[RISCV] Add a really basic cost model for SK_Splice.

While testing scalable vectors I found that if we generate a
vector splice intrinsic and run the code through the loop unroller,
we'll crash due to an invalid cost.

This adds a basic cost based on the 2 slide instructions used by the
lowering in D119303.

We probably need to factor LMUL into this, but that's true for
arithmetic instructions too. So I've ignored for the moment.

Reviewed By: ABataev

Differential Revision: https://reviews.llvm.org/D119316

diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
index 99e6774..a317501 100644 (file)
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
@@ -162,6 +162,24 @@ RISCVTTIImpl::getRegisterBitWidth(TargetTransformInfo::RegisterKind K) const {
   llvm_unreachable("Unsupported register kind");
 }
 
+InstructionCost RISCVTTIImpl::getSpliceCost(VectorType *Tp, int Index) {
+  std::pair<InstructionCost, MVT> LT = TLI->getTypeLegalizationCost(DL, Tp);
+
+  unsigned Cost = 2; // vslidedown+vslideup.
+  // TODO: LMUL should increase cost.
+  // TODO: Multiplying by LT.first implies this legalizes into multiple copies
+  // of similar code, but I think we expand through memory.
+  return Cost * LT.first;
+}
+
+InstructionCost RISCVTTIImpl::getShuffleCost(TTI::ShuffleKind Kind,
+                                             VectorType *Tp, ArrayRef<int> Mask,
+                                             int Index, VectorType *SubTp) {
+  if (Kind == TTI::SK_Splice && isa<ScalableVectorType>(Tp))
+    return getSpliceCost(Tp, Index);
+  return BaseT::getShuffleCost(Kind, Tp, Mask, Index, SubTp);
+}
+
 InstructionCost RISCVTTIImpl::getGatherScatterOpCost(
     unsigned Opcode, Type *DataTy, const Value *Ptr, bool VariableMask,
     Align Alignment, TTI::TargetCostKind CostKind, const Instruction *I) {

diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h
index e79c4f7..086c59c 100644 (file)
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h
@@ -73,6 +73,11 @@ public:
     return ST->useRVVForFixedLengthVectors() ? 16 : 0;
   }
 
+  InstructionCost getSpliceCost(VectorType *Tp, int Index);
+  InstructionCost getShuffleCost(TTI::ShuffleKind Kind, VectorType *Tp,
+                                 ArrayRef<int> Mask, int Index,
+                                 VectorType *SubTp);
+
   InstructionCost getGatherScatterOpCost(unsigned Opcode, Type *DataTy,
                                          const Value *Ptr, bool VariableMask,
                                          Align Alignment,

diff --git a/llvm/test/Analysis/CostModel/RISCV/rvv-shuffle.ll b/llvm/test/Analysis/CostModel/RISCV/rvv-shuffle.ll
index 497c16d..9371431 100644 (file)
--- a/llvm/test/Analysis/CostModel/RISCV/rvv-shuffle.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/rvv-shuffle.ll
@@ -106,20 +106,20 @@ declare <vscale x 2 x i1> @llvm.experimental.vector.reverse.nxv2i1(<vscale x 2 x
 
 define void @vector_splice() {
 ; CHECK-LABEL: 'vector_splice'
-; CHECK-NEXT:  Cost Model: Invalid cost for instruction: %splice_nxv16i8 = call <vscale x 16 x i8> @llvm.experimental.vector.splice.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i8> zeroinitializer, i32 1)
-; CHECK-NEXT:  Cost Model: Invalid cost for instruction: %splice_nxv32i8 = call <vscale x 32 x i8> @llvm.experimental.vector.splice.nxv32i8(<vscale x 32 x i8> zeroinitializer, <vscale x 32 x i8> zeroinitializer, i32 1)
-; CHECK-NEXT:  Cost Model: Invalid cost for instruction: %splice_nxv2i16 = call <vscale x 2 x i16> @llvm.experimental.vector.splice.nxv2i16(<vscale x 2 x i16> zeroinitializer, <vscale x 2 x i16> zeroinitializer, i32 1)
-; CHECK-NEXT:  Cost Model: Invalid cost for instruction: %splice_nxv4i16 = call <vscale x 4 x i16> @llvm.experimental.vector.splice.nxv4i16(<vscale x 4 x i16> zeroinitializer, <vscale x 4 x i16> zeroinitializer, i32 1)
-; CHECK-NEXT:  Cost Model: Invalid cost for instruction: %splice_nxv8i16 = call <vscale x 8 x i16> @llvm.experimental.vector.splice.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i16> zeroinitializer, i32 1)
