[llvm] [RISCV][CostModel] Correct the cost of some reductions (PR #118072)

Tue Dec 3 18:52:47 PST 2024

https://github.com/LiqinWeng updated https://github.com/llvm/llvm-project/pull/118072

>From 2a43729040bf71313d37e6de41062a23c3741814 Mon Sep 17 00:00:00 2001
From: LiqinWeng <liqin.weng at spacemit.com>
Date: Fri, 29 Nov 2024 17:15:50 +0800
Subject: [PATCH 1/4] [RISCV][CostModel] Correct the cost of some reductions

Reductions include: and/or/max/min
---
 .../Target/RISCV/RISCVTargetTransformInfo.cpp |  23 +-
 .../Analysis/CostModel/RISCV/cttz_elts.ll     |  16 +-
 .../Analysis/CostModel/RISCV/reduce-and.ll    | 128 ++++----
 .../Analysis/CostModel/RISCV/reduce-max.ll    | 256 ++++++++--------
 .../Analysis/CostModel/RISCV/reduce-min.ll    | 256 ++++++++--------
 .../Analysis/CostModel/RISCV/reduce-or.ll     | 128 ++++----
 .../CostModel/RISCV/reduce-scalable-fp.ll     | 144 ++++-----
 .../CostModel/RISCV/reduce-scalable-int.ll    | 288 +++++++++---------
 .../Analysis/CostModel/RISCV/vp-intrinsics.ll |  32 +-
 9 files changed, 639 insertions(+), 632 deletions(-)

diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
index 8f0ef69258b165..83994b1e445f75 100644
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
@@ -1470,27 +1470,27 @@ RISCVTTIImpl::getMinMaxReductionCost(Intrinsic::ID IID, VectorType *Ty,
     llvm_unreachable("Unsupported intrinsic");
   case Intrinsic::smax:
     SplitOp = RISCV::VMAX_VV;
-    Opcodes = {RISCV::VMV_S_X, RISCV::VREDMAX_VS, RISCV::VMV_X_S};
+    Opcodes = {RISCV::VREDMAX_VS, RISCV::VMV_X_S};
     break;
   case Intrinsic::smin:
     SplitOp = RISCV::VMIN_VV;
-    Opcodes = {RISCV::VMV_S_X, RISCV::VREDMIN_VS, RISCV::VMV_X_S};
+    Opcodes = {RISCV::VREDMIN_VS, RISCV::VMV_X_S};
     break;
   case Intrinsic::umax:
     SplitOp = RISCV::VMAXU_VV;
-    Opcodes = {RISCV::VMV_S_X, RISCV::VREDMAXU_VS, RISCV::VMV_X_S};
+    Opcodes = {RISCV::VREDMAXU_VS, RISCV::VMV_X_S};
     break;
   case Intrinsic::umin:
     SplitOp = RISCV::VMINU_VV;
-    Opcodes = {RISCV::VMV_S_X, RISCV::VREDMINU_VS, RISCV::VMV_X_S};
+    Opcodes = {RISCV::VREDMINU_VS, RISCV::VMV_X_S};
     break;
   case Intrinsic::maxnum:
     SplitOp = RISCV::VFMAX_VV;
-    Opcodes = {RISCV::VFMV_S_F, RISCV::VFREDMAX_VS, RISCV::VFMV_F_S};
+    Opcodes = {RISCV::VFREDMAX_VS, RISCV::VFMV_F_S};
     break;
   case Intrinsic::minnum:
     SplitOp = RISCV::VFMIN_VV;
-    Opcodes = {RISCV::VFMV_S_F, RISCV::VFREDMIN_VS, RISCV::VFMV_F_S};
+    Opcodes = {RISCV::VFREDMIN_VS, RISCV::VFMV_F_S};
     break;
   }
   // Add a cost for data larger than LMUL8
@@ -1534,6 +1534,13 @@ RISCVTTIImpl::getArithmeticReductionCost(unsigned Opcode, VectorType *Ty,
              getRISCVInstructionCost(Opcodes, LT.second, CostKind) +
              getCmpSelInstrCost(Instruction::ICmp, ElementTy, ElementTy,
                                 CmpInst::ICMP_EQ, CostKind);
+    } else if (ISD == ISD::XOR) {
+      // Example sequences:
+      //   vsetvli a0, zero, e8, mf8, ta, ma
+      //   vcpop.m a0, v0
+      //   andi a0, a0, 1
+      Opcodes = {RISCV::VCPOP_M};
+      return LT.first + getRISCVInstructionCost(Opcodes, LT.second, CostKind);
     } else {
       // Example sequences:
       //   vsetvli a0, zero, e8, mf8, ta, ma
@@ -1556,7 +1563,7 @@ RISCVTTIImpl::getArithmeticReductionCost(unsigned Opcode, VectorType *Ty,
     break;
   case ISD::OR:
     SplitOp = RISCV::VOR_VV;
-    Opcodes = {RISCV::VMV_S_X, RISCV::VREDOR_VS, RISCV::VMV_X_S};
+    Opcodes = {RISCV::VREDOR_VS, RISCV::VMV_X_S};
     break;
   case ISD::XOR:
     SplitOp = RISCV::VXOR_VV;
@@ -1564,7 +1571,7 @@ RISCVTTIImpl::getArithmeticReductionCost(unsigned Opcode, VectorType *Ty,
     break;
   case ISD::AND:
     SplitOp = RISCV::VAND_VV;
-    Opcodes = {RISCV::VMV_S_X, RISCV::VREDAND_VS, RISCV::VMV_X_S};
+    Opcodes = {RISCV::VREDAND_VS, RISCV::VMV_X_S};
     break;
   case ISD::FADD:
     // We can't promote f16/bf16 fadd reductions.
diff --git a/llvm/test/Analysis/CostModel/RISCV/cttz_elts.ll b/llvm/test/Analysis/CostModel/RISCV/cttz_elts.ll
index fce9f71d37634d..6d3dfc30613880 100644
--- a/llvm/test/Analysis/CostModel/RISCV/cttz_elts.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/cttz_elts.ll
@@ -9,28 +9,28 @@ define void @foo_no_vscale_range() {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i64.nxv16i1.zip = call i64 @llvm.experimental.cttz.elts.i64.nxv16i1(<vscale x 16 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i64.nxv32i1.zip = call i64 @llvm.experimental.cttz.elts.i64.nxv32i1(<vscale x 32 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i64.nxv64i1.zip = call i64 @llvm.experimental.cttz.elts.i64.nxv64i1(<vscale x 64 x i1> undef, i1 true)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 662 for instruction: %res.i64.nxv128i1.zip = call i64 @llvm.experimental.cttz.elts.i64.nxv128i1(<vscale x 128 x i1> undef, i1 true)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 661 for instruction: %res.i64.nxv128i1.zip = call i64 @llvm.experimental.cttz.elts.i64.nxv128i1(<vscale x 128 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i32.nxv2i1.zip = call i32 @llvm.experimental.cttz.elts.i32.nxv2i1(<vscale x 2 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i32.nxv4i1.zip = call i32 @llvm.experimental.cttz.elts.i32.nxv4i1(<vscale x 4 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i32.nxv8i1.zip = call i32 @llvm.experimental.cttz.elts.i32.nxv8i1(<vscale x 8 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i32.nxv16i1.zip = call i32 @llvm.experimental.cttz.elts.i32.nxv16i1(<vscale x 16 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i32.nxv32i1.zip = call i32 @llvm.experimental.cttz.elts.i32.nxv32i1(<vscale x 32 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i32.nxv64i1.zip = call i32 @llvm.experimental.cttz.elts.i32.nxv64i1(<vscale x 64 x i1> undef, i1 true)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 335 for instruction: %res.i32.nxv128i1.zip = call i32 @llvm.experimental.cttz.elts.i32.nxv128i1(<vscale x 128 x i1> undef, i1 true)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 334 for instruction: %res.i32.nxv128i1.zip = call i32 @llvm.experimental.cttz.elts.i32.nxv128i1(<vscale x 128 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i64.nxv2i1.nzip = call i64 @llvm.experimental.cttz.elts.i64.nxv2i1(<vscale x 2 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i64.nxv4i1.nzip = call i64 @llvm.experimental.cttz.elts.i64.nxv4i1(<vscale x 4 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i64.nxv8i1.nzip = call i64 @llvm.experimental.cttz.elts.i64.nxv8i1(<vscale x 8 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i64.nxv16i1.nzip = call i64 @llvm.experimental.cttz.elts.i64.nxv16i1(<vscale x 16 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i64.nxv32i1.nzip = call i64 @llvm.experimental.cttz.elts.i64.nxv32i1(<vscale x 32 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i64.nxv64i1.nzip = call i64 @llvm.experimental.cttz.elts.i64.nxv64i1(<vscale x 64 x i1> undef, i1 false)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 662 for instruction: %res.i64.nxv128i1.nzip = call i64 @llvm.experimental.cttz.elts.i64.nxv128i1(<vscale x 128 x i1> undef, i1 false)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 661 for instruction: %res.i64.nxv128i1.nzip = call i64 @llvm.experimental.cttz.elts.i64.nxv128i1(<vscale x 128 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i32.nxv2i1.nzip = call i32 @llvm.experimental.cttz.elts.i32.nxv2i1(<vscale x 2 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i32.nxv4i1.nzip = call i32 @llvm.experimental.cttz.elts.i32.nxv4i1(<vscale x 4 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i32.nxv8i1.nzip = call i32 @llvm.experimental.cttz.elts.i32.nxv8i1(<vscale x 8 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i32.nxv16i1.nzip = call i32 @llvm.experimental.cttz.elts.i32.nxv16i1(<vscale x 16 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i32.nxv32i1.nzip = call i32 @llvm.experimental.cttz.elts.i32.nxv32i1(<vscale x 32 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i32.nxv64i1.nzip = call i32 @llvm.experimental.cttz.elts.i32.nxv64i1(<vscale x 64 x i1> undef, i1 false)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 335 for instruction: %res.i32.nxv128i1.nzip = call i32 @llvm.experimental.cttz.elts.i32.nxv128i1(<vscale x 128 x i1> undef, i1 false)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 334 for instruction: %res.i32.nxv128i1.nzip = call i32 @llvm.experimental.cttz.elts.i32.nxv128i1(<vscale x 128 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
   %res.i64.nxv2i1.zip = call i64 @llvm.experimental.cttz.elts.i64.nxv2i1(<vscale x 2 x i1> undef, i1 true)
@@ -75,28 +75,28 @@ define void @foo_vscale_range_2_16() vscale_range(2,16) {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i64.nxv16i1.zip = call i64 @llvm.experimental.cttz.elts.i64.nxv16i1(<vscale x 16 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i64.nxv32i1.zip = call i64 @llvm.experimental.cttz.elts.i64.nxv32i1(<vscale x 32 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i64.nxv64i1.zip = call i64 @llvm.experimental.cttz.elts.i64.nxv64i1(<vscale x 64 x i1> undef, i1 true)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 172 for instruction: %res.i64.nxv128i1.zip = call i64 @llvm.experimental.cttz.elts.i64.nxv128i1(<vscale x 128 x i1> undef, i1 true)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 171 for instruction: %res.i64.nxv128i1.zip = call i64 @llvm.experimental.cttz.elts.i64.nxv128i1(<vscale x 128 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i32.nxv2i1.zip = call i32 @llvm.experimental.cttz.elts.i32.nxv2i1(<vscale x 2 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i32.nxv4i1.zip = call i32 @llvm.experimental.cttz.elts.i32.nxv4i1(<vscale x 4 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i32.nxv8i1.zip = call i32 @llvm.experimental.cttz.elts.i32.nxv8i1(<vscale x 8 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i32.nxv16i1.zip = call i32 @llvm.experimental.cttz.elts.i32.nxv16i1(<vscale x 16 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i32.nxv32i1.zip = call i32 @llvm.experimental.cttz.elts.i32.nxv32i1(<vscale x 32 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res.i32.nxv64i1.zip = call i32 @llvm.experimental.cttz.elts.i32.nxv64i1(<vscale x 64 x i1> undef, i1 true)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 172 for instruction: %res.i32.nxv128i1.zip = call i32 @llvm.experimental.cttz.elts.i32.nxv128i1(<vscale x 128 x i1> undef, i1 true)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 171 for instruction: %res.i32.nxv128i1.zip = call i32 @llvm.experimental.cttz.elts.i32.nxv128i1(<vscale x 128 x i1> undef, i1 true)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i64.nxv2i1.nzip = call i64 @llvm.experimental.cttz.elts.i64.nxv2i1(<vscale x 2 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i64.nxv4i1.nzip = call i64 @llvm.experimental.cttz.elts.i64.nxv4i1(<vscale x 4 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i64.nxv8i1.nzip = call i64 @llvm.experimental.cttz.elts.i64.nxv8i1(<vscale x 8 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i64.nxv16i1.nzip = call i64 @llvm.experimental.cttz.elts.i64.nxv16i1(<vscale x 16 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i64.nxv32i1.nzip = call i64 @llvm.experimental.cttz.elts.i64.nxv32i1(<vscale x 32 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i64.nxv64i1.nzip = call i64 @llvm.experimental.cttz.elts.i64.nxv64i1(<vscale x 64 x i1> undef, i1 false)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 172 for instruction: %res.i64.nxv128i1.nzip = call i64 @llvm.experimental.cttz.elts.i64.nxv128i1(<vscale x 128 x i1> undef, i1 false)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 171 for instruction: %res.i64.nxv128i1.nzip = call i64 @llvm.experimental.cttz.elts.i64.nxv128i1(<vscale x 128 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i32.nxv2i1.nzip = call i32 @llvm.experimental.cttz.elts.i32.nxv2i1(<vscale x 2 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i32.nxv4i1.nzip = call i32 @llvm.experimental.cttz.elts.i32.nxv4i1(<vscale x 4 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i32.nxv8i1.nzip = call i32 @llvm.experimental.cttz.elts.i32.nxv8i1(<vscale x 8 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i32.nxv16i1.nzip = call i32 @llvm.experimental.cttz.elts.i32.nxv16i1(<vscale x 16 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i32.nxv32i1.nzip = call i32 @llvm.experimental.cttz.elts.i32.nxv32i1(<vscale x 32 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res.i32.nxv64i1.nzip = call i32 @llvm.experimental.cttz.elts.i32.nxv64i1(<vscale x 64 x i1> undef, i1 false)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 172 for instruction: %res.i32.nxv128i1.nzip = call i32 @llvm.experimental.cttz.elts.i32.nxv128i1(<vscale x 128 x i1> undef, i1 false)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 171 for instruction: %res.i32.nxv128i1.nzip = call i32 @llvm.experimental.cttz.elts.i32.nxv128i1(<vscale x 128 x i1> undef, i1 false)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
   %res.i64.nxv2i1.zip = call i64 @llvm.experimental.cttz.elts.i64.nxv2i1(<vscale x 2 x i1> undef, i1 true)
diff --git a/llvm/test/Analysis/CostModel/RISCV/reduce-and.ll b/llvm/test/Analysis/CostModel/RISCV/reduce-and.ll
index a8eb4e9a280dd0..e4f1cf8ff418a8 100644
--- a/llvm/test/Analysis/CostModel/RISCV/reduce-and.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/reduce-and.ll
@@ -49,25 +49,25 @@ define i32 @reduce_i1(i32 %arg) {
 
 define i32 @reduce_i8(i32 %arg) {
 ; CHECK-LABEL: 'reduce_i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i8 @llvm.vector.reduce.and.v1i8(<1 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i8 @llvm.vector.reduce.and.v2i8(<2 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i8 @llvm.vector.reduce.and.v4i8(<4 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i8 @llvm.vector.reduce.and.v8i8(<8 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i8 @llvm.vector.reduce.and.v16i8(<16 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i8 @llvm.vector.reduce.and.v32i8(<32 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64 = call i8 @llvm.vector.reduce.and.v64i8(<64 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V128 = call i8 @llvm.vector.reduce.and.v128i8(<128 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i8 @llvm.vector.reduce.and.v1i8(<1 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i8 @llvm.vector.reduce.and.v2i8(<2 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i8 @llvm.vector.reduce.and.v4i8(<4 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i8 @llvm.vector.reduce.and.v8i8(<8 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i8 @llvm.vector.reduce.and.v16i8(<16 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i8 @llvm.vector.reduce.and.v32i8(<32 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64 = call i8 @llvm.vector.reduce.and.v64i8(<64 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128 = call i8 @llvm.vector.reduce.and.v128i8(<128 x i8> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i8 @llvm.vector.reduce.and.v1i8(<1 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i8 @llvm.vector.reduce.and.v2i8(<2 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i8 @llvm.vector.reduce.and.v4i8(<4 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i8 @llvm.vector.reduce.and.v8i8(<8 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i8 @llvm.vector.reduce.and.v16i8(<16 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i8 @llvm.vector.reduce.and.v32i8(<32 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i8 @llvm.vector.reduce.and.v64i8(<64 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i8 @llvm.vector.reduce.and.v128i8(<128 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i8 @llvm.vector.reduce.and.v1i8(<1 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i8 @llvm.vector.reduce.and.v2i8(<2 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i8 @llvm.vector.reduce.and.v4i8(<4 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i8 @llvm.vector.reduce.and.v8i8(<8 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i8 @llvm.vector.reduce.and.v16i8(<16 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i8 @llvm.vector.reduce.and.v32i8(<32 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i8 @llvm.vector.reduce.and.v64i8(<64 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = call i8 @llvm.vector.reduce.and.v128i8(<128 x i8> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i8 @llvm.vector.reduce.and.v1i8(<1 x i8> undef)
@@ -83,25 +83,25 @@ define i32 @reduce_i8(i32 %arg) {
 
