[ARM,MVE] Add intrinsics for vector comparisons.

[simon_tatham]

Summary

This adds the vcmp family of ACLE MVE intrinsics: vector/vector,
vector/scalar, and the predicated forms of both. All are represented
using standard existing IR: vector/scalar comparisons are represented
by making a vector out of the scalar first, and predicated forms are
represented by taking the bitwise AND of the input predicate and the
output of the comparison. Existing LLVM-side tests demonstrate that
ISel will pattern-match all of that back down to single MVE VCMPs.

The idiom of handling a vector/scalar operation by generating IR to
expand the scalar into a second vector is going to be needed for a lot
of MVE intrinsics, so to make that easy, I've provided a helper
function that automatically works out the element count.

The comparison intrinsics are the first ones that have to return a
predicate, in the user-facing mve_pred16_t format. This means we
have to use the arm_mve_pred_v2i low-level intrinsic to convert it
back from the logical <n x i1> form used in IR. I've done that
explicitly in the code gen specification for the builtins, because it
happens much more rarely in the ACLE API than passing a Predicate as
input, so it didn't seem worth automating in MveEmitter.

[dmgreen]

clang/lib/CodeGen/CGBuiltin.cpp

6798

Is this the same as or similar enough to CreateVectorSplat to use that instead?

llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp

3311 ↗

(On Diff #229477)

This can be a separate change.

[simon_tatham]

clang/lib/CodeGen/CGBuiltin.cpp

6798

Ah, thank you – I can see why I didn't find it under that name! I looked for all sorts of things like "dup" or "rep", and then gave up and painstakingly copied the long-winded IR I'd seen in an existing example. "splat" is one of those words you have no trouble finding after you know it's the word you're looking for...

Now I know there's an existing function that will do the job, I agree that I should use it (if only because it'll produce the most idiomatic one of the various synonymous IR representations). But I think I'll keep the wrapping helper function, because it's still useful to have it auto-compute the unique correct element count for an MVE vector – that will keep the call sites terse in arm_mve.td, and I expect there will be lots of those.

llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp

3311 ↗

(On Diff #229477)

OK, will do.

[simon_tatham]

Replaced my laborious manual vector splat with a call to the standard one; moved the InstCombine work out into a separate patch.

[simon_tatham]

D'oh, moved mve-vpt-from-intrinsics.ll into D70313.

[dmgreen]

LGTM

clang/include/clang/Basic/arm_mve.td

 def vcvt#half#q_m_f16: Intrinsic<
     VecOf<f16>, (args VecOf<f16>:$inactive, Vector:$a, PredOf<f32>:$pred),
     (IRInt<"vcvt_narrow_predicated"> $inactive, $a, halfconst, $pred)>;
 
 } // params = [f32], pnt = PNT_None
 
 } // loop over half = "b", "t"
 
+multiclass compare_with_pred<string condname, dag arguments,
+                             dag cmp, string suffix> {
+  // Make the predicated and unpredicated versions of a single comparison.
+  def: Intrinsic<Predicate, arguments,
+                 (u16 (IRInt<"pred_v2i", [Predicate]> cmp))>,
+       NameOverride<"vcmp" # condname # "q" # suffix>;
+  def: Intrinsic<Predicate, !con(arguments, (args Predicate:$inpred)),
+                 (u16 (IRInt<"pred_v2i", [Predicate]> (and $inpred, cmp)))>,
+       NameOverride<"vcmp" # condname # "q_m" # suffix>;
+}
+
+multiclass compare<string condname, IRBuilder cmpop> {
+  // Make all four variants of a comparison: the vector/vector and
+  // vector/scalar forms, each using compare_with_pred to make a
+  // predicated and unpredicated version.
+  defm: compare_with_pred<condname, (args Vector:$va, Vector:$vb),
+                          (cmpop $va, $vb), "">;
+  let pnt = PNT_NType in {
+    defm: compare_with_pred<condname, (args Vector:$va, unpromoted<Scalar>:$sb),
+                            (cmpop $va, (splat $sb)), "_n">;
+  }
+}
+let params = T.Int in {
+  defm: compare<"eq", icmp_eq>;
+  defm: compare<"ne", icmp_ne>;
+}
+let params = T.Signed in {
+  defm: compare<"gt", icmp_sgt>;
+  defm: compare<"ge", icmp_sge>;
+  defm: compare<"lt", icmp_slt>;
+  defm: compare<"le", icmp_sle>;
+}
+let params = T.Unsigned in {
+  defm: compare<"hi", icmp_ugt>;
+  defm: compare<"cs", icmp_uge>;
+}
+let params = T.Float in {
+  defm: compare<"eq", fcmp_eq>;
+  defm: compare<"ne", fcmp_ne>;
+  defm: compare<"gt", fcmp_gt>;
+  defm: compare<"ge", fcmp_ge>;
+  defm: compare<"lt", fcmp_lt>;
+  defm: compare<"le", fcmp_le>;
+}
+
 multiclass contiguous_load<string mnemonic, PrimitiveType memtype,
                            list<Type> same_size, list<Type> wider> {
   // Intrinsics named with explicit memory and element sizes that match:
   // vldrbq_?8, vldrhq_?16, vldrwq_?32.
   let params = same_size, pnt = PNT_None in {
     def: Intrinsic<Vector, (args CPtr<CopyKind<same_size[0], Scalar>>:$addr),
                    (load (address (CPtr<Vector> $addr), !srl(memtype.size,3)))>,
          NameOverride<mnemonic>;

