llvm/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-reduction-int.ll

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=riscv32 -mattr=+m,+v -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,RV32
; RUN: llc -mtriple=riscv64 -mattr=+m,+v -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,RV64

declare i8 @llvm.vector.reduce.add.v1i8(<1 x i8>)

define i8 @vreduce_add_v1i8(ptr %x) {
; CHECK-LABEL: vreduce_add_v1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.add.v1i8(<1 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.add.v2i8(<2 x i8>)

define i8 @vreduce_add_v2i8(ptr %x) {
; CHECK-LABEL: vreduce_add_v2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.add.v2i8(<2 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.add.v3i8(<3 x i8>)

define i8 @vreduce_add_v3i8(ptr %x) {
; CHECK-LABEL: vreduce_add_v3i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 3, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <3 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.add.v3i8(<3 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.add.v4i8(<4 x i8>)

define i8 @vreduce_add_v4i8(ptr %x) {
; CHECK-LABEL: vreduce_add_v4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.add.v4i8(<4 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.add.v8i8(<8 x i8>)

define i8 @vreduce_add_v8i8(ptr %x) {
; CHECK-LABEL: vreduce_add_v8i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e8, mf2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.add.v8i8(<8 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.add.v16i8(<16 x i8>)

define i8 @vreduce_add_v16i8(ptr %x) {
; CHECK-LABEL: vreduce_add_v16i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e8, m1, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.add.v16i8(<16 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.add.v32i8(<32 x i8>)

define i8 @vreduce_add_v32i8(ptr %x) {
; CHECK-LABEL: vreduce_add_v32i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e8, m2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v10, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v10
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.add.v32i8(<32 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.add.v64i8(<64 x i8>)

define i8 @vreduce_add_v64i8(ptr %x) {
; CHECK-LABEL: vreduce_add_v64i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e8, m4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v12, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v12
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.add.v64i8(<64 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.add.v128i8(<128 x i8>)

define i8 @vreduce_add_v128i8(ptr %x) {
; CHECK-LABEL: vreduce_add_v128i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v16, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.add.v128i8(<128 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.add.v256i8(<256 x i8>)

define i8 @vreduce_add_v256i8(ptr %x) {
; CHECK-LABEL: vreduce_add_v256i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle8.v v16, (a0)
; CHECK-NEXT:    vadd.vv v8, v8, v16
; CHECK-NEXT:    vmv.s.x v16, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <256 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.add.v256i8(<256 x i8> %v)
  ret i8 %red
}

declare i16 @llvm.vector.reduce.add.v1i16(<1 x i16>)

define i16 @vreduce_add_v1i16(ptr %x) {
; CHECK-LABEL: vreduce_add_v1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.add.v1i16(<1 x i16> %v)
  ret i16 %red
}

define i16 @vwreduce_add_v1i16(ptr %x) {
; CHECK-LABEL: vwreduce_add_v1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vsext.vf2 v9, v8
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %v = load <1 x i8>, ptr %x
  %e = sext <1 x i8> %v to <1 x i16>
  %red = call i16 @llvm.vector.reduce.add.v1i16(<1 x i16> %e)
  ret i16 %red
}

define i16 @vwreduce_uadd_v1i16(ptr %x) {
; CHECK-LABEL: vwreduce_uadd_v1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vzext.vf2 v9, v8
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %v = load <1 x i8>, ptr %x
  %e = zext <1 x i8> %v to <1 x i16>
  %red = call i16 @llvm.vector.reduce.add.v1i16(<1 x i16> %e)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.add.v2i16(<2 x i16>)

define i16 @vreduce_add_v2i16(ptr %x) {
; CHECK-LABEL: vreduce_add_v2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.add.v2i16(<2 x i16> %v)
  ret i16 %red
}

define i16 @vwreduce_add_v2i16(ptr %x) {
; CHECK-LABEL: vwreduce_add_v2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e16, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vsetvli zero, zero, e8, mf8, ta, ma
; CHECK-NEXT:    vwredsum.vs v8, v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e16, mf4, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i8>, ptr %x
  %e = sext <2 x i8> %v to <2 x i16>
  %red = call i16 @llvm.vector.reduce.add.v2i16(<2 x i16> %e)
  ret i16 %red
}

define i16 @vwreduce_uadd_v2i16(ptr %x) {
; CHECK-LABEL: vwreduce_uadd_v2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e16, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vsetvli zero, zero, e8, mf8, ta, ma
; CHECK-NEXT:    vwredsumu.vs v8, v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e16, mf4, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i8>, ptr %x
  %e = zext <2 x i8> %v to <2 x i16>
  %red = call i16 @llvm.vector.reduce.add.v2i16(<2 x i16> %e)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.add.v4i16(<4 x i16>)

define i16 @vreduce_add_v4i16(ptr %x) {
; CHECK-LABEL: vreduce_add_v4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.add.v4i16(<4 x i16> %v)
  ret i16 %red
}

define i16 @vwreduce_add_v4i16(ptr %x) {
; CHECK-LABEL: vwreduce_add_v4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vsetvli zero, zero, e8, mf4, ta, ma
; CHECK-NEXT:    vwredsum.vs v8, v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e16, mf2, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i8>, ptr %x
  %e = sext <4 x i8> %v to <4 x i16>
  %red = call i16 @llvm.vector.reduce.add.v4i16(<4 x i16> %e)
  ret i16 %red
}

define i16 @vwreduce_uadd_v4i16(ptr %x) {
; CHECK-LABEL: vwreduce_uadd_v4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vsetvli zero, zero, e8, mf4, ta, ma
; CHECK-NEXT:    vwredsumu.vs v8, v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e16, mf2, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i8>, ptr %x
  %e = zext <4 x i8> %v to <4 x i16>
  %red = call i16 @llvm.vector.reduce.add.v4i16(<4 x i16> %e)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.add.v8i16(<8 x i16>)

define i16 @vreduce_add_v8i16(ptr %x) {
; CHECK-LABEL: vreduce_add_v8i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> %v)
  ret i16 %red
}

define i16 @vwreduce_add_v8i16(ptr %x) {
; CHECK-LABEL: vwreduce_add_v8i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vsetvli zero, zero, e8, mf2, ta, ma
; CHECK-NEXT:    vwredsum.vs v8, v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i8>, ptr %x
  %e = sext <8 x i8> %v to <8 x i16>
  %red = call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> %e)
  ret i16 %red
}

define i16 @vwreduce_uadd_v8i16(ptr %x) {
; CHECK-LABEL: vwreduce_uadd_v8i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vsetvli zero, zero, e8, mf2, ta, ma
; CHECK-NEXT:    vwredsumu.vs v8, v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i8>, ptr %x
  %e = zext <8 x i8> %v to <8 x i16>
  %red = call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> %e)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.add.v16i16(<16 x i16>)

define i16 @vreduce_add_v16i16(ptr %x) {
; CHECK-LABEL: vreduce_add_v16i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e16, m2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v10, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v10
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.add.v16i16(<16 x i16> %v)
  ret i16 %red
}

define i16 @vwreduce_add_v16i16(ptr %x) {
; CHECK-LABEL: vwreduce_add_v16i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e16, m2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vsetvli zero, zero, e8, m1, ta, ma
; CHECK-NEXT:    vwredsum.vs v8, v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i8>, ptr %x
  %e = sext <16 x i8> %v to <16 x i16>
  %red = call i16 @llvm.vector.reduce.add.v16i16(<16 x i16> %e)
  ret i16 %red
}

define i16 @vwreduce_uadd_v16i16(ptr %x) {
; CHECK-LABEL: vwreduce_uadd_v16i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e16, m2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vsetvli zero, zero, e8, m1, ta, ma
; CHECK-NEXT:    vwredsumu.vs v8, v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i8>, ptr %x
  %e = zext <16 x i8> %v to <16 x i16>
  %red = call i16 @llvm.vector.reduce.add.v16i16(<16 x i16> %e)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.add.v32i16(<32 x i16>)

define i16 @vreduce_add_v32i16(ptr %x) {
; CHECK-LABEL: vreduce_add_v32i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e16, m4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v12, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v12
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.add.v32i16(<32 x i16> %v)
  ret i16 %red
}

define i16 @vwreduce_add_v32i16(ptr %x) {
; CHECK-LABEL: vwreduce_add_v32i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e16, m4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v10, zero
; CHECK-NEXT:    vsetvli zero, zero, e8, m2, ta, ma
; CHECK-NEXT:    vwredsum.vs v8, v8, v10
; CHECK-NEXT:    vsetvli zero, zero, e16, m4, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i8>, ptr %x
  %e = sext <32 x i8> %v to <32 x i16>
  %red = call i16 @llvm.vector.reduce.add.v32i16(<32 x i16> %e)
  ret i16 %red
}

define i16 @vwreduce_uadd_v32i16(ptr %x) {
; CHECK-LABEL: vwreduce_uadd_v32i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e16, m4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v10, zero
; CHECK-NEXT:    vsetvli zero, zero, e8, m2, ta, ma
; CHECK-NEXT:    vwredsumu.vs v8, v8, v10
; CHECK-NEXT:    vsetvli zero, zero, e16, m4, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i8>, ptr %x
  %e = zext <32 x i8> %v to <32 x i16>
  %red = call i16 @llvm.vector.reduce.add.v32i16(<32 x i16> %e)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.add.v64i16(<64 x i16>)

define i16 @vreduce_add_v64i16(ptr %x) {
; CHECK-LABEL: vreduce_add_v64i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v16, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.add.v64i16(<64 x i16> %v)
  ret i16 %red
}

define i16 @vwreduce_add_v64i16(ptr %x) {
; CHECK-LABEL: vwreduce_add_v64i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v12, zero
; CHECK-NEXT:    vsetvli zero, zero, e8, m4, ta, ma
; CHECK-NEXT:    vwredsum.vs v8, v8, v12
; CHECK-NEXT:    vsetvli zero, zero, e16, m8, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i8>, ptr %x
  %e = sext <64 x i8> %v to <64 x i16>
  %red = call i16 @llvm.vector.reduce.add.v64i16(<64 x i16> %e)
  ret i16 %red
}

define i16 @vwreduce_uadd_v64i16(ptr %x) {
; CHECK-LABEL: vwreduce_uadd_v64i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v12, zero
; CHECK-NEXT:    vsetvli zero, zero, e8, m4, ta, ma
; CHECK-NEXT:    vwredsumu.vs v8, v8, v12
; CHECK-NEXT:    vsetvli zero, zero, e16, m8, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i8>, ptr %x
  %e = zext <64 x i8> %v to <64 x i16>
  %red = call i16 @llvm.vector.reduce.add.v64i16(<64 x i16> %e)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.add.v128i16(<128 x i16>)

define i16 @vreduce_add_v128i16(ptr %x) {
; CHECK-LABEL: vreduce_add_v128i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle16.v v16, (a0)
; CHECK-NEXT:    vadd.vv v8, v8, v16
; CHECK-NEXT:    vmv.s.x v16, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.add.v128i16(<128 x i16> %v)
  ret i16 %red
}

define i16 @vwreduce_add_v128i16(ptr %x) {
; CHECK-LABEL: vwreduce_add_v128i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    li a0, 64
; CHECK-NEXT:    vsetvli zero, a0, e8, m8, ta, ma
; CHECK-NEXT:    vslidedown.vx v16, v8, a0
; CHECK-NEXT:    vsetvli zero, a0, e8, m4, ta, ma
; CHECK-NEXT:    vwadd.vv v24, v8, v16
; CHECK-NEXT:    vsetvli zero, zero, e16, m8, ta, ma
; CHECK-NEXT:    vmv.s.x v8, zero
; CHECK-NEXT:    vredsum.vs v8, v24, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i8>, ptr %x
  %e = sext <128 x i8> %v to <128 x i16>
  %red = call i16 @llvm.vector.reduce.add.v128i16(<128 x i16> %e)
  ret i16 %red
}

define i16 @vwreduce_uadd_v128i16(ptr %x) {
; CHECK-LABEL: vwreduce_uadd_v128i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    li a0, 64
; CHECK-NEXT:    vsetvli zero, a0, e8, m8, ta, ma
; CHECK-NEXT:    vslidedown.vx v16, v8, a0
; CHECK-NEXT:    vsetvli zero, a0, e8, m4, ta, ma
; CHECK-NEXT:    vwaddu.vv v24, v8, v16
; CHECK-NEXT:    vsetvli zero, zero, e16, m8, ta, ma
; CHECK-NEXT:    vmv.s.x v8, zero
; CHECK-NEXT:    vredsum.vs v8, v24, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i8>, ptr %x
  %e = zext <128 x i8> %v to <128 x i16>
  %red = call i16 @llvm.vector.reduce.add.v128i16(<128 x i16> %e)
  ret i16 %red
}

declare i32 @llvm.vector.reduce.add.v1i32(<1 x i32>)

define i32 @vreduce_add_v1i32(ptr %x) {
; CHECK-LABEL: vreduce_add_v1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.add.v1i32(<1 x i32> %v)
  ret i32 %red
}

define i32 @vwreduce_add_v1i32(ptr %x) {
; CHECK-LABEL: vwreduce_add_v1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vsext.vf2 v9, v8
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %v = load <1 x i16>, ptr %x
  %e = sext <1 x i16> %v to <1 x i32>
  %red = call i32 @llvm.vector.reduce.add.v1i32(<1 x i32> %e)
  ret i32 %red
}

define i32 @vwreduce_uadd_v1i32(ptr %x) {
; CHECK-LABEL: vwreduce_uadd_v1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vzext.vf2 v9, v8
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %v = load <1 x i16>, ptr %x
  %e = zext <1 x i16> %v to <1 x i32>
  %red = call i32 @llvm.vector.reduce.add.v1i32(<1 x i32> %e)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.add.v2i32(<2 x i32>)

define i32 @vreduce_add_v2i32(ptr %x) {
; CHECK-LABEL: vreduce_add_v2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.add.v2i32(<2 x i32> %v)
  ret i32 %red
}

define i32 @vwreduce_add_v2i32(ptr %x) {
; CHECK-LABEL: vwreduce_add_v2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e32, mf2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vsetvli zero, zero, e16, mf4, ta, ma
; CHECK-NEXT:    vwredsum.vs v8, v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i16>, ptr %x
  %e = sext <2 x i16> %v to <2 x i32>
  %red = call i32 @llvm.vector.reduce.add.v2i32(<2 x i32> %e)
  ret i32 %red
}

define i32 @vwreduce_uadd_v2i32(ptr %x) {
; CHECK-LABEL: vwreduce_uadd_v2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e32, mf2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vsetvli zero, zero, e16, mf4, ta, ma
; CHECK-NEXT:    vwredsumu.vs v8, v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i16>, ptr %x
  %e = zext <2 x i16> %v to <2 x i32>
  %red = call i32 @llvm.vector.reduce.add.v2i32(<2 x i32> %e)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.add.v4i32(<4 x i32>)

define i32 @vreduce_add_v4i32(ptr %x) {
; CHECK-LABEL: vreduce_add_v4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %v)
  ret i32 %red
}

define i32 @vwreduce_add_v4i32(ptr %x) {
; CHECK-LABEL: vwreduce_add_v4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vsetvli zero, zero, e16, mf2, ta, ma
; CHECK-NEXT:    vwredsum.vs v8, v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i16>, ptr %x
  %e = sext <4 x i16> %v to <4 x i32>
  %red = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %e)
  ret i32 %red
}

define i32 @vwreduce_uadd_v4i32(ptr %x) {
; CHECK-LABEL: vwreduce_uadd_v4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vsetvli zero, zero, e16, mf2, ta, ma
; CHECK-NEXT:    vwredsumu.vs v8, v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i16>, ptr %x
  %e = zext <4 x i16> %v to <4 x i32>
  %red = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %e)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.add.v8i32(<8 x i32>)

