[RISCV][TTI] Implement instruction cost for vp_merge.

[ElvisWang123]

This patch implement the instruction for vp_merge, which will generate similar instruction sequence to the select instruction.

[llvmbot]

@llvm/pr-subscribers-backend-risc-v

@llvm/pr-subscribers-llvm-analysis

Author: Elvis Wang (ElvisWang123)

Changes

This patch implement the instruction for vp_merge, which will generate similar instruction sequence to the select instruction.

---

Patch is 39.16 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/112327.diff

2 Files Affected:

• (modified) llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp (+5-1)
• (added) llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll (+240)

diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
index 8d18fd63e4a2e1..6147b977e857c9 100644
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
@@ -1132,6 +1132,10 @@ RISCVTTIImpl::getIntrinsicInstrCost(const IntrinsicCostAttributes &ICA,
     return getCmpSelInstrCost(*FOp, ICA.getReturnType(), ICA.getArgTypes()[0],
                               CmpInst::BAD_ICMP_PREDICATE, CostKind);
   }
+  case Intrinsic::vp_merge:
+    return getCmpSelInstrCost(Instruction::Select, ICA.getReturnType(),
+                              ICA.getArgTypes()[0], CmpInst::BAD_ICMP_PREDICATE,
+                              CostKind);
   }
 