clang/include/clang/Basic/arm_mve_defs.td

 def xelt_var: IRBuilder<"CreateExtractElement">;
 def trunc: IRBuilder<"CreateTrunc">;
 def bitcast: IRBuilder<"CreateBitCast">;
 def extend: CGHelperFn<"SignOrZeroExtend"> {
   let special_params = [IRBuilderIntParam<2, "bool">];
 }
 def zeroinit: IRFunction<"llvm::Constant::getNullValue">;
 def undef: IRFunction<"UndefValue::get">;
+def icmp_eq: IRBuilder<"CreateICmpEQ">;
+def icmp_ne: IRBuilder<"CreateICmpNE">;
+def icmp_ugt: IRBuilder<"CreateICmpUGT">;
+def icmp_uge: IRBuilder<"CreateICmpUGE">;
+def icmp_ult: IRBuilder<"CreateICmpULT">;
+def icmp_ule: IRBuilder<"CreateICmpULE">;
+def icmp_sgt: IRBuilder<"CreateICmpSGT">;
+def icmp_sge: IRBuilder<"CreateICmpSGE">;
+def icmp_slt: IRBuilder<"CreateICmpSLT">;
+def icmp_sle: IRBuilder<"CreateICmpSLE">;
+def fcmp_eq: IRBuilder<"CreateFCmpOEQ">;
+def fcmp_ne: IRBuilder<"CreateFCmpUNE">; // not O: it must return true on NaNs
+def fcmp_gt: IRBuilder<"CreateFCmpOGT">;
+def fcmp_ge: IRBuilder<"CreateFCmpOGE">;
+def fcmp_lt: IRBuilder<"CreateFCmpOLT">;
+def fcmp_le: IRBuilder<"CreateFCmpOLE">;
+def splat: CGHelperFn<"ARMMVEVectorSplat">;
 
 // A node that makes an Address out of a pointer-typed Value, by
 // providing an alignment as the second argument.
 def address;
 
 // Another node class you can use in the codegen dag. This one corresponds to
 // an IR intrinsic function, which has to be specialized to a particular list
 // of types.

clang/lib/CodeGen/CGBuiltin.cpp

 
 static llvm::Value *SignOrZeroExtend(CGBuilderTy &Builder, llvm::Value *V,
                                      llvm::Type *T, bool Unsigned) {
   // Helper function called by Tablegen-constructed ARM MVE builtin codegen,
   // which finds it convenient to specify signed/unsigned as a boolean flag.
   return Unsigned ? Builder.CreateZExt(V, T) : Builder.CreateSExt(V, T);
 }
 
+static llvm::Value *ARMMVEVectorSplat(CGBuilderTy &Builder, llvm::Value *V) {
+  // MVE-specific helper function for a vector splat, which infers the element
+  // count of the output vector by knowing that MVE vectors are all 128 bits
+  // wide.
+  unsigned Elements = 128 / V->getType()->getPrimitiveSizeInBits();
+  return Builder.CreateVectorSplat(Elements, V);
+}
+
 Value *CodeGenFunction::EmitARMMVEBuiltinExpr(unsigned BuiltinID,
                                               const CallExpr *E,
                                               ReturnValueSlot ReturnValue,
                                               llvm::Triple::ArchType Arch) {
   enum class CustomCodeGen { VLD24, VST24 } CustomCodeGenType;
   Intrinsic::ID IRIntr;
   unsigned NumVectors;
 

