_svst4_vnum_s64(svbool_t pg, int64_t *base, int64_t vnum, svint64x4_t d
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
 // CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 16)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 32)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 48)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], <vscale x 16 x i8> [[TMP5]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
 // CHECK-NEXT:    ret void
 //
@@ -414,10 +414,10 @@ void test_svst4_vnum_s64(svbool_t pg, int64_t *base, int64_t vnum, svint64x4_t d
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 32)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i64 48)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], <vscale x 16 x i8> [[TMP5]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -431,10 +431,10 @@ void test_svst4_vnum_u8(svbool_t pg, uint8_t *base, int64_t vnum, svuint8x4_t da
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 16)
+// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 24)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i16> [[TMP6]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -443,10 +443,10 @@ void test_svst4_vnum_u8(svbool_t pg, uint8_t *base, int64_t vnum, svuint8x4_t da
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i64 24)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i16> [[TMP6]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -460,10 +460,10 @@ void test_svst4_vnum_u16(svbool_t pg, uint16_t *base, int64_t vnum, svuint16x4_t
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 12)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i32> [[TMP5]], <vscale x 4 x i32> [[TMP6]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -472,10 +472,10 @@ void test_svst4_vnum_u16(svbool_t pg, uint16_t *base, int64_t vnum, svuint16x4_t
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i64 12)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i32> [[TMP5]], <vscale x 4 x i32> [[TMP6]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -489,10 +489,10 @@ void test_svst4_vnum_u32(svbool_t pg, uint32_t *base, int64_t vnum, svuint32x4_t
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 2)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 6)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i64> [[TMP6]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -501,10 +501,10 @@ void test_svst4_vnum_u32(svbool_t pg, uint32_t *base, int64_t vnum, svuint32x4_t
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i64 6)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i64> [[TMP6]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -518,10 +518,10 @@ void test_svst4_vnum_u64(svbool_t pg, uint64_t *base, int64_t vnum, svuint64x4_t
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i64 16)
+// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i64 24)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv8f16(<vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x half> [[TMP5]], <vscale x 8 x half> [[TMP6]], <vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -530,10 +530,10 @@ void test_svst4_vnum_u64(svbool_t pg, uint64_t *base, int64_t vnum, svuint64x4_t
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i64 16)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i64 24)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv8f16(<vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x half> [[TMP5]], <vscale x 8 x half> [[TMP6]], <vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -547,10 +547,10 @@ void test_svst4_vnum_f16(svbool_t pg, float16_t *base, int64_t vnum, svfloat16x4
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i64 8)
+// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i64 12)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv4f32(<vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], <vscale x 4 x float> [[TMP5]], <vscale x 4 x float> [[TMP6]], <vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -559,10 +559,10 @@ void test_svst4_vnum_f16(svbool_t pg, float16_t *base, int64_t vnum, svfloat16x4
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i64 8)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i64 12)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv4f32(<vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], <vscale x 4 x float> [[TMP5]], <vscale x 4 x float> [[TMP6]], <vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //
@@ -576,10 +576,10 @@ void test_svst4_vnum_f32(svbool_t pg, float32_t *base, int64_t vnum, svfloat32x4
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
 // CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 1)
-// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 2)
-// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 3)
+// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i64 2)
+// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i64 4)
+// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i64 6)
 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv2f64(<vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], <vscale x 2 x double> [[TMP5]], <vscale x 2 x double> [[TMP6]], <vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
 // CHECK-NEXT:    ret void
 //
@@ -588,10 +588,10 @@ void test_svst4_vnum_f32(svbool_t pg, float32_t *base, int64_t vnum, svfloat32x4
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
-// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 1)
-// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 2)
-// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i64 2)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i64 4)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i64 6)
 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.st4.nxv2f64(<vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], <vscale x 2 x double> [[TMP5]], <vscale x 2 x double> [[TMP6]], <vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
 // CPP-CHECK-NEXT:    ret void
 //

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2-bfloat.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2-bfloat.c
index 23f53823c673..f7bfc603fa1e 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2-bfloat.c
@@ -16,15 +16,15 @@
 
 // CHECK-LABEL: @test_svtbl2_bf16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i64 8)
 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tbl2.nxv8bf16(<vscale x 8 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x i16> [[INDICES:%.*]])
 // CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z16test_svtbl2_bf1614svbfloat16x2_tu12__SVUint16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.tbl2.nxv8bf16(<vscale x 8 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x i16> [[INDICES:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP2]]
 //

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2.c
index 8f815891a313..4ed342cca221 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2.c
@@ -17,15 +17,15 @@
 
 // CHECK-LABEL: @test_svtbl2_s8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i64 16)
 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tbl2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[INDICES:%.*]])
 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z14test_svtbl2_s810svint8x2_tu11__SVUint8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i64 16)
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tbl2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[INDICES:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
 //
@@ -36,15 +36,15 @@ svint8_t test_svtbl2_s8(svint8x2_t data, svuint8_t indices)
 