define i32 @vreduce_add_v8i32(ptr %x) {
; CHECK-LABEL: vreduce_add_v8i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v10, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v10
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %v)
  ret i32 %red
}

define i32 @vwreduce_add_v8i32(ptr %x) {
; CHECK-LABEL: vwreduce_add_v8i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
; CHECK-NEXT:    vwredsum.vs v8, v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i16>, ptr %x
  %e = sext <8 x i16> %v to <8 x i32>
  %red = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %e)
  ret i32 %red
}

define i32 @vwreduce_uadd_v8i32(ptr %x) {
; CHECK-LABEL: vwreduce_uadd_v8i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
; CHECK-NEXT:    vwredsumu.vs v8, v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i16>, ptr %x
  %e = zext <8 x i16> %v to <8 x i32>
  %red = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %e)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.add.v16i32(<16 x i32>)

define i32 @vreduce_add_v16i32(ptr %x) {
; CHECK-LABEL: vreduce_add_v16i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v12, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v12
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %v)
  ret i32 %red
}

define i32 @vwreduce_add_v16i32(ptr %x) {
; CHECK-LABEL: vwreduce_add_v16i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v10, zero
; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
; CHECK-NEXT:    vwredsum.vs v8, v8, v10
; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i16>, ptr %x
  %e = sext <16 x i16> %v to <16 x i32>
  %red = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %e)
  ret i32 %red
}

define i32 @vwreduce_uadd_v16i32(ptr %x) {
; CHECK-LABEL: vwreduce_uadd_v16i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v10, zero
; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
; CHECK-NEXT:    vwredsumu.vs v8, v8, v10
; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i16>, ptr %x
  %e = zext <16 x i16> %v to <16 x i32>
  %red = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %e)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.add.v32i32(<32 x i32>)

define i32 @vreduce_add_v32i32(ptr %x) {
; CHECK-LABEL: vreduce_add_v32i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v16, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.add.v32i32(<32 x i32> %v)
  ret i32 %red
}

define i32 @vwreduce_add_v32i32(ptr %x) {
; CHECK-LABEL: vwreduce_add_v32i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v12, zero
; CHECK-NEXT:    vsetvli zero, zero, e16, m4, ta, ma
; CHECK-NEXT:    vwredsum.vs v8, v8, v12
; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i16>, ptr %x
  %e = sext <32 x i16> %v to <32 x i32>
  %red = call i32 @llvm.vector.reduce.add.v32i32(<32 x i32> %e)
  ret i32 %red
}

define i32 @vwreduce_uadd_v32i32(ptr %x) {
; CHECK-LABEL: vwreduce_uadd_v32i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v12, zero
; CHECK-NEXT:    vsetvli zero, zero, e16, m4, ta, ma
; CHECK-NEXT:    vwredsumu.vs v8, v8, v12
; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i16>, ptr %x
  %e = zext <32 x i16> %v to <32 x i32>
  %red = call i32 @llvm.vector.reduce.add.v32i32(<32 x i32> %e)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.add.v64i32(<64 x i32>)

define i32 @vreduce_add_v64i32(ptr %x) {
; CHECK-LABEL: vreduce_add_v64i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle32.v v16, (a0)
; CHECK-NEXT:    vadd.vv v8, v8, v16
; CHECK-NEXT:    vmv.s.x v16, zero
; CHECK-NEXT:    vredsum.vs v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.add.v64i32(<64 x i32> %v)
  ret i32 %red
}

define i32 @vwreduce_add_v64i32(ptr %x) {
; CHECK-LABEL: vwreduce_add_v64i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    li a0, 32
; CHECK-NEXT:    vsetvli zero, a0, e16, m8, ta, ma
; CHECK-NEXT:    vslidedown.vx v16, v8, a0
; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
; CHECK-NEXT:    vwadd.vv v24, v8, v16
; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
; CHECK-NEXT:    vmv.s.x v8, zero
; CHECK-NEXT:    vredsum.vs v8, v24, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i16>, ptr %x
  %e = sext <64 x i16> %v to <64 x i32>
  %red = call i32 @llvm.vector.reduce.add.v64i32(<64 x i32> %e)
  ret i32 %red
}

define i32 @vwreduce_uadd_v64i32(ptr %x) {
; CHECK-LABEL: vwreduce_uadd_v64i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    li a0, 32
; CHECK-NEXT:    vsetvli zero, a0, e16, m8, ta, ma
; CHECK-NEXT:    vslidedown.vx v16, v8, a0
; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
; CHECK-NEXT:    vwaddu.vv v24, v8, v16
; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
; CHECK-NEXT:    vmv.s.x v8, zero
; CHECK-NEXT:    vredsum.vs v8, v24, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i16>, ptr %x
  %e = zext <64 x i16> %v to <64 x i32>
  %red = call i32 @llvm.vector.reduce.add.v64i32(<64 x i32> %e)
  ret i32 %red
}

declare i64 @llvm.vector.reduce.add.v1i64(<1 x i64>)

define i64 @vreduce_add_v1i64(ptr %x) {
; RV32-LABEL: vreduce_add_v1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsrl.vx v9, v8, a0
; RV32-NEXT:    vmv.x.s a1, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_add_v1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <1 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.add.v1i64(<1 x i64> %v)
  ret i64 %red
}

define i64 @vwreduce_add_v1i64(ptr %x) {
; RV32-LABEL: vwreduce_add_v1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vle32.v v8, (a0)
; RV32-NEXT:    vsext.vf2 v9, v8
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsrl.vx v8, v9, a0
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    ret
;
; RV64-LABEL: vwreduce_add_v1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vle32.v v8, (a0)
; RV64-NEXT:    vsext.vf2 v9, v8
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %v = load <1 x i32>, ptr %x
  %e = sext <1 x i32> %v to <1 x i64>
  %red = call i64 @llvm.vector.reduce.add.v1i64(<1 x i64> %e)
  ret i64 %red
}

define i64 @vwreduce_uadd_v1i64(ptr %x) {
; RV32-LABEL: vwreduce_uadd_v1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vle32.v v8, (a0)
; RV32-NEXT:    vzext.vf2 v9, v8
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsrl.vx v8, v9, a0
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    ret
;
; RV64-LABEL: vwreduce_uadd_v1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vle32.v v8, (a0)
; RV64-NEXT:    vzext.vf2 v9, v8
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %v = load <1 x i32>, ptr %x
  %e = zext <1 x i32> %v to <1 x i64>
  %red = call i64 @llvm.vector.reduce.add.v1i64(<1 x i64> %e)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.add.v2i64(<2 x i64>)

define i64 @vreduce_add_v2i64(ptr %x) {
; RV32-LABEL: vreduce_add_v2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vmv.s.x v9, zero
; RV32-NEXT:    vredsum.vs v8, v8, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_add_v2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.s.x v9, zero
; RV64-NEXT:    vredsum.vs v8, v8, v9
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <2 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %v)
  ret i64 %red
}

define i64 @vwreduce_add_v2i64(ptr %x) {
; RV32-LABEL: vwreduce_add_v2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV32-NEXT:    vle32.v v8, (a0)
; RV32-NEXT:    vmv.s.x v9, zero
; RV32-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
; RV32-NEXT:    vwredsum.vs v8, v8, v9
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vwreduce_add_v2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV64-NEXT:    vle32.v v8, (a0)
; RV64-NEXT:    vmv.s.x v9, zero
; RV64-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
; RV64-NEXT:    vwredsum.vs v8, v8, v9
; RV64-NEXT:    vsetvli zero, zero, e64, m1, ta, ma
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <2 x i32>, ptr %x
  %e = sext <2 x i32> %v to <2 x i64>
  %red = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %e)
  ret i64 %red
}

define i64 @vwreduce_uadd_v2i64(ptr %x) {
; RV32-LABEL: vwreduce_uadd_v2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV32-NEXT:    vle32.v v8, (a0)
; RV32-NEXT:    vmv.s.x v9, zero
; RV32-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
; RV32-NEXT:    vwredsumu.vs v8, v8, v9
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vwreduce_uadd_v2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV64-NEXT:    vle32.v v8, (a0)
; RV64-NEXT:    vmv.s.x v9, zero
; RV64-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
; RV64-NEXT:    vwredsumu.vs v8, v8, v9
; RV64-NEXT:    vsetvli zero, zero, e64, m1, ta, ma
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <2 x i32>, ptr %x
  %e = zext <2 x i32> %v to <2 x i64>
  %red = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %e)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.add.v4i64(<4 x i64>)

define i64 @vreduce_add_v4i64(ptr %x) {
; RV32-LABEL: vreduce_add_v4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vmv.s.x v10, zero
; RV32-NEXT:    vredsum.vs v8, v8, v10
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_add_v4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.s.x v10, zero
; RV64-NEXT:    vredsum.vs v8, v8, v10
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <4 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %v)
  ret i64 %red
}

define i64 @vwreduce_add_v4i64(ptr %x) {
; RV32-LABEL: vwreduce_add_v4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV32-NEXT:    vle32.v v8, (a0)
; RV32-NEXT:    vmv.s.x v9, zero
; RV32-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
; RV32-NEXT:    vwredsum.vs v8, v8, v9
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vwreduce_add_v4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV64-NEXT:    vle32.v v8, (a0)
; RV64-NEXT:    vmv.s.x v9, zero
; RV64-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
; RV64-NEXT:    vwredsum.vs v8, v8, v9
; RV64-NEXT:    vsetvli zero, zero, e64, m2, ta, ma
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <4 x i32>, ptr %x
  %e = sext <4 x i32> %v to <4 x i64>
  %red = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %e)
  ret i64 %red
}

define i64 @vwreduce_uadd_v4i64(ptr %x) {
; RV32-LABEL: vwreduce_uadd_v4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV32-NEXT:    vle32.v v8, (a0)
; RV32-NEXT:    vmv.s.x v9, zero
; RV32-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
; RV32-NEXT:    vwredsumu.vs v8, v8, v9
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vwreduce_uadd_v4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV64-NEXT:    vle32.v v8, (a0)
; RV64-NEXT:    vmv.s.x v9, zero
; RV64-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
; RV64-NEXT:    vwredsumu.vs v8, v8, v9
; RV64-NEXT:    vsetvli zero, zero, e64, m2, ta, ma
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <4 x i32>, ptr %x
  %e = zext <4 x i32> %v to <4 x i64>
  %red = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %e)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.add.v8i64(<8 x i64>)

define i64 @vreduce_add_v8i64(ptr %x) {
; RV32-LABEL: vreduce_add_v8i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vmv.s.x v12, zero
; RV32-NEXT:    vredsum.vs v8, v8, v12
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_add_v8i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.s.x v12, zero
; RV64-NEXT:    vredsum.vs v8, v8, v12
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <8 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %v)
  ret i64 %red
}

define i64 @vwreduce_add_v8i64(ptr %x) {
; RV32-LABEL: vwreduce_add_v8i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV32-NEXT:    vle32.v v8, (a0)
; RV32-NEXT:    vmv.s.x v10, zero
; RV32-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
; RV32-NEXT:    vwredsum.vs v8, v8, v10
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vwreduce_add_v8i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV64-NEXT:    vle32.v v8, (a0)
; RV64-NEXT:    vmv.s.x v10, zero
; RV64-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
; RV64-NEXT:    vwredsum.vs v8, v8, v10
; RV64-NEXT:    vsetvli zero, zero, e64, m4, ta, ma
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <8 x i32>, ptr %x
  %e = sext <8 x i32> %v to <8 x i64>
  %red = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %e)
  ret i64 %red
}

define i64 @vwreduce_uadd_v8i64(ptr %x) {
; RV32-LABEL: vwreduce_uadd_v8i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV32-NEXT:    vle32.v v8, (a0)
; RV32-NEXT:    vmv.s.x v10, zero
; RV32-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
; RV32-NEXT:    vwredsumu.vs v8, v8, v10
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vwreduce_uadd_v8i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV64-NEXT:    vle32.v v8, (a0)
; RV64-NEXT:    vmv.s.x v10, zero
; RV64-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
; RV64-NEXT:    vwredsumu.vs v8, v8, v10
; RV64-NEXT:    vsetvli zero, zero, e64, m4, ta, ma
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <8 x i32>, ptr %x
  %e = zext <8 x i32> %v to <8 x i64>
  %red = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %e)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.add.v16i64(<16 x i64>)

define i64 @vreduce_add_v16i64(ptr %x) {
; RV32-LABEL: vreduce_add_v16i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vmv.s.x v16, zero
; RV32-NEXT:    vredsum.vs v8, v8, v16
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_add_v16i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.s.x v16, zero
; RV64-NEXT:    vredsum.vs v8, v8, v16
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <16 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.add.v16i64(<16 x i64> %v)
  ret i64 %red
}

define i64 @vwreduce_add_v16i64(ptr %x) {
; RV32-LABEL: vwreduce_add_v16i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle32.v v8, (a0)
; RV32-NEXT:    vmv.s.x v12, zero
; RV32-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
; RV32-NEXT:    vwredsum.vs v8, v8, v12
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vwreduce_add_v16i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle32.v v8, (a0)
; RV64-NEXT:    vmv.s.x v12, zero
; RV64-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
; RV64-NEXT:    vwredsum.vs v8, v8, v12
; RV64-NEXT:    vsetvli zero, zero, e64, m8, ta, ma
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <16 x i32>, ptr %x
  %e = sext <16 x i32> %v to <16 x i64>
  %red = call i64 @llvm.vector.reduce.add.v16i64(<16 x i64> %e)
  ret i64 %red
}

define i64 @vwreduce_uadd_v16i64(ptr %x) {
; RV32-LABEL: vwreduce_uadd_v16i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle32.v v8, (a0)
; RV32-NEXT:    vmv.s.x v12, zero
; RV32-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
; RV32-NEXT:    vwredsumu.vs v8, v8, v12
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vwreduce_uadd_v16i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle32.v v8, (a0)
; RV64-NEXT:    vmv.s.x v12, zero
; RV64-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
; RV64-NEXT:    vwredsumu.vs v8, v8, v12
; RV64-NEXT:    vsetvli zero, zero, e64, m8, ta, ma
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <16 x i32>, ptr %x
  %e = zext <16 x i32> %v to <16 x i64>
  %red = call i64 @llvm.vector.reduce.add.v16i64(<16 x i64> %e)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.add.v32i64(<32 x i64>)

define i64 @vreduce_add_v32i64(ptr %x) {
; RV32-LABEL: vreduce_add_v32i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v16, (a0)
; RV32-NEXT:    vadd.vv v8, v8, v16
; RV32-NEXT:    vmv.s.x v16, zero
; RV32-NEXT:    vredsum.vs v8, v8, v16
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_add_v32i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v16, (a0)
; RV64-NEXT:    vadd.vv v8, v8, v16
; RV64-NEXT:    vmv.s.x v16, zero
; RV64-NEXT:    vredsum.vs v8, v8, v16
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <32 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.add.v32i64(<32 x i64> %v)
  ret i64 %red
}