clang/test/CodeGen/arm-mve-intrinsics/compare.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple thumbv8.1m.main-arm-none-eabi -target-feature +mve.fp -mfloat-abi hard -fallow-half-arguments-and-returns -O0 -disable-O0-optnone -S -emit-llvm -o - %s | opt -S -mem2reg -sroa | FileCheck %s
+// RUN: %clang_cc1 -triple thumbv8.1m.main-arm-none-eabi -target-feature +mve.fp -mfloat-abi hard -fallow-half-arguments-and-returns -O0 -disable-O0-optnone -DPOLYMORPHIC -S -emit-llvm -o - %s | opt -S -mem2reg -sroa | FileCheck %s
+
+#include <arm_mve.h>
+
+// CHECK-LABEL: @test_vcmpeqq_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp oeq <8 x half> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpeqq_f16(float16x8_t a, float16x8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpeqq_f16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp oeq <4 x float> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpeqq_f32(float32x4_t a, float32x4_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpeqq_f32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp eq <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpeqq_s8(int8x16_t a, int8x16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpeqq_s8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp eq <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpeqq_s16(int16x8_t a, int16x8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpeqq_s16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp eq <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpeqq_s32(int32x4_t a, int32x4_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpeqq_s32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp eq <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpeqq_u8(uint8x16_t a, uint8x16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpeqq_u8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp eq <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpeqq_u16(uint16x8_t a, uint16x8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpeqq_u16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp eq <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpeqq_u32(uint32x4_t a, uint32x4_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpeqq_u32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast float [[B_COERCE:%.*]] to i32
+// CHECK-NEXT:    [[TMP_0_EXTRACT_TRUNC:%.*]] = trunc i32 [[TMP0]] to i16
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16 [[TMP_0_EXTRACT_TRUNC]] to half
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x half> undef, half [[TMP1]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x half> [[DOTSPLATINSERT]], <8 x half> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp oeq <8 x half> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = trunc i32 [[TMP3]] to i16
+// CHECK-NEXT:    ret i16 [[TMP4]]
+//
+mve_pred16_t test_vcmpeqq_n_f16(float16x8_t a, float16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpeqq_n_f16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x float> undef, float [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x float> [[DOTSPLATINSERT]], <4 x float> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp oeq <4 x float> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpeqq_n_f32(float32x4_t a, float32_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpeqq_n_f32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp eq <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpeqq_n_s8(int8x16_t a, int8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpeqq_n_s8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp eq <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpeqq_n_s16(int16x8_t a, int16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpeqq_n_s16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp eq <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpeqq_n_s32(int32x4_t a, int32_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpeqq_n_s32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp eq <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpeqq_n_u8(uint8x16_t a, uint8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpeqq_n_u8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp eq <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpeqq_n_u16(uint16x8_t a, uint16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpeqq_n_u16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp eq <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpeqq_n_u32(uint32x4_t a, uint32_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpeqq_n_u32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_m_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp oeq <8 x half> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpeqq_m_f16(float16x8_t a, float16x8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpeqq_m_f16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_m_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp oeq <4 x float> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpeqq_m_f32(float32x4_t a, float32x4_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpeqq_m_f32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_m_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp eq <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpeqq_m_s8(int8x16_t a, int8x16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpeqq_m_s8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_m_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp eq <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpeqq_m_s16(int16x8_t a, int16x8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpeqq_m_s16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_m_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp eq <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpeqq_m_s32(int32x4_t a, int32x4_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpeqq_m_s32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_m_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp eq <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpeqq_m_u8(uint8x16_t a, uint8x16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpeqq_m_u8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_m_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp eq <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpeqq_m_u16(uint16x8_t a, uint16x8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpeqq_m_u16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_m_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp eq <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpeqq_m_u32(uint32x4_t a, uint32x4_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpeqq_m_u32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_m_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast float [[B_COERCE:%.*]] to i32
+// CHECK-NEXT:    [[TMP_0_EXTRACT_TRUNC:%.*]] = trunc i32 [[TMP0]] to i16
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16 [[TMP_0_EXTRACT_TRUNC]] to half
+// CHECK-NEXT:    [[TMP2:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP3:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP2]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x half> undef, half [[TMP1]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x half> [[DOTSPLATINSERT]], <8 x half> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP4:%.*]] = fcmp oeq <8 x half> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP5:%.*]] = and <8 x i1> [[TMP3]], [[TMP4]]
+// CHECK-NEXT:    [[TMP6:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP5]])
+// CHECK-NEXT:    [[TMP7:%.*]] = trunc i32 [[TMP6]] to i16
+// CHECK-NEXT:    ret i16 [[TMP7]]
+//