-; CHECK-NEXT:  Cost Model: Invalid cost for instruction: %splice_nxv16i16 = call <vscale x 16 x i16> @llvm.experimental.vector.splice.nxv16i16(<vscale x 16 x i16> zeroinitializer, <vscale x 16 x i16> zeroinitializer, i32 1)
-; CHECK-NEXT:  Cost Model: Invalid cost for instruction: %splice_nxv4i32 = call <vscale x 4 x i32> @llvm.experimental.vector.splice.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i32> zeroinitializer, i32 1)
-; CHECK-NEXT:  Cost Model: Invalid cost for instruction: %splice_nxv8i32 = call <vscale x 8 x i32> @llvm.experimental.vector.splice.nxv8i32(<vscale x 8 x i32> zeroinitializer, <vscale x 8 x i32> zeroinitializer, i32 1)
-; CHECK-NEXT:  Cost Model: Invalid cost for instruction: %splice_nxv2i64 = call <vscale x 2 x i64> @llvm.experimental.vector.splice.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i64> zeroinitializer, i32 1)
-; CHECK-NEXT:  Cost Model: Invalid cost for instruction: %splice_nxv4i64 = call <vscale x 4 x i64> @llvm.experimental.vector.splice.nxv4i64(<vscale x 4 x i64> zeroinitializer, <vscale x 4 x i64> zeroinitializer, i32 1)
-; CHECK-NEXT:  Cost Model: Invalid cost for instruction: %splice_nxv16i1 = call <vscale x 16 x i1> @llvm.experimental.vector.splice.nxv16i1(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i1> zeroinitializer, i32 1)
-; CHECK-NEXT:  Cost Model: Invalid cost for instruction: %splice_nxv8i1 = call <vscale x 8 x i1> @llvm.experimental.vector.splice.nxv8i1(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i1> zeroinitializer, i32 1)
-; CHECK-NEXT:  Cost Model: Invalid cost for instruction: %splice_nxv4i1 = call <vscale x 4 x i1> @llvm.experimental.vector.splice.nxv4i1(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i1> zeroinitializer, i32 1)
-; CHECK-NEXT:  Cost Model: Invalid cost for instruction: %splice_nxv2i1 = call <vscale x 2 x i1> @llvm.experimental.vector.splice.nxv2i1(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i1> zeroinitializer, i32 1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice_nxv16i8 = call <vscale x 16 x i8> @llvm.experimental.vector.splice.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i8> zeroinitializer, i32 1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice_nxv32i8 = call <vscale x 32 x i8> @llvm.experimental.vector.splice.nxv32i8(<vscale x 32 x i8> zeroinitializer, <vscale x 32 x i8> zeroinitializer, i32 1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice_nxv2i16 = call <vscale x 2 x i16> @llvm.experimental.vector.splice.nxv2i16(<vscale x 2 x i16> zeroinitializer, <vscale x 2 x i16> zeroinitializer, i32 1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice_nxv4i16 = call <vscale x 4 x i16> @llvm.experimental.vector.splice.nxv4i16(<vscale x 4 x i16> zeroinitializer, <vscale x 4 x i16> zeroinitializer, i32 1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice_nxv8i16 = call <vscale x 8 x i16> @llvm.experimental.vector.splice.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i16> zeroinitializer, i32 1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice_nxv16i16 = call <vscale x 16 x i16> @llvm.experimental.vector.splice.nxv16i16(<vscale x 16 x i16> zeroinitializer, <vscale x 16 x i16> zeroinitializer, i32 1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice_nxv4i32 = call <vscale x 4 x i32> @llvm.experimental.vector.splice.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i32> zeroinitializer, i32 1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice_nxv8i32 = call <vscale x 8 x i32> @llvm.experimental.vector.splice.nxv8i32(<vscale x 8 x i32> zeroinitializer, <vscale x 8 x i32> zeroinitializer, i32 1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice_nxv2i64 = call <vscale x 2 x i64> @llvm.experimental.vector.splice.