   if (ST->hasVInstructions() && RetTy->isVectorTy()) {
@@ -2431,4 +2435,4 @@ bool RISCVTTIImpl::isProfitableToSinkOperands(
     Ops.push_back(&OpIdx.value());
   }
   return true;
-}
\ No newline at end of file
+}
diff --git a/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll b/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll
new file mode 100644
index 00000000000000..57a81cc87fb6a3
--- /dev/null
+++ b/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll
@@ -0,0 +1,240 @@
+; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+v,+f,+d,+zfh,+zvfh -riscv-v-fixed-length-vector-lmul-max=1 < %s \
+; RUN: | FileCheck %s --check-prefixes=CHECK
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+v,+f,+d,+zfh,+zvfh -riscv-v-fixed-length-vector-lmul-max=1 \
+; RUN: --type-based-intrinsic-cost=true < %s | FileCheck %s --check-prefixes=TYPEBASED
+; Check that we don't crash querying costs when vectors are not enabled.
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64
+
+define void @vp_merge() {
+; CHECK-LABEL: 'vp_merge'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = call <1 x i8> @llvm.vp.merge.v1i8(<1 x i1> undef, <1 x i8> undef, <1 x i8> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %2 = call <2 x i8> @llvm.vp.merge.v2i8(<2 x i1> undef, <2 x i8> undef, <2 x i8> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %3 = call <4 x i8> @llvm.vp.merge.v4i8(<4 x i1> undef, <4 x i8> undef, <4 x i8> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %4 = call <8 x i8> @llvm.vp.merge.v8i8(<8 x i1> undef, <8 x i8> undef, <8 x i8> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %5 = call <16 x i8> @llvm.vp.merge.v16i8(<16 x i1> undef, <16 x i8> undef, <16 x i8> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %6 = call <32 x i8> @llvm.vp.merge.v32i8(<32 x i1> undef, <32 x i8> undef, <32 x i8> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %7 = call <vscale x 1 x i8> @llvm.vp.merge.nxv1i8(<vscale x 1 x i1> undef, <vscale x 1 x i8> undef, <vscale x 1 x i8> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %8 = call <vscale x 2 x i8> @llvm.vp.merge.nxv2i8(<vscale x 2 x i1> undef, <vscale x 2 x i8> undef, <vscale x 2 x i8> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %9 = call <vscale x 4 x i8> @llvm.vp.merge.nxv4i8(<vscale x 4 x i1> undef, <vscale x 4 x i8> undef, <vscale x 4 x i8> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %10 = call <vscale x 8 x i8> @llvm.vp.merge.nxv8i8(<vscale x 8 x i1> undef, <vscale x 8 x i8> undef, <vscale x 8 x i8> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %11 = call <vscale x 16 x i8> @llvm.vp.merge.nxv16i8(<vscale x 16 x i1> undef, <vscale x 16 x i8> undef, <vscale x 16 x i8> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %12 = call <vscale x 32 x i8> @llvm.vp.merge.nxv32i8(<vscale x 32 x i1> undef, <vscale x 32 x i8> undef, <vscale x 32 x i8> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %13 = call <1 x i16> @llvm.vp.merge.v1i16(<1 x i1> undef, <1 x i16> undef, <1 x i16> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %14 = call <2 x i16> @llvm.vp.merge.v2i16(<2 x i1> undef, <2 x i16> undef, <2 x i16> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %15 = call <4 x i16> @llvm.vp.merge.v4i16(<4 x i1> undef, <4 x i16> undef, <4 x i16> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %16 = call <8 x i16> @llvm.vp.merge.v8i16(<8 x i1> undef, <8 x i16> undef, <8 x i16> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %17 = call <16 x i16> @llvm.vp.merge.v16i16(<16 x i1> undef, <16 x i16> undef, <16 x i16> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %18 = call <32 x i16> @llvm.vp.merge.v32i16(<32 x i1> undef, <32 x i16> undef, <32 x i16> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %19 = call <vscale x 1 x i16> @llvm.vp.merge.nxv1i16(<vscale x 1 x i1> undef, <vscale x 1 x i16> undef, <vscale x 1 x i16> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %20 = call <vscale x 2 x i16> @llvm.vp.merge.nxv2i16(<vscale x 2 x i1> undef, <vscale x 2 x i16> undef, <vscale x 2 x i16> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %21 = call <vscale x 4 x i16> @llvm.vp.merge.nxv4i16(<vscale x 4 x i1> undef, <vscale x 4 x i16> undef, <vscale x 4 x i16> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %22 = call <vscale x 8 x i16> @llvm.vp.merge.nxv8i16(<vscale x 8 x i1> undef, <vscale x 8 x i16> undef, <vscale x 8 x i16> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %23 = call <vscale x 16 x i16> @llvm.vp.merge.nxv16i16(<vscale x 16 x i1> undef, <vscale x 16 x i16> undef, <vscale x 16 x i16> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %24 = call <vscale x 32 x i16> @llvm.vp.merge.nxv32i16(<vscale x 32 x i1> undef, <vscale x 32 x i16> undef, <vscale x 32 x i16> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %25 = call <1 x i32> @llvm.vp.merge.v1i32(<1 x i1> undef, <1 x i32> undef, <1 x i32> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %26 = call <2 x i32> @llvm.vp.merge.v2i32(<2 x i1> undef, <2 x i32> undef, <2 x i32> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %27 = call <4 x i32> @llvm.vp.merge.v4i32(<4 x i1> undef, <4 x i32> undef, <4 x i32> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %28 = call <8 x i32> @llvm.vp.merge.v8i32(<8 x i1> undef, <8 x i32> undef, <8 x i32> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %29 = call <16 x i32> @llvm.vp.merge.v16i32(<16 x i1> undef, <16 x i32> undef, <16 x i32> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %30 = call <32 x i32> @llvm.vp.merge.v32i32(<32 x i1> undef, <32 x i32> undef, <32 x i32> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %31 = call <vscale x 1 x i32> @llvm.vp.merge.nxv1i32(<vscale x 1 x i1> undef, <vscale x 1 x i32> undef, <vscale x 1 x i32> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %32 = call <vscale x 2 x i32> @llvm.vp.merge.nxv2i32(<vscale x 2 x i1> undef, <vscale x 2 x i32> undef, <vscale x 2 x i32> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %33 = call <vscale x 4 x i32> @llvm.vp.merge.nxv4i32(<vscale x 4 x i1> undef, <vscale x 4 x i32> undef, <vscale x 4 x i32> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %34 = call <vscale x 8 x i32> @llvm.vp.merge.nxv8i32(<vscale x 8 x i1> undef, <vscale x 8 x i32> undef, <vscale x 8 x i32> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %35 = call <vscale x 16 x i32> @llvm.vp.merge.nxv16i32(<vscale x 16 x i1> undef, <vscale x 16 x i32> undef, <vscale x 16 x i32> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %36 = call <vscale x 32 x i32> @llvm.vp.merge.nxv32i32(<vscale x 32 x i1> undef, <vscale x 32 x i32> undef, <vscale x 32 x i32> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %37 = call <1 x i64> @llvm.vp.merge.v1i64(<1 x i1> undef, <1 x i64> undef, <1 x i64> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %38 = call <2 x i64> @llvm.vp.merge.v2i64(<2 x i1> undef, <2 x i64> undef, <2 x i64> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %39 = call <4 x i64> @llvm.vp.merge.v4i64(<4 x i1> undef, <4 x i64> undef, <4 x i64> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %40 = call <8 x i64> @llvm.vp.merge.v8i64(<8 x i1> undef, <8 x i64> undef, <8 x i64> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %41 = call <16 x i64> @llvm.vp.merge.v16i64(<16 x i1> undef, <16 x i64> undef, <16 x i64> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %42 = call <32 x i64> @llvm.vp.merge.v32i64(<32 x i1> undef, <32 x i64> undef, <32 x i64> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %43 = call <vscale x 1 x i64> @llvm.vp.merge.nxv1i64(<vscale x 1 x i1> undef, <vscale x 1 x i64> undef, <vscale x 1 x i64> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %44 = call <vscale x 2 x i64> @llvm.vp.merge.nxv2i64(<vscale x 2 x i1> undef, <vscale x 2 x i64> undef, <vscale x 2 x i64> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %45 = call <vscale x 4 x i64> @llvm.vp.merge.nxv4i64(<vscale x 4 x i1> undef, <vscale x 4 x i64> undef, <vscale x 4 x i64> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %46 = call <vscale x 8 x i64> @llvm.vp.merge.nxv8i64(<vscale x 8 x i1> undef, <vscale x 8 x i64> undef, <vscale x 8 x i64> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %47 = call <vscale x 16 x i64> @llvm.vp.merge.nxv16i64(<vscale x 16 x i1> undef, <vscale x 16 x i64> undef, <vscale x 16 x i64> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %48 = call <vscale x 32 x i64> @llvm.vp.merge.nxv32i64(<vscale x 32 x i1> undef, <vscale x 32 x i64> undef, <vscale x 32 x i64> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %49 = call <1 x float> @llvm.vp.merge.v1f32(<1 x i1> undef, <1 x float> undef, <1 x float> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %50 = call <2 x float> @llvm.vp.merge.v2f32(<2 x i1> undef, <2 x float> undef, <2 x float> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %51 = call <4 x float> @llvm.vp.merge.v4f32(<4 x i1> undef, <4 x float> undef, <4 x float> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %52 = call <8 x float> @llvm.vp.merge.v8f32(<8 x i1> undef, <8 x float> undef, <8 x float> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %53 = call <16 x float> @llvm.vp.merge.v16f32(<16 x i1> undef, <16 x float> undef, <16 x float> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %54 = call <32 x float> @llvm.vp.merge.v32f32(<32 x i1> undef, <32 x float> undef, <32 x float> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %55 = call <vscale x 1 x float> @llvm.vp.merge.nxv1f32(<vscale x 1 x i1> undef, <vscale x 1 x float> undef, <vscale x 1 x float> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %56 = call <vscale x 2 x float> @llvm.vp.merge.nxv2f32(<vscale x 2 x i1> undef, <vscale x 2 x float> undef, <vscale x 2 x float> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %57 = call <vscale x 4 x float> @llvm.vp.merge.nxv4f32(<vscale x 4 x i1> undef, <vscale x 4 x float> undef, <vscale x 4 x float> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %58 = call <vscale x 8 x float> @llvm.vp.merge.nxv8f32(<vscale x 8 x i1> undef, <vscale x 8 x float> undef, <vscale x 8 x float> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %59 = call <vscale x 16 x float> @llvm.vp.merge.nxv16f32(<vscale x 16 x i1> undef, <vscale x 16 x float> undef, <vscale x 16 x float> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %60 = call <vscale x 32 x float> @llvm.vp.merge.nxv32f32(<vscale x 32 x i1> undef, <vscale x 32 x