+mve_pred16_t test_vcmpeqq_m_n_f16(float16x8_t a, float16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpeqq_m_n_f16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_m_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x float> undef, float [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x float> [[DOTSPLATINSERT]], <4 x float> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp oeq <4 x float> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpeqq_m_n_f32(float32x4_t a, float32_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpeqq_m_n_f32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_m_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp eq <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpeqq_m_n_s8(int8x16_t a, int8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpeqq_m_n_s8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_m_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp eq <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpeqq_m_n_s16(int16x8_t a, int16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpeqq_m_n_s16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_m_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp eq <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpeqq_m_n_s32(int32x4_t a, int32_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpeqq_m_n_s32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_m_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp eq <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpeqq_m_n_u8(uint8x16_t a, uint8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpeqq_m_n_u8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_m_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp eq <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpeqq_m_n_u16(uint16x8_t a, uint16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpeqq_m_n_u16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpeqq_m_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp eq <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpeqq_m_n_u32(uint32x4_t a, uint32_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpeqq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpeqq_m_n_u32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp une <8 x half> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpneq_f16(float16x8_t a, float16x8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpneq_f16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp une <4 x float> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpneq_f32(float32x4_t a, float32x4_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpneq_f32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ne <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpneq_s8(int8x16_t a, int8x16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpneq_s8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ne <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpneq_s16(int16x8_t a, int16x8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpneq_s16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ne <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpneq_s32(int32x4_t a, int32x4_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpneq_s32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ne <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpneq_u8(uint8x16_t a, uint8x16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpneq_u8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ne <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpneq_u16(uint16x8_t a, uint16x8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpneq_u16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ne <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpneq_u32(uint32x4_t a, uint32x4_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpneq_u32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast float [[B_COERCE:%.*]] to i32
+// CHECK-NEXT:    [[TMP_0_EXTRACT_TRUNC:%.*]] = trunc i32 [[TMP0]] to i16
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16 [[TMP_0_EXTRACT_TRUNC]] to half
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x half> undef, half [[TMP1]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x half> [[DOTSPLATINSERT]], <8 x half> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp une <8 x half> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = trunc i32 [[TMP3]] to i16
+// CHECK-NEXT:    ret i16 [[TMP4]]
+//
+mve_pred16_t test_vcmpneq_n_f16(float16x8_t a, float16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpneq_n_f16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x float> undef, float [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x float> [[DOTSPLATINSERT]], <4 x float> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp une <4 x float> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpneq_n_f32(float32x4_t a, float32_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpneq_n_f32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ne <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpneq_n_s8(int8x16_t a, int8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpneq_n_s8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ne <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpneq_n_s16(int16x8_t a, int16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpneq_n_s16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ne <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpneq_n_s32(int32x4_t a, int32_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpneq_n_s32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ne <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpneq_n_u8(uint8x16_t a, uint8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpneq_n_u8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ne <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpneq_n_u16(uint16x8_t a, uint16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpneq_n_u16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ne <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpneq_n_u32(uint32x4_t a, uint32_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpneq_n_u32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_m_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp une <8 x half> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpneq_m_f16(float16x8_t a, float16x8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpneq_m_f16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_m_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp une <4 x float> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpneq_m_f32(float32x4_t a, float32x4_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpneq_m_f32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_m_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ne <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpneq_m_s8(int8x16_t a, int8x16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpneq_m_s8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_m_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ne <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpneq_m_s16(int16x8_t a, int16x8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpneq_m_s16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_m_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ne <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpneq_m_s32(int32x4_t a, int32x4_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpneq_m_s32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_m_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ne <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpneq_m_u8(uint8x16_t a, uint8x16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpneq_m_u8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_m_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ne <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpneq_m_u16(uint16x8_t a, uint16x8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpneq_m_u16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_m_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ne <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpneq_m_u32(uint32x4_t a, uint32x4_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpneq_m_u32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_m_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast float [[B_COERCE:%.*]] to i32
+// CHECK-NEXT:    [[TMP_0_EXTRACT_TRUNC:%.*]] = trunc i32 [[TMP0]] to i16
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16 [[TMP_0_EXTRACT_TRUNC]] to half
+// CHECK-NEXT:    [[TMP2:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP3:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP2]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x half> undef, half [[TMP1]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x half> [[DOTSPLATINSERT]], <8 x half> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP4:%.*]] = fcmp une <8 x half> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP5:%.*]] = and <8 x i1> [[TMP3]], [[TMP4]]
+// CHECK-NEXT:    [[TMP6:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP5]])
+// CHECK-NEXT:    [[TMP7:%.*]] = trunc i32 [[TMP6]] to i16
+// CHECK-NEXT:    ret i16 [[TMP7]]
+//