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i64> zeroinitializer, i32 1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice_nxv4i64 = call <vscale x 4 x i64> @llvm.experimental.vector.splice.nxv4i64(<vscale x 4 x i64> zeroinitializer, <vscale x 4 x i64> zeroinitializer, i32 1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice_nxv16i1 = call <vscale x 16 x i1> @llvm.experimental.vector.splice.nxv16i1(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i1> zeroinitializer, i32 1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice_nxv8i1 = call <vscale x 8 x i1> @llvm.experimental.vector.splice.nxv8i1(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i1> zeroinitializer, i32 1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice_nxv4i1 = call <vscale x 4 x i1> @llvm.experimental.vector.splice.nxv4i1(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i1> zeroinitializer, i32 1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice_nxv2i1 = call <vscale x 2 x i1> @llvm.experimental.vector.splice.nxv2i1(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i1> zeroinitializer, i32 1)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %splice_nxv16i8 = call <vscale x 16 x i8> @llvm.experimental.vector.splice.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i8> zeroinitializer, i32 1)

diff --git a/llvm/test/Analysis/CostModel/RISCV/splice.ll b/llvm/test/Analysis/CostModel/RISCV/splice.ll
new file mode 100644 (file)
index 0000000..1ab316b
--- /dev/null
+++ b/llvm/test/Analysis/CostModel/RISCV/splice.ll
@@ -0,0 +1,170 @@
+; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
+; RUN: opt < %s -analyze -cost-model -S -mtriple=riscv64 -mattr=+v,+f,+d,+zfh | FileCheck %s
+
+define void @vector_splice() {
+; CHECK-LABEL: 'vector_splice'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv1i8 = call <vscale x 1 x i8> @llvm.experimental.vector.splice.nxv1i8(<vscale x 1 x i8> zeroinitializer, <vscale x 1 x i8> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv2i8 = call <vscale x 2 x i8> @llvm.experimental.vector.splice.nxv2i8(<vscale x 2 x i8> zeroinitializer, <vscale x 2 x i8> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv4i8 = call <vscale x 4 x i8> @llvm.experimental.vector.splice.nxv4i8(<vscale x 4 x i8> zeroinitializer, <vscale x 4 x i8> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv8i8 = call <vscale x 8 x i8> @llvm.experimental.vector.splice.nxv8i8(<vscale x 8 x i8> zeroinitializer, <vscale x 8 x i8> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv16i8 = call <vscale x 16 x i8> @llvm.experimental.vector.splice.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i8> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv32i8 = call <vscale x 32 x i8> @llvm.experimental.vector.splice.nxv32i8(<vscale x 32 x i8> zeroinitializer, <vscale x 32 x i8> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv64i8 = call <vscale x 64 x i8> @llvm.experimental.vector.splice.nxv64i8(<vscale x 64 x i8> zeroinitializer, <vscale x 64 x i8> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv1i16 = call <vscale x 1 x i16> @llvm.experimental.vector.splice.nxv1i16(<vscale x 1 x i16> zeroinitializer, <vscale x 1 x i16> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv2i16 = call <vscale x 2 x i16> @llvm.experimental.vector.splice.nxv2i16(<vscale x 2 x i16> zeroinitializer, <vscale x 2 x i16> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv4i16 = call <vscale x 4 x i16> @llvm.experimental.vector.splice.nxv4i16(<vscale x 4 x i16> zeroinitializer, <vscale x 4 x i16> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv8i16 = call <vscale x 8 x i16> @llvm.experimental.vector.splice.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i16> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv16i16 = call <vscale x 16 x i16> @llvm.experimental.vector.splice.nxv16i16(<vscale x 16 x i16> zeroinitializer, <vscale x 16 x i16> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv32i16 = call <vscale x 32 x i16> @llvm.experimental.vector.splice.nxv32i16(<vscale x 32 x i16> zeroinitializer, <vscale x 32 x i16> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %splice.nxv64i16 = call <vscale x 64 x i16> @llvm.experimental.vector.splice.nxv64i16(<vscale x 64 x i16> zeroinitializer, <vscale x 64 x i16> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv1i32 = call <vscale x 1 x i32> @llvm.experimental.vector.splice.nxv1i32(<vscale x 1 x i32> zeroinitializer, <vscale x 1 x i32> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv2i32 = call <vscale x 2 x i32> @llvm.experimental.vector.splice.nxv2i32(<vscale x 2 x i32> zeroinitializer, <vscale x 2 x i32> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv4i32 = call <vscale x 4 x i32> @llvm.experimental.vector.splice.