define i64 @vwreduce_add_v32i64(ptr %x) {
; RV32-LABEL: vwreduce_add_v32i64:
; RV32:       # %bb.0:
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; RV32-NEXT:    vle32.v v8, (a0)
; RV32-NEXT:    vsetivli zero, 16, e32, m8, ta, ma
; RV32-NEXT:    vslidedown.vi v16, v8, 16
; RV32-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; RV32-NEXT:    vwadd.vv v24, v8, v16
; RV32-NEXT:    vsetvli zero, zero, e64, m8, ta, ma
; RV32-NEXT:    vmv.s.x v8, zero
; RV32-NEXT:    vredsum.vs v8, v24, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vwreduce_add_v32i64:
; RV64:       # %bb.0:
; RV64-NEXT:    li a1, 32
; RV64-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; RV64-NEXT:    vle32.v v8, (a0)
; RV64-NEXT:    vsetivli zero, 16, e32, m8, ta, ma
; RV64-NEXT:    vslidedown.vi v16, v8, 16
; RV64-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; RV64-NEXT:    vwadd.vv v24, v8, v16
; RV64-NEXT:    vsetvli zero, zero, e64, m8, ta, ma
; RV64-NEXT:    vmv.s.x v8, zero
; RV64-NEXT:    vredsum.vs v8, v24, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <32 x i32>, ptr %x
  %e = sext <32 x i32> %v to <32 x i64>
  %red = call i64 @llvm.vector.reduce.add.v32i64(<32 x i64> %e)
  ret i64 %red
}

define i64 @vwreduce_uadd_v32i64(ptr %x) {
; RV32-LABEL: vwreduce_uadd_v32i64:
; RV32:       # %bb.0:
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; RV32-NEXT:    vle32.v v8, (a0)
; RV32-NEXT:    vsetivli zero, 16, e32, m8, ta, ma
; RV32-NEXT:    vslidedown.vi v16, v8, 16
; RV32-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; RV32-NEXT:    vwaddu.vv v24, v8, v16
; RV32-NEXT:    vsetvli zero, zero, e64, m8, ta, ma
; RV32-NEXT:    vmv.s.x v8, zero
; RV32-NEXT:    vredsum.vs v8, v24, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vwreduce_uadd_v32i64:
; RV64:       # %bb.0:
; RV64-NEXT:    li a1, 32
; RV64-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; RV64-NEXT:    vle32.v v8, (a0)
; RV64-NEXT:    vsetivli zero, 16, e32, m8, ta, ma
; RV64-NEXT:    vslidedown.vi v16, v8, 16
; RV64-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; RV64-NEXT:    vwaddu.vv v24, v8, v16
; RV64-NEXT:    vsetvli zero, zero, e64, m8, ta, ma
; RV64-NEXT:    vmv.s.x v8, zero
; RV64-NEXT:    vredsum.vs v8, v24, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <32 x i32>, ptr %x
  %e = zext <32 x i32> %v to <32 x i64>
  %red = call i64 @llvm.vector.reduce.add.v32i64(<32 x i64> %e)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.add.v64i64(<64 x i64>)

define i64 @vreduce_add_v64i64(ptr %x) nounwind {
; RV32-LABEL: vreduce_add_v64i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a1, a0, 384
; RV32-NEXT:    vle64.v v16, (a1)
; RV32-NEXT:    addi a1, a0, 256
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v24, (a0)
; RV32-NEXT:    vle64.v v0, (a1)
; RV32-NEXT:    vadd.vv v16, v24, v16
; RV32-NEXT:    vadd.vv v8, v8, v0
; RV32-NEXT:    vadd.vv v8, v8, v16
; RV32-NEXT:    vmv.s.x v16, zero
; RV32-NEXT:    vredsum.vs v8, v8, v16
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_add_v64i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a1, a0, 384
; RV64-NEXT:    vle64.v v16, (a1)
; RV64-NEXT:    addi a1, a0, 256
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v24, (a0)
; RV64-NEXT:    vle64.v v0, (a1)
; RV64-NEXT:    vadd.vv v16, v24, v16
; RV64-NEXT:    vadd.vv v8, v8, v0
; RV64-NEXT:    vadd.vv v8, v8, v16
; RV64-NEXT:    vmv.s.x v16, zero
; RV64-NEXT:    vredsum.vs v8, v8, v16
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <64 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.add.v64i64(<64 x i64> %v)
  ret i64 %red
}

define i64 @vwreduce_add_v64i64(ptr %x) {
; RV32-LABEL: vwreduce_add_v64i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    csrr a1, vlenb
; RV32-NEXT:    slli a1, a1, 4
; RV32-NEXT:    sub sp, sp, a1
; RV32-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x10, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 16 * vlenb
; RV32-NEXT:    addi a1, a0, 128
; RV32-NEXT:    li a2, 32
; RV32-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
; RV32-NEXT:    vle32.v v8, (a0)
; RV32-NEXT:    addi a0, sp, 16
; RV32-NEXT:    vs8r.v v8, (a0) # Unknown-size Folded Spill
; RV32-NEXT:    vle32.v v16, (a1)
; RV32-NEXT:    vsetivli zero, 16, e32, m8, ta, ma
; RV32-NEXT:    vslidedown.vi v24, v8, 16
; RV32-NEXT:    vslidedown.vi v0, v16, 16
; RV32-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; RV32-NEXT:    vwadd.vv v8, v24, v0
; RV32-NEXT:    csrr a0, vlenb
; RV32-NEXT:    slli a0, a0, 3
; RV32-NEXT:    add a0, sp, a0
; RV32-NEXT:    addi a0, a0, 16
; RV32-NEXT:    vs8r.v v8, (a0) # Unknown-size Folded Spill
; RV32-NEXT:    addi a0, sp, 16
; RV32-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
; RV32-NEXT:    vwadd.vv v0, v8, v16
; RV32-NEXT:    csrr a0, vlenb
; RV32-NEXT:    slli a0, a0, 3
; RV32-NEXT:    add a0, sp, a0
; RV32-NEXT:    addi a0, a0, 16
; RV32-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
; RV32-NEXT:    vsetvli zero, zero, e64, m8, ta, ma
; RV32-NEXT:    vadd.vv v8, v0, v8
; RV32-NEXT:    vmv.s.x v16, zero
; RV32-NEXT:    vredsum.vs v8, v8, v16
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a2
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    csrr a2, vlenb
; RV32-NEXT:    slli a2, a2, 4
; RV32-NEXT:    add sp, sp, a2
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vwreduce_add_v64i64:
; RV64:       # %bb.0:
; RV64-NEXT:    addi sp, sp, -16
; RV64-NEXT:    .cfi_def_cfa_offset 16
; RV64-NEXT:    csrr a1, vlenb
; RV64-NEXT:    slli a1, a1, 4
; RV64-NEXT:    sub sp, sp, a1
; RV64-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x10, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 16 * vlenb
; RV64-NEXT:    addi a1, a0, 128
; RV64-NEXT:    li a2, 32
; RV64-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
; RV64-NEXT:    vle32.v v8, (a0)
; RV64-NEXT:    addi a0, sp, 16
; RV64-NEXT:    vs8r.v v8, (a0) # Unknown-size Folded Spill
; RV64-NEXT:    vle32.v v16, (a1)
; RV64-NEXT:    vsetivli zero, 16, e32, m8, ta, ma
; RV64-NEXT:    vslidedown.vi v24, v8, 16
; RV64-NEXT:    vslidedown.vi v0, v16, 16
; RV64-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; RV64-NEXT:    vwadd.vv v8, v24, v0
; RV64-NEXT:    csrr a0, vlenb
; RV64-NEXT:    slli a0, a0, 3
; RV64-NEXT:    add a0, sp, a0
; RV64-NEXT:    addi a0, a0, 16
; RV64-NEXT:    vs8r.v v8, (a0) # Unknown-size Folded Spill
; RV64-NEXT:    addi a0, sp, 16
; RV64-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
; RV64-NEXT:    vwadd.vv v0, v8, v16
; RV64-NEXT:    csrr a0, vlenb
; RV64-NEXT:    slli a0, a0, 3
; RV64-NEXT:    add a0, sp, a0
; RV64-NEXT:    addi a0, a0, 16
; RV64-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
; RV64-NEXT:    vsetvli zero, zero, e64, m8, ta, ma
; RV64-NEXT:    vadd.vv v8, v0, v8
; RV64-NEXT:    vmv.s.x v16, zero
; RV64-NEXT:    vredsum.vs v8, v8, v16
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    csrr a1, vlenb
; RV64-NEXT:    slli a1, a1, 4
; RV64-NEXT:    add sp, sp, a1
; RV64-NEXT:    addi sp, sp, 16
; RV64-NEXT:    ret
  %v = load <64 x i32>, ptr %x
  %e = sext <64 x i32> %v to <64 x i64>
  %red = call i64 @llvm.vector.reduce.add.v64i64(<64 x i64> %e)
  ret i64 %red
}

define i64 @vwreduce_uadd_v64i64(ptr %x) {
; RV32-LABEL: vwreduce_uadd_v64i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    csrr a1, vlenb
; RV32-NEXT:    slli a1, a1, 4
; RV32-NEXT:    sub sp, sp, a1
; RV32-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x10, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 16 * vlenb
; RV32-NEXT:    addi a1, a0, 128
; RV32-NEXT:    li a2, 32
; RV32-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
; RV32-NEXT:    vle32.v v8, (a0)
; RV32-NEXT:    addi a0, sp, 16
; RV32-NEXT:    vs8r.v v8, (a0) # Unknown-size Folded Spill
; RV32-NEXT:    vle32.v v16, (a1)
; RV32-NEXT:    vsetivli zero, 16, e32, m8, ta, ma
; RV32-NEXT:    vslidedown.vi v24, v8, 16
; RV32-NEXT:    vslidedown.vi v0, v16, 16
; RV32-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; RV32-NEXT:    vwaddu.vv v8, v24, v0
; RV32-NEXT:    csrr a0, vlenb
; RV32-NEXT:    slli a0, a0, 3
; RV32-NEXT:    add a0, sp, a0
; RV32-NEXT:    addi a0, a0, 16
; RV32-NEXT:    vs8r.v v8, (a0) # Unknown-size Folded Spill
; RV32-NEXT:    addi a0, sp, 16
; RV32-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
; RV32-NEXT:    vwaddu.vv v0, v8, v16
; RV32-NEXT:    csrr a0, vlenb
; RV32-NEXT:    slli a0, a0, 3
; RV32-NEXT:    add a0, sp, a0
; RV32-NEXT:    addi a0, a0, 16
; RV32-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
; RV32-NEXT:    vsetvli zero, zero, e64, m8, ta, ma
; RV32-NEXT:    vadd.vv v8, v0, v8
; RV32-NEXT:    vmv.s.x v16, zero
; RV32-NEXT:    vredsum.vs v8, v8, v16
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a2
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    csrr a2, vlenb
; RV32-NEXT:    slli a2, a2, 4
; RV32-NEXT:    add sp, sp, a2
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vwreduce_uadd_v64i64:
; RV64:       # %bb.0:
; RV64-NEXT:    addi sp, sp, -16
; RV64-NEXT:    .cfi_def_cfa_offset 16
; RV64-NEXT:    csrr a1, vlenb
; RV64-NEXT:    slli a1, a1, 4
; RV64-NEXT:    sub sp, sp, a1
; RV64-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x10, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 16 * vlenb
; RV64-NEXT:    addi a1, a0, 128
; RV64-NEXT:    li a2, 32
; RV64-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
; RV64-NEXT:    vle32.v v8, (a0)
; RV64-NEXT:    addi a0, sp, 16
; RV64-NEXT:    vs8r.v v8, (a0) # Unknown-size Folded Spill
; RV64-NEXT:    vle32.v v16, (a1)
; RV64-NEXT:    vsetivli zero, 16, e32, m8, ta, ma
; RV64-NEXT:    vslidedown.vi v24, v8, 16
; RV64-NEXT:    vslidedown.vi v0, v16, 16
; RV64-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; RV64-NEXT:    vwaddu.vv v8, v24, v0
; RV64-NEXT:    csrr a0, vlenb
; RV64-NEXT:    slli a0, a0, 3
; RV64-NEXT:    add a0, sp, a0
; RV64-NEXT:    addi a0, a0, 16
; RV64-NEXT:    vs8r.v v8, (a0) # Unknown-size Folded Spill
; RV64-NEXT:    addi a0, sp, 16
; RV64-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
; RV64-NEXT:    vwaddu.vv v0, v8, v16
; RV64-NEXT:    csrr a0, vlenb
; RV64-NEXT:    slli a0, a0, 3
; RV64-NEXT:    add a0, sp, a0
; RV64-NEXT:    addi a0, a0, 16
; RV64-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
; RV64-NEXT:    vsetvli zero, zero, e64, m8, ta, ma
; RV64-NEXT:    vadd.vv v8, v0, v8
; RV64-NEXT:    vmv.s.x v16, zero
; RV64-NEXT:    vredsum.vs v8, v8, v16
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    csrr a1, vlenb
; RV64-NEXT:    slli a1, a1, 4
; RV64-NEXT:    add sp, sp, a1
; RV64-NEXT:    addi sp, sp, 16
; RV64-NEXT:    ret
  %v = load <64 x i32>, ptr %x
  %e = zext <64 x i32> %v to <64 x i64>
  %red = call i64 @llvm.vector.reduce.add.v64i64(<64 x i64> %e)
  ret i64 %red
}

declare i8 @llvm.vector.reduce.and.v1i8(<1 x i8>)

define i8 @vreduce_and_v1i8(ptr %x) {
; CHECK-LABEL: vreduce_and_v1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.and.v1i8(<1 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.and.v2i8(<2 x i8>)

define i8 @vreduce_and_v2i8(ptr %x) {
; CHECK-LABEL: vreduce_and_v2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.and.v2i8(<2 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.and.v3i8(<3 x i8>)

define i8 @vreduce_and_v3i8(ptr %x) {
; CHECK-LABEL: vreduce_and_v3i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 3, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    li a0, -1
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vredand.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <3 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.and.v3i8(<3 x i8> %v)
  ret i8 %red
}


declare i8 @llvm.vector.reduce.and.v4i8(<4 x i8>)

define i8 @vreduce_and_v4i8(ptr %x) {
; CHECK-LABEL: vreduce_and_v4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.and.v4i8(<4 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.and.v8i8(<8 x i8>)

define i8 @vreduce_and_v8i8(ptr %x) {
; CHECK-LABEL: vreduce_and_v8i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e8, mf2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.and.v8i8(<8 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.and.v16i8(<16 x i8>)

define i8 @vreduce_and_v16i8(ptr %x) {
; CHECK-LABEL: vreduce_and_v16i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e8, m1, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.and.v16i8(<16 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.and.v32i8(<32 x i8>)

define i8 @vreduce_and_v32i8(ptr %x) {
; CHECK-LABEL: vreduce_and_v32i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e8, m2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.and.v32i8(<32 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.and.v64i8(<64 x i8>)

define i8 @vreduce_and_v64i8(ptr %x) {
; CHECK-LABEL: vreduce_and_v64i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e8, m4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.and.v64i8(<64 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.and.v128i8(<128 x i8>)

define i8 @vreduce_and_v128i8(ptr %x) {
; CHECK-LABEL: vreduce_and_v128i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.and.v128i8(<128 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.and.v256i8(<256 x i8>)

define i8 @vreduce_and_v256i8(ptr %x) {
; CHECK-LABEL: vreduce_and_v256i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle8.v v16, (a0)
; CHECK-NEXT:    vand.vv v8, v8, v16
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <256 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.and.v256i8(<256 x i8> %v)
  ret i8 %red
}

declare i16 @llvm.vector.reduce.and.v1i16(<1 x i16>)

define i16 @vreduce_and_v1i16(ptr %x) {
; CHECK-LABEL: vreduce_and_v1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.and.v1i16(<1 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.and.v2i16(<2 x i16>)

define i16 @vreduce_and_v2i16(ptr %x) {
; CHECK-LABEL: vreduce_and_v2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.and.v2i16(<2 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.and.v4i16(<4 x i16>)