+mve_pred16_t test_vcmpneq_m_n_f16(float16x8_t a, float16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpneq_m_n_f16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_m_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x float> undef, float [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x float> [[DOTSPLATINSERT]], <4 x float> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp une <4 x float> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpneq_m_n_f32(float32x4_t a, float32_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpneq_m_n_f32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_m_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ne <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpneq_m_n_s8(int8x16_t a, int8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpneq_m_n_s8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_m_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ne <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpneq_m_n_s16(int16x8_t a, int16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpneq_m_n_s16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_m_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ne <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpneq_m_n_s32(int32x4_t a, int32_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpneq_m_n_s32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_m_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ne <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpneq_m_n_u8(uint8x16_t a, uint8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpneq_m_n_u8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_m_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ne <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpneq_m_n_u16(uint16x8_t a, uint16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpneq_m_n_u16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpneq_m_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ne <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpneq_m_n_u32(uint32x4_t a, uint32_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpneq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpneq_m_n_u32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp oge <8 x half> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgeq_f16(float16x8_t a, float16x8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgeq_f16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp oge <4 x float> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgeq_f32(float32x4_t a, float32x4_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgeq_f32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sge <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgeq_s8(int8x16_t a, int8x16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgeq_s8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sge <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgeq_s16(int16x8_t a, int16x8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgeq_s16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sge <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgeq_s32(int32x4_t a, int32x4_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgeq_s32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpcsq_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp uge <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpcsq_u8(uint8x16_t a, uint8x16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpcsq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpcsq_u8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpcsq_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp uge <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpcsq_u16(uint16x8_t a, uint16x8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpcsq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpcsq_u16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpcsq_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp uge <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpcsq_u32(uint32x4_t a, uint32x4_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpcsq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpcsq_u32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast float [[B_COERCE:%.*]] to i32
+// CHECK-NEXT:    [[TMP_0_EXTRACT_TRUNC:%.*]] = trunc i32 [[TMP0]] to i16
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16 [[TMP_0_EXTRACT_TRUNC]] to half
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x half> undef, half [[TMP1]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x half> [[DOTSPLATINSERT]], <8 x half> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp oge <8 x half> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = trunc i32 [[TMP3]] to i16
+// CHECK-NEXT:    ret i16 [[TMP4]]
+//
+mve_pred16_t test_vcmpgeq_n_f16(float16x8_t a, float16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgeq_n_f16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x float> undef, float [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x float> [[DOTSPLATINSERT]], <4 x float> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp oge <4 x float> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgeq_n_f32(float32x4_t a, float32_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgeq_n_f32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sge <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgeq_n_s8(int8x16_t a, int8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgeq_n_s8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sge <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgeq_n_s16(int16x8_t a, int16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgeq_n_s16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sge <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgeq_n_s32(int32x4_t a, int32_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgeq_n_s32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpcsq_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp uge <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpcsq_n_u8(uint8x16_t a, uint8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpcsq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpcsq_n_u8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpcsq_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp uge <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpcsq_n_u16(uint16x8_t a, uint16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpcsq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpcsq_n_u16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpcsq_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp uge <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpcsq_n_u32(uint32x4_t a, uint32_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpcsq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpcsq_n_u32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_m_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp oge <8 x half> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgeq_m_f16(float16x8_t a, float16x8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgeq_m_f16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_m_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp oge <4 x float> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgeq_m_f32(float32x4_t a, float32x4_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgeq_m_f32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_m_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sge <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgeq_m_s8(int8x16_t a, int8x16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgeq_m_s8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_m_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sge <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgeq_m_s16(int16x8_t a, int16x8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgeq_m_s16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_m_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sge <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgeq_m_s32(int32x4_t a, int32x4_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgeq_m_s32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpcsq_m_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp uge <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpcsq_m_u8(uint8x16_t a, uint8x16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpcsq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpcsq_m_u8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpcsq_m_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp uge <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpcsq_m_u16(uint16x8_t a, uint16x8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpcsq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpcsq_m_u16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpcsq_m_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp uge <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpcsq_m_u32(uint32x4_t a, uint32x4_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpcsq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpcsq_m_u32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_m_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast float [[B_COERCE:%.*]] to i32
+// CHECK-NEXT:    [[TMP_0_EXTRACT_TRUNC:%.*]] = trunc i32 [[TMP0]] to i16
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16 [[TMP_0_EXTRACT_TRUNC]] to half
+// CHECK-NEXT:    [[TMP2:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP3:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP2]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x half> undef, half [[TMP1]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x half> [[DOTSPLATINSERT]], <8 x half> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP4:%.*]] = fcmp oge <8 x half> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP5:%.*]] = and <8 x i1> [[TMP3]], [[TMP4]]
+// CHECK-NEXT:    [[TMP6:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP5]])
+// CHECK-NEXT:    [[TMP7:%.*]] = trunc i32 [[TMP6]] to i16