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i32> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv8i32 = call <vscale x 8 x i32> @llvm.experimental.vector.splice.nxv8i32(<vscale x 8 x i32> zeroinitializer, <vscale x 8 x i32> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv16i32 = call <vscale x 16 x i32> @llvm.experimental.vector.splice.nxv16i32(<vscale x 16 x i32> zeroinitializer, <vscale x 16 x i32> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %splice.nxv32i32 = call <vscale x 32 x i32> @llvm.experimental.vector.splice.nxv32i32(<vscale x 32 x i32> zeroinitializer, <vscale x 32 x i32> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %splice.nxv64i32 = call <vscale x 64 x i32> @llvm.experimental.vector.splice.nxv64i32(<vscale x 64 x i32> zeroinitializer, <vscale x 64 x i32> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv1i64 = call <vscale x 1 x i64> @llvm.experimental.vector.splice.nxv1i64(<vscale x 1 x i64> zeroinitializer, <vscale x 1 x i64> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv2i64 = call <vscale x 2 x i64> @llvm.experimental.vector.splice.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i64> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv4i64 = call <vscale x 4 x i64> @llvm.experimental.vector.splice.nxv4i64(<vscale x 4 x i64> zeroinitializer, <vscale x 4 x i64> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv8i64 = call <vscale x 8 x i64> @llvm.experimental.vector.splice.nxv8i64(<vscale x 8 x i64> zeroinitializer, <vscale x 8 x i64> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %splice.nxv16i64 = call <vscale x 16 x i64> @llvm.experimental.vector.splice.nxv16i64(<vscale x 16 x i64> zeroinitializer, <vscale x 16 x i64> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %splice.nxv32i64 = call <vscale x 32 x i64> @llvm.experimental.vector.splice.nxv32i64(<vscale x 32 x i64> zeroinitializer, <vscale x 32 x i64> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %splice.nxv64i64 = call <vscale x 64 x i64> @llvm.experimental.vector.splice.nxv64i64(<vscale x 64 x i64> zeroinitializer, <vscale x 64 x i64> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv1f16 = call <vscale x 1 x half> @llvm.experimental.vector.splice.nxv1f16(<vscale x 1 x half> zeroinitializer, <vscale x 1 x half> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv2f16 = call <vscale x 2 x half> @llvm.experimental.vector.splice.nxv2f16(<vscale x 2 x half> zeroinitializer, <vscale x 2 x half> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv4f16 = call <vscale x 4 x half> @llvm.experimental.vector.splice.nxv4f16(<vscale x 4 x half> zeroinitializer, <vscale x 4 x half> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv8f16 = call <vscale x 8 x half> @llvm.experimental.vector.splice.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x half> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv16f16 = call <vscale x 16 x half> @llvm.experimental.vector.splice.nxv16f16(<vscale x 16 x half> zeroinitializer, <vscale x 16 x half> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv32f16 = call <vscale x 32 x half> @llvm.experimental.vector.splice.nxv32f16(<vscale x 32 x half> zeroinitializer, <vscale x 32 x half> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %splice.nxv64f16 = call <vscale x 64 x half> @llvm.experimental.vector.splice.nxv64f16(<vscale x 64 x half> zeroinitializer, <vscale x 64 x half> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv1f32 = call <vscale x 1 x float> @llvm.experimental.vector.splice.nxv1f32(<vscale x 1 x float> zeroinitializer, <vscale x 1 x float> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv2f32 = call <vscale x 2 x float> @llvm.experimental.vector.splice.nxv2f32(<vscale x 2 x float> zeroinitializer, <vscale x 2 x float> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv4f32 = call <vscale x 4 x float> @llvm.experimental.vector.splice.