 define i32 @reduce_i16(i32 %arg) {
 ; CHECK-LABEL: 'reduce_i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i16 @llvm.vector.reduce.and.v1i16(<1 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i16 @llvm.vector.reduce.and.v2i16(<2 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i16 @llvm.vector.reduce.and.v4i16(<4 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i16 @llvm.vector.reduce.and.v8i16(<8 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i16 @llvm.vector.reduce.and.v16i16(<16 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i16 @llvm.vector.reduce.and.v32i16(<32 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64 = call i16 @llvm.vector.reduce.and.v64i16(<64 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V128 = call i16 @llvm.vector.reduce.and.v128i16(<128 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i16 @llvm.vector.reduce.and.v1i16(<1 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i16 @llvm.vector.reduce.and.v2i16(<2 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i16 @llvm.vector.reduce.and.v4i16(<4 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i16 @llvm.vector.reduce.and.v8i16(<8 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i16 @llvm.vector.reduce.and.v16i16(<16 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i16 @llvm.vector.reduce.and.v32i16(<32 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64 = call i16 @llvm.vector.reduce.and.v64i16(<64 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V128 = call i16 @llvm.vector.reduce.and.v128i16(<128 x i16> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i16 @llvm.vector.reduce.and.v1i16(<1 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i16 @llvm.vector.reduce.and.v2i16(<2 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i16 @llvm.vector.reduce.and.v4i16(<4 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i16 @llvm.vector.reduce.and.v8i16(<8 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i16 @llvm.vector.reduce.and.v16i16(<16 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i16 @llvm.vector.reduce.and.v32i16(<32 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i16 @llvm.vector.reduce.and.v64i16(<64 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128 = call i16 @llvm.vector.reduce.and.v128i16(<128 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i16 @llvm.vector.reduce.and.v1i16(<1 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i16 @llvm.vector.reduce.and.v2i16(<2 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i16 @llvm.vector.reduce.and.v4i16(<4 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i16 @llvm.vector.reduce.and.v8i16(<8 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i16 @llvm.vector.reduce.and.v16i16(<16 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i16 @llvm.vector.reduce.and.v32i16(<32 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i16 @llvm.vector.reduce.and.v64i16(<64 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i16 @llvm.vector.reduce.and.v128i16(<128 x i16> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i16 @llvm.vector.reduce.and.v1i16(<1 x i16> undef)
@@ -117,25 +117,25 @@ define i32 @reduce_i16(i32 %arg) {
 
 define i32 @reduce_i32(i32 %arg) {
 ; CHECK-LABEL: 'reduce_i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i32 @llvm.vector.reduce.and.v1i32(<1 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i32 @llvm.vector.reduce.and.v2i32(<2 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i32 @llvm.vector.reduce.and.v4i32(<4 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i32 @llvm.vector.reduce.and.v8i32(<8 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i32 @llvm.vector.reduce.and.v16i32(<16 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i32 @llvm.vector.reduce.and.v32i32(<32 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64 = call i32 @llvm.vector.reduce.and.v64i32(<64 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 31 for instruction: %V128 = call i32 @llvm.vector.reduce.and.v128i32(<128 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i32 @llvm.vector.reduce.and.v1i32(<1 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i32 @llvm.vector.reduce.and.v2i32(<2 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i32 @llvm.vector.reduce.and.v4i32(<4 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i32 @llvm.vector.reduce.and.v8i32(<8 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i32 @llvm.vector.reduce.and.v16i32(<16 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i32 @llvm.vector.reduce.and.v32i32(<32 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V64 = call i32 @llvm.vector.reduce.and.v64i32(<64 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V128 = call i32 @llvm.vector.reduce.and.v128i32(<128 x i32> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i32 @llvm.vector.reduce.and.v1i32(<1 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i32 @llvm.vector.reduce.and.v2i32(<2 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i32 @llvm.vector.reduce.and.v4i32(<4 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i32 @llvm.vector.reduce.and.v8i32(<8 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i32 @llvm.vector.reduce.and.v16i32(<16 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i32 @llvm.vector.reduce.and.v32i32(<32 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64 = call i32 @llvm.vector.reduce.and.v64i32(<64 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V128 = call i32 @llvm.vector.reduce.and.v128i32(<128 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i32 @llvm.vector.reduce.and.v1i32(<1 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i32 @llvm.vector.reduce.and.v2i32(<2 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i32 @llvm.vector.reduce.and.v4i32(<4 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i32 @llvm.vector.reduce.and.v8i32(<8 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i32 @llvm.vector.reduce.and.v16i32(<16 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i32 @llvm.vector.reduce.and.v32i32(<32 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i32 @llvm.vector.reduce.and.v64i32(<64 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V128 = call i32 @llvm.vector.reduce.and.v128i32(<128 x i32> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i32 @llvm.vector.reduce.and.v1i32(<1 x i32> undef)
@@ -151,25 +151,25 @@ define i32 @reduce_i32(i32 %arg) {
 
 define i32 @reduce_i64(i32 %arg) {
 ; CHECK-LABEL: 'reduce_i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i64 @llvm.vector.reduce.and.v1i64(<1 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i64 @llvm.vector.reduce.and.v2i64(<2 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i64 @llvm.vector.reduce.and.v4i64(<4 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i64 @llvm.vector.reduce.and.v8i64(<8 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i64 @llvm.vector.reduce.and.v16i64(<16 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V32 = call i64 @llvm.vector.reduce.and.v32i64(<32 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V64 = call i64 @llvm.vector.reduce.and.v64i64(<64 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 62 for instruction: %V128 = call i64 @llvm.vector.reduce.and.v128i64(<128 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i64 @llvm.vector.reduce.and.v1i64(<1 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i64 @llvm.vector.reduce.and.v2i64(<2 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i64 @llvm.vector.reduce.and.v4i64(<4 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i64 @llvm.vector.reduce.and.v8i64(<8 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i64 @llvm.vector.reduce.and.v16i64(<16 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V32 = call i64 @llvm.vector.reduce.and.v32i64(<32 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V64 = call i64 @llvm.vector.reduce.and.v64i64(<64 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 61 for instruction: %V128 = call i64 @llvm.vector.reduce.and.v128i64(<128 x i64> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i64 @llvm.vector.reduce.and.v1i64(<1 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i64 @llvm.vector.reduce.and.v2i64(<2 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i64 @llvm.vector.reduce.and.v4i64(<4 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i64 @llvm.vector.reduce.and.v8i64(<8 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i64 @llvm.vector.reduce.and.v16i64(<16 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32 = call i64 @llvm.vector.reduce.and.v32i64(<32 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V64 = call i64 @llvm.vector.reduce.and.v64i64(<64 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V128 = call i64 @llvm.vector.reduce.and.v128i64(<128 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i64 @llvm.vector.reduce.and.v1i64(<1 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i64 @llvm.vector.reduce.and.v2i64(<2 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i64 @llvm.vector.reduce.and.v4i64(<4 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i64 @llvm.vector.reduce.and.v8i64(<8 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i64 @llvm.vector.reduce.and.v16i64(<16 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i64 @llvm.vector.reduce.and.v32i64(<32 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V64 = call i64 @llvm.vector.reduce.and.v64i64(<64 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V128 = call i64 @llvm.vector.reduce.and.v128i64(<128 x i64> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i64 @llvm.vector.reduce.and.v1i64(<1 x i64> undef)
diff --git a/llvm/test/Analysis/CostModel/RISCV/reduce-max.ll b/llvm/test/Analysis/CostModel/RISCV/reduce-max.ll
index c5d677e574c13c..f11e9f2b5ae837 100644
--- a/llvm/test/Analysis/CostModel/RISCV/reduce-max.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/reduce-max.ll
@@ -40,25 +40,25 @@ define i32 @reduce_umin_i1(i32 %arg) {
 
 define i32 @reduce_umax_i8(i32 %arg) {
 ; CHECK-LABEL: 'reduce_umax_i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i8 @llvm.vector.reduce.umax.v1i8(<1 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i8 @llvm.vector.reduce.umax.v2i8(<2 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i8 @llvm.vector.reduce.umax.v4i8(<4 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i8 @llvm.vector.reduce.umax.v8i8(<8 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i8 @llvm.vector.reduce.umax.v16i8(<16 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i8 @llvm.vector.reduce.umax.v32i8(<32 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64 = call i8 @llvm.vector.reduce.umax.v64i8(<64 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V128 = call i8 @llvm.vector.reduce.umax.v128i8(<128 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i8 @llvm.vector.reduce.umax.v1i8(<1 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i8 @llvm.vector.reduce.umax.v2i8(<2 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i8 @llvm.vector.reduce.umax.v4i8(<4 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i8 @llvm.vector.reduce.umax.v8i8(<8 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i8 @llvm.vector.reduce.umax.v16i8(<16 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i8 @llvm.vector.reduce.umax.v32i8(<32 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64 = call i8 @llvm.vector.reduce.umax.v64i8(<64 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128 = call i8 @llvm.vector.reduce.umax.v128i8(<128 x i8> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_umax_i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i8 @llvm.vector.reduce.umax.v1i8(<1 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i8 @llvm.vector.reduce.umax.v2i8(<2 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i8 @llvm.vector.reduce.umax.v4i8(<4 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i8 @llvm.vector.reduce.umax.v8i8(<8 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i8 @llvm.vector.reduce.umax.v16i8(<16 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i8 @llvm.vector.reduce.umax.v32i8(<32 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i8 @llvm.vector.reduce.umax.v64i8(<64 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i8 @llvm.vector.reduce.umax.v128i8(<128 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i8 @llvm.vector.reduce.umax.v1i8(<1 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i8 @llvm.vector.reduce.umax.v2i8(<2 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i8 @llvm.vector.reduce.umax.v4i8(<4 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i8 @llvm.vector.reduce.umax.v8i8(<8 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i8 @llvm.vector.reduce.umax.v16i8(<16 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i8 @llvm.vector.reduce.umax.v32i8(<32 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i8 @llvm.vector.reduce.umax.v64i8(<64 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = call i8 @llvm.vector.reduce.umax.v128i8(<128 x i8> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i8 @llvm.vector.reduce.umax.v1i8(<1 x i8> undef)
@@ -74,25 +74,25 @@ define i32 @reduce_umax_i8(i32 %arg) {
 
 define i32 @reduce_umax_i16(i32 %arg) {
 ; CHECK-LABEL: 'reduce_umax_i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i16 @llvm.vector.reduce.umax.v1i16(<1 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i16 @llvm.vector.reduce.umax.v2i16(<2 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i16 @llvm.vector.reduce.umax.v4i16(<4 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i16 @llvm.vector.reduce.umax.v8i16(<8 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i16 @llvm.vector.reduce.umax.v16i16(<16 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i16 @llvm.vector.reduce.umax.v32i16(<32 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64 = call i16 @llvm.vector.reduce.umax.v64i16(<64 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V128 = call i16 @llvm.vector.reduce.umax.v128i16(<128 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i16 @llvm.vector.reduce.umax.v1i16(<1 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i16 @llvm.vector.reduce.umax.v2i16(<2 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i16 @llvm.vector.reduce.umax.v4i16(<4 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i16 @llvm.vector.reduce.umax.v8i16(<8 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i16 @llvm.vector.reduce.umax.v16i16(<16 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i16 @llvm.vector.reduce.umax.v32i16(<32 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64 = call i16 @llvm.vector.reduce.umax.v64i16(<64 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V128 = call i16 @llvm.vector.reduce.umax.v128i16(<128 x i16> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_umax_i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i16 @llvm.vector.reduce.umax.v1i16(<1 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i16 @llvm.vector.reduce.umax.v2i16(<2 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i16 @llvm.vector.reduce.umax.v4i16(<4 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i16 @llvm.vector.reduce.umax.v8i16(<8 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i16 @llvm.vector.reduce.umax.v16i16(<16 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i16 @llvm.vector.reduce.umax.v32i16(<32 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i16 @llvm.vector.reduce.umax.v64i16(<64 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128 = call i16 @llvm.vector.reduce.umax.v128i16(<128 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i16 @llvm.vector.reduce.umax.v1i16(<1 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i16 @llvm.vector.reduce.umax.v2i16(<2 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i16 @llvm.vector.reduce.umax.v4i16(<4 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i16 @llvm.vector.reduce.umax.v8i16(<8 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i16 @llvm.vector.reduce.umax.v16i16(<16 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i16 @llvm.vector.reduce.umax.v32i16(<32 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i16 @llvm.vector.reduce.umax.v64i16(<64 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i16 @llvm.vector.reduce.umax.v128i16(<128 x i16> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i16 @llvm.vector.reduce.umax.v1i16(<1 x i16> undef)
@@ -108,25 +108,25 @@ define i32 @reduce_umax_i16(i32 %arg) {
 
 define i32 @reduce_umax_i32(i32 %arg) {
 ; CHECK-LABEL: 'reduce_umax_i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i32 @llvm.vector.reduce.umax.v1i32(<1 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i32 @llvm.vector.reduce.umax.v2i32(<2 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i32 @llvm.vector.reduce.umax.v4i32(<4 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i32 @llvm.vector.reduce.umax.v8i32(<8 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i32 @llvm.vector.reduce.umax.v16i32(<16 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i32 @llvm.vector.reduce.umax.v32i32(<32 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64 = call i32 @llvm.vector.reduce.umax.v64i32(<64 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 31 for instruction: %V128 = call i32 @llvm.vector.reduce.umax.v128i32(<128 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i32 @llvm.vector.reduce.umax.v1i32(<1 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i32 @llvm.vector.reduce.umax.v2i32(<2 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i32 @llvm.vector.reduce.umax.v4i32(<4 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i32 @llvm.vector.reduce.umax.v8i32(<8 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i32 @llvm.vector.reduce.umax.v16i32(<16 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i32 @llvm.vector.reduce.umax.v32i32(<32 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V64 = call i32 @llvm.vector.reduce.umax.v64i32(<64 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V128 = call i32 @llvm.vector.reduce.umax.v128i32(<128 x i32> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_umax_i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i32 @llvm.vector.reduce.umax.v1i32(<1 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i32 @llvm.vector.reduce.umax.v2i32(<2 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i32 @llvm.vector.reduce.umax.v4i32(<4 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i32 @llvm.vector.reduce.umax.v8i32(<8 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i32 @llvm.vector.reduce.umax.v16i32(<16 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i32 @llvm.vector.reduce.umax.v32i32(<32 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64 = call i32 @llvm.vector.reduce.umax.v64i32(<64 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V128 = call i32 @llvm.vector.reduce.umax.v128i32(<128 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i32 @llvm.vector.reduce.umax.v1i32(<1 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i32 @llvm.vector.reduce.umax.v2i32(<2 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i32 @llvm.vector.reduce.umax.v4i32(<4 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i32 @llvm.vector.reduce.umax.v8i32(<8 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i32 @llvm.vector.reduce.umax.v16i32(<16 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i32 @llvm.vector.reduce.umax.v32i32(<32 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i32 @llvm.vector.reduce.umax.v64i32(<64 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V128 = call i32 @llvm.vector.reduce.umax.v128i32(<128 x i32> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i32 @llvm.vector.reduce.umax.v1i32(<1 x i32> undef)
@@ -142,25 +142,25 @@ define i32 @reduce_umax_i32(i32 %arg) {
 
 define i32 @reduce_umax_i64(i32 %arg) {
 ; CHECK-LABEL: 'reduce_umax_i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i64 @llvm.vector.reduce.umax.v1i64(<1 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i64 @llvm.vector.reduce.umax.v2i64(<2 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i64 @llvm.vector.reduce.umax.v4i64(<4 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i64 @llvm.vector.reduce.umax.v8i64(<8 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i64 @llvm.vector.reduce.umax.v16i64(<16 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V32 = call i64 @llvm.vector.reduce.umax.v32i64(<32 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V64 = call i64 @llvm.vector.reduce.umax.v64i64(<64 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 62 for instruction: %V128 = call i64 @llvm.vector.reduce.umax.v128i64(<128 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i64 @llvm.vector.reduce.umax.v1i64(<1 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i64 @llvm.vector.reduce.umax.v2i64(<2 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i64 @llvm.vector.reduce.umax.v4i64(<4 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i64 @llvm.vector.reduce.umax.v8i64(<8 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i64 @llvm.vector.reduce.umax.v16i64(<16 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V32 = call i64 @llvm.vector.reduce.umax.v32i64(<32 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V64 = call i64 @llvm.vector.reduce.umax.v64i64(<64 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 61 for instruction: %V128 = call i64 @llvm.vector.reduce.umax.v128i64(<128 x i64> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_umax_i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i64 @llvm.vector.reduce.umax.v1i64(<1 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i64 @llvm.vector.reduce.umax.v2i64(<2 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i64 @llvm.vector.reduce.umax.v4i64(<4 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i64 @llvm.vector.reduce.umax.v8i64(<8 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i64 @llvm.vector.reduce.umax.v16i64(<16 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32 = call i64 @llvm.vector.reduce.umax.v32i64(<32 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V64 = call i64 @llvm.vector.reduce.umax.v64i64(<64 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V128 = call i64 @llvm.vector.reduce.umax.v128i64(<128 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i64 @llvm.vector.reduce.umax.v1i64(<1 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i64 @llvm.vector.reduce.umax.v2i64(<2 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i64 @llvm.vector.reduce.umax.v4i64(<4 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i64 @llvm.vector.reduce.umax.v8i64(<8 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i64 @llvm.vector.reduce.umax.v16i64(<16 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i64 @llvm.vector.reduce.umax.v32i64(<32 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V64 = call i64 @llvm.vector.reduce.umax.v64i64(<64 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V128 = call i64 @llvm.vector.reduce.umax.v128i64(<128 x i64> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i64 @llvm.vector.reduce.umax.v1i64(<1 x i64> undef)
@@ -210,25 +210,25 @@ define i32 @reduce_smin_i1(i32 %arg) {
 