+// CHECK-NEXT:    ret i16 [[TMP7]]
+//
+mve_pred16_t test_vcmpgeq_m_n_f16(float16x8_t a, float16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgeq_m_n_f16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_m_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x float> undef, float [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x float> [[DOTSPLATINSERT]], <4 x float> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp oge <4 x float> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgeq_m_n_f32(float32x4_t a, float32_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgeq_m_n_f32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_m_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sge <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgeq_m_n_s8(int8x16_t a, int8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgeq_m_n_s8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_m_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sge <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgeq_m_n_s16(int16x8_t a, int16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgeq_m_n_s16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgeq_m_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sge <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgeq_m_n_s32(int32x4_t a, int32_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgeq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgeq_m_n_s32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpcsq_m_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp uge <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpcsq_m_n_u8(uint8x16_t a, uint8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpcsq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpcsq_m_n_u8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpcsq_m_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp uge <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpcsq_m_n_u16(uint16x8_t a, uint16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpcsq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpcsq_m_n_u16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpcsq_m_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp uge <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpcsq_m_n_u32(uint32x4_t a, uint32_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpcsq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpcsq_m_n_u32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp ogt <8 x half> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgtq_f16(float16x8_t a, float16x8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgtq_f16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp ogt <4 x float> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgtq_f32(float32x4_t a, float32x4_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgtq_f32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sgt <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgtq_s8(int8x16_t a, int8x16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgtq_s8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sgt <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgtq_s16(int16x8_t a, int16x8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgtq_s16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sgt <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgtq_s32(int32x4_t a, int32x4_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgtq_s32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmphiq_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ugt <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmphiq_u8(uint8x16_t a, uint8x16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmphiq(a, b);
+#else /* POLYMORPHIC */
+    return vcmphiq_u8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmphiq_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ugt <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmphiq_u16(uint16x8_t a, uint16x8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmphiq(a, b);
+#else /* POLYMORPHIC */
+    return vcmphiq_u16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmphiq_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ugt <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmphiq_u32(uint32x4_t a, uint32x4_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmphiq(a, b);
+#else /* POLYMORPHIC */
+    return vcmphiq_u32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast float [[B_COERCE:%.*]] to i32
+// CHECK-NEXT:    [[TMP_0_EXTRACT_TRUNC:%.*]] = trunc i32 [[TMP0]] to i16
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16 [[TMP_0_EXTRACT_TRUNC]] to half
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x half> undef, half [[TMP1]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x half> [[DOTSPLATINSERT]], <8 x half> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp ogt <8 x half> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = trunc i32 [[TMP3]] to i16
+// CHECK-NEXT:    ret i16 [[TMP4]]
+//
+mve_pred16_t test_vcmpgtq_n_f16(float16x8_t a, float16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgtq_n_f16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x float> undef, float [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x float> [[DOTSPLATINSERT]], <4 x float> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp ogt <4 x float> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgtq_n_f32(float32x4_t a, float32_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgtq_n_f32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sgt <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgtq_n_s8(int8x16_t a, int8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgtq_n_s8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sgt <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgtq_n_s16(int16x8_t a, int16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgtq_n_s16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sgt <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpgtq_n_s32(int32x4_t a, int32_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpgtq_n_s32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmphiq_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ugt <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmphiq_n_u8(uint8x16_t a, uint8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmphiq(a, b);
+#else /* POLYMORPHIC */
+    return vcmphiq_n_u8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmphiq_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ugt <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmphiq_n_u16(uint16x8_t a, uint16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmphiq(a, b);
+#else /* POLYMORPHIC */
+    return vcmphiq_n_u16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmphiq_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp ugt <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmphiq_n_u32(uint32x4_t a, uint32_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmphiq(a, b);
+#else /* POLYMORPHIC */
+    return vcmphiq_n_u32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_m_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp ogt <8 x half> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgtq_m_f16(float16x8_t a, float16x8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgtq_m_f16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_m_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp ogt <4 x float> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgtq_m_f32(float32x4_t a, float32x4_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgtq_m_f32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_m_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sgt <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgtq_m_s8(int8x16_t a, int8x16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgtq_m_s8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_m_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sgt <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgtq_m_s16(int16x8_t a, int16x8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgtq_m_s16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_m_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sgt <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgtq_m_s32(int32x4_t a, int32x4_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgtq_m_s32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmphiq_m_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ugt <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmphiq_m_u8(uint8x16_t a, uint8x16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmphiq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmphiq_m_u8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmphiq_m_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ugt <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmphiq_m_u16(uint16x8_t a, uint16x8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmphiq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmphiq_m_u16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmphiq_m_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ugt <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmphiq_m_u32(uint32x4_t a, uint32x4_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmphiq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmphiq_m_u32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_m_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast float [[B_COERCE:%.*]] to i32