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x float> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv8f32 = call <vscale x 8 x float> @llvm.experimental.vector.splice.nxv8f32(<vscale x 8 x float> zeroinitializer, <vscale x 8 x float> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv16f32 = call <vscale x 16 x float> @llvm.experimental.vector.splice.nxv16f32(<vscale x 16 x float> zeroinitializer, <vscale x 16 x float> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %splice.nxv32f32 = call <vscale x 32 x float> @llvm.experimental.vector.splice.nxv32f32(<vscale x 32 x float> zeroinitializer, <vscale x 32 x float> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %splice.nxv64f32 = call <vscale x 64 x float> @llvm.experimental.vector.splice.nxv64f32(<vscale x 64 x float> zeroinitializer, <vscale x 64 x float> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv1f64 = call <vscale x 1 x double> @llvm.experimental.vector.splice.nxv1f64(<vscale x 1 x double> zeroinitializer, <vscale x 1 x double> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv2f64 = call <vscale x 2 x double> @llvm.experimental.vector.splice.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x double> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv4f64 = call <vscale x 4 x double> @llvm.experimental.vector.splice.nxv4f64(<vscale x 4 x double> zeroinitializer, <vscale x 4 x double> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %splice.nxv8f64 = call <vscale x 8 x double> @llvm.experimental.vector.splice.nxv8f64(<vscale x 8 x double> zeroinitializer, <vscale x 8 x double> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %splice.nxv16f64 = call <vscale x 16 x double> @llvm.experimental.vector.splice.nxv16f64(<vscale x 16 x double> zeroinitializer, <vscale x 16 x double> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %splice.nxv32f64 = call <vscale x 32 x double> @llvm.experimental.vector.splice.nxv32f64(<vscale x 32 x double> zeroinitializer, <vscale x 32 x double> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %splice.nxv64f64 = call <vscale x 64 x double> @llvm.experimental.vector.splice.nxv64f64(<vscale x 64 x double> zeroinitializer, <vscale x 64 x double> zeroinitializer, i32 -1)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+  %splice.nxv1i8 = call <vscale x 1 x i8> @llvm.experimental.vector.splice.nxv1i8(<vscale x 1 x i8> zeroinitializer, <vscale x 1 x i8> zeroinitializer, i32 -1)
+  %splice.nxv2i8 = call <vscale x 2 x i8> @llvm.experimental.vector.splice.nxv2i8(<vscale x 2 x i8> zeroinitializer, <vscale x 2 x i8> zeroinitializer, i32 -1)
+  %splice.nxv4i8 = call <vscale x 4 x i8> @llvm.experimental.vector.splice.nxv4i8(<vscale x 4 x i8> zeroinitializer, <vscale x 4 x i8> zeroinitializer, i32 -1)
+  %splice.nxv8i8 = call <vscale x 8 x i8> @llvm.experimental.vector.splice.nxv8i8(<vscale x 8 x i8> zeroinitializer, <vscale x 8 x i8> zeroinitializer, i32 -1)
+  %splice.nxv16i8 = call <vscale x 16 x i8> @llvm.experimental.vector.splice.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i8> zeroinitializer, i32 -1)
+  %splice.nxv32i8 = call <vscale x 32 x i8> @llvm.experimental.vector.splice.nxv32i8(<vscale x 32 x i8> zeroinitializer, <vscale x 32 x i8> zeroinitializer, i32 -1)
+  %splice.nxv64i8 = call <vscale x 64 x i8> @llvm.experimental.vector.splice.nxv64i8(<vscale x 64 x i8> zeroinitializer, <vscale x 64 x i8> zeroinitializer, i32 -1)
+
+  %splice.nxv1i16 = call <vscale x 1 x i16> @llvm.experimental.vector.splice.nxv1i16(<vscale x 1 x i16> zeroinitializer, <vscale x 1 x i16> zeroinitializer, i32 -1)
+  %splice.nxv2i16 = call <vscale x 2 x i16> @llvm.experimental.vector.splice.nxv2i16(<vscale x 2 x i16> zeroinitializer, <vscale x 2 x i16> zeroinitializer, i32 -1)
+  %splice.nxv4i16 = call <vscale x 4 x i16> @llvm.experimental.vector.splice.nxv4i16(<vscale x 4 x i16> zeroinitializer, <vscale x 4 x i16> zeroinitializer, i32 -1)
+  %splice.nxv8i16 = call <vscale x 8 x i16> @llvm.experimental.vector.splice.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i16> zeroinitializer, i32 -1)