 define i32 @reduce_smax_i8(i32 %arg) {
 ; CHECK-LABEL: 'reduce_smax_i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i8 @llvm.vector.reduce.smax.v1i8(<1 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i8 @llvm.vector.reduce.smax.v2i8(<2 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i8 @llvm.vector.reduce.smax.v4i8(<4 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i8 @llvm.vector.reduce.smax.v8i8(<8 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i8 @llvm.vector.reduce.smax.v16i8(<16 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i8 @llvm.vector.reduce.smax.v32i8(<32 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64 = call i8 @llvm.vector.reduce.smax.v64i8(<64 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V128 = call i8 @llvm.vector.reduce.smax.v128i8(<128 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i8 @llvm.vector.reduce.smax.v1i8(<1 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i8 @llvm.vector.reduce.smax.v2i8(<2 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i8 @llvm.vector.reduce.smax.v4i8(<4 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i8 @llvm.vector.reduce.smax.v8i8(<8 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i8 @llvm.vector.reduce.smax.v16i8(<16 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i8 @llvm.vector.reduce.smax.v32i8(<32 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64 = call i8 @llvm.vector.reduce.smax.v64i8(<64 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128 = call i8 @llvm.vector.reduce.smax.v128i8(<128 x i8> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_smax_i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i8 @llvm.vector.reduce.smax.v1i8(<1 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i8 @llvm.vector.reduce.smax.v2i8(<2 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i8 @llvm.vector.reduce.smax.v4i8(<4 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i8 @llvm.vector.reduce.smax.v8i8(<8 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i8 @llvm.vector.reduce.smax.v16i8(<16 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i8 @llvm.vector.reduce.smax.v32i8(<32 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i8 @llvm.vector.reduce.smax.v64i8(<64 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i8 @llvm.vector.reduce.smax.v128i8(<128 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i8 @llvm.vector.reduce.smax.v1i8(<1 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i8 @llvm.vector.reduce.smax.v2i8(<2 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i8 @llvm.vector.reduce.smax.v4i8(<4 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i8 @llvm.vector.reduce.smax.v8i8(<8 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i8 @llvm.vector.reduce.smax.v16i8(<16 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i8 @llvm.vector.reduce.smax.v32i8(<32 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i8 @llvm.vector.reduce.smax.v64i8(<64 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = call i8 @llvm.vector.reduce.smax.v128i8(<128 x i8> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i8 @llvm.vector.reduce.smax.v1i8(<1 x i8> undef)
@@ -244,25 +244,25 @@ define i32 @reduce_smax_i8(i32 %arg) {
 
 define i32 @reduce_smax_i16(i32 %arg) {
 ; CHECK-LABEL: 'reduce_smax_i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i16 @llvm.vector.reduce.smax.v1i16(<1 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i16 @llvm.vector.reduce.smax.v2i16(<2 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i16 @llvm.vector.reduce.smax.v4i16(<4 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i16 @llvm.vector.reduce.smax.v8i16(<8 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i16 @llvm.vector.reduce.smax.v16i16(<16 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i16 @llvm.vector.reduce.smax.v32i16(<32 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64 = call i16 @llvm.vector.reduce.smax.v64i16(<64 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V128 = call i16 @llvm.vector.reduce.smax.v128i16(<128 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i16 @llvm.vector.reduce.smax.v1i16(<1 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i16 @llvm.vector.reduce.smax.v2i16(<2 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i16 @llvm.vector.reduce.smax.v4i16(<4 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i16 @llvm.vector.reduce.smax.v8i16(<8 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i16 @llvm.vector.reduce.smax.v16i16(<16 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i16 @llvm.vector.reduce.smax.v32i16(<32 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64 = call i16 @llvm.vector.reduce.smax.v64i16(<64 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V128 = call i16 @llvm.vector.reduce.smax.v128i16(<128 x i16> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_smax_i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i16 @llvm.vector.reduce.smax.v1i16(<1 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i16 @llvm.vector.reduce.smax.v2i16(<2 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i16 @llvm.vector.reduce.smax.v4i16(<4 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i16 @llvm.vector.reduce.smax.v8i16(<8 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i16 @llvm.vector.reduce.smax.v16i16(<16 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i16 @llvm.vector.reduce.smax.v32i16(<32 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i16 @llvm.vector.reduce.smax.v64i16(<64 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128 = call i16 @llvm.vector.reduce.smax.v128i16(<128 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i16 @llvm.vector.reduce.smax.v1i16(<1 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i16 @llvm.vector.reduce.smax.v2i16(<2 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i16 @llvm.vector.reduce.smax.v4i16(<4 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i16 @llvm.vector.reduce.smax.v8i16(<8 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i16 @llvm.vector.reduce.smax.v16i16(<16 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i16 @llvm.vector.reduce.smax.v32i16(<32 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i16 @llvm.vector.reduce.smax.v64i16(<64 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i16 @llvm.vector.reduce.smax.v128i16(<128 x i16> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i16 @llvm.vector.reduce.smax.v1i16(<1 x i16> undef)
@@ -278,25 +278,25 @@ define i32 @reduce_smax_i16(i32 %arg) {
 
 define i32 @reduce_smax_i32(i32 %arg) {
 ; CHECK-LABEL: 'reduce_smax_i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i32 @llvm.vector.reduce.smax.v1i32(<1 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i32 @llvm.vector.reduce.smax.v2i32(<2 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i32 @llvm.vector.reduce.smax.v4i32(<4 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i32 @llvm.vector.reduce.smax.v8i32(<8 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i32 @llvm.vector.reduce.smax.v16i32(<16 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i32 @llvm.vector.reduce.smax.v32i32(<32 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64 = call i32 @llvm.vector.reduce.smax.v64i32(<64 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 31 for instruction: %V128 = call i32 @llvm.vector.reduce.smax.v128i32(<128 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i32 @llvm.vector.reduce.smax.v1i32(<1 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i32 @llvm.vector.reduce.smax.v2i32(<2 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i32 @llvm.vector.reduce.smax.v4i32(<4 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i32 @llvm.vector.reduce.smax.v8i32(<8 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i32 @llvm.vector.reduce.smax.v16i32(<16 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i32 @llvm.vector.reduce.smax.v32i32(<32 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V64 = call i32 @llvm.vector.reduce.smax.v64i32(<64 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V128 = call i32 @llvm.vector.reduce.smax.v128i32(<128 x i32> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_smax_i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i32 @llvm.vector.reduce.smax.v1i32(<1 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i32 @llvm.vector.reduce.smax.v2i32(<2 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i32 @llvm.vector.reduce.smax.v4i32(<4 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i32 @llvm.vector.reduce.smax.v8i32(<8 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i32 @llvm.vector.reduce.smax.v16i32(<16 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i32 @llvm.vector.reduce.smax.v32i32(<32 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64 = call i32 @llvm.vector.reduce.smax.v64i32(<64 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V128 = call i32 @llvm.vector.reduce.smax.v128i32(<128 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i32 @llvm.vector.reduce.smax.v1i32(<1 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i32 @llvm.vector.reduce.smax.v2i32(<2 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i32 @llvm.vector.reduce.smax.v4i32(<4 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i32 @llvm.vector.reduce.smax.v8i32(<8 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i32 @llvm.vector.reduce.smax.v16i32(<16 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i32 @llvm.vector.reduce.smax.v32i32(<32 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i32 @llvm.vector.reduce.smax.v64i32(<64 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V128 = call i32 @llvm.vector.reduce.smax.v128i32(<128 x i32> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i32 @llvm.vector.reduce.smax.v1i32(<1 x i32> undef)
@@ -312,25 +312,25 @@ define i32 @reduce_smax_i32(i32 %arg) {
 
 define i32 @reduce_smax_i64(i32 %arg) {
 ; CHECK-LABEL: 'reduce_smax_i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i64 @llvm.vector.reduce.smax.v1i64(<1 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i64 @llvm.vector.reduce.smax.v2i64(<2 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i64 @llvm.vector.reduce.smax.v4i64(<4 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i64 @llvm.vector.reduce.smax.v8i64(<8 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i64 @llvm.vector.reduce.smax.v16i64(<16 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V32 = call i64 @llvm.vector.reduce.smax.v32i64(<32 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V64 = call i64 @llvm.vector.reduce.smax.v64i64(<64 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 62 for instruction: %V128 = call i64 @llvm.vector.reduce.smax.v128i64(<128 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i64 @llvm.vector.reduce.smax.v1i64(<1 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i64 @llvm.vector.reduce.smax.v2i64(<2 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i64 @llvm.vector.reduce.smax.v4i64(<4 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i64 @llvm.vector.reduce.smax.v8i64(<8 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i64 @llvm.vector.reduce.smax.v16i64(<16 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V32 = call i64 @llvm.vector.reduce.smax.v32i64(<32 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V64 = call i64 @llvm.vector.reduce.smax.v64i64(<64 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 61 for instruction: %V128 = call i64 @llvm.vector.reduce.smax.v128i64(<128 x i64> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_smax_i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i64 @llvm.vector.reduce.smax.v1i64(<1 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i64 @llvm.vector.reduce.smax.v2i64(<2 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i64 @llvm.vector.reduce.smax.v4i64(<4 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i64 @llvm.vector.reduce.smax.v8i64(<8 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i64 @llvm.vector.reduce.smax.v16i64(<16 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32 = call i64 @llvm.vector.reduce.smax.v32i64(<32 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V64 = call i64 @llvm.vector.reduce.smax.v64i64(<64 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V128 = call i64 @llvm.vector.reduce.smax.v128i64(<128 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i64 @llvm.vector.reduce.smax.v1i64(<1 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i64 @llvm.vector.reduce.smax.v2i64(<2 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i64 @llvm.vector.reduce.smax.v4i64(<4 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i64 @llvm.vector.reduce.smax.v8i64(<8 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i64 @llvm.vector.reduce.smax.v16i64(<16 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i64 @llvm.vector.reduce.smax.v32i64(<32 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V64 = call i64 @llvm.vector.reduce.smax.v64i64(<64 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V128 = call i64 @llvm.vector.reduce.smax.v128i64(<128 x i64> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i64 @llvm.vector.reduce.smax.v1i64(<1 x i64> undef)
diff --git a/llvm/test/Analysis/CostModel/RISCV/reduce-min.ll b/llvm/test/Analysis/CostModel/RISCV/reduce-min.ll
index 89bff381384156..457fdbe46f73b5 100644
--- a/llvm/test/Analysis/CostModel/RISCV/reduce-min.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/reduce-min.ll
@@ -40,25 +40,25 @@ define i32 @reduce_umin_i1(i32 %arg) {
 
 define i32 @reduce_umin_i8(i32 %arg) {
 ; CHECK-LABEL: 'reduce_umin_i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i8 @llvm.vector.reduce.umin.v1i8(<1 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i8 @llvm.vector.reduce.umin.v2i8(<2 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i8 @llvm.vector.reduce.umin.v4i8(<4 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i8 @llvm.vector.reduce.umin.v8i8(<8 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i8 @llvm.vector.reduce.umin.v16i8(<16 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i8 @llvm.vector.reduce.umin.v32i8(<32 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64 = call i8 @llvm.vector.reduce.umin.v64i8(<64 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V128 = call i8 @llvm.vector.reduce.umin.v128i8(<128 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i8 @llvm.vector.reduce.umin.v1i8(<1 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i8 @llvm.vector.reduce.umin.v2i8(<2 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i8 @llvm.vector.reduce.umin.v4i8(<4 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i8 @llvm.vector.reduce.umin.v8i8(<8 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i8 @llvm.vector.reduce.umin.v16i8(<16 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i8 @llvm.vector.reduce.umin.v32i8(<32 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64 = call i8 @llvm.vector.reduce.umin.v64i8(<64 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128 = call i8 @llvm.vector.reduce.umin.v128i8(<128 x i8> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_umin_i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i8 @llvm.vector.reduce.umin.v1i8(<1 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i8 @llvm.vector.reduce.umin.v2i8(<2 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i8 @llvm.vector.reduce.umin.v4i8(<4 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i8 @llvm.vector.reduce.umin.v8i8(<8 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i8 @llvm.vector.reduce.umin.v16i8(<16 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i8 @llvm.vector.reduce.umin.v32i8(<32 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i8 @llvm.vector.reduce.umin.v64i8(<64 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i8 @llvm.vector.reduce.umin.v128i8(<128 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i8 @llvm.vector.reduce.umin.v1i8(<1 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i8 @llvm.vector.reduce.umin.v2i8(<2 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i8 @llvm.vector.reduce.umin.v4i8(<4 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i8 @llvm.vector.reduce.umin.v8i8(<8 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i8 @llvm.vector.reduce.umin.v16i8(<16 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i8 @llvm.vector.reduce.umin.v32i8(<32 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i8 @llvm.vector.reduce.umin.v64i8(<64 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = call i8 @llvm.vector.reduce.umin.v128i8(<128 x i8> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i8 @llvm.vector.reduce.umin.v1i8(<1 x i8> undef)
@@ -74,25 +74,25 @@ define i32 @reduce_umin_i8(i32 %arg) {
 
 define i32 @reduce_umin_i16(i32 %arg) {
 ; CHECK-LABEL: 'reduce_umin_i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i16 @llvm.vector.reduce.umin.v1i16(<1 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i16 @llvm.vector.reduce.umin.v2i16(<2 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i16 @llvm.vector.reduce.umin.v4i16(<4 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i16 @llvm.vector.reduce.umin.v8i16(<8 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i16 @llvm.vector.reduce.umin.v16i16(<16 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i16 @llvm.vector.reduce.umin.v32i16(<32 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64 = call i16 @llvm.vector.reduce.umin.v64i16(<64 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V128 = call i16 @llvm.vector.reduce.umin.v128i16(<128 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i16 @llvm.vector.reduce.umin.v1i16(<1 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i16 @llvm.vector.reduce.umin.v2i16(<2 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i16 @llvm.vector.reduce.umin.v4i16(<4 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i16 @llvm.vector.reduce.umin.v8i16(<8 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i16 @llvm.vector.reduce.umin.v16i16(<16 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i16 @llvm.vector.reduce.umin.v32i16(<32 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64 = call i16 @llvm.vector.reduce.umin.v64i16(<64 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V128 = call i16 @llvm.vector.reduce.umin.v128i16(<128 x i16> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_umin_i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i16 @llvm.vector.reduce.umin.v1i16(<1 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i16 @llvm.vector.reduce.umin.v2i16(<2 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i16 @llvm.vector.reduce.umin.v4i16(<4 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i16 @llvm.vector.reduce.umin.v8i16(<8 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i16 @llvm.vector.reduce.umin.v16i16(<16 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i16 @llvm.vector.reduce.umin.v32i16(<32 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i16 @llvm.vector.reduce.umin.v64i16(<64 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128 = call i16 @llvm.vector.reduce.umin.v128i16(<128 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i16 @llvm.vector.reduce.umin.v1i16(<1 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i16 @llvm.vector.reduce.umin.v2i16(<2 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i16 @llvm.vector.reduce.umin.v4i16(<4 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i16 @llvm.vector.reduce.umin.v8i16(<8 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i16 @llvm.vector.reduce.umin.v16i16(<16 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i16 @llvm.vector.reduce.umin.v32i16(<32 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i16 @llvm.vector.reduce.umin.v64i16(<64 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i16 @llvm.vector.reduce.umin.v128i16(<128 x i16> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i16 @llvm.vector.reduce.umin.v1i16(<1 x i16> undef)
@@ -108,25 +108,25 @@ define i32 @reduce_umin_i16(i32 %arg) {
 