define i16 @vreduce_and_v4i16(ptr %x) {
; CHECK-LABEL: vreduce_and_v4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.and.v4i16(<4 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.and.v8i16(<8 x i16>)

define i16 @vreduce_and_v8i16(ptr %x) {
; CHECK-LABEL: vreduce_and_v8i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.and.v8i16(<8 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.and.v16i16(<16 x i16>)

define i16 @vreduce_and_v16i16(ptr %x) {
; CHECK-LABEL: vreduce_and_v16i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e16, m2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.and.v16i16(<16 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.and.v32i16(<32 x i16>)

define i16 @vreduce_and_v32i16(ptr %x) {
; CHECK-LABEL: vreduce_and_v32i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e16, m4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.and.v32i16(<32 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.and.v64i16(<64 x i16>)

define i16 @vreduce_and_v64i16(ptr %x) {
; CHECK-LABEL: vreduce_and_v64i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.and.v64i16(<64 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.and.v128i16(<128 x i16>)

define i16 @vreduce_and_v128i16(ptr %x) {
; CHECK-LABEL: vreduce_and_v128i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle16.v v16, (a0)
; CHECK-NEXT:    vand.vv v8, v8, v16
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.and.v128i16(<128 x i16> %v)
  ret i16 %red
}

declare i32 @llvm.vector.reduce.and.v1i32(<1 x i32>)

define i32 @vreduce_and_v1i32(ptr %x) {
; CHECK-LABEL: vreduce_and_v1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.and.v1i32(<1 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.and.v2i32(<2 x i32>)

define i32 @vreduce_and_v2i32(ptr %x) {
; CHECK-LABEL: vreduce_and_v2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.and.v2i32(<2 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.and.v4i32(<4 x i32>)

define i32 @vreduce_and_v4i32(ptr %x) {
; CHECK-LABEL: vreduce_and_v4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.and.v4i32(<4 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.and.v8i32(<8 x i32>)

define i32 @vreduce_and_v8i32(ptr %x) {
; CHECK-LABEL: vreduce_and_v8i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.and.v8i32(<8 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.and.v16i32(<16 x i32>)

define i32 @vreduce_and_v16i32(ptr %x) {
; CHECK-LABEL: vreduce_and_v16i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.and.v16i32(<16 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.and.v32i32(<32 x i32>)

define i32 @vreduce_and_v32i32(ptr %x) {
; CHECK-LABEL: vreduce_and_v32i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.and.v32i32(<32 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.and.v64i32(<64 x i32>)

define i32 @vreduce_and_v64i32(ptr %x) {
; CHECK-LABEL: vreduce_and_v64i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle32.v v16, (a0)
; CHECK-NEXT:    vand.vv v8, v8, v16
; CHECK-NEXT:    vredand.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.and.v64i32(<64 x i32> %v)
  ret i32 %red
}

declare i64 @llvm.vector.reduce.and.v1i64(<1 x i64>)

define i64 @vreduce_and_v1i64(ptr %x) {
; RV32-LABEL: vreduce_and_v1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsrl.vx v9, v8, a0
; RV32-NEXT:    vmv.x.s a1, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_and_v1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <1 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.and.v1i64(<1 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.and.v2i64(<2 x i64>)

define i64 @vreduce_and_v2i64(ptr %x) {
; RV32-LABEL: vreduce_and_v2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredand.vs v8, v8, v8
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v9, v8, a0
; RV32-NEXT:    vmv.x.s a1, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_and_v2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredand.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <2 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.and.v2i64(<2 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.and.v4i64(<4 x i64>)

define i64 @vreduce_and_v4i64(ptr %x) {
; RV32-LABEL: vreduce_and_v4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredand.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_and_v4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredand.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <4 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.and.v4i64(<4 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.and.v8i64(<8 x i64>)

define i64 @vreduce_and_v8i64(ptr %x) {
; RV32-LABEL: vreduce_and_v8i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredand.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_and_v8i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredand.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <8 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.and.v8i64(<8 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.and.v16i64(<16 x i64>)

define i64 @vreduce_and_v16i64(ptr %x) {
; RV32-LABEL: vreduce_and_v16i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredand.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_and_v16i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredand.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <16 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.and.v16i64(<16 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.and.v32i64(<32 x i64>)

define i64 @vreduce_and_v32i64(ptr %x) {
; RV32-LABEL: vreduce_and_v32i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v16, (a0)
; RV32-NEXT:    vand.vv v8, v8, v16
; RV32-NEXT:    vredand.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_and_v32i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v16, (a0)
; RV64-NEXT:    vand.vv v8, v8, v16
; RV64-NEXT:    vredand.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <32 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.and.v32i64(<32 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.and.v64i64(<64 x i64>)

define i64 @vreduce_and_v64i64(ptr %x) nounwind {
; RV32-LABEL: vreduce_and_v64i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a1, a0, 384
; RV32-NEXT:    vle64.v v16, (a1)
; RV32-NEXT:    addi a1, a0, 256
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v24, (a0)
; RV32-NEXT:    vle64.v v0, (a1)
; RV32-NEXT:    vand.vv v16, v24, v16
; RV32-NEXT:    vand.vv v8, v8, v0
; RV32-NEXT:    vand.vv v8, v8, v16
; RV32-NEXT:    vredand.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_and_v64i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a1, a0, 256
; RV64-NEXT:    addi a2, a0, 384
; RV64-NEXT:    vle64.v v16, (a2)
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v24, (a0)
; RV64-NEXT:    vle64.v v0, (a1)
; RV64-NEXT:    vand.vv v16, v24, v16
; RV64-NEXT:    vand.vv v8, v8, v0
; RV64-NEXT:    vand.vv v8, v8, v16
; RV64-NEXT:    vredand.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <64 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.and.v64i64(<64 x i64> %v)
  ret i64 %red
}

declare i8 @llvm.vector.reduce.or.v1i8(<1 x i8>)

define i8 @vreduce_or_v1i8(ptr %x) {
; CHECK-LABEL: vreduce_or_v1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.or.v1i8(<1 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.or.v2i8(<2 x i8>)

define i8 @vreduce_or_v2i8(ptr %x) {
; CHECK-LABEL: vreduce_or_v2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.or.v2i8(<2 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.or.v3i8(<3 x i8>)

define i8 @vreduce_or_v3i8(ptr %x) {
; CHECK-LABEL: vreduce_or_v3i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 3, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredor.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <3 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.or.v3i8(<3 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.or.v4i8(<4 x i8>)

define i8 @vreduce_or_v4i8(ptr %x) {
; CHECK-LABEL: vreduce_or_v4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.or.v4i8(<4 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.or.v8i8(<8 x i8>)

define i8 @vreduce_or_v8i8(ptr %x) {
; CHECK-LABEL: vreduce_or_v8i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e8, mf2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.or.v8i8(<8 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.or.v16i8(<16 x i8>)

define i8 @vreduce_or_v16i8(ptr %x) {
; CHECK-LABEL: vreduce_or_v16i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e8, m1, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.or.v16i8(<16 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.or.v32i8(<32 x i8>)

define i8 @vreduce_or_v32i8(ptr %x) {
; CHECK-LABEL: vreduce_or_v32i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e8, m2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.or.v32i8(<32 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.or.v64i8(<64 x i8>)

define i8 @vreduce_or_v64i8(ptr %x) {
; CHECK-LABEL: vreduce_or_v64i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e8, m4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.or.v64i8(<64 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.or.v128i8(<128 x i8>)

define i8 @vreduce_or_v128i8(ptr %x) {
; CHECK-LABEL: vreduce_or_v128i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.or.v128i8(<128 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.or.v256i8(<256 x i8>)

define i8 @vreduce_or_v256i8(ptr %x) {
; CHECK-LABEL: vreduce_or_v256i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle8.v v16, (a0)
; CHECK-NEXT:    vor.vv v8, v8, v16
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <256 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.or.v256i8(<256 x i8> %v)
  ret i8 %red
}

declare i16 @llvm.vector.reduce.or.v1i16(<1 x i16>)

define i16 @vreduce_or_v1i16(ptr %x) {
; CHECK-LABEL: vreduce_or_v1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.or.v1i16(<1 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.or.v2i16(<2 x i16>)

define i16 @vreduce_or_v2i16(ptr %x) {
; CHECK-LABEL: vreduce_or_v2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.or.v2i16(<2 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.or.v4i16(<4 x i16>)

define i16 @vreduce_or_v4i16(ptr %x) {
; CHECK-LABEL: vreduce_or_v4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.or.v4i16(<4 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.or.v8i16(<8 x i16>)

define i16 @vreduce_or_v8i16(ptr %x) {
; CHECK-LABEL: vreduce_or_v8i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.or.v8i16(<8 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.or.v16i16(<16 x i16>)

define i16 @vreduce_or_v16i16(ptr %x) {
; CHECK-LABEL: vreduce_or_v16i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e16, m2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.or.v16i16(<16 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.or.v32i16(<32 x i16>)

define i16 @vreduce_or_v32i16(ptr %x) {
; CHECK-LABEL: vreduce_or_v32i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e16, m4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.or.v32i16(<32 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.or.v64i16(<64 x i16>)

define i16 @vreduce_or_v64i16(ptr %x) {
; CHECK-LABEL: vreduce_or_v64i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.or.v64i16(<64 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.or.v128i16(<128 x i16>)

define i16 @vreduce_or_v128i16(ptr %x) {
; CHECK-LABEL: vreduce_or_v128i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle16.v v16, (a0)
; CHECK-NEXT:    vor.vv v8, v8, v16
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.or.v128i16(<128 x i16> %v)
  ret i16 %red
}

declare i32 @llvm.vector.reduce.or.v1i32(<1 x i32>)

define i32 @vreduce_or_v1i32(ptr %x) {
; CHECK-LABEL: vreduce_or_v1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.or.v1i32(<1 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.or.v2i32(<2 x i32>)

define i32 @vreduce_or_v2i32(ptr %x) {
; CHECK-LABEL: vreduce_or_v2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.or.v2i32(<2 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.or.v4i32(<4 x i32>)

define i32 @vreduce_or_v4i32(ptr %x) {
; CHECK-LABEL: vreduce_or_v4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.or.v4i32(<4 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.or.v8i32(<8 x i32>)

define i32 @vreduce_or_v8i32(ptr %x) {
; CHECK-LABEL: vreduce_or_v8i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.or.v8i32(<8 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.or.v16i32(<16 x i32>)

define i32 @vreduce_or_v16i32(ptr %x) {
; CHECK-LABEL: vreduce_or_v16i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.or.v16i32(<16 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.or.v32i32(<32 x i32>)

define i32 @vreduce_or_v32i32(ptr %x) {
; CHECK-LABEL: vreduce_or_v32i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.or.v32i32(<32 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.or.v64i32(<64 x i32>)

define i32 @vreduce_or_v64i32(ptr %x) {
; CHECK-LABEL: vreduce_or_v64i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle32.v v16, (a0)
; CHECK-NEXT:    vor.vv v8, v8, v16
; CHECK-NEXT:    vredor.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.or.v64i32(<64 x i32> %v)
  ret i32 %red
}

declare i64 @llvm.vector.reduce.or.v1i64(<1 x i64>)

define i64 @vreduce_or_v1i64(ptr %x) {
; RV32-LABEL: vreduce_or_v1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsrl.vx v9, v8, a0
; RV32-NEXT:    vmv.x.s a1, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_or_v1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <1 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.or.v1i64(<1 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.or.v2i64(<2 x i64>)

define i64 @vreduce_or_v2i64(ptr %x) {
; RV32-LABEL: vreduce_or_v2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredor.vs v8, v8, v8
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v9, v8, a0
; RV32-NEXT:    vmv.x.s a1, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_or_v2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredor.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <2 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.or.v4i64(<4 x i64>)

define i64 @vreduce_or_v4i64(ptr %x) {
; RV32-LABEL: vreduce_or_v4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredor.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_or_v4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredor.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <4 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.or.v4i64(<4 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.or.v8i64(<8 x i64>)

define i64 @vreduce_or_v8i64(ptr %x) {
; RV32-LABEL: vreduce_or_v8i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredor.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_or_v8i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredor.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <8 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.or.v8i64(<8 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.or.v16i64(<16 x i64>)

define i64 @vreduce_or_v16i64(ptr %x) {
; RV32-LABEL: vreduce_or_v16i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredor.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_or_v16i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredor.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <16 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.or.v16i64(<16 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.or.v32i64(<32 x i64>)

define i64 @vreduce_or_v32i64(ptr %x) {
; RV32-LABEL: vreduce_or_v32i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v16, (a0)
; RV32-NEXT:    vor.vv v8, v8, v16
; RV32-NEXT:    vredor.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_or_v32i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v16, (a0)
; RV64-NEXT:    vor.vv v8, v8, v16
; RV64-NEXT:    vredor.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <32 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.or.v32i64(<32 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.or.v64i64(<64 x i64>)

define i64 @vreduce_or_v64i64(ptr %x) nounwind {
; RV32-LABEL: vreduce_or_v64i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a1, a0, 384
; RV32-NEXT:    vle64.v v16, (a1)
; RV32-NEXT:    addi a1, a0, 256
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v24, (a0)
; RV32-NEXT:    vle64.v v0, (a1)
; RV32-NEXT:    vor.vv v16, v24, v16
; RV32-NEXT:    vor.vv v8, v8, v0
; RV32-NEXT:    vor.vv v8, v8, v16
; RV32-NEXT:    vredor.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_or_v64i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a1, a0, 256
; RV64-NEXT:    addi a2, a0, 384
; RV64-NEXT:    vle64.v v16, (a2)
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v24, (a0)
; RV64-NEXT:    vle64.v v0, (a1)
; RV64-NEXT:    vor.vv v16, v24, v16
; RV64-NEXT:    vor.vv v8, v8, v0
; RV64-NEXT:    vor.vv v8, v8, v16
; RV64-NEXT:    vredor.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <64 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.or.v64i64(<64 x i64> %v)
  ret i64 %red
}

declare i8 @llvm.vector.reduce.xor.v1i8(<1 x i8>)

define i8 @vreduce_xor_v1i8(ptr %x) {
; CHECK-LABEL: vreduce_xor_v1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.xor.v1i8(<1 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.xor.v2i8(<2 x i8>)

define i8 @vreduce_xor_v2i8(ptr %x) {
; CHECK-LABEL: vreduce_xor_v2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.xor.v2i8(<2 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.xor.v3i8(<3 x i8>)

define i8 @vreduce_xor_v3i8(ptr %x) {
; CHECK-LABEL: vreduce_xor_v3i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 3, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <3 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.xor.v3i8(<3 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.xor.v4i8(<4 x i8>)

define i8 @vreduce_xor_v4i8(ptr %x) {
; CHECK-LABEL: vreduce_xor_v4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.xor.v4i8(<4 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.xor.v8i8(<8 x i8>)

define i8 @vreduce_xor_v8i8(ptr %x) {
; CHECK-LABEL: vreduce_xor_v8i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e8, mf2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.xor.v8i8(<8 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.xor.v16i8(<16 x i8>)

define i8 @vreduce_xor_v16i8(ptr %x) {
; CHECK-LABEL: vreduce_xor_v16i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e8, m1, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.xor.v16i8(<16 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.xor.v32i8(<32 x i8>)

define i8 @vreduce_xor_v32i8(ptr %x) {
; CHECK-LABEL: vreduce_xor_v32i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e8, m2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v10, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v10
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.xor.v32i8(<32 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.xor.v64i8(<64 x i8>)