+  %splice.nxv16i16 = call <vscale x 16 x i16> @llvm.experimental.vector.splice.nxv16i16(<vscale x 16 x i16> zeroinitializer, <vscale x 16 x i16> zeroinitializer, i32 -1)
+  %splice.nxv32i16 = call <vscale x 32 x i16> @llvm.experimental.vector.splice.nxv32i16(<vscale x 32 x i16> zeroinitializer, <vscale x 32 x i16> zeroinitializer, i32 -1)
+  %splice.nxv64i16 = call <vscale x 64 x i16> @llvm.experimental.vector.splice.nxv64i16(<vscale x 64 x i16> zeroinitializer, <vscale x 64 x i16> zeroinitializer, i32 -1)
+
+  %splice.nxv1i32 = call <vscale x 1 x i32> @llvm.experimental.vector.splice.nxv1i32(<vscale x 1 x i32> zeroinitializer, <vscale x 1 x i32> zeroinitializer, i32 -1)
+  %splice.nxv2i32 = call <vscale x 2 x i32> @llvm.experimental.vector.splice.nxv2i32(<vscale x 2 x i32> zeroinitializer, <vscale x 2 x i32> zeroinitializer, i32 -1)
+  %splice.nxv4i32 = call <vscale x 4 x i32> @llvm.experimental.vector.splice.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i32> zeroinitializer, i32 -1)
+  %splice.nxv8i32 = call <vscale x 8 x i32> @llvm.experimental.vector.splice.nxv8i32(<vscale x 8 x i32> zeroinitializer, <vscale x 8 x i32> zeroinitializer, i32 -1)
+  %splice.nxv16i32 = call <vscale x 16 x i32> @llvm.experimental.vector.splice.nxv16i32(<vscale x 16 x i32> zeroinitializer, <vscale x 16 x i32> zeroinitializer, i32 -1)
+  %splice.nxv32i32 = call <vscale x 32 x i32> @llvm.experimental.vector.splice.nxv32i32(<vscale x 32 x i32> zeroinitializer, <vscale x 32 x i32> zeroinitializer, i32 -1)
+  %splice.nxv64i32 = call <vscale x 64 x i32> @llvm.experimental.vector.splice.nxv64i32(<vscale x 64 x i32> zeroinitializer, <vscale x 64 x i32> zeroinitializer, i32 -1)
+
+  %splice.nxv1i64 = call <vscale x 1 x i64> @llvm.experimental.vector.splice.nxv1i64(<vscale x 1 x i64> zeroinitializer, <vscale x 1 x i64> zeroinitializer, i32 -1)
+  %splice.nxv2i64 = call <vscale x 2 x i64> @llvm.experimental.vector.splice.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i64> zeroinitializer, i32 -1)
+  %splice.nxv4i64 = call <vscale x 4 x i64> @llvm.experimental.vector.splice.nxv4i64(<vscale x 4 x i64> zeroinitializer, <vscale x 4 x i64> zeroinitializer, i32 -1)
+  %splice.nxv8i64 = call <vscale x 8 x i64> @llvm.experimental.vector.splice.nxv8i64(<vscale x 8 x i64> zeroinitializer, <vscale x 8 x i64> zeroinitializer, i32 -1)
+  %splice.nxv16i64 = call <vscale x 16 x i64> @llvm.experimental.vector.splice.nxv16i64(<vscale x 16 x i64> zeroinitializer, <vscale x 16 x i64> zeroinitializer, i32 -1)
+  %splice.nxv32i64 = call <vscale x 32 x i64> @llvm.experimental.vector.splice.nxv32i64(<vscale x 32 x i64> zeroinitializer, <vscale x 32 x i64> zeroinitializer, i32 -1)
+  %splice.nxv64i64 = call <vscale x 64 x i64> @llvm.experimental.vector.splice.nxv64i64(<vscale x 64 x i64> zeroinitializer, <vscale x 64 x i64> zeroinitializer, i32 -1)
+
+  %splice.nxv1f16 = call <vscale x 1 x half> @llvm.experimental.vector.splice.nxv1f16(<vscale x 1 x half> zeroinitializer, <vscale x 1 x half> zeroinitializer, i32 -1)
+  %splice.nxv2f16 = call <vscale x 2 x half> @llvm.experimental.vector.splice.nxv2f16(<vscale x 2 x half> zeroinitializer, <vscale x 2 x half> zeroinitializer, i32 -1)
+  %splice.nxv4f16 = call <vscale x 4 x half> @llvm.experimental.vector.splice.nxv4f16(<vscale x 4 x half> zeroinitializer, <vscale x 4 x half> zeroinitializer, i32 -1)
+  %splice.nxv8f16 = call <vscale x 8 x half> @llvm.experimental.vector.splice.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x half> zeroinitializer, i32 -1)
+  %splice.nxv16f16 = call <vscale x 16 x half> @llvm.experimental.vector.splice.nxv16f16(<vscale x 16 x half> zeroinitializer, <vscale x 16 x half> zeroinitializer, i32 -1)
+  %splice.nxv32f16 = call <vscale x 32 x half> @llvm.experimental.vector.splice.nxv32f16(<vscale x 32 x half> zeroinitializer, <vscale x 32 x half> zeroinitializer, i32 -1)
+  %splice.nxv64f16 = call <vscale x 64 x half> @llvm.experimental.vector.splice.nxv64f16(<vscale x 64 x half> zeroinitializer, <vscale x 64 x half> zeroinitializer, i32 -1)
+