 define i32 @reduce_umin_i32(i32 %arg) {
 ; CHECK-LABEL: 'reduce_umin_i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i32 @llvm.vector.reduce.umin.v1i32(<1 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i32 @llvm.vector.reduce.umin.v2i32(<2 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i32 @llvm.vector.reduce.umin.v4i32(<4 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i32 @llvm.vector.reduce.umin.v8i32(<8 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i32 @llvm.vector.reduce.umin.v16i32(<16 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i32 @llvm.vector.reduce.umin.v32i32(<32 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64 = call i32 @llvm.vector.reduce.umin.v64i32(<64 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 31 for instruction: %V128 = call i32 @llvm.vector.reduce.umin.v128i32(<128 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i32 @llvm.vector.reduce.umin.v1i32(<1 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i32 @llvm.vector.reduce.umin.v2i32(<2 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i32 @llvm.vector.reduce.umin.v4i32(<4 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i32 @llvm.vector.reduce.umin.v8i32(<8 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i32 @llvm.vector.reduce.umin.v16i32(<16 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i32 @llvm.vector.reduce.umin.v32i32(<32 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V64 = call i32 @llvm.vector.reduce.umin.v64i32(<64 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V128 = call i32 @llvm.vector.reduce.umin.v128i32(<128 x i32> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_umin_i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i32 @llvm.vector.reduce.umin.v1i32(<1 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i32 @llvm.vector.reduce.umin.v2i32(<2 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i32 @llvm.vector.reduce.umin.v4i32(<4 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i32 @llvm.vector.reduce.umin.v8i32(<8 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i32 @llvm.vector.reduce.umin.v16i32(<16 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i32 @llvm.vector.reduce.umin.v32i32(<32 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64 = call i32 @llvm.vector.reduce.umin.v64i32(<64 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V128 = call i32 @llvm.vector.reduce.umin.v128i32(<128 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i32 @llvm.vector.reduce.umin.v1i32(<1 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i32 @llvm.vector.reduce.umin.v2i32(<2 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i32 @llvm.vector.reduce.umin.v4i32(<4 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i32 @llvm.vector.reduce.umin.v8i32(<8 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i32 @llvm.vector.reduce.umin.v16i32(<16 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i32 @llvm.vector.reduce.umin.v32i32(<32 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i32 @llvm.vector.reduce.umin.v64i32(<64 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V128 = call i32 @llvm.vector.reduce.umin.v128i32(<128 x i32> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i32 @llvm.vector.reduce.umin.v1i32(<1 x i32> undef)
@@ -142,25 +142,25 @@ define i32 @reduce_umin_i32(i32 %arg) {
 
 define i32 @reduce_umin_i64(i32 %arg) {
 ; CHECK-LABEL: 'reduce_umin_i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i64 @llvm.vector.reduce.umin.v1i64(<1 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i64 @llvm.vector.reduce.umin.v2i64(<2 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i64 @llvm.vector.reduce.umin.v4i64(<4 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i64 @llvm.vector.reduce.umin.v8i64(<8 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i64 @llvm.vector.reduce.umin.v16i64(<16 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V32 = call i64 @llvm.vector.reduce.umin.v32i64(<32 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V64 = call i64 @llvm.vector.reduce.umin.v64i64(<64 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 62 for instruction: %V128 = call i64 @llvm.vector.reduce.umin.v128i64(<128 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i64 @llvm.vector.reduce.umin.v1i64(<1 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i64 @llvm.vector.reduce.umin.v2i64(<2 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i64 @llvm.vector.reduce.umin.v4i64(<4 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i64 @llvm.vector.reduce.umin.v8i64(<8 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i64 @llvm.vector.reduce.umin.v16i64(<16 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V32 = call i64 @llvm.vector.reduce.umin.v32i64(<32 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V64 = call i64 @llvm.vector.reduce.umin.v64i64(<64 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 61 for instruction: %V128 = call i64 @llvm.vector.reduce.umin.v128i64(<128 x i64> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_umin_i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i64 @llvm.vector.reduce.umin.v1i64(<1 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i64 @llvm.vector.reduce.umin.v2i64(<2 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i64 @llvm.vector.reduce.umin.v4i64(<4 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i64 @llvm.vector.reduce.umin.v8i64(<8 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i64 @llvm.vector.reduce.umin.v16i64(<16 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32 = call i64 @llvm.vector.reduce.umin.v32i64(<32 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V64 = call i64 @llvm.vector.reduce.umin.v64i64(<64 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V128 = call i64 @llvm.vector.reduce.umin.v128i64(<128 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i64 @llvm.vector.reduce.umin.v1i64(<1 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i64 @llvm.vector.reduce.umin.v2i64(<2 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i64 @llvm.vector.reduce.umin.v4i64(<4 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i64 @llvm.vector.reduce.umin.v8i64(<8 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i64 @llvm.vector.reduce.umin.v16i64(<16 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i64 @llvm.vector.reduce.umin.v32i64(<32 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V64 = call i64 @llvm.vector.reduce.umin.v64i64(<64 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V128 = call i64 @llvm.vector.reduce.umin.v128i64(<128 x i64> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i64 @llvm.vector.reduce.umin.v1i64(<1 x i64> undef)
@@ -210,25 +210,25 @@ define i32 @reduce_smin_i1(i32 %arg) {
 
 define i32 @reduce_smin_i8(i32 %arg) {
 ; CHECK-LABEL: 'reduce_smin_i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i8 @llvm.vector.reduce.smin.v1i8(<1 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i8 @llvm.vector.reduce.smin.v2i8(<2 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i8 @llvm.vector.reduce.smin.v4i8(<4 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i8 @llvm.vector.reduce.smin.v8i8(<8 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i8 @llvm.vector.reduce.smin.v16i8(<16 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i8 @llvm.vector.reduce.smin.v32i8(<32 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64 = call i8 @llvm.vector.reduce.smin.v64i8(<64 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V128 = call i8 @llvm.vector.reduce.smin.v128i8(<128 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i8 @llvm.vector.reduce.smin.v1i8(<1 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i8 @llvm.vector.reduce.smin.v2i8(<2 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i8 @llvm.vector.reduce.smin.v4i8(<4 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i8 @llvm.vector.reduce.smin.v8i8(<8 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i8 @llvm.vector.reduce.smin.v16i8(<16 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i8 @llvm.vector.reduce.smin.v32i8(<32 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64 = call i8 @llvm.vector.reduce.smin.v64i8(<64 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128 = call i8 @llvm.vector.reduce.smin.v128i8(<128 x i8> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_smin_i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i8 @llvm.vector.reduce.smin.v1i8(<1 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i8 @llvm.vector.reduce.smin.v2i8(<2 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i8 @llvm.vector.reduce.smin.v4i8(<4 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i8 @llvm.vector.reduce.smin.v8i8(<8 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i8 @llvm.vector.reduce.smin.v16i8(<16 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i8 @llvm.vector.reduce.smin.v32i8(<32 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i8 @llvm.vector.reduce.smin.v64i8(<64 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i8 @llvm.vector.reduce.smin.v128i8(<128 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i8 @llvm.vector.reduce.smin.v1i8(<1 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i8 @llvm.vector.reduce.smin.v2i8(<2 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i8 @llvm.vector.reduce.smin.v4i8(<4 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i8 @llvm.vector.reduce.smin.v8i8(<8 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i8 @llvm.vector.reduce.smin.v16i8(<16 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i8 @llvm.vector.reduce.smin.v32i8(<32 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i8 @llvm.vector.reduce.smin.v64i8(<64 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = call i8 @llvm.vector.reduce.smin.v128i8(<128 x i8> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i8 @llvm.vector.reduce.smin.v1i8(<1 x i8> undef)
@@ -244,25 +244,25 @@ define i32 @reduce_smin_i8(i32 %arg) {
 
 define i32 @reduce_smin_i16(i32 %arg) {
 ; CHECK-LABEL: 'reduce_smin_i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i16 @llvm.vector.reduce.smin.v1i16(<1 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i16 @llvm.vector.reduce.smin.v2i16(<2 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i16 @llvm.vector.reduce.smin.v4i16(<4 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i16 @llvm.vector.reduce.smin.v8i16(<8 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i16 @llvm.vector.reduce.smin.v16i16(<16 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i16 @llvm.vector.reduce.smin.v32i16(<32 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64 = call i16 @llvm.vector.reduce.smin.v64i16(<64 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V128 = call i16 @llvm.vector.reduce.smin.v128i16(<128 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i16 @llvm.vector.reduce.smin.v1i16(<1 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i16 @llvm.vector.reduce.smin.v2i16(<2 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i16 @llvm.vector.reduce.smin.v4i16(<4 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i16 @llvm.vector.reduce.smin.v8i16(<8 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i16 @llvm.vector.reduce.smin.v16i16(<16 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i16 @llvm.vector.reduce.smin.v32i16(<32 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64 = call i16 @llvm.vector.reduce.smin.v64i16(<64 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V128 = call i16 @llvm.vector.reduce.smin.v128i16(<128 x i16> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_smin_i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i16 @llvm.vector.reduce.smin.v1i16(<1 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i16 @llvm.vector.reduce.smin.v2i16(<2 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i16 @llvm.vector.reduce.smin.v4i16(<4 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i16 @llvm.vector.reduce.smin.v8i16(<8 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i16 @llvm.vector.reduce.smin.v16i16(<16 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i16 @llvm.vector.reduce.smin.v32i16(<32 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i16 @llvm.vector.reduce.smin.v64i16(<64 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128 = call i16 @llvm.vector.reduce.smin.v128i16(<128 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i16 @llvm.vector.reduce.smin.v1i16(<1 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i16 @llvm.vector.reduce.smin.v2i16(<2 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i16 @llvm.vector.reduce.smin.v4i16(<4 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i16 @llvm.vector.reduce.smin.v8i16(<8 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i16 @llvm.vector.reduce.smin.v16i16(<16 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i16 @llvm.vector.reduce.smin.v32i16(<32 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i16 @llvm.vector.reduce.smin.v64i16(<64 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i16 @llvm.vector.reduce.smin.v128i16(<128 x i16> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i16 @llvm.vector.reduce.smin.v1i16(<1 x i16> undef)
@@ -278,25 +278,25 @@ define i32 @reduce_smin_i16(i32 %arg) {
 
 define i32 @reduce_smin_i32(i32 %arg) {
 ; CHECK-LABEL: 'reduce_smin_i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i32 @llvm.vector.reduce.smin.v1i32(<1 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i32 @llvm.vector.reduce.smin.v2i32(<2 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i32 @llvm.vector.reduce.smin.v4i32(<4 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i32 @llvm.vector.reduce.smin.v8i32(<8 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i32 @llvm.vector.reduce.smin.v16i32(<16 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i32 @llvm.vector.reduce.smin.v32i32(<32 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64 = call i32 @llvm.vector.reduce.smin.v64i32(<64 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 31 for instruction: %V128 = call i32 @llvm.vector.reduce.smin.v128i32(<128 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i32 @llvm.vector.reduce.smin.v1i32(<1 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i32 @llvm.vector.reduce.smin.v2i32(<2 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i32 @llvm.vector.reduce.smin.v4i32(<4 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i32 @llvm.vector.reduce.smin.v8i32(<8 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i32 @llvm.vector.reduce.smin.v16i32(<16 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i32 @llvm.vector.reduce.smin.v32i32(<32 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V64 = call i32 @llvm.vector.reduce.smin.v64i32(<64 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V128 = call i32 @llvm.vector.reduce.smin.v128i32(<128 x i32> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_smin_i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i32 @llvm.vector.reduce.smin.v1i32(<1 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i32 @llvm.vector.reduce.smin.v2i32(<2 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i32 @llvm.vector.reduce.smin.v4i32(<4 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i32 @llvm.vector.reduce.smin.v8i32(<8 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i32 @llvm.vector.reduce.smin.v16i32(<16 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i32 @llvm.vector.reduce.smin.v32i32(<32 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64 = call i32 @llvm.vector.reduce.smin.v64i32(<64 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V128 = call i32 @llvm.vector.reduce.smin.v128i32(<128 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i32 @llvm.vector.reduce.smin.v1i32(<1 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i32 @llvm.vector.reduce.smin.v2i32(<2 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i32 @llvm.vector.reduce.smin.v4i32(<4 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i32 @llvm.vector.reduce.smin.v8i32(<8 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i32 @llvm.vector.reduce.smin.v16i32(<16 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i32 @llvm.vector.reduce.smin.v32i32(<32 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i32 @llvm.vector.reduce.smin.v64i32(<64 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V128 = call i32 @llvm.vector.reduce.smin.v128i32(<128 x i32> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i32 @llvm.vector.reduce.smin.v1i32(<1 x i32> undef)
@@ -312,25 +312,25 @@ define i32 @reduce_smin_i32(i32 %arg) {
 
 define i32 @reduce_smin_i64(i32 %arg) {
 ; CHECK-LABEL: 'reduce_smin_i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i64 @llvm.vector.reduce.smin.v1i64(<1 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i64 @llvm.vector.reduce.smin.v2i64(<2 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i64 @llvm.vector.reduce.smin.v4i64(<4 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i64 @llvm.vector.reduce.smin.v8i64(<8 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i64 @llvm.vector.reduce.smin.v16i64(<16 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V32 = call i64 @llvm.vector.reduce.smin.v32i64(<32 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V64 = call i64 @llvm.vector.reduce.smin.v64i64(<64 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 62 for instruction: %V128 = call i64 @llvm.vector.reduce.smin.v128i64(<128 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i64 @llvm.vector.reduce.smin.v1i64(<1 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i64 @llvm.vector.reduce.smin.v2i64(<2 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i64 @llvm.vector.reduce.smin.v4i64(<4 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i64 @llvm.vector.reduce.smin.v8i64(<8 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i64 @llvm.vector.reduce.smin.v16i64(<16 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V32 = call i64 @llvm.vector.reduce.smin.v32i64(<32 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V64 = call i64 @llvm.vector.reduce.smin.v64i64(<64 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 61 for instruction: %V128 = call i64 @llvm.vector.reduce.smin.v128i64(<128 x i64> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_smin_i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i64 @llvm.vector.reduce.smin.v1i64(<1 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i64 @llvm.vector.reduce.smin.v2i64(<2 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i64 @llvm.vector.reduce.smin.v4i64(<4 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i64 @llvm.vector.reduce.smin.v8i64(<8 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i64 @llvm.vector.reduce.smin.v16i64(<16 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32 = call i64 @llvm.vector.reduce.smin.v32i64(<32 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V64 = call i64 @llvm.vector.reduce.smin.v64i64(<64 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V128 = call i64 @llvm.vector.reduce.smin.v128i64(<128 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i64 @llvm.vector.reduce.smin.v1i64(<1 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i64 @llvm.vector.reduce.smin.v2i64(<2 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i64 @llvm.vector.reduce.smin.v4i64(<4 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i64 @llvm.vector.reduce.smin.v8i64(<8 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i64 @llvm.vector.reduce.smin.v16i64(<16 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i64 @llvm.vector.reduce.smin.v32i64(<32 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V64 = call i64 @llvm.vector.reduce.smin.v64i64(<64 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V128 = call i64 @llvm.vector.reduce.smin.v128i64(<128 x i64> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i64 @llvm.vector.reduce.smin.v1i64(<1 x i64> undef)
diff --git a/llvm/test/Analysis/CostModel/RISCV/reduce-or.ll b/llvm/test/Analysis/CostModel/RISCV/reduce-or.ll
index d75a95f3fadd97..b0e90253b2810d 100644
--- a/llvm/test/Analysis/CostModel/RISCV/reduce-or.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/reduce-or.ll
@@ -49,25 +49,25 @@ define i32 @reduce_i1(i32 %arg) {
 
 define i32 @reduce_i8(i32 %arg) {
 ; CHECK-LABEL: 'reduce_i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i8 @llvm.vector.reduce.or.v1i8(<1 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i8 @llvm.vector.reduce.or.v2i8(<2 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i8 @llvm.vector.reduce.or.v4i8(<4 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i8 @llvm.vector.reduce.or.v8i8(<8 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i8 @llvm.vector.reduce.or.v16i8(<16 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i8 @llvm.vector.reduce.or.v32i8(<32 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64 = call i8 @llvm.vector.reduce.or.v64i8(<64 x i8> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V128 = call i8 @llvm.vector.reduce.or.v128i8(<128 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i8 @llvm.vector.reduce.or.v1i8(<1 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i8 @llvm.vector.reduce.or.v2i8(<2 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i8 @llvm.vector.reduce.or.v4i8(<4 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i8 @llvm.vector.reduce.or.v8i8(<8 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i8 @llvm.vector.reduce.or.v16i8(<16 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i8 @llvm.vector.reduce.or.v32i8(<32 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64 = call i8 @llvm.vector.reduce.or.v64i8(<64 x i8> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128 = call i8 @llvm.vector.reduce.or.v128i8(<128 x i8> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i8 @llvm.vector.reduce.or.v1i8(<1 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i8 @llvm.vector.reduce.or.v2i8(<2 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i8 @llvm.vector.reduce.or.v4i8(<4 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i8 @llvm.vector.reduce.or.v8i8(<8 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i8 @llvm.vector.reduce.or.v16i8(<16 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i8 @llvm.vector.reduce.or.v32i8(<32 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i8 @llvm.vector.reduce.or.v64i8(<64 x i8> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i8 @llvm.vector.reduce.or.v128i8(<128 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i8 @llvm.vector.reduce.or.v1i8(<1 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i8 @llvm.vector.reduce.or.v2i8(<2 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i8 @llvm.vector.reduce.or.v4i8(<4 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i8 @llvm.vector.reduce.or.v8i8(<8 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i8 @llvm.vector.reduce.or.v16i8(<16 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i8 @llvm.vector.reduce.or.v32i8(<32 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i8 @llvm.vector.reduce.or.v64i8(<64 x i8> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = call i8 @llvm.vector.reduce.or.v128i8(<128 x i8> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i8 @llvm.vector.reduce.or.v1i8(<1 x i8> undef)
@@ -83,25 +83,25 @@ define i32 @reduce_i8(i32 %arg) {
 