define i8 @vreduce_xor_v64i8(ptr %x) {
; CHECK-LABEL: vreduce_xor_v64i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e8, m4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v12, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v12
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.xor.v64i8(<64 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.xor.v128i8(<128 x i8>)

define i8 @vreduce_xor_v128i8(ptr %x) {
; CHECK-LABEL: vreduce_xor_v128i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v16, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.xor.v128i8(<128 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.xor.v256i8(<256 x i8>)

define i8 @vreduce_xor_v256i8(ptr %x) {
; CHECK-LABEL: vreduce_xor_v256i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle8.v v16, (a0)
; CHECK-NEXT:    vxor.vv v8, v8, v16
; CHECK-NEXT:    vmv.s.x v16, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <256 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.xor.v256i8(<256 x i8> %v)
  ret i8 %red
}

declare i16 @llvm.vector.reduce.xor.v1i16(<1 x i16>)

define i16 @vreduce_xor_v1i16(ptr %x) {
; CHECK-LABEL: vreduce_xor_v1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.xor.v1i16(<1 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.xor.v2i16(<2 x i16>)

define i16 @vreduce_xor_v2i16(ptr %x) {
; CHECK-LABEL: vreduce_xor_v2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.xor.v2i16(<2 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.xor.v4i16(<4 x i16>)

define i16 @vreduce_xor_v4i16(ptr %x) {
; CHECK-LABEL: vreduce_xor_v4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.xor.v4i16(<4 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.xor.v8i16(<8 x i16>)

define i16 @vreduce_xor_v8i16(ptr %x) {
; CHECK-LABEL: vreduce_xor_v8i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.xor.v8i16(<8 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.xor.v16i16(<16 x i16>)

define i16 @vreduce_xor_v16i16(ptr %x) {
; CHECK-LABEL: vreduce_xor_v16i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e16, m2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v10, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v10
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.xor.v16i16(<16 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.xor.v32i16(<32 x i16>)

define i16 @vreduce_xor_v32i16(ptr %x) {
; CHECK-LABEL: vreduce_xor_v32i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e16, m4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v12, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v12
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.xor.v32i16(<32 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.xor.v64i16(<64 x i16>)

define i16 @vreduce_xor_v64i16(ptr %x) {
; CHECK-LABEL: vreduce_xor_v64i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v16, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.xor.v64i16(<64 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.xor.v128i16(<128 x i16>)

define i16 @vreduce_xor_v128i16(ptr %x) {
; CHECK-LABEL: vreduce_xor_v128i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle16.v v16, (a0)
; CHECK-NEXT:    vxor.vv v8, v8, v16
; CHECK-NEXT:    vmv.s.x v16, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.xor.v128i16(<128 x i16> %v)
  ret i16 %red
}

declare i32 @llvm.vector.reduce.xor.v1i32(<1 x i32>)

define i32 @vreduce_xor_v1i32(ptr %x) {
; CHECK-LABEL: vreduce_xor_v1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.xor.v1i32(<1 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.xor.v2i32(<2 x i32>)

define i32 @vreduce_xor_v2i32(ptr %x) {
; CHECK-LABEL: vreduce_xor_v2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.xor.v2i32(<2 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.xor.v4i32(<4 x i32>)

define i32 @vreduce_xor_v4i32(ptr %x) {
; CHECK-LABEL: vreduce_xor_v4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.xor.v4i32(<4 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.xor.v8i32(<8 x i32>)

define i32 @vreduce_xor_v8i32(ptr %x) {
; CHECK-LABEL: vreduce_xor_v8i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v10, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v10
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.xor.v8i32(<8 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.xor.v16i32(<16 x i32>)

define i32 @vreduce_xor_v16i32(ptr %x) {
; CHECK-LABEL: vreduce_xor_v16i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v12, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v12
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.xor.v16i32(<16 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.xor.v32i32(<32 x i32>)

define i32 @vreduce_xor_v32i32(ptr %x) {
; CHECK-LABEL: vreduce_xor_v32i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v16, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.xor.v32i32(<32 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.xor.v64i32(<64 x i32>)

define i32 @vreduce_xor_v64i32(ptr %x) {
; CHECK-LABEL: vreduce_xor_v64i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle32.v v16, (a0)
; CHECK-NEXT:    vxor.vv v8, v8, v16
; CHECK-NEXT:    vmv.s.x v16, zero
; CHECK-NEXT:    vredxor.vs v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.xor.v64i32(<64 x i32> %v)
  ret i32 %red
}

declare i64 @llvm.vector.reduce.xor.v1i64(<1 x i64>)

define i64 @vreduce_xor_v1i64(ptr %x) {
; RV32-LABEL: vreduce_xor_v1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsrl.vx v9, v8, a0
; RV32-NEXT:    vmv.x.s a1, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_xor_v1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <1 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.xor.v1i64(<1 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.xor.v2i64(<2 x i64>)

define i64 @vreduce_xor_v2i64(ptr %x) {
; RV32-LABEL: vreduce_xor_v2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vmv.s.x v9, zero
; RV32-NEXT:    vredxor.vs v8, v8, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_xor_v2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.s.x v9, zero
; RV64-NEXT:    vredxor.vs v8, v8, v9
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <2 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.xor.v2i64(<2 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.xor.v4i64(<4 x i64>)

define i64 @vreduce_xor_v4i64(ptr %x) {
; RV32-LABEL: vreduce_xor_v4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vmv.s.x v10, zero
; RV32-NEXT:    vredxor.vs v8, v8, v10
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_xor_v4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.s.x v10, zero
; RV64-NEXT:    vredxor.vs v8, v8, v10
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <4 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.xor.v4i64(<4 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.xor.v8i64(<8 x i64>)

define i64 @vreduce_xor_v8i64(ptr %x) {
; RV32-LABEL: vreduce_xor_v8i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vmv.s.x v12, zero
; RV32-NEXT:    vredxor.vs v8, v8, v12
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_xor_v8i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.s.x v12, zero
; RV64-NEXT:    vredxor.vs v8, v8, v12
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <8 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.xor.v8i64(<8 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.xor.v16i64(<16 x i64>)

define i64 @vreduce_xor_v16i64(ptr %x) {
; RV32-LABEL: vreduce_xor_v16i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vmv.s.x v16, zero
; RV32-NEXT:    vredxor.vs v8, v8, v16
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_xor_v16i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.s.x v16, zero
; RV64-NEXT:    vredxor.vs v8, v8, v16
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <16 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.xor.v16i64(<16 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.xor.v32i64(<32 x i64>)

define i64 @vreduce_xor_v32i64(ptr %x) {
; RV32-LABEL: vreduce_xor_v32i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v16, (a0)
; RV32-NEXT:    vxor.vv v8, v8, v16
; RV32-NEXT:    vmv.s.x v16, zero
; RV32-NEXT:    vredxor.vs v8, v8, v16
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_xor_v32i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v16, (a0)
; RV64-NEXT:    vxor.vv v8, v8, v16
; RV64-NEXT:    vmv.s.x v16, zero
; RV64-NEXT:    vredxor.vs v8, v8, v16
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <32 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.xor.v32i64(<32 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.xor.v64i64(<64 x i64>)

define i64 @vreduce_xor_v64i64(ptr %x) nounwind {
; RV32-LABEL: vreduce_xor_v64i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a1, a0, 384
; RV32-NEXT:    vle64.v v16, (a1)
; RV32-NEXT:    addi a1, a0, 256
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v24, (a0)
; RV32-NEXT:    vle64.v v0, (a1)
; RV32-NEXT:    vxor.vv v16, v24, v16
; RV32-NEXT:    vxor.vv v8, v8, v0
; RV32-NEXT:    vxor.vv v8, v8, v16
; RV32-NEXT:    vmv.s.x v16, zero
; RV32-NEXT:    vredxor.vs v8, v8, v16
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_xor_v64i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a1, a0, 384
; RV64-NEXT:    vle64.v v16, (a1)
; RV64-NEXT:    addi a1, a0, 256
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v24, (a0)
; RV64-NEXT:    vle64.v v0, (a1)
; RV64-NEXT:    vxor.vv v16, v24, v16
; RV64-NEXT:    vxor.vv v8, v8, v0
; RV64-NEXT:    vxor.vv v8, v8, v16
; RV64-NEXT:    vmv.s.x v16, zero
; RV64-NEXT:    vredxor.vs v8, v8, v16
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <64 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.xor.v64i64(<64 x i64> %v)
  ret i64 %red
}

declare i8 @llvm.vector.reduce.smin.v1i8(<1 x i8>)

define i8 @vreduce_smin_v1i8(ptr %x) {
; CHECK-LABEL: vreduce_smin_v1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smin.v1i8(<1 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smin.v2i8(<2 x i8>)

define i8 @vreduce_smin_v2i8(ptr %x) {
; CHECK-LABEL: vreduce_smin_v2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smin.v2i8(<2 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smin.v3i8(<3 x i8>)

define i8 @vreduce_smin_v3i8(ptr %x) {
; CHECK-LABEL: vreduce_smin_v3i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 3, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    li a0, 127
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vredmin.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <3 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smin.v3i8(<3 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smin.v4i8(<4 x i8>)

define i8 @vreduce_smin_v4i8(ptr %x) {
; CHECK-LABEL: vreduce_smin_v4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smin.v4i8(<4 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smin.v8i8(<8 x i8>)

define i8 @vreduce_smin_v8i8(ptr %x) {
; CHECK-LABEL: vreduce_smin_v8i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e8, mf2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smin.v8i8(<8 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smin.v16i8(<16 x i8>)

define i8 @vreduce_smin_v16i8(ptr %x) {
; CHECK-LABEL: vreduce_smin_v16i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e8, m1, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smin.v16i8(<16 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smin.v32i8(<32 x i8>)

define i8 @vreduce_smin_v32i8(ptr %x) {
; CHECK-LABEL: vreduce_smin_v32i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e8, m2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smin.v32i8(<32 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smin.v64i8(<64 x i8>)

define i8 @vreduce_smin_v64i8(ptr %x) {
; CHECK-LABEL: vreduce_smin_v64i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e8, m4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smin.v64i8(<64 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smin.v128i8(<128 x i8>)

define i8 @vreduce_smin_v128i8(ptr %x) {
; CHECK-LABEL: vreduce_smin_v128i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smin.v128i8(<128 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smin.v256i8(<256 x i8>)

define i8 @vreduce_smin_v256i8(ptr %x) {
; CHECK-LABEL: vreduce_smin_v256i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle8.v v16, (a0)
; CHECK-NEXT:    vmin.vv v8, v8, v16
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <256 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smin.v256i8(<256 x i8> %v)
  ret i8 %red
}

declare i16 @llvm.vector.reduce.smin.v1i16(<1 x i16>)

define i16 @vreduce_smin_v1i16(ptr %x) {
; CHECK-LABEL: vreduce_smin_v1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.smin.v1i16(<1 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.smin.v2i16(<2 x i16>)

define i16 @vreduce_smin_v2i16(ptr %x) {
; CHECK-LABEL: vreduce_smin_v2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.smin.v2i16(<2 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.smin.v4i16(<4 x i16>)

define i16 @vreduce_smin_v4i16(ptr %x) {
; CHECK-LABEL: vreduce_smin_v4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.smin.v4i16(<4 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.smin.v8i16(<8 x i16>)

define i16 @vreduce_smin_v8i16(ptr %x) {
; CHECK-LABEL: vreduce_smin_v8i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.smin.v8i16(<8 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.smin.v16i16(<16 x i16>)

define i16 @vreduce_smin_v16i16(ptr %x) {
; CHECK-LABEL: vreduce_smin_v16i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e16, m2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.smin.v16i16(<16 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.smin.v32i16(<32 x i16>)

define i16 @vreduce_smin_v32i16(ptr %x) {
; CHECK-LABEL: vreduce_smin_v32i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e16, m4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.smin.v32i16(<32 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.smin.v64i16(<64 x i16>)

define i16 @vreduce_smin_v64i16(ptr %x) {
; CHECK-LABEL: vreduce_smin_v64i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.smin.v64i16(<64 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.smin.v128i16(<128 x i16>)

define i16 @vreduce_smin_v128i16(ptr %x) {
; CHECK-LABEL: vreduce_smin_v128i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle16.v v16, (a0)
; CHECK-NEXT:    vmin.vv v8, v8, v16
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.smin.v128i16(<128 x i16> %v)
  ret i16 %red
}

declare i32 @llvm.vector.reduce.smin.v1i32(<1 x i32>)

define i32 @vreduce_smin_v1i32(ptr %x) {
; CHECK-LABEL: vreduce_smin_v1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.smin.v1i32(<1 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.smin.v2i32(<2 x i32>)

define i32 @vreduce_smin_v2i32(ptr %x) {
; CHECK-LABEL: vreduce_smin_v2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.smin.v2i32(<2 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.smin.v4i32(<4 x i32>)

define i32 @vreduce_smin_v4i32(ptr %x) {
; CHECK-LABEL: vreduce_smin_v4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.smin.v4i32(<4 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.smin.v8i32(<8 x i32>)

define i32 @vreduce_smin_v8i32(ptr %x) {
; CHECK-LABEL: vreduce_smin_v8i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.smin.v8i32(<8 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.smin.v16i32(<16 x i32>)

define i32 @vreduce_smin_v16i32(ptr %x) {
; CHECK-LABEL: vreduce_smin_v16i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.smin.v16i32(<16 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.smin.v32i32(<32 x i32>)

define i32 @vreduce_smin_v32i32(ptr %x) {
; CHECK-LABEL: vreduce_smin_v32i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.smin.v32i32(<32 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.smin.v64i32(<64 x i32>)

define i32 @vreduce_smin_v64i32(ptr %x) {
; CHECK-LABEL: vreduce_smin_v64i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle32.v v16, (a0)
; CHECK-NEXT:    vmin.vv v8, v8, v16
; CHECK-NEXT:    vredmin.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.smin.v64i32(<64 x i32> %v)
  ret i32 %red
}

declare i64 @llvm.vector.reduce.smin.v1i64(<1 x i64>)

define i64 @vreduce_smin_v1i64(ptr %x) {
; RV32-LABEL: vreduce_smin_v1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsrl.vx v9, v8, a0
; RV32-NEXT:    vmv.x.s a1, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_smin_v1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <1 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.smin.v1i64(<1 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.smin.v2i64(<2 x i64>)

define i64 @vreduce_smin_v2i64(ptr %x) {
; RV32-LABEL: vreduce_smin_v2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredmin.vs v8, v8, v8
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v9, v8, a0
; RV32-NEXT:    vmv.x.s a1, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_smin_v2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredmin.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <2 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.smin.v2i64(<2 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.smin.v4i64(<4 x i64>)

define i64 @vreduce_smin_v4i64(ptr %x) {
; RV32-LABEL: vreduce_smin_v4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredmin.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_smin_v4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredmin.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <4 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.smin.v4i64(<4 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.smin.v8i64(<8 x i64>)

define i64 @vreduce_smin_v8i64(ptr %x) {
; RV32-LABEL: vreduce_smin_v8i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredmin.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_smin_v8i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredmin.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <8 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.smin.v8i64(<8 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.smin.v16i64(<16 x i64>)

define i64 @vreduce_smin_v16i64(ptr %x) {
; RV32-LABEL: vreduce_smin_v16i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredmin.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_smin_v16i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredmin.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <16 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.smin.v16i64(<16 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.smin.v32i64(<32 x i64>)

define i64 @vreduce_smin_v32i64(ptr %x) {
; RV32-LABEL: vreduce_smin_v32i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v16, (a0)
; RV32-NEXT:    vmin.vv v8, v8, v16
; RV32-NEXT:    vredmin.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_smin_v32i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v16, (a0)
; RV64-NEXT:    vmin.vv v8, v8, v16
; RV64-NEXT:    vredmin.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <32 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.smin.v32i64(<32 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.smin.v64i64(<64 x i64>)