 // CHECK-LABEL: @test_svtbl2_s16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i64 8)
 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tbl2.nxv8i16(<vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[INDICES:%.*]])
 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svtbl2_s1611svint16x2_tu12__SVUint16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tbl2.nxv8i16(<vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[INDICES:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
 //
@@ -55,15 +55,15 @@ svint16_t test_svtbl2_s16(svint16x2_t data, svuint16_t indices)
 
 // CHECK-LABEL: @test_svtbl2_s32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i64 4)
 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tbl2.nxv4i32(<vscale x 4 x i32> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[INDICES:%.*]])
 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svtbl2_s3211svint32x2_tu12__SVUint32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i64 4)
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tbl2.nxv4i32(<vscale x 4 x i32> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[INDICES:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
 //
@@ -74,15 +74,15 @@ svint32_t test_svtbl2_s32(svint32x2_t data, svuint32_t indices)
 
 // CHECK-LABEL: @test_svtbl2_s64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i64 2)
 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tbl2.nxv2i64(<vscale x 2 x i64> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[INDICES:%.*]])
 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svtbl2_s6411svint64x2_tu12__SVUint64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i64 2)
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tbl2.nxv2i64(<vscale x 2 x i64> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[INDICES:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
 //
@@ -93,15 +93,15 @@ svint64_t test_svtbl2_s64(svint64x2_t data, svuint64_t indices)
 
 // CHECK-LABEL: @test_svtbl2_u8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i64 16)
 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tbl2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[INDICES:%.*]])
 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z14test_svtbl2_u811svuint8x2_tu11__SVUint8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i64 16)
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.tbl2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[INDICES:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
 //
@@ -112,15 +112,15 @@ svuint8_t test_svtbl2_u8(svuint8x2_t data, svuint8_t indices)
 
 // CHECK-LABEL: @test_svtbl2_u16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i64 8)
 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tbl2.nxv8i16(<vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[INDICES:%.*]])
 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svtbl2_u1612svuint16x2_tu12__SVUint16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.tbl2.nxv8i16(<vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[INDICES:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
 //
@@ -131,15 +131,15 @@ svuint16_t test_svtbl2_u16(svuint16x2_t data, svuint16_t indices)
 
 // CHECK-LABEL: @test_svtbl2_u32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i64 4)
 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tbl2.nxv4i32(<vscale x 4 x i32> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[INDICES:%.*]])
 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svtbl2_u3212svuint32x2_tu12__SVUint32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i64 4)
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.tbl2.nxv4i32(<vscale x 4 x i32> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[INDICES:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
 //
@@ -150,15 +150,15 @@ svuint32_t test_svtbl2_u32(svuint32x2_t data, svuint32_t indices)
 
 // CHECK-LABEL: @test_svtbl2_u64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i64 2)
 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tbl2.nxv2i64(<vscale x 2 x i64> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[INDICES:%.*]])
 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svtbl2_u6412svuint64x2_tu12__SVUint64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i64 2)
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.tbl2.nxv2i64(<vscale x 2 x i64> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[INDICES:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
 //
@@ -169,15 +169,15 @@ svuint64_t test_svtbl2_u64(svuint64x2_t data, svuint64_t indices)
 
 // CHECK-LABEL: @test_svtbl2_f16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i64 8)
 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tbl2.nxv8f16(<vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x i16> [[INDICES:%.*]])
 // CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svtbl2_f1613svfloat16x2_tu12__SVUint16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i64 8)
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.tbl2.nxv8f16(<vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x i16> [[INDICES:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
 //
@@ -188,15 +188,15 @@ svfloat16_t test_svtbl2_f16(svfloat16x2_t data, svuint16_t indices)
 
 // CHECK-LABEL: @test_svtbl2_f32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i64 4)
 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tbl2.nxv4f32(<vscale x 4 x float> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x i32> [[INDICES:%.*]])
 // CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svtbl2_f3213svfloat32x2_tu12__SVUint32_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i64 4)
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.tbl2.nxv4f32(<vscale x 4 x float> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x i32> [[INDICES:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
 //
@@ -207,15 +207,15 @@ svfloat32_t test_svtbl2_f32(svfloat32x2_t data, svuint32_t indices)
 
 // CHECK-LABEL: @test_svtbl2_f64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i32 0)
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i64 2)
 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tbl2.nxv2f64(<vscale x 2 x double> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x i64> [[INDICES:%.*]])
 // CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
 //
 // CPP-CHECK-LABEL: @_Z15test_svtbl2_f6413svfloat64x2_tu12__SVUint64_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i32 0)
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i64 2)
 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.tbl2.nxv2f64(<vscale x 2 x double> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x i64> [[INDICES:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
 //