 define i32 @reduce_i16(i32 %arg) {
 ; CHECK-LABEL: 'reduce_i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i16 @llvm.vector.reduce.or.v1i16(<1 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i16 @llvm.vector.reduce.or.v2i16(<2 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i16 @llvm.vector.reduce.or.v4i16(<4 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i16 @llvm.vector.reduce.or.v8i16(<8 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i16 @llvm.vector.reduce.or.v16i16(<16 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i16 @llvm.vector.reduce.or.v32i16(<32 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64 = call i16 @llvm.vector.reduce.or.v64i16(<64 x i16> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V128 = call i16 @llvm.vector.reduce.or.v128i16(<128 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i16 @llvm.vector.reduce.or.v1i16(<1 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i16 @llvm.vector.reduce.or.v2i16(<2 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i16 @llvm.vector.reduce.or.v4i16(<4 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i16 @llvm.vector.reduce.or.v8i16(<8 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i16 @llvm.vector.reduce.or.v16i16(<16 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i16 @llvm.vector.reduce.or.v32i16(<32 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64 = call i16 @llvm.vector.reduce.or.v64i16(<64 x i16> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V128 = call i16 @llvm.vector.reduce.or.v128i16(<128 x i16> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i16 @llvm.vector.reduce.or.v1i16(<1 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i16 @llvm.vector.reduce.or.v2i16(<2 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i16 @llvm.vector.reduce.or.v4i16(<4 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i16 @llvm.vector.reduce.or.v8i16(<8 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i16 @llvm.vector.reduce.or.v16i16(<16 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i16 @llvm.vector.reduce.or.v32i16(<32 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i16 @llvm.vector.reduce.or.v64i16(<64 x i16> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128 = call i16 @llvm.vector.reduce.or.v128i16(<128 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i16 @llvm.vector.reduce.or.v1i16(<1 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i16 @llvm.vector.reduce.or.v2i16(<2 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i16 @llvm.vector.reduce.or.v4i16(<4 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i16 @llvm.vector.reduce.or.v8i16(<8 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i16 @llvm.vector.reduce.or.v16i16(<16 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i16 @llvm.vector.reduce.or.v32i16(<32 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i16 @llvm.vector.reduce.or.v64i16(<64 x i16> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i16 @llvm.vector.reduce.or.v128i16(<128 x i16> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i16 @llvm.vector.reduce.or.v1i16(<1 x i16> undef)
@@ -117,25 +117,25 @@ define i32 @reduce_i16(i32 %arg) {
 
 define i32 @reduce_i32(i32 %arg) {
 ; CHECK-LABEL: 'reduce_i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i32 @llvm.vector.reduce.or.v1i32(<1 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i32 @llvm.vector.reduce.or.v2i32(<2 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i32 @llvm.vector.reduce.or.v4i32(<4 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i32 @llvm.vector.reduce.or.v8i32(<8 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i32 @llvm.vector.reduce.or.v16i32(<16 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i32 @llvm.vector.reduce.or.v32i32(<32 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V64 = call i32 @llvm.vector.reduce.or.v64i32(<64 x i32> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 31 for instruction: %V128 = call i32 @llvm.vector.reduce.or.v128i32(<128 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i32 @llvm.vector.reduce.or.v1i32(<1 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i32 @llvm.vector.reduce.or.v2i32(<2 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i32 @llvm.vector.reduce.or.v4i32(<4 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i32 @llvm.vector.reduce.or.v8i32(<8 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i32 @llvm.vector.reduce.or.v16i32(<16 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V32 = call i32 @llvm.vector.reduce.or.v32i32(<32 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V64 = call i32 @llvm.vector.reduce.or.v64i32(<64 x i32> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V128 = call i32 @llvm.vector.reduce.or.v128i32(<128 x i32> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i32 @llvm.vector.reduce.or.v1i32(<1 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i32 @llvm.vector.reduce.or.v2i32(<2 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i32 @llvm.vector.reduce.or.v4i32(<4 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i32 @llvm.vector.reduce.or.v8i32(<8 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i32 @llvm.vector.reduce.or.v16i32(<16 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i32 @llvm.vector.reduce.or.v32i32(<32 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64 = call i32 @llvm.vector.reduce.or.v64i32(<64 x i32> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V128 = call i32 @llvm.vector.reduce.or.v128i32(<128 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i32 @llvm.vector.reduce.or.v1i32(<1 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i32 @llvm.vector.reduce.or.v2i32(<2 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i32 @llvm.vector.reduce.or.v4i32(<4 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i32 @llvm.vector.reduce.or.v8i32(<8 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i32 @llvm.vector.reduce.or.v16i32(<16 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i32 @llvm.vector.reduce.or.v32i32(<32 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i32 @llvm.vector.reduce.or.v64i32(<64 x i32> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V128 = call i32 @llvm.vector.reduce.or.v128i32(<128 x i32> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i32 @llvm.vector.reduce.or.v1i32(<1 x i32> undef)
@@ -151,25 +151,25 @@ define i32 @reduce_i32(i32 %arg) {
 
 define i32 @reduce_i64(i32 %arg) {
 ; CHECK-LABEL: 'reduce_i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i64 @llvm.vector.reduce.or.v1i64(<1 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4 = call i64 @llvm.vector.reduce.or.v4i64(<4 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i64 @llvm.vector.reduce.or.v8i64(<8 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i64 @llvm.vector.reduce.or.v16i64(<16 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V32 = call i64 @llvm.vector.reduce.or.v32i64(<32 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V64 = call i64 @llvm.vector.reduce.or.v64i64(<64 x i64> undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 62 for instruction: %V128 = call i64 @llvm.vector.reduce.or.v128i64(<128 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V1 = call i64 @llvm.vector.reduce.or.v1i64(<1 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i64 @llvm.vector.reduce.or.v4i64(<4 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8 = call i64 @llvm.vector.reduce.or.v8i64(<8 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V16 = call i64 @llvm.vector.reduce.or.v16i64(<16 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V32 = call i64 @llvm.vector.reduce.or.v32i64(<32 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V64 = call i64 @llvm.vector.reduce.or.v64i64(<64 x i64> undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 61 for instruction: %V128 = call i64 @llvm.vector.reduce.or.v128i64(<128 x i64> undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; SIZE-LABEL: 'reduce_i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i64 @llvm.vector.reduce.or.v1i64(<1 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i64 @llvm.vector.reduce.or.v4i64(<4 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i64 @llvm.vector.reduce.or.v8i64(<8 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i64 @llvm.vector.reduce.or.v16i64(<16 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32 = call i64 @llvm.vector.reduce.or.v32i64(<32 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V64 = call i64 @llvm.vector.reduce.or.v64i64(<64 x i64> undef)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V128 = call i64 @llvm.vector.reduce.or.v128i64(<128 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i64 @llvm.vector.reduce.or.v1i64(<1 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i64 @llvm.vector.reduce.or.v4i64(<4 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i64 @llvm.vector.reduce.or.v8i64(<8 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i64 @llvm.vector.reduce.or.v16i64(<16 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i64 @llvm.vector.reduce.or.v32i64(<32 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V64 = call i64 @llvm.vector.reduce.or.v64i64(<64 x i64> undef)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V128 = call i64 @llvm.vector.reduce.or.v128i64(<128 x i64> undef)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %V1   = call i64 @llvm.vector.reduce.or.v1i64(<1 x i64> undef)
diff --git a/llvm/test/Analysis/CostModel/RISCV/reduce-scalable-fp.ll b/llvm/test/Analysis/CostModel/RISCV/reduce-scalable-fp.ll
index 4f6e0ba074ed81..8ae387f48ccaac 100644
--- a/llvm/test/Analysis/CostModel/RISCV/reduce-scalable-fp.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/reduce-scalable-fp.ll
@@ -452,11 +452,11 @@ declare half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half>)
 
 define half @vreduce_fmin_nxv1f16(<vscale x 1 x half> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv1f16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret half %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv1f16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret half %red
 ;
   %red = call half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
@@ -465,11 +465,11 @@ define half @vreduce_fmin_nxv1f16(<vscale x 1 x half> %v) {
 
 define half @vreduce_fmin_nxv1f16_nonans(<vscale x 1 x half> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv1f16_nonans'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret half %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv1f16_nonans'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret half %red
 ;
   %red = call nnan half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
@@ -478,11 +478,11 @@ define half @vreduce_fmin_nxv1f16_nonans(<vscale x 1 x half> %v) {
 
 define half @vreduce_fmin_nxv1f16_nonans_noinfs(<vscale x 1 x half> %v) #1 {
 ; CHECK-LABEL: 'vreduce_fmin_nxv1f16_nonans_noinfs'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret half %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv1f16_nonans_noinfs'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret half %red
 ;
   %red = call nnan ninf half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
@@ -493,11 +493,11 @@ declare half @llvm.vector.reduce.fmin.nxv2f16(<vscale x 2 x half>)
 
 define half @vreduce_fmin_nxv2f16(<vscale x 2 x half> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv2f16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv2f16(<vscale x 2 x half> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv2f16(<vscale x 2 x half> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret half %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv2f16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv2f16(<vscale x 2 x half> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv2f16(<vscale x 2 x half> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret half %red
 ;
   %red = call half @llvm.vector.reduce.fmin.nxv2f16(<vscale x 2 x half> %v)
@@ -508,11 +508,11 @@ declare half @llvm.vector.reduce.fmin.nxv4f16(<vscale x 4 x half>)
 
 define half @vreduce_fmin_nxv4f16(<vscale x 4 x half> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv4f16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv4f16(<vscale x 4 x half> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv4f16(<vscale x 4 x half> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret half %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv4f16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv4f16(<vscale x 4 x half> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv4f16(<vscale x 4 x half> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret half %red
 ;
   %red = call half @llvm.vector.reduce.fmin.nxv4f16(<vscale x 4 x half> %v)
@@ -523,11 +523,11 @@ declare half @llvm.vector.reduce.fmin.nxv64f16(<vscale x 64 x half>)
 
 define half @vreduce_fmin_nxv64f16(<vscale x 64 x half> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv64f16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv64f16(<vscale x 64 x half> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv64f16(<vscale x 64 x half> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret half %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv64f16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv64f16(<vscale x 64 x half> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv64f16(<vscale x 64 x half> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret half %red
 ;
   %red = call half @llvm.vector.reduce.fmin.nxv64f16(<vscale x 64 x half> %v)
@@ -538,11 +538,11 @@ declare float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float>)
 
 define float @vreduce_fmin_nxv1f32(<vscale x 1 x float> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv1f32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv1f32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret float %red
 ;
   %red = call float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
@@ -551,11 +551,11 @@ define float @vreduce_fmin_nxv1f32(<vscale x 1 x float> %v) {
 
 define float @vreduce_fmin_nxv1f32_nonans(<vscale x 1 x float> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv1f32_nonans'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv1f32_nonans'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret float %red
 ;
   %red = call nnan float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
@@ -564,11 +564,11 @@ define float @vreduce_fmin_nxv1f32_nonans(<vscale x 1 x float> %v) {
 
 define float @vreduce_fmin_nxv1f32_nonans_noinfs(<vscale x 1 x float> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv1f32_nonans_noinfs'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv1f32_nonans_noinfs'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret float %red
 ;
   %red = call nnan ninf float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
@@ -579,11 +579,11 @@ declare float @llvm.vector.reduce.fmin.nxv2f32(<vscale x 2 x float>)
 
 define float @vreduce_fmin_nxv2f32(<vscale x 2 x float> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv2f32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv2f32(<vscale x 2 x float> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv2f32(<vscale x 2 x float> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv2f32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv2f32(<vscale x 2 x float> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv2f32(<vscale x 2 x float> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret float %red
 ;
   %red = call float @llvm.vector.reduce.fmin.nxv2f32(<vscale x 2 x float> %v)
@@ -594,11 +594,11 @@ declare float @llvm.vector.reduce.fmin.nxv4f32(<vscale x 4 x float>)
 
 define float @vreduce_fmin_nxv4f32(<vscale x 4 x float> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv4f32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv4f32(<vscale x 4 x float> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv4f32(<vscale x 4 x float> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv4f32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv4f32(<vscale x 4 x float> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv4f32(<vscale x 4 x float> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret float %red
 ;
   %red = call float @llvm.vector.reduce.fmin.nxv4f32(<vscale x 4 x float> %v)
@@ -609,11 +609,11 @@ declare float @llvm.vector.reduce.fmin.nxv32f32(<vscale x 32 x float>)
 
 define float @vreduce_fmin_nxv32f32(<vscale x 32 x float> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv32f32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv32f32(<vscale x 32 x float> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv32f32(<vscale x 32 x float> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv32f32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv32f32(<vscale x 32 x float> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv32f32(<vscale x 32 x float> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret float %red
 ;
   %red = call float @llvm.vector.reduce.fmin.nxv32f32(<vscale x 32 x float> %v)
@@ -624,11 +624,11 @@ declare double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double>)
 
 define double @vreduce_fmin_nxv1f64(<vscale x 1 x double> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv1f64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv1f64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret double %red
 ;
   %red = call double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
@@ -637,11 +637,11 @@ define double @vreduce_fmin_nxv1f64(<vscale x 1 x double> %v) {
 
 define double @vreduce_fmin_nxv1f64_nonans(<vscale x 1 x double> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv1f64_nonans'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv1f64_nonans'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret double %red
 ;
   %red = call nnan double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
@@ -650,11 +650,11 @@ define double @vreduce_fmin_nxv1f64_nonans(<vscale x 1 x double> %v) {
 
 define double @vreduce_fmin_nxv1f64_nonans_noinfs(<vscale x 1 x double> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv1f64_nonans_noinfs'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv1f64_nonans_noinfs'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret double %red
 ;
   %red = call nnan ninf double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
@@ -665,11 +665,11 @@ declare double @llvm.vector.reduce.fmin.nxv2f64(<vscale x 2 x double>)
 
 define double @vreduce_fmin_nxv2f64(<vscale x 2 x double> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv2f64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv2f64(<vscale x 2 x double> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv2f64(<vscale x 2 x double> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv2f64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv2f64(<vscale x 2 x double> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv2f64(<vscale x 2 x double> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret double %red
 ;
   %red = call double @llvm.vector.reduce.fmin.nxv2f64(<vscale x 2 x double> %v)
@@ -680,11 +680,11 @@ declare double @llvm.vector.reduce.fmin.nxv4f64(<vscale x 4 x double>)
 
 define double @vreduce_fmin_nxv4f64(<vscale x 4 x double> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv4f64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv4f64(<vscale x 4 x double> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv4f64(<vscale x 4 x double> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv4f64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv4f64(<vscale x 4 x double> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv4f64(<vscale x 4 x double> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret double %red
 ;
   %red = call double @llvm.vector.reduce.fmin.nxv4f64(<vscale x 4 x double> %v)
@@ -695,11 +695,11 @@ declare double @llvm.vector.reduce.fmin.nxv16f64(<vscale x 16 x double>)
 
 define double @vreduce_fmin_nxv16f64(<vscale x 16 x double> %v) {
 ; CHECK-LABEL: 'vreduce_fmin_nxv16f64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv16f64(<vscale x 16 x double> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv16f64(<vscale x 16 x double> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %red
 ;
 ; SIZE-LABEL: 'vreduce_fmin_nxv16f64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv16f64(<vscale x 16 x double> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv16f64(<vscale x 16 x double> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret double %red
 ;
   %red = call double @llvm.vector.reduce.fmin.nxv16f64(<vscale x 16 x double> %v)
@@ -710,11 +710,11 @@ declare half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half>)
 
 define half @vreduce_fmax_nxv1f16(<vscale x 1 x half> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv1f16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret half %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv1f16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret half %red
 ;
   %red = call half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
@@ -723,11 +723,11 @@ define half @vreduce_fmax_nxv1f16(<vscale x 1 x half> %v) {
 
 define half @vreduce_fmax_nxv1f16_nonans(<vscale x 1 x half> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv1f16_nonans'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret half %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv1f16_nonans'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret half %red
 ;
   %red = call nnan half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
@@ -736,11 +736,11 @@ define half @vreduce_fmax_nxv1f16_nonans(<vscale x 1 x half> %v) {
 
 define half @vreduce_fmax_nxv1f16_nonans_noinfs(<vscale x 1 x half> %v) #1 {
 ; CHECK-LABEL: 'vreduce_fmax_nxv1f16_nonans_noinfs'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret half %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv1f16_nonans_noinfs'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret half %red
 ;
   %red = call nnan ninf half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
@@ -751,11 +751,11 @@ declare half @llvm.vector.reduce.fmax.nxv2f16(<vscale x 2 x half>)
 
 define half @vreduce_fmax_nxv2f16(<vscale x 2 x half> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv2f16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv2f16(<vscale x 2 x half> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv2f16(<vscale x 2 x half> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret half %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv2f16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv2f16(<vscale x 2 x half> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv2f16(<vscale x 2 x half> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret half %red
 ;
   %red = call half @llvm.vector.reduce.fmax.nxv2f16(<vscale x 2 x half> %v)
@@ -766,11 +766,11 @@ declare half @llvm.vector.reduce.fmax.nxv4f16(<vscale x 4 x half>)
 
 define half @vreduce_fmax_nxv4f16(<vscale x 4 x half> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv4f16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv4f16(<vscale x 4 x half> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv4f16(<vscale x 4 x half> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret half %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv4f16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv4f16(<vscale x 4 x half> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv4f16(<vscale x 4 x half> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret half %red
 ;
   %red = call half @llvm.vector.reduce.fmax.nxv4f16(<vscale x 4 x half> %v)
@@ -781,11 +781,11 @@ declare half @llvm.vector.reduce.fmax.nxv64f16(<vscale x 64 x half>)
 