define i64 @vreduce_smin_v64i64(ptr %x) nounwind {
; RV32-LABEL: vreduce_smin_v64i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a1, a0, 384
; RV32-NEXT:    vle64.v v16, (a1)
; RV32-NEXT:    addi a1, a0, 256
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v24, (a0)
; RV32-NEXT:    vle64.v v0, (a1)
; RV32-NEXT:    vmin.vv v16, v24, v16
; RV32-NEXT:    vmin.vv v8, v8, v0
; RV32-NEXT:    vmin.vv v8, v8, v16
; RV32-NEXT:    vredmin.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_smin_v64i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a1, a0, 256
; RV64-NEXT:    addi a2, a0, 384
; RV64-NEXT:    vle64.v v16, (a2)
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v24, (a0)
; RV64-NEXT:    vle64.v v0, (a1)
; RV64-NEXT:    vmin.vv v16, v24, v16
; RV64-NEXT:    vmin.vv v8, v8, v0
; RV64-NEXT:    vmin.vv v8, v8, v16
; RV64-NEXT:    vredmin.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <64 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.smin.v64i64(<64 x i64> %v)
  ret i64 %red
}

declare i8 @llvm.vector.reduce.smax.v1i8(<1 x i8>)

define i8 @vreduce_smax_v1i8(ptr %x) {
; CHECK-LABEL: vreduce_smax_v1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smax.v1i8(<1 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smax.v2i8(<2 x i8>)

define i8 @vreduce_smax_v2i8(ptr %x) {
; CHECK-LABEL: vreduce_smax_v2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smax.v2i8(<2 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smax.v3i8(<3 x i8>)

define i8 @vreduce_smax_v3i8(ptr %x) {
; CHECK-LABEL: vreduce_smax_v3i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 3, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    li a0, -128
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vredmax.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <3 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smax.v3i8(<3 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smax.v4i8(<4 x i8>)

define i8 @vreduce_smax_v4i8(ptr %x) {
; CHECK-LABEL: vreduce_smax_v4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smax.v4i8(<4 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smax.v8i8(<8 x i8>)

define i8 @vreduce_smax_v8i8(ptr %x) {
; CHECK-LABEL: vreduce_smax_v8i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e8, mf2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smax.v8i8(<8 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smax.v16i8(<16 x i8>)

define i8 @vreduce_smax_v16i8(ptr %x) {
; CHECK-LABEL: vreduce_smax_v16i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e8, m1, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smax.v16i8(<16 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smax.v32i8(<32 x i8>)

define i8 @vreduce_smax_v32i8(ptr %x) {
; CHECK-LABEL: vreduce_smax_v32i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e8, m2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smax.v32i8(<32 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smax.v64i8(<64 x i8>)

define i8 @vreduce_smax_v64i8(ptr %x) {
; CHECK-LABEL: vreduce_smax_v64i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e8, m4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smax.v64i8(<64 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smax.v128i8(<128 x i8>)

define i8 @vreduce_smax_v128i8(ptr %x) {
; CHECK-LABEL: vreduce_smax_v128i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smax.v128i8(<128 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.smax.v256i8(<256 x i8>)

define i8 @vreduce_smax_v256i8(ptr %x) {
; CHECK-LABEL: vreduce_smax_v256i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle8.v v16, (a0)
; CHECK-NEXT:    vmax.vv v8, v8, v16
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <256 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.smax.v256i8(<256 x i8> %v)
  ret i8 %red
}

declare i16 @llvm.vector.reduce.smax.v1i16(<1 x i16>)

define i16 @vreduce_smax_v1i16(ptr %x) {
; CHECK-LABEL: vreduce_smax_v1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.smax.v1i16(<1 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.smax.v2i16(<2 x i16>)

define i16 @vreduce_smax_v2i16(ptr %x) {
; CHECK-LABEL: vreduce_smax_v2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.smax.v2i16(<2 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.smax.v4i16(<4 x i16>)

define i16 @vreduce_smax_v4i16(ptr %x) {
; CHECK-LABEL: vreduce_smax_v4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.smax.v4i16(<4 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.smax.v8i16(<8 x i16>)

define i16 @vreduce_smax_v8i16(ptr %x) {
; CHECK-LABEL: vreduce_smax_v8i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.smax.v8i16(<8 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.smax.v16i16(<16 x i16>)

define i16 @vreduce_smax_v16i16(ptr %x) {
; CHECK-LABEL: vreduce_smax_v16i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e16, m2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.smax.v16i16(<16 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.smax.v32i16(<32 x i16>)

define i16 @vreduce_smax_v32i16(ptr %x) {
; CHECK-LABEL: vreduce_smax_v32i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e16, m4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.smax.v32i16(<32 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.smax.v64i16(<64 x i16>)

define i16 @vreduce_smax_v64i16(ptr %x) {
; CHECK-LABEL: vreduce_smax_v64i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.smax.v64i16(<64 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.smax.v128i16(<128 x i16>)

define i16 @vreduce_smax_v128i16(ptr %x) {
; CHECK-LABEL: vreduce_smax_v128i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle16.v v16, (a0)
; CHECK-NEXT:    vmax.vv v8, v8, v16
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.smax.v128i16(<128 x i16> %v)
  ret i16 %red
}

declare i32 @llvm.vector.reduce.smax.v1i32(<1 x i32>)

define i32 @vreduce_smax_v1i32(ptr %x) {
; CHECK-LABEL: vreduce_smax_v1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.smax.v1i32(<1 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.smax.v2i32(<2 x i32>)

define i32 @vreduce_smax_v2i32(ptr %x) {
; CHECK-LABEL: vreduce_smax_v2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.smax.v2i32(<2 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.smax.v4i32(<4 x i32>)

define i32 @vreduce_smax_v4i32(ptr %x) {
; CHECK-LABEL: vreduce_smax_v4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.smax.v4i32(<4 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.smax.v8i32(<8 x i32>)

define i32 @vreduce_smax_v8i32(ptr %x) {
; CHECK-LABEL: vreduce_smax_v8i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.smax.v8i32(<8 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.smax.v16i32(<16 x i32>)

define i32 @vreduce_smax_v16i32(ptr %x) {
; CHECK-LABEL: vreduce_smax_v16i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.smax.v16i32(<16 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.smax.v32i32(<32 x i32>)

define i32 @vreduce_smax_v32i32(ptr %x) {
; CHECK-LABEL: vreduce_smax_v32i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.smax.v32i32(<32 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.smax.v64i32(<64 x i32>)

define i32 @vreduce_smax_v64i32(ptr %x) {
; CHECK-LABEL: vreduce_smax_v64i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle32.v v16, (a0)
; CHECK-NEXT:    vmax.vv v8, v8, v16
; CHECK-NEXT:    vredmax.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.smax.v64i32(<64 x i32> %v)
  ret i32 %red
}

declare i64 @llvm.vector.reduce.smax.v1i64(<1 x i64>)

define i64 @vreduce_smax_v1i64(ptr %x) {
; RV32-LABEL: vreduce_smax_v1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsrl.vx v9, v8, a0
; RV32-NEXT:    vmv.x.s a1, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_smax_v1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <1 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.smax.v1i64(<1 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.smax.v2i64(<2 x i64>)

define i64 @vreduce_smax_v2i64(ptr %x) {
; RV32-LABEL: vreduce_smax_v2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredmax.vs v8, v8, v8
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v9, v8, a0
; RV32-NEXT:    vmv.x.s a1, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_smax_v2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredmax.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <2 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.smax.v2i64(<2 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.smax.v4i64(<4 x i64>)

define i64 @vreduce_smax_v4i64(ptr %x) {
; RV32-LABEL: vreduce_smax_v4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredmax.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_smax_v4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredmax.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <4 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.smax.v4i64(<4 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.smax.v8i64(<8 x i64>)

define i64 @vreduce_smax_v8i64(ptr %x) {
; RV32-LABEL: vreduce_smax_v8i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredmax.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_smax_v8i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredmax.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <8 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.smax.v8i64(<8 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.smax.v16i64(<16 x i64>)

define i64 @vreduce_smax_v16i64(ptr %x) {
; RV32-LABEL: vreduce_smax_v16i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredmax.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_smax_v16i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredmax.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <16 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.smax.v16i64(<16 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.smax.v32i64(<32 x i64>)

define i64 @vreduce_smax_v32i64(ptr %x) {
; RV32-LABEL: vreduce_smax_v32i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v16, (a0)
; RV32-NEXT:    vmax.vv v8, v8, v16
; RV32-NEXT:    vredmax.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_smax_v32i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v16, (a0)
; RV64-NEXT:    vmax.vv v8, v8, v16
; RV64-NEXT:    vredmax.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <32 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.smax.v32i64(<32 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.smax.v64i64(<64 x i64>)

define i64 @vreduce_smax_v64i64(ptr %x) nounwind {
; RV32-LABEL: vreduce_smax_v64i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a1, a0, 384
; RV32-NEXT:    vle64.v v16, (a1)
; RV32-NEXT:    addi a1, a0, 256
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v24, (a0)
; RV32-NEXT:    vle64.v v0, (a1)
; RV32-NEXT:    vmax.vv v16, v24, v16
; RV32-NEXT:    vmax.vv v8, v8, v0
; RV32-NEXT:    vmax.vv v8, v8, v16
; RV32-NEXT:    vredmax.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_smax_v64i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a1, a0, 256
; RV64-NEXT:    addi a2, a0, 384
; RV64-NEXT:    vle64.v v16, (a2)
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v24, (a0)
; RV64-NEXT:    vle64.v v0, (a1)
; RV64-NEXT:    vmax.vv v16, v24, v16
; RV64-NEXT:    vmax.vv v8, v8, v0
; RV64-NEXT:    vmax.vv v8, v8, v16
; RV64-NEXT:    vredmax.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <64 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.smax.v64i64(<64 x i64> %v)
  ret i64 %red
}

declare i8 @llvm.vector.reduce.umin.v1i8(<1 x i8>)

define i8 @vreduce_umin_v1i8(ptr %x) {
; CHECK-LABEL: vreduce_umin_v1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umin.v1i8(<1 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umin.v2i8(<2 x i8>)

define i8 @vreduce_umin_v2i8(ptr %x) {
; CHECK-LABEL: vreduce_umin_v2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umin.v2i8(<2 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umin.v3i8(<3 x i8>)

define i8 @vreduce_umin_v3i8(ptr %x) {
; CHECK-LABEL: vreduce_umin_v3i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 3, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    li a0, -1
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vredminu.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <3 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umin.v3i8(<3 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umin.v4i8(<4 x i8>)

define i8 @vreduce_umin_v4i8(ptr %x) {
; CHECK-LABEL: vreduce_umin_v4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umin.v4i8(<4 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umin.v8i8(<8 x i8>)

define i8 @vreduce_umin_v8i8(ptr %x) {
; CHECK-LABEL: vreduce_umin_v8i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e8, mf2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umin.v8i8(<8 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umin.v16i8(<16 x i8>)

define i8 @vreduce_umin_v16i8(ptr %x) {
; CHECK-LABEL: vreduce_umin_v16i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e8, m1, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umin.v16i8(<16 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umin.v32i8(<32 x i8>)

define i8 @vreduce_umin_v32i8(ptr %x) {
; CHECK-LABEL: vreduce_umin_v32i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e8, m2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umin.v32i8(<32 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umin.v64i8(<64 x i8>)

define i8 @vreduce_umin_v64i8(ptr %x) {
; CHECK-LABEL: vreduce_umin_v64i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e8, m4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umin.v64i8(<64 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umin.v128i8(<128 x i8>)

define i8 @vreduce_umin_v128i8(ptr %x) {
; CHECK-LABEL: vreduce_umin_v128i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umin.v128i8(<128 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umin.v256i8(<256 x i8>)

define i8 @vreduce_umin_v256i8(ptr %x) {
; CHECK-LABEL: vreduce_umin_v256i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle8.v v16, (a0)
; CHECK-NEXT:    vminu.vv v8, v8, v16
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <256 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umin.v256i8(<256 x i8> %v)
  ret i8 %red
}

declare i16 @llvm.vector.reduce.umin.v1i16(<1 x i16>)

define i16 @vreduce_umin_v1i16(ptr %x) {
; CHECK-LABEL: vreduce_umin_v1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.umin.v1i16(<1 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.umin.v2i16(<2 x i16>)

define i16 @vreduce_umin_v2i16(ptr %x) {
; CHECK-LABEL: vreduce_umin_v2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.umin.v2i16(<2 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.umin.v4i16(<4 x i16>)

define i16 @vreduce_umin_v4i16(ptr %x) {
; CHECK-LABEL: vreduce_umin_v4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.umin.v4i16(<4 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.umin.v8i16(<8 x i16>)

define i16 @vreduce_umin_v8i16(ptr %x) {
; CHECK-LABEL: vreduce_umin_v8i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.umin.v8i16(<8 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.umin.v16i16(<16 x i16>)

define i16 @vreduce_umin_v16i16(ptr %x) {
; CHECK-LABEL: vreduce_umin_v16i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e16, m2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.umin.v16i16(<16 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.umin.v32i16(<32 x i16>)

define i16 @vreduce_umin_v32i16(ptr %x) {
; CHECK-LABEL: vreduce_umin_v32i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e16, m4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.umin.v32i16(<32 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.umin.v64i16(<64 x i16>)

define i16 @vreduce_umin_v64i16(ptr %x) {
; CHECK-LABEL: vreduce_umin_v64i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.umin.v64i16(<64 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.umin.v128i16(<128 x i16>)

define i16 @vreduce_umin_v128i16(ptr %x) {
; CHECK-LABEL: vreduce_umin_v128i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle16.v v16, (a0)
; CHECK-NEXT:    vminu.vv v8, v8, v16
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.umin.v128i16(<128 x i16> %v)
  ret i16 %red
}

declare i32 @llvm.vector.reduce.umin.v1i32(<1 x i32>)

define i32 @vreduce_umin_v1i32(ptr %x) {
; CHECK-LABEL: vreduce_umin_v1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.umin.v1i32(<1 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.umin.v2i32(<2 x i32>)

define i32 @vreduce_umin_v2i32(ptr %x) {
; CHECK-LABEL: vreduce_umin_v2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.umin.v2i32(<2 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.umin.v4i32(<4 x i32>)

define i32 @vreduce_umin_v4i32(ptr %x) {
; CHECK-LABEL: vreduce_umin_v4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.umin.v4i32(<4 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.umin.v8i32(<8 x i32>)

define i32 @vreduce_umin_v8i32(ptr %x) {
; CHECK-LABEL: vreduce_umin_v8i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.umin.v8i32(<8 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.umin.v16i32(<16 x i32>)

define i32 @vreduce_umin_v16i32(ptr %x) {
; CHECK-LABEL: vreduce_umin_v16i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.umin.v16i32(<16 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.umin.v32i32(<32 x i32>)

define i32 @vreduce_umin_v32i32(ptr %x) {
; CHECK-LABEL: vreduce_umin_v32i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.umin.v32i32(<32 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.umin.v64i32(<64 x i32>)

define i32 @vreduce_umin_v64i32(ptr %x) {
; CHECK-LABEL: vreduce_umin_v64i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle32.v v16, (a0)
; CHECK-NEXT:    vminu.vv v8, v8, v16
; CHECK-NEXT:    vredminu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.umin.v64i32(<64 x i32> %v)
  ret i32 %red
}

declare i64 @llvm.vector.reduce.umin.v1i64(<1 x i64>)