float> undef, <vscale x 32 x float> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %61 = call <1 x double> @llvm.vp.merge.v1f64(<1 x i1> undef, <1 x double> undef, <1 x double> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %62 = call <2 x double> @llvm.vp.merge.v2f64(<2 x i1> undef, <2 x double> undef, <2 x double> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %63 = call <4 x double> @llvm.vp.merge.v4f64(<4 x i1> undef, <4 x double> undef, <4 x double> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %64 = call <8 x double> @llvm.vp.merge.v8f64(<8 x i1> undef, <8 x double> undef, <8 x double> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %65 = call <16 x double> @llvm.vp.merge.v16f64(<16 x i1> undef, <16 x double> undef, <16 x double> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %66 = call <32 x double> @llvm.vp.merge.v32f64(<32 x i1> undef, <32 x double> undef, <32 x double> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %67 = call <vscale x 1 x double> @llvm.vp.merge.nxv1f64(<vscale x 1 x i1> undef, <vscale x 1 x double> undef, <vscale x 1 x double> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %68 = call <vscale x 2 x double> @llvm.vp.merge.nxv2f64(<vscale x 2 x i1> undef, <vscale x 2 x double> undef, <vscale x 2 x double> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %69 = call <vscale x 4 x double> @llvm.vp.merge.nxv4f64(<vscale x 4 x i1> undef, <vscale x 4 x double> undef, <vscale x 4 x double> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %70 = call <vscale x 8 x double> @llvm.vp.merge.nxv8f64(<vscale x 8 x i1> undef, <vscale x 8 x double> undef, <vscale x 8 x double> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %71 = call <vscale x 16 x double> @llvm.vp.merge.nxv16f64(<vscale x 16 x i1> undef, <vscale x 16 x double> undef, <vscale x 16 x double> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %72 = call <vscale x 32 x double> @llvm.vp.merge.nxv32f64(<vscale x 32 x i1> undef, <vscale x 32 x double> undef, <vscale x 32 x double> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; TYPEBASED-LABEL: 'vp_merge'
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = call <1 x i8> @llvm.vp.merge.v1i8(<1 x i1> undef, <1 x i8> undef, <1 x i8> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %2 = call <2 x i8> @llvm.vp.merge.v2i8(<2 x i1> undef, <2 x i8> undef, <2 x i8> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %3 = call <4 x i8> @llvm.vp.merge.v4i8(<4 x i1> undef, <4 x i8> undef, <4 x i8> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %4 = call <8 x i8> @llvm.vp.merge.v8i8(<8 x i1> undef, <8 x i8> undef, <8 x i8> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %5 = call <16 x i8> @llvm.vp.merge.v16i8(<16 x i1> undef, <16 x i8> undef, <16 x i8> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %6 = call <32 x i8> @llvm.vp.merge.v32i8(<32 x i1> undef, <32 x i8> undef, <32 x i8> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %7 = call <vscale x 1 x i8> @llvm.vp.merge.nxv1i8(<vscale x 1 x i1> undef, <vscale x 1 x i8> undef, <vscale x 1 x i8> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %8 = call <vscale x 2 x i8> @llvm.vp.merge.nxv2i8(<vscale x 2 x i1> undef, <vscale x 2 x i8> undef, <vscale x 2 x i8> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %9 = call <vscale x 4 x i8> @llvm.vp.merge.nxv4i8(<vscale x 4 x i1> undef, <vscale x 4 x i8> undef, <vscale x 4 x i8> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %10 = call <vscale x 8 x i8> @llvm.vp.merge.nxv8i8(<vscale x 8 x i1> undef, <vscale x 8 x i8> undef, <vscale x 8 x i8> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %11 = call <vscale x 16 x i8> @llvm.vp.merge.nxv16i8(<vscale x 16 x i1> undef, <vscale x 16 x i8> undef, <vscale x 16 x i8> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %12 = call <vscale x 32 x i8> @llvm.vp.merge.nxv32i8(<vscale x 32 x i1> undef, <vscale x 32 x i8> undef, <vscale x 32 x i8> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %13 = call <1 x i16> @llvm.vp.merge.v1i16(<1 x i1> undef, <1 x i16> undef, <1 x i16> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %14 = call <2 x i16> @llvm.vp.merge.v2i16(<2 x i1> undef, <2 x i16> undef, <2 x i16> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %15 = call <4 x i16> @llvm.vp.merge.v4i16(<4 x i1> undef, <4 x i16> undef, <4 x i16> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %16 = call <8 x i16> @llvm.vp.merge.v8i16(<8 x i1> undef, <8 x i16> undef, <8 x i16> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %17 = call <16 x i16> @llvm.vp.merge.v16i16(<16 x i1> undef, <16 x i16> undef, <16 x i16> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %18 = call <32 x i16> @llvm.vp.merge.v32i16(<32 x i1> undef, <32 x i16> undef, <32 x i16> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %19 = call <vscale x 1 x i16> @llvm.vp.merge.nxv1i16(<vscale x 1 x i1> undef, <vscale x 1 x i16> undef, <vscale x 1 x i16> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %20 = call <vscale x 2 x i16> @llvm.vp.merge.nxv2i16(<vscale x 2 x i1> undef, <vscale x 2 x i16> undef, <vscale x 2 x i16> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %21 = call <vscale x 4 x i16> @llvm.vp.merge.nxv4i16(<vscal...
[truncated]