 define half @vreduce_fmax_nxv64f16(<vscale x 64 x half> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv64f16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv64f16(<vscale x 64 x half> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv64f16(<vscale x 64 x half> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret half %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv64f16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv64f16(<vscale x 64 x half> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv64f16(<vscale x 64 x half> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret half %red
 ;
   %red = call half @llvm.vector.reduce.fmax.nxv64f16(<vscale x 64 x half> %v)
@@ -796,11 +796,11 @@ declare float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float>)
 
 define float @vreduce_fmax_nxv1f32(<vscale x 1 x float> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv1f32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv1f32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret float %red
 ;
   %red = call float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
@@ -809,11 +809,11 @@ define float @vreduce_fmax_nxv1f32(<vscale x 1 x float> %v) {
 
 define float @vreduce_fmax_nxv1f32_nonans(<vscale x 1 x float> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv1f32_nonans'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv1f32_nonans'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret float %red
 ;
   %red = call nnan float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
@@ -822,11 +822,11 @@ define float @vreduce_fmax_nxv1f32_nonans(<vscale x 1 x float> %v) {
 
 define float @vreduce_fmax_nxv1f32_nonans_noinfs(<vscale x 1 x float> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv1f32_nonans_noinfs'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv1f32_nonans_noinfs'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret float %red
 ;
   %red = call nnan ninf float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
@@ -837,11 +837,11 @@ declare float @llvm.vector.reduce.fmax.nxv2f32(<vscale x 2 x float>)
 
 define float @vreduce_fmax_nxv2f32(<vscale x 2 x float> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv2f32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv2f32(<vscale x 2 x float> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv2f32(<vscale x 2 x float> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv2f32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv2f32(<vscale x 2 x float> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv2f32(<vscale x 2 x float> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret float %red
 ;
   %red = call float @llvm.vector.reduce.fmax.nxv2f32(<vscale x 2 x float> %v)
@@ -852,11 +852,11 @@ declare float @llvm.vector.reduce.fmax.nxv4f32(<vscale x 4 x float>)
 
 define float @vreduce_fmax_nxv4f32(<vscale x 4 x float> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv4f32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv4f32(<vscale x 4 x float> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv4f32(<vscale x 4 x float> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv4f32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv4f32(<vscale x 4 x float> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv4f32(<vscale x 4 x float> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret float %red
 ;
   %red = call float @llvm.vector.reduce.fmax.nxv4f32(<vscale x 4 x float> %v)
@@ -867,11 +867,11 @@ declare float @llvm.vector.reduce.fmax.nxv32f32(<vscale x 32 x float>)
 
 define float @vreduce_fmax_nxv32f32(<vscale x 32 x float> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv32f32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv32f32(<vscale x 32 x float> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv32f32(<vscale x 32 x float> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv32f32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv32f32(<vscale x 32 x float> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv32f32(<vscale x 32 x float> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret float %red
 ;
   %red = call float @llvm.vector.reduce.fmax.nxv32f32(<vscale x 32 x float> %v)
@@ -882,11 +882,11 @@ declare double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double>)
 
 define double @vreduce_fmax_nxv1f64(<vscale x 1 x double> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv1f64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv1f64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret double %red
 ;
   %red = call double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
@@ -895,11 +895,11 @@ define double @vreduce_fmax_nxv1f64(<vscale x 1 x double> %v) {
 
 define double @vreduce_fmax_nxv1f64_nonans(<vscale x 1 x double> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv1f64_nonans'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv1f64_nonans'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret double %red
 ;
   %red = call nnan double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
@@ -908,11 +908,11 @@ define double @vreduce_fmax_nxv1f64_nonans(<vscale x 1 x double> %v) {
 
 define double @vreduce_fmax_nxv1f64_nonans_noinfs(<vscale x 1 x double> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv1f64_nonans_noinfs'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv1f64_nonans_noinfs'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret double %red
 ;
   %red = call nnan ninf double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
@@ -923,11 +923,11 @@ declare double @llvm.vector.reduce.fmax.nxv2f64(<vscale x 2 x double>)
 
 define double @vreduce_fmax_nxv2f64(<vscale x 2 x double> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv2f64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv2f64(<vscale x 2 x double> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv2f64(<vscale x 2 x double> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv2f64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv2f64(<vscale x 2 x double> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv2f64(<vscale x 2 x double> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret double %red
 ;
   %red = call double @llvm.vector.reduce.fmax.nxv2f64(<vscale x 2 x double> %v)
@@ -938,11 +938,11 @@ declare double @llvm.vector.reduce.fmax.nxv4f64(<vscale x 4 x double>)
 
 define double @vreduce_fmax_nxv4f64(<vscale x 4 x double> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv4f64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv4f64(<vscale x 4 x double> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv4f64(<vscale x 4 x double> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv4f64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv4f64(<vscale x 4 x double> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv4f64(<vscale x 4 x double> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret double %red
 ;
   %red = call double @llvm.vector.reduce.fmax.nxv4f64(<vscale x 4 x double> %v)
@@ -953,11 +953,11 @@ declare double @llvm.vector.reduce.fmax.nxv16f64(<vscale x 16 x double>)
 
 define double @vreduce_fmax_nxv16f64(<vscale x 16 x double> %v) {
 ; CHECK-LABEL: 'vreduce_fmax_nxv16f64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv16f64(<vscale x 16 x double> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv16f64(<vscale x 16 x double> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %red
 ;
 ; SIZE-LABEL: 'vreduce_fmax_nxv16f64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv16f64(<vscale x 16 x double> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv16f64(<vscale x 16 x double> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret double %red
 ;
   %red = call double @llvm.vector.reduce.fmax.nxv16f64(<vscale x 16 x double> %v)
diff --git a/llvm/test/Analysis/CostModel/RISCV/reduce-scalable-int.ll b/llvm/test/Analysis/CostModel/RISCV/reduce-scalable-int.ll
index 2807f7526760f8..a1a22b56358a37 100644
--- a/llvm/test/Analysis/CostModel/RISCV/reduce-scalable-int.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/reduce-scalable-int.ll
@@ -23,11 +23,11 @@ declare i8 @llvm.vector.reduce.umax.nxv1i8(<vscale x 1 x i8>)
 
 define signext i8 @vreduce_umax_nxv1i8(<vscale x 1 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_umax_nxv1i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv1i8(<vscale x 1 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv1i8(<vscale x 1 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_umax_nxv1i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv1i8(<vscale x 1 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv1i8(<vscale x 1 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.umax.nxv1i8(<vscale x 1 x i8> %v)
@@ -38,11 +38,11 @@ declare i8 @llvm.vector.reduce.smax.nxv1i8(<vscale x 1 x i8>)
 
 define signext i8 @vreduce_smax_nxv1i8(<vscale x 1 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_smax_nxv1i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv1i8(<vscale x 1 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv1i8(<vscale x 1 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_smax_nxv1i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv1i8(<vscale x 1 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv1i8(<vscale x 1 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.smax.nxv1i8(<vscale x 1 x i8> %v)
@@ -53,11 +53,11 @@ declare i8 @llvm.vector.reduce.umin.nxv1i8(<vscale x 1 x i8>)
 
 define signext i8 @vreduce_umin_nxv1i8(<vscale x 1 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_umin_nxv1i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv1i8(<vscale x 1 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv1i8(<vscale x 1 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_umin_nxv1i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv1i8(<vscale x 1 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv1i8(<vscale x 1 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.umin.nxv1i8(<vscale x 1 x i8> %v)
@@ -68,11 +68,11 @@ declare i8 @llvm.vector.reduce.smin.nxv1i8(<vscale x 1 x i8>)
 
 define signext i8 @vreduce_smin_nxv1i8(<vscale x 1 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_smin_nxv1i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv1i8(<vscale x 1 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv1i8(<vscale x 1 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_smin_nxv1i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv1i8(<vscale x 1 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv1i8(<vscale x 1 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.smin.nxv1i8(<vscale x 1 x i8> %v)
@@ -83,11 +83,11 @@ declare i8 @llvm.vector.reduce.and.nxv1i8(<vscale x 1 x i8>)
 
 define signext i8 @vreduce_and_nxv1i8(<vscale x 1 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_and_nxv1i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.and.nxv1i8(<vscale x 1 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.and.nxv1i8(<vscale x 1 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_and_nxv1i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.and.nxv1i8(<vscale x 1 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.and.nxv1i8(<vscale x 1 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.and.nxv1i8(<vscale x 1 x i8> %v)
@@ -98,11 +98,11 @@ declare i8 @llvm.vector.reduce.or.nxv1i8(<vscale x 1 x i8>)
 
 define signext i8 @vreduce_or_nxv1i8(<vscale x 1 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_or_nxv1i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.or.nxv1i8(<vscale x 1 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.or.nxv1i8(<vscale x 1 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_or_nxv1i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.or.nxv1i8(<vscale x 1 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.or.nxv1i8(<vscale x 1 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.or.nxv1i8(<vscale x 1 x i8> %v)
@@ -143,11 +143,11 @@ declare i8 @llvm.vector.reduce.umax.nxv2i8(<vscale x 2 x i8>)
 
 define signext i8 @vreduce_umax_nxv2i8(<vscale x 2 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_umax_nxv2i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv2i8(<vscale x 2 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv2i8(<vscale x 2 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_umax_nxv2i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv2i8(<vscale x 2 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv2i8(<vscale x 2 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.umax.nxv2i8(<vscale x 2 x i8> %v)
@@ -158,11 +158,11 @@ declare i8 @llvm.vector.reduce.smax.nxv2i8(<vscale x 2 x i8>)
 
 define signext i8 @vreduce_smax_nxv2i8(<vscale x 2 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_smax_nxv2i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv2i8(<vscale x 2 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv2i8(<vscale x 2 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_smax_nxv2i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv2i8(<vscale x 2 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv2i8(<vscale x 2 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.smax.nxv2i8(<vscale x 2 x i8> %v)
@@ -173,11 +173,11 @@ declare i8 @llvm.vector.reduce.umin.nxv2i8(<vscale x 2 x i8>)
 
 define signext i8 @vreduce_umin_nxv2i8(<vscale x 2 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_umin_nxv2i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv2i8(<vscale x 2 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv2i8(<vscale x 2 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_umin_nxv2i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv2i8(<vscale x 2 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv2i8(<vscale x 2 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.umin.nxv2i8(<vscale x 2 x i8> %v)
@@ -188,11 +188,11 @@ declare i8 @llvm.vector.reduce.smin.nxv2i8(<vscale x 2 x i8>)
 
 define signext i8 @vreduce_smin_nxv2i8(<vscale x 2 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_smin_nxv2i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv2i8(<vscale x 2 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv2i8(<vscale x 2 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_smin_nxv2i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv2i8(<vscale x 2 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv2i8(<vscale x 2 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.smin.nxv2i8(<vscale x 2 x i8> %v)
@@ -203,11 +203,11 @@ declare i8 @llvm.vector.reduce.and.nxv2i8(<vscale x 2 x i8>)
 
 define signext i8 @vreduce_and_nxv2i8(<vscale x 2 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_and_nxv2i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i8 @llvm.vector.reduce.and.nxv2i8(<vscale x 2 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.and.nxv2i8(<vscale x 2 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_and_nxv2i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.and.nxv2i8(<vscale x 2 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.and.nxv2i8(<vscale x 2 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.and.nxv2i8(<vscale x 2 x i8> %v)
@@ -218,11 +218,11 @@ declare i8 @llvm.vector.reduce.or.nxv2i8(<vscale x 2 x i8>)
 
 define signext i8 @vreduce_or_nxv2i8(<vscale x 2 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_or_nxv2i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i8 @llvm.vector.reduce.or.nxv2i8(<vscale x 2 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.or.nxv2i8(<vscale x 2 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_or_nxv2i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.or.nxv2i8(<vscale x 2 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.or.nxv2i8(<vscale x 2 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.or.nxv2i8(<vscale x 2 x i8> %v)
@@ -263,11 +263,11 @@ declare i8 @llvm.vector.reduce.umax.nxv4i8(<vscale x 4 x i8>)
 
 define signext i8 @vreduce_umax_nxv4i8(<vscale x 4 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_umax_nxv4i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv4i8(<vscale x 4 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv4i8(<vscale x 4 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_umax_nxv4i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv4i8(<vscale x 4 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv4i8(<vscale x 4 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.umax.nxv4i8(<vscale x 4 x i8> %v)
@@ -278,11 +278,11 @@ declare i8 @llvm.vector.reduce.smax.nxv4i8(<vscale x 4 x i8>)
 
 define signext i8 @vreduce_smax_nxv4i8(<vscale x 4 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_smax_nxv4i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv4i8(<vscale x 4 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv4i8(<vscale x 4 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_smax_nxv4i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv4i8(<vscale x 4 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv4i8(<vscale x 4 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.smax.nxv4i8(<vscale x 4 x i8> %v)
@@ -293,11 +293,11 @@ declare i8 @llvm.vector.reduce.umin.nxv4i8(<vscale x 4 x i8>)
 
 define signext i8 @vreduce_umin_nxv4i8(<vscale x 4 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_umin_nxv4i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv4i8(<vscale x 4 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv4i8(<vscale x 4 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_umin_nxv4i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv4i8(<vscale x 4 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv4i8(<vscale x 4 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.umin.nxv4i8(<vscale x 4 x i8> %v)
@@ -308,11 +308,11 @@ declare i8 @llvm.vector.reduce.smin.nxv4i8(<vscale x 4 x i8>)
 
 define signext i8 @vreduce_smin_nxv4i8(<vscale x 4 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_smin_nxv4i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv4i8(<vscale x 4 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv4i8(<vscale x 4 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_smin_nxv4i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv4i8(<vscale x 4 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv4i8(<vscale x 4 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.smin.nxv4i8(<vscale x 4 x i8> %v)
@@ -323,11 +323,11 @@ declare i8 @llvm.vector.reduce.and.nxv4i8(<vscale x 4 x i8>)
 
 define signext i8 @vreduce_and_nxv4i8(<vscale x 4 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_and_nxv4i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i8 @llvm.vector.reduce.and.nxv4i8(<vscale x 4 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i8 @llvm.vector.reduce.and.nxv4i8(<vscale x 4 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_and_nxv4i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.and.nxv4i8(<vscale x 4 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.and.nxv4i8(<vscale x 4 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.and.nxv4i8(<vscale x 4 x i8> %v)
@@ -338,11 +338,11 @@ declare i8 @llvm.vector.reduce.or.nxv4i8(<vscale x 4 x i8>)
 
 define signext i8 @vreduce_or_nxv4i8(<vscale x 4 x i8> %v) {
 ; CHECK-LABEL: 'vreduce_or_nxv4i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i8 @llvm.vector.reduce.or.nxv4i8(<vscale x 4 x i8> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i8 @llvm.vector.reduce.or.nxv4i8(<vscale x 4 x i8> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
 ;
 ; SIZE-LABEL: 'vreduce_or_nxv4i8'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.or.nxv4i8(<vscale x 4 x i8> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.or.nxv4i8(<vscale x 4 x i8> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
 ;
   %red = call i8 @llvm.vector.reduce.or.nxv4i8(<vscale x 4 x i8> %v)
@@ -415,11 +415,11 @@ declare i16 @llvm.vector.reduce.umax.nxv1i16(<vscale x 1 x i16>)
 
 define signext i16 @vreduce_umax_nxv1i16(<vscale x 1 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_umax_nxv1i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv1i16(<vscale x 1 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv1i16(<vscale x 1 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_umax_nxv1i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv1i16(<vscale x 1 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv1i16(<vscale x 1 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.umax.nxv1i16(<vscale x 1 x i16> %v)
@@ -430,11 +430,11 @@ declare i16 @llvm.vector.reduce.smax.nxv1i16(<vscale x 1 x i16>)
 
 define signext i16 @vreduce_smax_nxv1i16(<vscale x 1 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_smax_nxv1i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv1i16(<vscale x 1 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv1i16(<vscale x 1 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_smax_nxv1i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv1i16(<vscale x 1 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv1i16(<vscale x 1 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.smax.nxv1i16(<vscale x 1 x i16> %v)
@@ -445,11 +445,11 @@ declare i16 @llvm.vector.reduce.umin.nxv1i16(<vscale x 1 x i16>)
 
 define signext i16 @vreduce_umin_nxv1i16(<vscale x 1 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_umin_nxv1i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv1i16(<vscale x 1 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv1i16(<vscale x 1 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_umin_nxv1i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv1i16(<vscale x 1 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv1i16(<vscale x 1 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.umin.nxv1i16(<vscale x 1 x i16> %v)
@@ -460,11 +460,11 @@ declare i16 @llvm.vector.reduce.smin.nxv1i16(<vscale x 1 x i16>)
 
 define signext i16 @vreduce_smin_nxv1i16(<vscale x 1 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_smin_nxv1i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv1i16(<vscale x 1 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv1i16(<vscale x 1 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_smin_nxv1i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv1i16(<vscale x 1 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv1i16(<vscale x 1 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.smin.nxv1i16(<vscale x 1 x i16> %v)
@@ -475,11 +475,11 @@ declare i16 @llvm.vector.reduce.and.nxv1i16(<vscale x 1 x i16>)
 