define i64 @vreduce_umin_v1i64(ptr %x) {
; RV32-LABEL: vreduce_umin_v1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsrl.vx v9, v8, a0
; RV32-NEXT:    vmv.x.s a1, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_umin_v1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <1 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.umin.v1i64(<1 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.umin.v2i64(<2 x i64>)

define i64 @vreduce_umin_v2i64(ptr %x) {
; RV32-LABEL: vreduce_umin_v2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredminu.vs v8, v8, v8
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v9, v8, a0
; RV32-NEXT:    vmv.x.s a1, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_umin_v2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredminu.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <2 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.umin.v2i64(<2 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.umin.v4i64(<4 x i64>)

define i64 @vreduce_umin_v4i64(ptr %x) {
; RV32-LABEL: vreduce_umin_v4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredminu.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_umin_v4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredminu.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <4 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.umin.v4i64(<4 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.umin.v8i64(<8 x i64>)

define i64 @vreduce_umin_v8i64(ptr %x) {
; RV32-LABEL: vreduce_umin_v8i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredminu.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_umin_v8i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredminu.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <8 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.umin.v8i64(<8 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.umin.v16i64(<16 x i64>)

define i64 @vreduce_umin_v16i64(ptr %x) {
; RV32-LABEL: vreduce_umin_v16i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredminu.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_umin_v16i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredminu.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <16 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.umin.v16i64(<16 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.umin.v32i64(<32 x i64>)

define i64 @vreduce_umin_v32i64(ptr %x) {
; RV32-LABEL: vreduce_umin_v32i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v16, (a0)
; RV32-NEXT:    vminu.vv v8, v8, v16
; RV32-NEXT:    vredminu.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_umin_v32i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v16, (a0)
; RV64-NEXT:    vminu.vv v8, v8, v16
; RV64-NEXT:    vredminu.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <32 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.umin.v32i64(<32 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.umin.v64i64(<64 x i64>)

define i64 @vreduce_umin_v64i64(ptr %x) nounwind {
; RV32-LABEL: vreduce_umin_v64i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a1, a0, 384
; RV32-NEXT:    vle64.v v16, (a1)
; RV32-NEXT:    addi a1, a0, 256
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v24, (a0)
; RV32-NEXT:    vle64.v v0, (a1)
; RV32-NEXT:    vminu.vv v16, v24, v16
; RV32-NEXT:    vminu.vv v8, v8, v0
; RV32-NEXT:    vminu.vv v8, v8, v16
; RV32-NEXT:    vredminu.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_umin_v64i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a1, a0, 256
; RV64-NEXT:    addi a2, a0, 384
; RV64-NEXT:    vle64.v v16, (a2)
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v24, (a0)
; RV64-NEXT:    vle64.v v0, (a1)
; RV64-NEXT:    vminu.vv v16, v24, v16
; RV64-NEXT:    vminu.vv v8, v8, v0
; RV64-NEXT:    vminu.vv v8, v8, v16
; RV64-NEXT:    vredminu.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <64 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.umin.v64i64(<64 x i64> %v)
  ret i64 %red
}

declare i8 @llvm.vector.reduce.umax.v1i8(<1 x i8>)

define i8 @vreduce_umax_v1i8(ptr %x) {
; CHECK-LABEL: vreduce_umax_v1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umax.v1i8(<1 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umax.v2i8(<2 x i8>)

define i8 @vreduce_umax_v2i8(ptr %x) {
; CHECK-LABEL: vreduce_umax_v2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umax.v2i8(<2 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umax.v3i8(<3 x i8>)

define i8 @vreduce_umax_v3i8(ptr %x) {
; CHECK-LABEL: vreduce_umax_v3i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 3, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.s.x v9, zero
; CHECK-NEXT:    vredmaxu.vs v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <3 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umax.v3i8(<3 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umax.v4i8(<4 x i8>)

define i8 @vreduce_umax_v4i8(ptr %x) {
; CHECK-LABEL: vreduce_umax_v4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umax.v4i8(<4 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umax.v8i8(<8 x i8>)

define i8 @vreduce_umax_v8i8(ptr %x) {
; CHECK-LABEL: vreduce_umax_v8i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e8, mf2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umax.v8i8(<8 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umax.v16i8(<16 x i8>)

define i8 @vreduce_umax_v16i8(ptr %x) {
; CHECK-LABEL: vreduce_umax_v16i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e8, m1, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umax.v16i8(<16 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umax.v32i8(<32 x i8>)

define i8 @vreduce_umax_v32i8(ptr %x) {
; CHECK-LABEL: vreduce_umax_v32i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e8, m2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umax.v32i8(<32 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umax.v64i8(<64 x i8>)

define i8 @vreduce_umax_v64i8(ptr %x) {
; CHECK-LABEL: vreduce_umax_v64i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e8, m4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umax.v64i8(<64 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umax.v128i8(<128 x i8>)

define i8 @vreduce_umax_v128i8(ptr %x) {
; CHECK-LABEL: vreduce_umax_v128i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umax.v128i8(<128 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.umax.v256i8(<256 x i8>)

define i8 @vreduce_umax_v256i8(ptr %x) {
; CHECK-LABEL: vreduce_umax_v256i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle8.v v16, (a0)
; CHECK-NEXT:    vmaxu.vv v8, v8, v16
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <256 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.umax.v256i8(<256 x i8> %v)
  ret i8 %red
}

declare i16 @llvm.vector.reduce.umax.v1i16(<1 x i16>)

define i16 @vreduce_umax_v1i16(ptr %x) {
; CHECK-LABEL: vreduce_umax_v1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.umax.v1i16(<1 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.umax.v2i16(<2 x i16>)

define i16 @vreduce_umax_v2i16(ptr %x) {
; CHECK-LABEL: vreduce_umax_v2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.umax.v2i16(<2 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.umax.v4i16(<4 x i16>)

define i16 @vreduce_umax_v4i16(ptr %x) {
; CHECK-LABEL: vreduce_umax_v4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.umax.v4i16(<4 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.umax.v8i16(<8 x i16>)

define i16 @vreduce_umax_v8i16(ptr %x) {
; CHECK-LABEL: vreduce_umax_v8i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.umax.v8i16(<8 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.umax.v16i16(<16 x i16>)

define i16 @vreduce_umax_v16i16(ptr %x) {
; CHECK-LABEL: vreduce_umax_v16i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e16, m2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.umax.v16i16(<16 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.umax.v32i16(<32 x i16>)

define i16 @vreduce_umax_v32i16(ptr %x) {
; CHECK-LABEL: vreduce_umax_v32i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e16, m4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.umax.v32i16(<32 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.umax.v64i16(<64 x i16>)

define i16 @vreduce_umax_v64i16(ptr %x) {
; CHECK-LABEL: vreduce_umax_v64i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.umax.v64i16(<64 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.umax.v128i16(<128 x i16>)

define i16 @vreduce_umax_v128i16(ptr %x) {
; CHECK-LABEL: vreduce_umax_v128i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle16.v v16, (a0)
; CHECK-NEXT:    vmaxu.vv v8, v8, v16
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.umax.v128i16(<128 x i16> %v)
  ret i16 %red
}

declare i32 @llvm.vector.reduce.umax.v1i32(<1 x i32>)

define i32 @vreduce_umax_v1i32(ptr %x) {
; CHECK-LABEL: vreduce_umax_v1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.umax.v1i32(<1 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.umax.v2i32(<2 x i32>)

define i32 @vreduce_umax_v2i32(ptr %x) {
; CHECK-LABEL: vreduce_umax_v2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.umax.v2i32(<2 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.umax.v4i32(<4 x i32>)

define i32 @vreduce_umax_v4i32(ptr %x) {
; CHECK-LABEL: vreduce_umax_v4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.umax.v4i32(<4 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.umax.v8i32(<8 x i32>)

define i32 @vreduce_umax_v8i32(ptr %x) {
; CHECK-LABEL: vreduce_umax_v8i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.umax.v8i32(<8 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.umax.v16i32(<16 x i32>)

define i32 @vreduce_umax_v16i32(ptr %x) {
; CHECK-LABEL: vreduce_umax_v16i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.umax.v16i32(<16 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.umax.v32i32(<32 x i32>)

define i32 @vreduce_umax_v32i32(ptr %x) {
; CHECK-LABEL: vreduce_umax_v32i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.umax.v32i32(<32 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.umax.v64i32(<64 x i32>)

define i32 @vreduce_umax_v64i32(ptr %x) {
; CHECK-LABEL: vreduce_umax_v64i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle32.v v16, (a0)
; CHECK-NEXT:    vmaxu.vv v8, v8, v16
; CHECK-NEXT:    vredmaxu.vs v8, v8, v8
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.umax.v64i32(<64 x i32> %v)
  ret i32 %red
}

declare i64 @llvm.vector.reduce.umax.v1i64(<1 x i64>)

define i64 @vreduce_umax_v1i64(ptr %x) {
; RV32-LABEL: vreduce_umax_v1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsrl.vx v9, v8, a0
; RV32-NEXT:    vmv.x.s a1, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_umax_v1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <1 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.umax.v1i64(<1 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.umax.v2i64(<2 x i64>)

define i64 @vreduce_umax_v2i64(ptr %x) {
; RV32-LABEL: vreduce_umax_v2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredmaxu.vs v8, v8, v8
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v9, v8, a0
; RV32-NEXT:    vmv.x.s a1, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_umax_v2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredmaxu.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <2 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.umax.v2i64(<2 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.umax.v4i64(<4 x i64>)

define i64 @vreduce_umax_v4i64(ptr %x) {
; RV32-LABEL: vreduce_umax_v4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredmaxu.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_umax_v4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredmaxu.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <4 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.umax.v4i64(<4 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.umax.v8i64(<8 x i64>)

define i64 @vreduce_umax_v8i64(ptr %x) {
; RV32-LABEL: vreduce_umax_v8i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredmaxu.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_umax_v8i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredmaxu.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <8 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.umax.v8i64(<8 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.umax.v16i64(<16 x i64>)

define i64 @vreduce_umax_v16i64(ptr %x) {
; RV32-LABEL: vreduce_umax_v16i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vredmaxu.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_umax_v16i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vredmaxu.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <16 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.umax.v16i64(<16 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.umax.v32i64(<32 x i64>)

define i64 @vreduce_umax_v32i64(ptr %x) {
; RV32-LABEL: vreduce_umax_v32i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v16, (a0)
; RV32-NEXT:    vmaxu.vv v8, v8, v16
; RV32-NEXT:    vredmaxu.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_umax_v32i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v16, (a0)
; RV64-NEXT:    vmaxu.vv v8, v8, v16
; RV64-NEXT:    vredmaxu.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <32 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.umax.v32i64(<32 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.umax.v64i64(<64 x i64>)

define i64 @vreduce_umax_v64i64(ptr %x) nounwind {
; RV32-LABEL: vreduce_umax_v64i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a1, a0, 384
; RV32-NEXT:    vle64.v v16, (a1)
; RV32-NEXT:    addi a1, a0, 256
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v24, (a0)
; RV32-NEXT:    vle64.v v0, (a1)
; RV32-NEXT:    vmaxu.vv v16, v24, v16
; RV32-NEXT:    vmaxu.vv v8, v8, v0
; RV32-NEXT:    vmaxu.vv v8, v8, v16
; RV32-NEXT:    vredmaxu.vs v8, v8, v8
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_umax_v64i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a1, a0, 256
; RV64-NEXT:    addi a2, a0, 384
; RV64-NEXT:    vle64.v v16, (a2)
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v24, (a0)
; RV64-NEXT:    vle64.v v0, (a1)
; RV64-NEXT:    vmaxu.vv v16, v24, v16
; RV64-NEXT:    vmaxu.vv v8, v8, v0
; RV64-NEXT:    vmaxu.vv v8, v8, v16
; RV64-NEXT:    vredmaxu.vs v8, v8, v8
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <64 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.umax.v64i64(<64 x i64> %v)
  ret i64 %red
}

declare i8 @llvm.vector.reduce.mul.v1i8(<1 x i8>)

define i8 @vreduce_mul_v1i8(ptr %x) {
; CHECK-LABEL: vreduce_mul_v1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.mul.v1i8(<1 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.mul.v2i8(<2 x i8>)

define i8 @vreduce_mul_v2i8(ptr %x) {
; CHECK-LABEL: vreduce_mul_v2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e8, mf8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    lbu a0, 1(a0)
; CHECK-NEXT:    vmul.vx v8, v8, a0
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.mul.v2i8(<2 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.mul.v3i8(<3 x i8>)

define i8 @vreduce_mul_v3i8(ptr %x) {
; CHECK-LABEL: vreduce_mul_v3i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 3, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vsetivli zero, 4, e8, mf4, ta, ma
; CHECK-NEXT:    vmv.v.i v9, 1
; CHECK-NEXT:    vslideup.vi v8, v9, 3
; CHECK-NEXT:    vsetivli zero, 2, e8, mf4, ta, ma
; CHECK-NEXT:    vslidedown.vi v9, v8, 2
; CHECK-NEXT:    vsetivli zero, 2, e8, mf8, ta, ma
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vslidedown.vi v9, v8, 1
; CHECK-NEXT:    vsetivli zero, 1, e8, mf8, ta, ma
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <3 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.mul.v3i8(<3 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.mul.v4i8(<4 x i8>)

define i8 @vreduce_mul_v4i8(ptr %x) {
; CHECK-LABEL: vreduce_mul_v4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e8, mf4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vslidedown.vi v9, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vrgather.vi v9, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.mul.v4i8(<4 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.mul.v8i8(<8 x i8>)

define i8 @vreduce_mul_v8i8(ptr %x) {
; CHECK-LABEL: vreduce_mul_v8i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e8, mf2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vslidedown.vi v9, v8, 4
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vslidedown.vi v9, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vrgather.vi v9, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.mul.v8i8(<8 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.mul.v16i8(<16 x i8>)

define i8 @vreduce_mul_v16i8(ptr %x) {
; CHECK-LABEL: vreduce_mul_v16i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e8, m1, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vslidedown.vi v9, v8, 8
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vslidedown.vi v9, v8, 4
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vslidedown.vi v9, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vrgather.vi v9, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.mul.v16i8(<16 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.mul.v32i8(<32 x i8>)

define i8 @vreduce_mul_v32i8(ptr %x) {
; CHECK-LABEL: vreduce_mul_v32i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e8, m2, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    vslidedown.vi v10, v8, 16
; CHECK-NEXT:    vmul.vv v8, v8, v10
; CHECK-NEXT:    vslidedown.vi v10, v8, 8
; CHECK-NEXT:    vmul.vv v8, v8, v10
; CHECK-NEXT:    vslidedown.vi v10, v8, 4
; CHECK-NEXT:    vmul.vv v8, v8, v10
; CHECK-NEXT:    vslidedown.vi v10, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v10
; CHECK-NEXT:    vrgather.vi v10, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v10
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.mul.v32i8(<32 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.mul.v64i8(<64 x i8>)

define i8 @vreduce_mul_v64i8(ptr %x) {
; CHECK-LABEL: vreduce_mul_v64i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e8, m4, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    li a0, 32
; CHECK-NEXT:    vslidedown.vx v12, v8, a0
; CHECK-NEXT:    vmul.vv v8, v8, v12
; CHECK-NEXT:    vslidedown.vi v12, v8, 16
; CHECK-NEXT:    vmul.vv v8, v8, v12
; CHECK-NEXT:    vslidedown.vi v12, v8, 8
; CHECK-NEXT:    vmul.vv v8, v8, v12
; CHECK-NEXT:    vslidedown.vi v12, v8, 4
; CHECK-NEXT:    vmul.vv v8, v8, v12
; CHECK-NEXT:    vslidedown.vi v12, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v12
; CHECK-NEXT:    vrgather.vi v12, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v12
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.mul.v64i8(<64 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.mul.v128i8(<128 x i8>)