+// CHECK-NEXT:    [[TMP_0_EXTRACT_TRUNC:%.*]] = trunc i32 [[TMP0]] to i16
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16 [[TMP_0_EXTRACT_TRUNC]] to half
+// CHECK-NEXT:    [[TMP2:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP3:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP2]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x half> undef, half [[TMP1]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x half> [[DOTSPLATINSERT]], <8 x half> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP4:%.*]] = fcmp ogt <8 x half> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP5:%.*]] = and <8 x i1> [[TMP3]], [[TMP4]]
+// CHECK-NEXT:    [[TMP6:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP5]])
+// CHECK-NEXT:    [[TMP7:%.*]] = trunc i32 [[TMP6]] to i16
+// CHECK-NEXT:    ret i16 [[TMP7]]
+//
+mve_pred16_t test_vcmpgtq_m_n_f16(float16x8_t a, float16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgtq_m_n_f16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_m_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x float> undef, float [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x float> [[DOTSPLATINSERT]], <4 x float> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp ogt <4 x float> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgtq_m_n_f32(float32x4_t a, float32_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgtq_m_n_f32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_m_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sgt <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgtq_m_n_s8(int8x16_t a, int8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgtq_m_n_s8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_m_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sgt <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgtq_m_n_s16(int16x8_t a, int16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgtq_m_n_s16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpgtq_m_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sgt <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpgtq_m_n_s32(int32x4_t a, int32_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpgtq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpgtq_m_n_s32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmphiq_m_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ugt <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmphiq_m_n_u8(uint8x16_t a, uint8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmphiq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmphiq_m_n_u8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmphiq_m_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ugt <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmphiq_m_n_u16(uint16x8_t a, uint16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmphiq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmphiq_m_n_u16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmphiq_m_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp ugt <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmphiq_m_n_u32(uint32x4_t a, uint32_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmphiq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmphiq_m_n_u32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp ole <8 x half> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpleq_f16(float16x8_t a, float16x8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpleq_f16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp ole <4 x float> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpleq_f32(float32x4_t a, float32x4_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpleq_f32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sle <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpleq_s8(int8x16_t a, int8x16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpleq_s8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sle <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpleq_s16(int16x8_t a, int16x8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpleq_s16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sle <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpleq_s32(int32x4_t a, int32x4_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpleq_s32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast float [[B_COERCE:%.*]] to i32
+// CHECK-NEXT:    [[TMP_0_EXTRACT_TRUNC:%.*]] = trunc i32 [[TMP0]] to i16
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16 [[TMP_0_EXTRACT_TRUNC]] to half
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x half> undef, half [[TMP1]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x half> [[DOTSPLATINSERT]], <8 x half> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp ole <8 x half> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = trunc i32 [[TMP3]] to i16
+// CHECK-NEXT:    ret i16 [[TMP4]]
+//
+mve_pred16_t test_vcmpleq_n_f16(float16x8_t a, float16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpleq_n_f16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x float> undef, float [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x float> [[DOTSPLATINSERT]], <4 x float> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp ole <4 x float> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpleq_n_f32(float32x4_t a, float32_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpleq_n_f32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sle <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpleq_n_s8(int8x16_t a, int8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpleq_n_s8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sle <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpleq_n_s16(int16x8_t a, int16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpleq_n_s16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp sle <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpleq_n_s32(int32x4_t a, int32_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpleq_n_s32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_m_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp ole <8 x half> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpleq_m_f16(float16x8_t a, float16x8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpleq_m_f16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_m_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp ole <4 x float> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpleq_m_f32(float32x4_t a, float32x4_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpleq_m_f32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_m_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sle <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpleq_m_s8(int8x16_t a, int8x16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpleq_m_s8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_m_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sle <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpleq_m_s16(int16x8_t a, int16x8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpleq_m_s16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_m_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sle <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpleq_m_s32(int32x4_t a, int32x4_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpleq_m_s32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_m_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast float [[B_COERCE:%.*]] to i32
+// CHECK-NEXT:    [[TMP_0_EXTRACT_TRUNC:%.*]] = trunc i32 [[TMP0]] to i16
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16 [[TMP_0_EXTRACT_TRUNC]] to half
+// CHECK-NEXT:    [[TMP2:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP3:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP2]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x half> undef, half [[TMP1]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x half> [[DOTSPLATINSERT]], <8 x half> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP4:%.*]] = fcmp ole <8 x half> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP5:%.*]] = and <8 x i1> [[TMP3]], [[TMP4]]
+// CHECK-NEXT:    [[TMP6:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP5]])
+// CHECK-NEXT:    [[TMP7:%.*]] = trunc i32 [[TMP6]] to i16
+// CHECK-NEXT:    ret i16 [[TMP7]]
+//