 define signext i16 @vreduce_and_nxv1i16(<vscale x 1 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_and_nxv1i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.and.nxv1i16(<vscale x 1 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.and.nxv1i16(<vscale x 1 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_and_nxv1i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.and.nxv1i16(<vscale x 1 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.and.nxv1i16(<vscale x 1 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.and.nxv1i16(<vscale x 1 x i16> %v)
@@ -490,11 +490,11 @@ declare i16 @llvm.vector.reduce.or.nxv1i16(<vscale x 1 x i16>)
 
 define signext i16 @vreduce_or_nxv1i16(<vscale x 1 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_or_nxv1i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.or.nxv1i16(<vscale x 1 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.or.nxv1i16(<vscale x 1 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_or_nxv1i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.or.nxv1i16(<vscale x 1 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.or.nxv1i16(<vscale x 1 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.or.nxv1i16(<vscale x 1 x i16> %v)
@@ -567,11 +567,11 @@ declare i16 @llvm.vector.reduce.umax.nxv2i16(<vscale x 2 x i16>)
 
 define signext i16 @vreduce_umax_nxv2i16(<vscale x 2 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_umax_nxv2i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv2i16(<vscale x 2 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv2i16(<vscale x 2 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_umax_nxv2i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv2i16(<vscale x 2 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv2i16(<vscale x 2 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.umax.nxv2i16(<vscale x 2 x i16> %v)
@@ -582,11 +582,11 @@ declare i16 @llvm.vector.reduce.smax.nxv2i16(<vscale x 2 x i16>)
 
 define signext i16 @vreduce_smax_nxv2i16(<vscale x 2 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_smax_nxv2i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv2i16(<vscale x 2 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv2i16(<vscale x 2 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_smax_nxv2i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv2i16(<vscale x 2 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv2i16(<vscale x 2 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.smax.nxv2i16(<vscale x 2 x i16> %v)
@@ -597,11 +597,11 @@ declare i16 @llvm.vector.reduce.umin.nxv2i16(<vscale x 2 x i16>)
 
 define signext i16 @vreduce_umin_nxv2i16(<vscale x 2 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_umin_nxv2i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv2i16(<vscale x 2 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv2i16(<vscale x 2 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_umin_nxv2i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv2i16(<vscale x 2 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv2i16(<vscale x 2 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.umin.nxv2i16(<vscale x 2 x i16> %v)
@@ -612,11 +612,11 @@ declare i16 @llvm.vector.reduce.smin.nxv2i16(<vscale x 2 x i16>)
 
 define signext i16 @vreduce_smin_nxv2i16(<vscale x 2 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_smin_nxv2i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv2i16(<vscale x 2 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv2i16(<vscale x 2 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_smin_nxv2i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv2i16(<vscale x 2 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv2i16(<vscale x 2 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.smin.nxv2i16(<vscale x 2 x i16> %v)
@@ -627,11 +627,11 @@ declare i16 @llvm.vector.reduce.and.nxv2i16(<vscale x 2 x i16>)
 
 define signext i16 @vreduce_and_nxv2i16(<vscale x 2 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_and_nxv2i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i16 @llvm.vector.reduce.and.nxv2i16(<vscale x 2 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.and.nxv2i16(<vscale x 2 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_and_nxv2i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.and.nxv2i16(<vscale x 2 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.and.nxv2i16(<vscale x 2 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.and.nxv2i16(<vscale x 2 x i16> %v)
@@ -642,11 +642,11 @@ declare i16 @llvm.vector.reduce.or.nxv2i16(<vscale x 2 x i16>)
 
 define signext i16 @vreduce_or_nxv2i16(<vscale x 2 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_or_nxv2i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i16 @llvm.vector.reduce.or.nxv2i16(<vscale x 2 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.or.nxv2i16(<vscale x 2 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_or_nxv2i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.or.nxv2i16(<vscale x 2 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.or.nxv2i16(<vscale x 2 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.or.nxv2i16(<vscale x 2 x i16> %v)
@@ -719,11 +719,11 @@ declare i16 @llvm.vector.reduce.umax.nxv4i16(<vscale x 4 x i16>)
 
 define signext i16 @vreduce_umax_nxv4i16(<vscale x 4 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_umax_nxv4i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv4i16(<vscale x 4 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv4i16(<vscale x 4 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_umax_nxv4i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv4i16(<vscale x 4 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv4i16(<vscale x 4 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.umax.nxv4i16(<vscale x 4 x i16> %v)
@@ -734,11 +734,11 @@ declare i16 @llvm.vector.reduce.smax.nxv4i16(<vscale x 4 x i16>)
 
 define signext i16 @vreduce_smax_nxv4i16(<vscale x 4 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_smax_nxv4i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv4i16(<vscale x 4 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv4i16(<vscale x 4 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_smax_nxv4i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv4i16(<vscale x 4 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv4i16(<vscale x 4 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.smax.nxv4i16(<vscale x 4 x i16> %v)
@@ -749,11 +749,11 @@ declare i16 @llvm.vector.reduce.umin.nxv4i16(<vscale x 4 x i16>)
 
 define signext i16 @vreduce_umin_nxv4i16(<vscale x 4 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_umin_nxv4i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv4i16(<vscale x 4 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv4i16(<vscale x 4 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_umin_nxv4i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv4i16(<vscale x 4 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv4i16(<vscale x 4 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.umin.nxv4i16(<vscale x 4 x i16> %v)
@@ -764,11 +764,11 @@ declare i16 @llvm.vector.reduce.smin.nxv4i16(<vscale x 4 x i16>)
 
 define signext i16 @vreduce_smin_nxv4i16(<vscale x 4 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_smin_nxv4i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv4i16(<vscale x 4 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv4i16(<vscale x 4 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_smin_nxv4i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv4i16(<vscale x 4 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv4i16(<vscale x 4 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.smin.nxv4i16(<vscale x 4 x i16> %v)
@@ -779,11 +779,11 @@ declare i16 @llvm.vector.reduce.and.nxv4i16(<vscale x 4 x i16>)
 
 define signext i16 @vreduce_and_nxv4i16(<vscale x 4 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_and_nxv4i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i16 @llvm.vector.reduce.and.nxv4i16(<vscale x 4 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i16 @llvm.vector.reduce.and.nxv4i16(<vscale x 4 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_and_nxv4i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.and.nxv4i16(<vscale x 4 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.and.nxv4i16(<vscale x 4 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.and.nxv4i16(<vscale x 4 x i16> %v)
@@ -794,11 +794,11 @@ declare i16 @llvm.vector.reduce.or.nxv4i16(<vscale x 4 x i16>)
 
 define signext i16 @vreduce_or_nxv4i16(<vscale x 4 x i16> %v) {
 ; CHECK-LABEL: 'vreduce_or_nxv4i16'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i16 @llvm.vector.reduce.or.nxv4i16(<vscale x 4 x i16> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i16 @llvm.vector.reduce.or.nxv4i16(<vscale x 4 x i16> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
 ;
 ; SIZE-LABEL: 'vreduce_or_nxv4i16'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.or.nxv4i16(<vscale x 4 x i16> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.or.nxv4i16(<vscale x 4 x i16> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
 ;
   %red = call i16 @llvm.vector.reduce.or.nxv4i16(<vscale x 4 x i16> %v)
@@ -871,11 +871,11 @@ declare i32 @llvm.vector.reduce.umax.nxv1i32(<vscale x 1 x i32>)
 
 define signext i32 @vreduce_umax_nxv1i32(<vscale x 1 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_umax_nxv1i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv1i32(<vscale x 1 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv1i32(<vscale x 1 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_umax_nxv1i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv1i32(<vscale x 1 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv1i32(<vscale x 1 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.umax.nxv1i32(<vscale x 1 x i32> %v)
@@ -886,11 +886,11 @@ declare i32 @llvm.vector.reduce.smax.nxv1i32(<vscale x 1 x i32>)
 
 define signext i32 @vreduce_smax_nxv1i32(<vscale x 1 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_smax_nxv1i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv1i32(<vscale x 1 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv1i32(<vscale x 1 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_smax_nxv1i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv1i32(<vscale x 1 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv1i32(<vscale x 1 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.smax.nxv1i32(<vscale x 1 x i32> %v)
@@ -901,11 +901,11 @@ declare i32 @llvm.vector.reduce.umin.nxv1i32(<vscale x 1 x i32>)
 
 define signext i32 @vreduce_umin_nxv1i32(<vscale x 1 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_umin_nxv1i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv1i32(<vscale x 1 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv1i32(<vscale x 1 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_umin_nxv1i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv1i32(<vscale x 1 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv1i32(<vscale x 1 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.umin.nxv1i32(<vscale x 1 x i32> %v)
@@ -916,11 +916,11 @@ declare i32 @llvm.vector.reduce.smin.nxv1i32(<vscale x 1 x i32>)
 
 define signext i32 @vreduce_smin_nxv1i32(<vscale x 1 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_smin_nxv1i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv1i32(<vscale x 1 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv1i32(<vscale x 1 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_smin_nxv1i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv1i32(<vscale x 1 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv1i32(<vscale x 1 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.smin.nxv1i32(<vscale x 1 x i32> %v)
@@ -931,11 +931,11 @@ declare i32 @llvm.vector.reduce.and.nxv1i32(<vscale x 1 x i32>)
 
 define signext i32 @vreduce_and_nxv1i32(<vscale x 1 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_and_nxv1i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.and.nxv1i32(<vscale x 1 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.and.nxv1i32(<vscale x 1 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_and_nxv1i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.and.nxv1i32(<vscale x 1 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.and.nxv1i32(<vscale x 1 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.and.nxv1i32(<vscale x 1 x i32> %v)
@@ -946,11 +946,11 @@ declare i32 @llvm.vector.reduce.or.nxv1i32(<vscale x 1 x i32>)
 
 define signext i32 @vreduce_or_nxv1i32(<vscale x 1 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_or_nxv1i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.or.nxv1i32(<vscale x 1 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.or.nxv1i32(<vscale x 1 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_or_nxv1i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.or.nxv1i32(<vscale x 1 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.or.nxv1i32(<vscale x 1 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.or.nxv1i32(<vscale x 1 x i32> %v)
@@ -1023,11 +1023,11 @@ declare i32 @llvm.vector.reduce.umax.nxv2i32(<vscale x 2 x i32>)
 
 define signext i32 @vreduce_umax_nxv2i32(<vscale x 2 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_umax_nxv2i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv2i32(<vscale x 2 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv2i32(<vscale x 2 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_umax_nxv2i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv2i32(<vscale x 2 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv2i32(<vscale x 2 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.umax.nxv2i32(<vscale x 2 x i32> %v)
@@ -1038,11 +1038,11 @@ declare i32 @llvm.vector.reduce.smax.nxv2i32(<vscale x 2 x i32>)
 
 define signext i32 @vreduce_smax_nxv2i32(<vscale x 2 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_smax_nxv2i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv2i32(<vscale x 2 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv2i32(<vscale x 2 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_smax_nxv2i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv2i32(<vscale x 2 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv2i32(<vscale x 2 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.smax.nxv2i32(<vscale x 2 x i32> %v)
@@ -1053,11 +1053,11 @@ declare i32 @llvm.vector.reduce.umin.nxv2i32(<vscale x 2 x i32>)
 
 define signext i32 @vreduce_umin_nxv2i32(<vscale x 2 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_umin_nxv2i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv2i32(<vscale x 2 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv2i32(<vscale x 2 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_umin_nxv2i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv2i32(<vscale x 2 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv2i32(<vscale x 2 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.umin.nxv2i32(<vscale x 2 x i32> %v)
@@ -1068,11 +1068,11 @@ declare i32 @llvm.vector.reduce.smin.nxv2i32(<vscale x 2 x i32>)
 
 define signext i32 @vreduce_smin_nxv2i32(<vscale x 2 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_smin_nxv2i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv2i32(<vscale x 2 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv2i32(<vscale x 2 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_smin_nxv2i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv2i32(<vscale x 2 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv2i32(<vscale x 2 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.smin.nxv2i32(<vscale x 2 x i32> %v)
@@ -1083,11 +1083,11 @@ declare i32 @llvm.vector.reduce.and.nxv2i32(<vscale x 2 x i32>)
 
 define signext i32 @vreduce_and_nxv2i32(<vscale x 2 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_and_nxv2i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i32 @llvm.vector.reduce.and.nxv2i32(<vscale x 2 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.and.nxv2i32(<vscale x 2 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_and_nxv2i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.and.nxv2i32(<vscale x 2 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.and.nxv2i32(<vscale x 2 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.and.nxv2i32(<vscale x 2 x i32> %v)
@@ -1098,11 +1098,11 @@ declare i32 @llvm.vector.reduce.or.nxv2i32(<vscale x 2 x i32>)
 
 define signext i32 @vreduce_or_nxv2i32(<vscale x 2 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_or_nxv2i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i32 @llvm.vector.reduce.or.nxv2i32(<vscale x 2 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.or.nxv2i32(<vscale x 2 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_or_nxv2i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.or.nxv2i32(<vscale x 2 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.or.nxv2i32(<vscale x 2 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.or.nxv2i32(<vscale x 2 x i32> %v)
@@ -1175,11 +1175,11 @@ declare i32 @llvm.vector.reduce.umax.nxv4i32(<vscale x 4 x i32>)
 
 define signext i32 @vreduce_umax_nxv4i32(<vscale x 4 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_umax_nxv4i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv4i32(<vscale x 4 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv4i32(<vscale x 4 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_umax_nxv4i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv4i32(<vscale x 4 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv4i32(<vscale x 4 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.umax.nxv4i32(<vscale x 4 x i32> %v)
@@ -1190,11 +1190,11 @@ declare i32 @llvm.vector.reduce.smax.nxv4i32(<vscale x 4 x i32>)
 
 define signext i32 @vreduce_smax_nxv4i32(<vscale x 4 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_smax_nxv4i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv4i32(<vscale x 4 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv4i32(<vscale x 4 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_smax_nxv4i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv4i32(<vscale x 4 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv4i32(<vscale x 4 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.smax.nxv4i32(<vscale x 4 x i32> %v)
@@ -1205,11 +1205,11 @@ declare i32 @llvm.vector.reduce.umin.nxv4i32(<vscale x 4 x i32>)
 
 define signext i32 @vreduce_umin_nxv4i32(<vscale x 4 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_umin_nxv4i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv4i32(<vscale x 4 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv4i32(<vscale x 4 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_umin_nxv4i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv4i32(<vscale x 4 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv4i32(<vscale x 4 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.umin.nxv4i32(<vscale x 4 x i32> %v)
@@ -1220,11 +1220,11 @@ declare i32 @llvm.vector.reduce.smin.nxv4i32(<vscale x 4 x i32>)
 
 define signext i32 @vreduce_smin_nxv4i32(<vscale x 4 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_smin_nxv4i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv4i32(<vscale x 4 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv4i32(<vscale x 4 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_smin_nxv4i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv4i32(<vscale x 4 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv4i32(<vscale x 4 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.smin.nxv4i32(<vscale x 4 x i32> %v)
@@ -1235,11 +1235,11 @@ declare i32 @llvm.vector.reduce.and.nxv4i32(<vscale x 4 x i32>)
 
 define signext i32 @vreduce_and_nxv4i32(<vscale x 4 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_and_nxv4i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i32 @llvm.vector.reduce.and.nxv4i32(<vscale x 4 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i32 @llvm.vector.reduce.and.nxv4i32(<vscale x 4 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_and_nxv4i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.and.nxv4i32(<vscale x 4 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.and.nxv4i32(<vscale x 4 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.and.nxv4i32(<vscale x 4 x i32> %v)
@@ -1250,11 +1250,11 @@ declare i32 @llvm.vector.reduce.or.nxv4i32(<vscale x 4 x i32>)
 
 define signext i32 @vreduce_or_nxv4i32(<vscale x 4 x i32> %v) {
 ; CHECK-LABEL: 'vreduce_or_nxv4i32'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i32 @llvm.vector.reduce.or.nxv4i32(<vscale x 4 x i32> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i32 @llvm.vector.reduce.or.nxv4i32(<vscale x 4 x i32> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
 ;
 ; SIZE-LABEL: 'vreduce_or_nxv4i32'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.or.nxv4i32(<vscale x 4 x i32> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.or.nxv4i32(<vscale x 4 x i32> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
 ;
   %red = call i32 @llvm.vector.reduce.or.nxv4i32(<vscale x 4 x i32> %v)
@@ -1327,11 +1327,11 @@ declare i64 @llvm.vector.reduce.umax.nxv1i64(<vscale x 1 x i64>)
 
 define i64 @vreduce_umax_nxv1i64(<vscale x 1 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_umax_nxv1i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv1i64(<vscale x 1 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv1i64(<vscale x 1 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_umax_nxv1i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv1i64(<vscale x 1 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv1i64(<vscale x 1 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.umax.nxv1i64(<vscale x 1 x i64> %v)
@@ -1342,11 +1342,11 @@ declare i64 @llvm.vector.reduce.smax.nxv1i64(<vscale x 1 x i64>)
 
 define i64 @vreduce_smax_nxv1i64(<vscale x 1 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_smax_nxv1i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv1i64(<vscale x 1 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv1i64(<vscale x 1 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_smax_nxv1i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv1i64(<vscale x 1 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv1i64(<vscale x 1 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.smax.nxv1i64(<vscale x 1 x i64> %v)
@@ -1357,11 +1357,11 @@ declare i64 @llvm.vector.reduce.umin.nxv1i64(<vscale x 1 x i64>)
 