+  %splice.nxv1f32 = call <vscale x 1 x float> @llvm.experimental.vector.splice.nxv1f32(<vscale x 1 x float> zeroinitializer, <vscale x 1 x float> zeroinitializer, i32 -1)
+  %splice.nxv2f32 = call <vscale x 2 x float> @llvm.experimental.vector.splice.nxv2f32(<vscale x 2 x float> zeroinitializer, <vscale x 2 x float> zeroinitializer, i32 -1)
+  %splice.nxv4f32 = call <vscale x 4 x float> @llvm.experimental.vector.splice.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x float> zeroinitializer, i32 -1)
+  %splice.nxv8f32 = call <vscale x 8 x float> @llvm.experimental.vector.splice.nxv8f32(<vscale x 8 x float> zeroinitializer, <vscale x 8 x float> zeroinitializer, i32 -1)
+  %splice.nxv16f32 = call <vscale x 16 x float> @llvm.experimental.vector.splice.nxv16f32(<vscale x 16 x float> zeroinitializer, <vscale x 16 x float> zeroinitializer, i32 -1)
+  %splice.nxv32f32 = call <vscale x 32 x float> @llvm.experimental.vector.splice.nxv32f32(<vscale x 32 x float> zeroinitializer, <vscale x 32 x float> zeroinitializer, i32 -1)
+  %splice.nxv64f32 = call <vscale x 64 x float> @llvm.experimental.vector.splice.nxv64f32(<vscale x 64 x float> zeroinitializer, <vscale x 64 x float> zeroinitializer, i32 -1)
+
+  %splice.nxv1f64 = call <vscale x 1 x double> @llvm.experimental.vector.splice.nxv1f64(<vscale x 1 x double> zeroinitializer, <vscale x 1 x double> zeroinitializer, i32 -1)
+  %splice.nxv2f64 = call <vscale x 2 x double> @llvm.experimental.vector.splice.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x double> zeroinitializer, i32 -1)
+  %splice.nxv4f64 = call <vscale x 4 x double> @llvm.experimental.vector.splice.nxv4f64(<vscale x 4 x double> zeroinitializer, <vscale x 4 x double> zeroinitializer, i32 -1)
+  %splice.nxv8f64 = call <vscale x 8 x double> @llvm.experimental.vector.splice.nxv8f64(<vscale x 8 x double> zeroinitializer, <vscale x 8 x double> zeroinitializer, i32 -1)
+  %splice.nxv16f64 = call <vscale x 16 x double> @llvm.experimental.vector.splice.nxv16f64(<vscale x 16 x double> zeroinitializer, <vscale x 16 x double> zeroinitializer, i32 -1)
+  %splice.nxv32f64 = call <vscale x 32 x double> @llvm.experimental.vector.splice.nxv32f64(<vscale x 32 x double> zeroinitializer, <vscale x 32 x double> zeroinitializer, i32 -1)
+  %splice.nxv64f64 = call <vscale x 64 x double> @llvm.experimental.vector.splice.nxv64f64(<vscale x 64 x double> zeroinitializer, <vscale x 64 x double> zeroinitializer, i32 -1)
+
+  ret void
+}
+
+declare <vscale x 1 x i8> @llvm.experimental.vector.splice.nxv1i8(<vscale x 1 x i8>, <vscale x 1 x i8>, i32)
+declare <vscale x 2 x i8> @llvm.experimental.vector.splice.nxv2i8(<vscale x 2 x i8>, <vscale x 2 x i8>, i32)
+declare <vscale x 4 x i8> @llvm.experimental.vector.splice.nxv4i8(<vscale x 4 x i8>, <vscale x 4 x i8>, i32)
+declare <vscale x 8 x i8> @llvm.experimental.vector.splice.nxv8i8(<vscale x 8 x i8>, <vscale x 8 x i8>, i32)
+declare <vscale x 16 x i8> @llvm.experimental.vector.splice.nxv16i8(<vscale x 16 x i8>, <vscale x 16 x i8>, i32)
+declare <vscale x 32 x i8> @llvm.experimental.vector.splice.nxv32i8(<vscale x 32 x i8>, <vscale x 32 x i8>, i32)
+declare <vscale x 64 x i8> @llvm.experimental.vector.splice.nxv64i8(<vscale x 64 x i8>, <vscale x 64 x i8>, i32)
+
+declare <vscale x 1 x i16> @llvm.experimental.vector.splice.nxv1i16(<vscale x 1 x i16>, <vscale x 1 x i16>, i32)
+declare <vscale x 2 x i16> @llvm.experimental.vector.splice.nxv2i16(<vscale x 2 x i16>, <vscale x 2 x i16>, i32)
+declare <vscale x 4 x i16> @llvm.experimental.vector.splice.nxv4i16(<vscale x 4 x i16>, <vscale x 4 x i16>, i32)
+declare <vscale x 8 x i16> @llvm.experimental.vector.splice.nxv8i16(<vscale x 8 x i16>, <vscale x 8 x i16>, i32)
+declare <vscale x 16 x i16> @llvm.experimental.vector.splice.nxv16i16(<vscale x 16 x i16>, <vscale x 16 x i16>, i32)
+declare <vscale x 32 x i16> @llvm.experimental.vector.splice.nxv32i16(<vscale x 32 x i16>, <vscale x 32 x i16>, i32)
+declare <vscale x 64 x i16> @llvm.experimental.vector.splice.nxv64i16(<vscale x 64 x i16>, <vscale x 64 x i16>, i32)
+
+declare <vscale x 1 x i32> @llvm.experimental.vector.splice.nxv1i32(<vscale x 1 x i32>, <vscale x 1 x i32>, i32)