+mve_pred16_t test_vcmpleq_m_n_f16(float16x8_t a, float16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpleq_m_n_f16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_m_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x float> undef, float [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x float> [[DOTSPLATINSERT]], <4 x float> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp ole <4 x float> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpleq_m_n_f32(float32x4_t a, float32_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpleq_m_n_f32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_m_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sle <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpleq_m_n_s8(int8x16_t a, int8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpleq_m_n_s8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_m_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sle <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpleq_m_n_s16(int16x8_t a, int16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpleq_m_n_s16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpleq_m_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp sle <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpleq_m_n_s32(int32x4_t a, int32_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpleq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpleq_m_n_s32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp olt <8 x half> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpltq_f16(float16x8_t a, float16x8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpltq_f16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp olt <4 x float> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpltq_f32(float32x4_t a, float32x4_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpltq_f32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp slt <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpltq_s8(int8x16_t a, int8x16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpltq_s8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp slt <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpltq_s16(int16x8_t a, int16x8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpltq_s16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp slt <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpltq_s32(int32x4_t a, int32x4_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpltq_s32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast float [[B_COERCE:%.*]] to i32
+// CHECK-NEXT:    [[TMP_0_EXTRACT_TRUNC:%.*]] = trunc i32 [[TMP0]] to i16
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16 [[TMP_0_EXTRACT_TRUNC]] to half
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x half> undef, half [[TMP1]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x half> [[DOTSPLATINSERT]], <8 x half> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp olt <8 x half> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = trunc i32 [[TMP3]] to i16
+// CHECK-NEXT:    ret i16 [[TMP4]]
+//
+mve_pred16_t test_vcmpltq_n_f16(float16x8_t a, float16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpltq_n_f16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x float> undef, float [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x float> [[DOTSPLATINSERT]], <4 x float> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = fcmp olt <4 x float> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpltq_n_f32(float32x4_t a, float32_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpltq_n_f32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp slt <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpltq_n_s8(int8x16_t a, int8_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpltq_n_s8(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp slt <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpltq_n_s16(int16x8_t a, int16_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpltq_n_s16(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP0:%.*]] = icmp slt <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = trunc i32 [[TMP1]] to i16
+// CHECK-NEXT:    ret i16 [[TMP2]]
+//
+mve_pred16_t test_vcmpltq_n_s32(int32x4_t a, int32_t b)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq(a, b);
+#else /* POLYMORPHIC */
+    return vcmpltq_n_s32(a, b);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_m_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp olt <8 x half> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpltq_m_f16(float16x8_t a, float16x8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpltq_m_f16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_m_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp olt <4 x float> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpltq_m_f32(float32x4_t a, float32x4_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpltq_m_f32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_m_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp slt <16 x i8> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpltq_m_s8(int8x16_t a, int8x16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpltq_m_s8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_m_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp slt <8 x i16> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpltq_m_s16(int16x8_t a, int16x8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpltq_m_s16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_m_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp slt <4 x i32> [[A:%.*]], [[B:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpltq_m_s32(int32x4_t a, int32x4_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpltq_m_s32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_m_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast float [[B_COERCE:%.*]] to i32
+// CHECK-NEXT:    [[TMP_0_EXTRACT_TRUNC:%.*]] = trunc i32 [[TMP0]] to i16
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16 [[TMP_0_EXTRACT_TRUNC]] to half
+// CHECK-NEXT:    [[TMP2:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP3:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP2]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x half> undef, half [[TMP1]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x half> [[DOTSPLATINSERT]], <8 x half> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP4:%.*]] = fcmp olt <8 x half> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP5:%.*]] = and <8 x i1> [[TMP3]], [[TMP4]]
+// CHECK-NEXT:    [[TMP6:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP5]])
+// CHECK-NEXT:    [[TMP7:%.*]] = trunc i32 [[TMP6]] to i16
+// CHECK-NEXT:    ret i16 [[TMP7]]
+//
+mve_pred16_t test_vcmpltq_m_n_f16(float16x8_t a, float16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpltq_m_n_f16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_m_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x float> undef, float [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x float> [[DOTSPLATINSERT]], <4 x float> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = fcmp olt <4 x float> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpltq_m_n_f32(float32x4_t a, float32_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpltq_m_n_f32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_m_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <16 x i8> undef, i8 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <16 x i8> [[DOTSPLATINSERT]], <16 x i8> undef, <16 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp slt <16 x i8> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <16 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v16i1(<16 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpltq_m_n_s8(int8x16_t a, int8_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpltq_m_n_s8(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_m_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <8 x i16> undef, i16 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <8 x i16> [[DOTSPLATINSERT]], <8 x i16> undef, <8 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp slt <8 x i16> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <8 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v8i1(<8 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpltq_m_n_s16(int16x8_t a, int16_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpltq_m_n_s16(a, b, p);
+#endif /* POLYMORPHIC */
+}
+
+// CHECK-LABEL: @test_vcmpltq_m_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]])
+// CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> undef, i32 [[B:%.*]], i32 0
+// CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK-NEXT:    [[TMP2:%.*]] = icmp slt <4 x i32> [[A:%.*]], [[DOTSPLAT]]
+// CHECK-NEXT:    [[TMP3:%.*]] = and <4 x i1> [[TMP1]], [[TMP2]]
+// CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.arm.mve.pred.v2i.v4i1(<4 x i1> [[TMP3]])
+// CHECK-NEXT:    [[TMP5:%.*]] = trunc i32 [[TMP4]] to i16
+// CHECK-NEXT:    ret i16 [[TMP5]]
+//
+mve_pred16_t test_vcmpltq_m_n_s32(int32x4_t a, int32_t b, mve_pred16_t p)
+{
+#ifdef POLYMORPHIC
+    return vcmpltq_m(a, b, p);
+#else /* POLYMORPHIC */
+    return vcmpltq_m_n_s32(a, b, p);
+#endif /* POLYMORPHIC */
+}
+