 define i64 @vreduce_umin_nxv1i64(<vscale x 1 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_umin_nxv1i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv1i64(<vscale x 1 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv1i64(<vscale x 1 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_umin_nxv1i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv1i64(<vscale x 1 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv1i64(<vscale x 1 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.umin.nxv1i64(<vscale x 1 x i64> %v)
@@ -1372,11 +1372,11 @@ declare i64 @llvm.vector.reduce.smin.nxv1i64(<vscale x 1 x i64>)
 
 define i64 @vreduce_smin_nxv1i64(<vscale x 1 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_smin_nxv1i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv1i64(<vscale x 1 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv1i64(<vscale x 1 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_smin_nxv1i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv1i64(<vscale x 1 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv1i64(<vscale x 1 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.smin.nxv1i64(<vscale x 1 x i64> %v)
@@ -1387,11 +1387,11 @@ declare i64 @llvm.vector.reduce.and.nxv1i64(<vscale x 1 x i64>)
 
 define i64 @vreduce_and_nxv1i64(<vscale x 1 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_and_nxv1i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.and.nxv1i64(<vscale x 1 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.and.nxv1i64(<vscale x 1 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_and_nxv1i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.and.nxv1i64(<vscale x 1 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.and.nxv1i64(<vscale x 1 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.and.nxv1i64(<vscale x 1 x i64> %v)
@@ -1402,11 +1402,11 @@ declare i64 @llvm.vector.reduce.or.nxv1i64(<vscale x 1 x i64>)
 
 define i64 @vreduce_or_nxv1i64(<vscale x 1 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_or_nxv1i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.or.nxv1i64(<vscale x 1 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.or.nxv1i64(<vscale x 1 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_or_nxv1i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.or.nxv1i64(<vscale x 1 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.or.nxv1i64(<vscale x 1 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.or.nxv1i64(<vscale x 1 x i64> %v)
@@ -1479,11 +1479,11 @@ declare i64 @llvm.vector.reduce.umax.nxv2i64(<vscale x 2 x i64>)
 
 define i64 @vreduce_umax_nxv2i64(<vscale x 2 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_umax_nxv2i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv2i64(<vscale x 2 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv2i64(<vscale x 2 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_umax_nxv2i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv2i64(<vscale x 2 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv2i64(<vscale x 2 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.umax.nxv2i64(<vscale x 2 x i64> %v)
@@ -1494,11 +1494,11 @@ declare i64 @llvm.vector.reduce.smax.nxv2i64(<vscale x 2 x i64>)
 
 define i64 @vreduce_smax_nxv2i64(<vscale x 2 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_smax_nxv2i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv2i64(<vscale x 2 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv2i64(<vscale x 2 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_smax_nxv2i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv2i64(<vscale x 2 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv2i64(<vscale x 2 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.smax.nxv2i64(<vscale x 2 x i64> %v)
@@ -1509,11 +1509,11 @@ declare i64 @llvm.vector.reduce.umin.nxv2i64(<vscale x 2 x i64>)
 
 define i64 @vreduce_umin_nxv2i64(<vscale x 2 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_umin_nxv2i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv2i64(<vscale x 2 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv2i64(<vscale x 2 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_umin_nxv2i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv2i64(<vscale x 2 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv2i64(<vscale x 2 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.umin.nxv2i64(<vscale x 2 x i64> %v)
@@ -1524,11 +1524,11 @@ declare i64 @llvm.vector.reduce.smin.nxv2i64(<vscale x 2 x i64>)
 
 define i64 @vreduce_smin_nxv2i64(<vscale x 2 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_smin_nxv2i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv2i64(<vscale x 2 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv2i64(<vscale x 2 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_smin_nxv2i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv2i64(<vscale x 2 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv2i64(<vscale x 2 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.smin.nxv2i64(<vscale x 2 x i64> %v)
@@ -1539,11 +1539,11 @@ declare i64 @llvm.vector.reduce.and.nxv2i64(<vscale x 2 x i64>)
 
 define i64 @vreduce_and_nxv2i64(<vscale x 2 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_and_nxv2i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i64 @llvm.vector.reduce.and.nxv2i64(<vscale x 2 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.and.nxv2i64(<vscale x 2 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_and_nxv2i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.and.nxv2i64(<vscale x 2 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.and.nxv2i64(<vscale x 2 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.and.nxv2i64(<vscale x 2 x i64> %v)
@@ -1554,11 +1554,11 @@ declare i64 @llvm.vector.reduce.or.nxv2i64(<vscale x 2 x i64>)
 
 define i64 @vreduce_or_nxv2i64(<vscale x 2 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_or_nxv2i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i64 @llvm.vector.reduce.or.nxv2i64(<vscale x 2 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.or.nxv2i64(<vscale x 2 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_or_nxv2i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.or.nxv2i64(<vscale x 2 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.or.nxv2i64(<vscale x 2 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.or.nxv2i64(<vscale x 2 x i64> %v)
@@ -1631,11 +1631,11 @@ declare i64 @llvm.vector.reduce.umax.nxv4i64(<vscale x 4 x i64>)
 
 define i64 @vreduce_umax_nxv4i64(<vscale x 4 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_umax_nxv4i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv4i64(<vscale x 4 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv4i64(<vscale x 4 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_umax_nxv4i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv4i64(<vscale x 4 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv4i64(<vscale x 4 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.umax.nxv4i64(<vscale x 4 x i64> %v)
@@ -1646,11 +1646,11 @@ declare i64 @llvm.vector.reduce.smax.nxv4i64(<vscale x 4 x i64>)
 
 define i64 @vreduce_smax_nxv4i64(<vscale x 4 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_smax_nxv4i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv4i64(<vscale x 4 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv4i64(<vscale x 4 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_smax_nxv4i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv4i64(<vscale x 4 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv4i64(<vscale x 4 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.smax.nxv4i64(<vscale x 4 x i64> %v)
@@ -1661,11 +1661,11 @@ declare i64 @llvm.vector.reduce.umin.nxv4i64(<vscale x 4 x i64>)
 
 define i64 @vreduce_umin_nxv4i64(<vscale x 4 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_umin_nxv4i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv4i64(<vscale x 4 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv4i64(<vscale x 4 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_umin_nxv4i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv4i64(<vscale x 4 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv4i64(<vscale x 4 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.umin.nxv4i64(<vscale x 4 x i64> %v)
@@ -1676,11 +1676,11 @@ declare i64 @llvm.vector.reduce.smin.nxv4i64(<vscale x 4 x i64>)
 
 define i64 @vreduce_smin_nxv4i64(<vscale x 4 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_smin_nxv4i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv4i64(<vscale x 4 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv4i64(<vscale x 4 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_smin_nxv4i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv4i64(<vscale x 4 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv4i64(<vscale x 4 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.smin.nxv4i64(<vscale x 4 x i64> %v)
@@ -1691,11 +1691,11 @@ declare i64 @llvm.vector.reduce.and.nxv4i64(<vscale x 4 x i64>)
 
 define i64 @vreduce_and_nxv4i64(<vscale x 4 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_and_nxv4i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i64 @llvm.vector.reduce.and.nxv4i64(<vscale x 4 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i64 @llvm.vector.reduce.and.nxv4i64(<vscale x 4 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_and_nxv4i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.and.nxv4i64(<vscale x 4 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.and.nxv4i64(<vscale x 4 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.and.nxv4i64(<vscale x 4 x i64> %v)
@@ -1706,11 +1706,11 @@ declare i64 @llvm.vector.reduce.or.nxv4i64(<vscale x 4 x i64>)
 
 define i64 @vreduce_or_nxv4i64(<vscale x 4 x i64> %v) {
 ; CHECK-LABEL: 'vreduce_or_nxv4i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %red = call i64 @llvm.vector.reduce.or.nxv4i64(<vscale x 4 x i64> %v)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %red = call i64 @llvm.vector.reduce.or.nxv4i64(<vscale x 4 x i64> %v)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
 ;
 ; SIZE-LABEL: 'vreduce_or_nxv4i64'
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.or.nxv4i64(<vscale x 4 x i64> %v)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.or.nxv4i64(<vscale x 4 x i64> %v)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
 ;
   %red = call i64 @llvm.vector.reduce.or.nxv4i64(<vscale x 4 x i64> %v)
diff --git a/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll b/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll
index c7cd845a0a03f8..5126a6a0a3cbcd 100644
--- a/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll
@@ -1822,35 +1822,35 @@ define void @reduce_fadd() {
 define void @reduce_other() {
 ; CHECK-LABEL: 'reduce_other'
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = call i32 @llvm.vp.reduce.xor.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %2 = call i32 @llvm.vp.reduce.and.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %3 = call i32 @llvm.vp.reduce.or.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %2 = call i32 @llvm.vp.reduce.and.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %3 = call i32 @llvm.vp.reduce.or.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %4 = call i32 @llvm.vp.reduce.mul.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %5 = call i32 @llvm.vp.reduce.smin.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %6 = call i32 @llvm.vp.reduce.smax.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %7 = call i32 @llvm.vp.reduce.umin.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %8 = call i32 @llvm.vp.reduce.umax.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %5 = call i32 @llvm.vp.reduce.smin.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %6 = call i32 @llvm.vp.reduce.smax.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %7 = call i32 @llvm.vp.reduce.umin.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %8 = call i32 @llvm.vp.reduce.umax.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
 ; CHECK-NEXT:  Cost Model: Invalid cost for instruction: %9 = call i32 @llvm.vp.reduce.mul.nxv4i32(i32 undef, <vscale x 4 x i32> undef, <vscale x 4 x i1> undef, i32 undef)
 ; CHECK-NEXT:  Cost Model: Invalid cost for instruction: %10 = call float @llvm.vp.reduce.fmul.nxv4f32(float undef, <vscale x 4 x float> undef, <vscale x 4 x i1> undef, i32 undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %11 = call float @llvm.vp.reduce.fmin.nxv4f32(float undef, <vscale x 4 x float> undef, <vscale x 4 x i1> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %11 = call float @llvm.vp.reduce.fmin.nxv4f32(float undef, <vscale x 4 x float> undef, <vscale x 4 x i1> undef, i32 undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %12 = call float @llvm.vp.reduce.fminimum.nxv4f32(float undef, <vscale x 4 x float> undef, <vscale x 4 x i1> undef, i32 undef)
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %13 = call float @llvm.vp.reduce.fmax.nxv4f32(float undef, <vscale x 4 x float> undef, <vscale x 4 x i1> undef, i32 undef)
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %13 = call float @llvm.vp.reduce.fmax.nxv4f32(float undef, <vscale x 4 x float> undef, <vscale x 4 x i1> undef, i32 undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %14 = call float @llvm.vp.reduce.fmaximum.nxv4f32(float undef, <vscale x 4 x float> undef, <vscale x 4 x i1> undef, i32 undef)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; TYPEBASED-LABEL: 'reduce_other'
 ; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = call i32 @llvm.vp.reduce.xor.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
-; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %2 = call i32 @llvm.vp.reduce.and.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
-; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %3 = call i32 @llvm.vp.reduce.or.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %2 = call i32 @llvm.vp.reduce.and.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %3 = call i32 @llvm.vp.reduce.or.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
 ; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %4 = call i32 @llvm.vp.reduce.mul.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
-; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %5 = call i32 @llvm.vp.reduce.smin.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
-; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %6 = call i32 @llvm.vp.reduce.smax.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
-; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %7 = call i32 @llvm.vp.reduce.umin.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
-; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %8 = call i32 @llvm.vp.reduce.umax.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %5 = call i32 @llvm.vp.reduce.smin.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %6 = call i32 @llvm.vp.reduce.smax.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %7 = call i32 @llvm.vp.reduce.umin.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %8 = call i32 @llvm.vp.reduce.umax.v4i32(i32 undef, <4 x i32> undef, <4 x i1> undef, i32 undef)
 ; TYPEBASED-NEXT:  Cost Model: Invalid cost for instruction: %9 = call i32 @llvm.vp.reduce.mul.nxv4i32(i32 undef, <vscale x 4 x i32> undef, <vscale x 4 x i1> undef, i32 undef)
 ; TYPEBASED-NEXT:  Cost Model: Invalid cost for instruction: %10 = call float @llvm.vp.reduce.fmul.nxv4f32(float undef, <vscale x 4 x float> undef, <vscale x 4 x i1> undef, i32 undef)
-; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %11 = call float @llvm.vp.reduce.fmin.nxv4f32(float undef, <vscale x 4 x float> undef, <vscale x 4 x i1> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %11 = call float @llvm.vp.reduce.fmin.nxv4f32(float undef, <vscale x 4 x float> undef, <vscale x 4 x i1> undef, i32 undef)
 ; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %12 = call float @llvm.vp.reduce.fminimum.nxv4f32(float undef, <vscale x 4 x float> undef, <vscale x 4 x i1> undef, i32 undef)
-; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %13 = call float @llvm.vp.reduce.fmax.nxv4f32(float undef, <vscale x 4 x float> undef, <vscale x 4 x i1> undef, i32 undef)
+; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %13 = call float @llvm.vp.reduce.fmax.nxv4f32(float undef, <vscale x 4 x float> undef, <vscale x 4 x i1> undef, i32 undef)
 ; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %14 = call float @llvm.vp.reduce.fmaximum.nxv4f32(float undef, <vscale x 4 x float> undef, <vscale x 4 x i1> undef, i32 undef)
 ; TYPEBASED-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;

>From 1588e926e5dfdbcc0227ec44c6789bbe0e97e023 Mon Sep 17 00:00:00 2001
From: LiqinWeng <liqin.weng at spacemit.com>
Date: Sun, 1 Dec 2024 11:45:01 +0800
Subject: [PATCH 2/4] fix the comments

---
 llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
index 83994b1e445f75..b3c72bb9d348e6 100644
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
@@ -1462,7 +1462,7 @@ RISCVTTIImpl::getMinMaxReductionCost(Intrinsic::ID IID, VectorType *Ty,
     return ExtraCost + getRISCVInstructionCost(Opcodes, LT.second, CostKind);
   }
 
-  // IR Reduction is composed by two vmv and one rvv reduction instruction.
+  // IR Reduction is composed by one rvv reduction instruction and vmv
   unsigned SplitOp;
   SmallVector<unsigned, 3> Opcodes;
   switch (IID) {
@@ -1554,7 +1554,9 @@ RISCVTTIImpl::getArithmeticReductionCost(unsigned Opcode, VectorType *Ty,
     }
   }
 
-  // IR Reduction is composed by two vmv and one rvv reduction instruction.
+  // IR Reduction of or/and is composed by one vmv and one rvv reduction
+  // instruction, and others is composed by two vmv and one rvv reduction
+  // instruction
   unsigned SplitOp;
   switch (ISD) {
   case ISD::ADD:

>From c3c17c92fadbdd8198dbc870167a6526630170d4 Mon Sep 17 00:00:00 2001
From: LiqinWeng <liqin.weng at spacemit.com>
Date: Mon, 2 Dec 2024 10:29:28 +0800
Subject: [PATCH 3/4] Remove the vxi1 of XOR implement

---
 llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp | 7 -------
 1 file changed, 7 deletions(-)

diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
index b3c72bb9d348e6..6d834db00f5039 100644
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
@@ -1534,13 +1534,6 @@ RISCVTTIImpl::getArithmeticReductionCost(unsigned Opcode, VectorType *Ty,
              getRISCVInstructionCost(Opcodes, LT.second, CostKind) +
              getCmpSelInstrCost(Instruction::ICmp, ElementTy, ElementTy,
                                 CmpInst::ICMP_EQ, CostKind);
-    } else if (ISD == ISD::XOR) {
-      // Example sequences:
-      //   vsetvli a0, zero, e8, mf8, ta, ma
-      //   vcpop.m a0, v0
-      //   andi a0, a0, 1
-      Opcodes = {RISCV::VCPOP_M};
-      return LT.first + getRISCVInstructionCost(Opcodes, LT.second, CostKind);
     } else {
       // Example sequences:
       //   vsetvli a0, zero, e8, mf8, ta, ma

>From 72c9cfcd19e2be069d08c4dc4207c8bebe19d3ff Mon Sep 17 00:00:00 2001
From: LiqinWeng <liqin.weng at spacemit.com>
Date: Tue, 3 Dec 2024 14:51:08 +0800
Subject: [PATCH 4/4] add cost of  llvm.vector.reduce.xor with nxvxi1 or vxi1

---
 llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
index 8064e244090e46..918b901ddcb4d9 100644
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
@@ -1548,6 +1548,15 @@ RISCVTTIImpl::getArithmeticReductionCost(unsigned Opcode, VectorType *Ty,
              getRISCVInstructionCost(RISCV::VCPOP_M, LT.second, CostKind) +
              getCmpSelInstrCost(Instruction::ICmp, ElementTy, ElementTy,
                                 CmpInst::ICMP_EQ, CostKind);
+    } else if (ISD == ISD::XOR) {
+      // Example sequences:
+      //   vsetvli a0, zero, e8, mf8, ta, ma
+      //   vmxor.mm v8, v0, v8 ; needed every time type is split
+      //   vcpop.m a0, v8
+      //   andi a0, a0, 1
+      return (LT.first - 1) *
+                 getRISCVInstructionCost(RISCV::VMXOR_MM, LT.second, CostKind) +
+             getRISCVInstructionCost(RISCV::VCPOP_M, LT.second, CostKind) + 1;
     } else {
       // Example sequences:
       //   vsetvli a0, zero, e8, mf8, ta, ma