+declare <vscale x 2 x i32> @llvm.experimental.vector.splice.nxv2i32(<vscale x 2 x i32>, <vscale x 2 x i32>, i32)
+declare <vscale x 4 x i32> @llvm.experimental.vector.splice.nxv4i32(<vscale x 4 x i32>, <vscale x 4 x i32>, i32)
+declare <vscale x 8 x i32> @llvm.experimental.vector.splice.nxv8i32(<vscale x 8 x i32>, <vscale x 8 x i32>, i32)
+declare <vscale x 16 x i32> @llvm.experimental.vector.splice.nxv16i32(<vscale x 16 x i32>, <vscale x 16 x i32>, i32)
+declare <vscale x 32 x i32> @llvm.experimental.vector.splice.nxv32i32(<vscale x 32 x i32>, <vscale x 32 x i32>, i32)
+declare <vscale x 64 x i32> @llvm.experimental.vector.splice.nxv64i32(<vscale x 64 x i32>, <vscale x 64 x i32>, i32)
+
+declare <vscale x 1 x i64> @llvm.experimental.vector.splice.nxv1i64(<vscale x 1 x i64>, <vscale x 1 x i64>, i32)
+declare <vscale x 2 x i64> @llvm.experimental.vector.splice.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i64>, i32)
+declare <vscale x 4 x i64> @llvm.experimental.vector.splice.nxv4i64(<vscale x 4 x i64>, <vscale x 4 x i64>, i32)
+declare <vscale x 8 x i64> @llvm.experimental.vector.splice.nxv8i64(<vscale x 8 x i64>, <vscale x 8 x i64>, i32)
+declare <vscale x 16 x i64> @llvm.experimental.vector.splice.nxv16i64(<vscale x 16 x i64>, <vscale x 16 x i64>, i32)
+declare <vscale x 32 x i64> @llvm.experimental.vector.splice.nxv32i64(<vscale x 32 x i64>, <vscale x 32 x i64>, i32)
+declare <vscale x 64 x i64> @llvm.experimental.vector.splice.nxv64i64(<vscale x 64 x i64>, <vscale x 64 x i64>, i32)
+
+declare <vscale x 1 x half> @llvm.experimental.vector.splice.nxv1f16(<vscale x 1 x half>, <vscale x 1 x half>, i32)
+declare <vscale x 2 x half> @llvm.experimental.vector.splice.nxv2f16(<vscale x 2 x half>, <vscale x 2 x half>, i32)
+declare <vscale x 4 x half> @llvm.experimental.vector.splice.nxv4f16(<vscale x 4 x half>, <vscale x 4 x half>, i32)
+declare <vscale x 8 x half> @llvm.experimental.vector.splice.nxv8f16(<vscale x 8 x half>, <vscale x 8 x half>, i32)
+declare <vscale x 16 x half> @llvm.experimental.vector.splice.nxv16f16(<vscale x 16 x half>, <vscale x 16 x half>, i32)
+declare <vscale x 32 x half> @llvm.experimental.vector.splice.nxv32f16(<vscale x 32 x half>, <vscale x 32 x half>, i32)
+declare <vscale x 64 x half> @llvm.experimental.vector.splice.nxv64f16(<vscale x 64 x half>, <vscale x 64 x half>, i32)
+
+declare <vscale x 1 x float> @llvm.experimental.vector.splice.nxv1f32(<vscale x 1 x float>, <vscale x 1 x float>, i32)
+declare <vscale x 2 x float> @llvm.experimental.vector.splice.nxv2f32(<vscale x 2 x float>, <vscale x 2 x float>, i32)
+declare <vscale x 4 x float> @llvm.experimental.vector.splice.nxv4f32(<vscale x 4 x float>, <vscale x 4 x float>, i32)
+declare <vscale x 8 x float> @llvm.experimental.vector.splice.nxv8f32(<vscale x 8 x float>, <vscale x 8 x float>, i32)
+declare <vscale x 16 x float> @llvm.experimental.vector.splice.nxv16f32(<vscale x 16 x float>, <vscale x 16 x float>, i32)
+declare <vscale x 32 x float> @llvm.experimental.vector.splice.nxv32f32(<vscale x 32 x float>, <vscale x 32 x float>, i32)
+declare <vscale x 64 x float> @llvm.experimental.vector.splice.nxv64f32(<vscale x 64 x float>, <vscale x 64 x float>, i32)
+
+declare <vscale x 1 x double> @llvm.experimental.vector.splice.nxv1f64(<vscale x 1 x double>, <vscale x 1 x double>, i32)
+declare <vscale x 2 x double> @llvm.experimental.vector.splice.nxv2f64(<vscale x 2 x double>, <vscale x 2 x double>, i32)
+declare <vscale x 4 x double> @llvm.experimental.vector.splice.nxv4f64(<vscale x 4 x double>, <vscale x 4 x double>, i32)
+declare <vscale x 8 x double> @llvm.experimental.vector.splice.nxv8f64(<vscale x 8 x double>, <vscale x 8 x double>, i32)
+declare <vscale x 16 x double> @llvm.experimental.vector.splice.nxv16f64(<vscale x 16 x double>, <vscale x 16 x double>, i32)
+declare <vscale x 32 x double> @llvm.experimental.vector.splice.nxv32f64(<vscale x 32 x double>, <vscale x 32 x double>, i32)
+declare <vscale x 64 x double> @llvm.experimental.vector.splice.nxv64f64(<vscale x 64 x double>, <vscale x 64 x double>, i32)