define i8 @vreduce_mul_v128i8(ptr %x) {
; CHECK-LABEL: vreduce_mul_v128i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    li a0, 64
; CHECK-NEXT:    vslidedown.vx v16, v8, a0
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    li a0, 32
; CHECK-NEXT:    vslidedown.vx v16, v8, a0
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 16
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 8
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 4
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vrgather.vi v16, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.mul.v128i8(<128 x i8> %v)
  ret i8 %red
}

declare i8 @llvm.vector.reduce.mul.v256i8(<256 x i8>)

define i8 @vreduce_mul_v256i8(ptr %x) {
; CHECK-LABEL: vreduce_mul_v256i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 128
; CHECK-NEXT:    vsetvli zero, a1, e8, m8, ta, ma
; CHECK-NEXT:    vle8.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle8.v v16, (a0)
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    li a0, 64
; CHECK-NEXT:    vslidedown.vx v16, v8, a0
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    li a0, 32
; CHECK-NEXT:    vslidedown.vx v16, v8, a0
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 16
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 8
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 4
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vrgather.vi v16, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <256 x i8>, ptr %x
  %red = call i8 @llvm.vector.reduce.mul.v256i8(<256 x i8> %v)
  ret i8 %red
}

declare i16 @llvm.vector.reduce.mul.v1i16(<1 x i16>)

define i16 @vreduce_mul_v1i16(ptr %x) {
; CHECK-LABEL: vreduce_mul_v1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.mul.v1i16(<1 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.mul.v2i16(<2 x i16>)

define i16 @vreduce_mul_v2i16(ptr %x) {
; CHECK-LABEL: vreduce_mul_v2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e16, mf4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    lh a0, 2(a0)
; CHECK-NEXT:    vmul.vx v8, v8, a0
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.mul.v2i16(<2 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.mul.v4i16(<4 x i16>)

define i16 @vreduce_mul_v4i16(ptr %x) {
; CHECK-LABEL: vreduce_mul_v4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vslidedown.vi v9, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vrgather.vi v9, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.mul.v4i16(<4 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.mul.v8i16(<8 x i16>)

define i16 @vreduce_mul_v8i16(ptr %x) {
; CHECK-LABEL: vreduce_mul_v8i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vslidedown.vi v9, v8, 4
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vslidedown.vi v9, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vrgather.vi v9, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.mul.v8i16(<8 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.mul.v16i16(<16 x i16>)

define i16 @vreduce_mul_v16i16(ptr %x) {
; CHECK-LABEL: vreduce_mul_v16i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e16, m2, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vslidedown.vi v10, v8, 8
; CHECK-NEXT:    vmul.vv v8, v8, v10
; CHECK-NEXT:    vslidedown.vi v10, v8, 4
; CHECK-NEXT:    vmul.vv v8, v8, v10
; CHECK-NEXT:    vslidedown.vi v10, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v10
; CHECK-NEXT:    vrgather.vi v10, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v10
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.mul.v16i16(<16 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.mul.v32i16(<32 x i16>)

define i16 @vreduce_mul_v32i16(ptr %x) {
; CHECK-LABEL: vreduce_mul_v32i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e16, m4, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    vslidedown.vi v12, v8, 16
; CHECK-NEXT:    vmul.vv v8, v8, v12
; CHECK-NEXT:    vslidedown.vi v12, v8, 8
; CHECK-NEXT:    vmul.vv v8, v8, v12
; CHECK-NEXT:    vslidedown.vi v12, v8, 4
; CHECK-NEXT:    vmul.vv v8, v8, v12
; CHECK-NEXT:    vslidedown.vi v12, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v12
; CHECK-NEXT:    vrgather.vi v12, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v12
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.mul.v32i16(<32 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.mul.v64i16(<64 x i16>)

define i16 @vreduce_mul_v64i16(ptr %x) {
; CHECK-LABEL: vreduce_mul_v64i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    li a0, 32
; CHECK-NEXT:    vslidedown.vx v16, v8, a0
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 16
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 8
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 4
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vrgather.vi v16, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.mul.v64i16(<64 x i16> %v)
  ret i16 %red
}

declare i16 @llvm.vector.reduce.mul.v128i16(<128 x i16>)

define i16 @vreduce_mul_v128i16(ptr %x) {
; CHECK-LABEL: vreduce_mul_v128i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 64
; CHECK-NEXT:    vsetvli zero, a1, e16, m8, ta, ma
; CHECK-NEXT:    vle16.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle16.v v16, (a0)
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    li a0, 32
; CHECK-NEXT:    vslidedown.vx v16, v8, a0
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 16
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 8
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 4
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vrgather.vi v16, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <128 x i16>, ptr %x
  %red = call i16 @llvm.vector.reduce.mul.v128i16(<128 x i16> %v)
  ret i16 %red
}

declare i32 @llvm.vector.reduce.mul.v1i32(<1 x i32>)

define i32 @vreduce_mul_v1i32(ptr %x) {
; CHECK-LABEL: vreduce_mul_v1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <1 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.mul.v1i32(<1 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.mul.v2i32(<2 x i32>)

define i32 @vreduce_mul_v2i32(ptr %x) {
; CHECK-LABEL: vreduce_mul_v2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 2, e32, mf2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    lw a0, 4(a0)
; CHECK-NEXT:    vmul.vx v8, v8, a0
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <2 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.mul.v2i32(<2 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.mul.v4i32(<4 x i32>)

define i32 @vreduce_mul_v4i32(ptr %x) {
; CHECK-LABEL: vreduce_mul_v4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vslidedown.vi v9, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vrgather.vi v9, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v9
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <4 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.mul.v4i32(<4 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.mul.v8i32(<8 x i32>)

define i32 @vreduce_mul_v8i32(ptr %x) {
; CHECK-LABEL: vreduce_mul_v8i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 8, e32, m2, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vslidedown.vi v10, v8, 4
; CHECK-NEXT:    vmul.vv v8, v8, v10
; CHECK-NEXT:    vslidedown.vi v10, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v10
; CHECK-NEXT:    vrgather.vi v10, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v10
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <8 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.mul.v8i32(<8 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.mul.v16i32(<16 x i32>)

define i32 @vreduce_mul_v16i32(ptr %x) {
; CHECK-LABEL: vreduce_mul_v16i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vslidedown.vi v12, v8, 8
; CHECK-NEXT:    vmul.vv v8, v8, v12
; CHECK-NEXT:    vslidedown.vi v12, v8, 4
; CHECK-NEXT:    vmul.vv v8, v8, v12
; CHECK-NEXT:    vslidedown.vi v12, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v12
; CHECK-NEXT:    vrgather.vi v12, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v12
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <16 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.mul.v16i32(<16 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.mul.v32i32(<32 x i32>)

define i32 @vreduce_mul_v32i32(ptr %x) {
; CHECK-LABEL: vreduce_mul_v32i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    vslidedown.vi v16, v8, 16
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 8
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 4
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vrgather.vi v16, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <32 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.mul.v32i32(<32 x i32> %v)
  ret i32 %red
}

declare i32 @llvm.vector.reduce.mul.v64i32(<64 x i32>)

define i32 @vreduce_mul_v64i32(ptr %x) {
; CHECK-LABEL: vreduce_mul_v64i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    li a1, 32
; CHECK-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; CHECK-NEXT:    vle32.v v8, (a0)
; CHECK-NEXT:    addi a0, a0, 128
; CHECK-NEXT:    vle32.v v16, (a0)
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 16
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 8
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 4
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vslidedown.vi v16, v8, 2
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vrgather.vi v16, v8, 1
; CHECK-NEXT:    vmul.vv v8, v8, v16
; CHECK-NEXT:    vmv.x.s a0, v8
; CHECK-NEXT:    ret
  %v = load <64 x i32>, ptr %x
  %red = call i32 @llvm.vector.reduce.mul.v64i32(<64 x i32> %v)
  ret i32 %red
}

declare i64 @llvm.vector.reduce.mul.v1i64(<1 x i64>)

define i64 @vreduce_mul_v1i64(ptr %x) {
; RV32-LABEL: vreduce_mul_v1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    li a0, 32
; RV32-NEXT:    vsrl.vx v9, v8, a0
; RV32-NEXT:    vmv.x.s a1, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_mul_v1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <1 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.mul.v1i64(<1 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.mul.v2i64(<2 x i64>)

define i64 @vreduce_mul_v2i64(ptr %x) {
; RV32-LABEL: vreduce_mul_v2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a0, a0, 8
; RV32-NEXT:    vlse64.v v9, (a0), zero
; RV32-NEXT:    vmul.vv v8, v8, v9
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_mul_v2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    ld a0, 8(a0)
; RV64-NEXT:    vmul.vx v8, v8, a0
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <2 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.mul.v2i64(<2 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.mul.v4i64(<4 x i64>)

define i64 @vreduce_mul_v4i64(ptr %x) {
; RV32-LABEL: vreduce_mul_v4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vslidedown.vi v10, v8, 2
; RV32-NEXT:    vmul.vv v8, v8, v10
; RV32-NEXT:    vrgather.vi v10, v8, 1
; RV32-NEXT:    vmul.vv v8, v8, v10
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m2, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_mul_v4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 4, e64, m2, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vslidedown.vi v10, v8, 2
; RV64-NEXT:    vmul.vv v8, v8, v10
; RV64-NEXT:    vrgather.vi v10, v8, 1
; RV64-NEXT:    vmul.vv v8, v8, v10
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <4 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.mul.v4i64(<4 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.mul.v8i64(<8 x i64>)

define i64 @vreduce_mul_v8i64(ptr %x) {
; RV32-LABEL: vreduce_mul_v8i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vslidedown.vi v12, v8, 4
; RV32-NEXT:    vmul.vv v8, v8, v12
; RV32-NEXT:    vslidedown.vi v12, v8, 2
; RV32-NEXT:    vmul.vv v8, v8, v12
; RV32-NEXT:    vrgather.vi v12, v8, 1
; RV32-NEXT:    vmul.vv v8, v8, v12
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m4, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_mul_v8i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 8, e64, m4, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vslidedown.vi v12, v8, 4
; RV64-NEXT:    vmul.vv v8, v8, v12
; RV64-NEXT:    vslidedown.vi v12, v8, 2
; RV64-NEXT:    vmul.vv v8, v8, v12
; RV64-NEXT:    vrgather.vi v12, v8, 1
; RV64-NEXT:    vmul.vv v8, v8, v12
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <8 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.mul.v8i64(<8 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.mul.v16i64(<16 x i64>)

define i64 @vreduce_mul_v16i64(ptr %x) {
; RV32-LABEL: vreduce_mul_v16i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    vslidedown.vi v16, v8, 8
; RV32-NEXT:    vmul.vv v8, v8, v16
; RV32-NEXT:    vslidedown.vi v16, v8, 4
; RV32-NEXT:    vmul.vv v8, v8, v16
; RV32-NEXT:    vslidedown.vi v16, v8, 2
; RV32-NEXT:    vmul.vv v8, v8, v16
; RV32-NEXT:    vrgather.vi v16, v8, 1
; RV32-NEXT:    vmul.vv v8, v8, v16
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m8, ta, ma
; RV32-NEXT:    vsrl.vx v8, v8, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_mul_v16i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    vslidedown.vi v16, v8, 8
; RV64-NEXT:    vmul.vv v8, v8, v16
; RV64-NEXT:    vslidedown.vi v16, v8, 4
; RV64-NEXT:    vmul.vv v8, v8, v16
; RV64-NEXT:    vslidedown.vi v16, v8, 2
; RV64-NEXT:    vmul.vv v8, v8, v16
; RV64-NEXT:    vrgather.vi v16, v8, 1
; RV64-NEXT:    vmul.vv v8, v8, v16
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <16 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.mul.v16i64(<16 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.mul.v32i64(<32 x i64>)

define i64 @vreduce_mul_v32i64(ptr %x) {
; RV32-LABEL: vreduce_mul_v32i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v16, (a0)
; RV32-NEXT:    vmul.vv v8, v8, v16
; RV32-NEXT:    vslidedown.vi v16, v8, 8
; RV32-NEXT:    vmul.vv v8, v8, v16
; RV32-NEXT:    vslidedown.vi v16, v8, 4
; RV32-NEXT:    vmul.vv v8, v8, v16
; RV32-NEXT:    vslidedown.vi v16, v8, 2
; RV32-NEXT:    vmul.vv v8, v8, v16
; RV32-NEXT:    vrgather.vi v16, v8, 1
; RV32-NEXT:    vmul.vv v8, v8, v16
; RV32-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    vslidedown.vi v8, v8, 1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_mul_v32i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v16, (a0)
; RV64-NEXT:    vmul.vv v8, v8, v16
; RV64-NEXT:    vslidedown.vi v16, v8, 8
; RV64-NEXT:    vmul.vv v8, v8, v16
; RV64-NEXT:    vslidedown.vi v16, v8, 4
; RV64-NEXT:    vmul.vv v8, v8, v16
; RV64-NEXT:    vslidedown.vi v16, v8, 2
; RV64-NEXT:    vmul.vv v8, v8, v16
; RV64-NEXT:    vrgather.vi v16, v8, 1
; RV64-NEXT:    vmul.vv v8, v8, v16
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <32 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.mul.v32i64(<32 x i64> %v)
  ret i64 %red
}

declare i64 @llvm.vector.reduce.mul.v64i64(<64 x i64>)

define i64 @vreduce_mul_v64i64(ptr %x) nounwind {
; RV32-LABEL: vreduce_mul_v64i64:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV32-NEXT:    vle64.v v8, (a0)
; RV32-NEXT:    addi a1, a0, 384
; RV32-NEXT:    vle64.v v16, (a1)
; RV32-NEXT:    addi a1, a0, 256
; RV32-NEXT:    addi a0, a0, 128
; RV32-NEXT:    vle64.v v24, (a0)
; RV32-NEXT:    vle64.v v0, (a1)
; RV32-NEXT:    vmul.vv v16, v24, v16
; RV32-NEXT:    vmul.vv v8, v8, v0
; RV32-NEXT:    vmul.vv v8, v8, v16
; RV32-NEXT:    vslidedown.vi v16, v8, 8
; RV32-NEXT:    vmul.vv v8, v8, v16
; RV32-NEXT:    vslidedown.vi v16, v8, 4
; RV32-NEXT:    vmul.vv v8, v8, v16
; RV32-NEXT:    vslidedown.vi v16, v8, 2
; RV32-NEXT:    vmul.vv v8, v8, v16
; RV32-NEXT:    vrgather.vi v16, v8, 1
; RV32-NEXT:    vmul.vv v8, v8, v16
; RV32-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; RV32-NEXT:    vmv.x.s a0, v8
; RV32-NEXT:    vslidedown.vi v8, v8, 1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    ret
;
; RV64-LABEL: vreduce_mul_v64i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 16, e64, m8, ta, ma
; RV64-NEXT:    vle64.v v8, (a0)
; RV64-NEXT:    addi a1, a0, 384
; RV64-NEXT:    vle64.v v16, (a1)
; RV64-NEXT:    addi a1, a0, 256
; RV64-NEXT:    addi a0, a0, 128
; RV64-NEXT:    vle64.v v24, (a0)
; RV64-NEXT:    vle64.v v0, (a1)
; RV64-NEXT:    vmul.vv v16, v24, v16
; RV64-NEXT:    vmul.vv v8, v8, v0
; RV64-NEXT:    vmul.vv v8, v8, v16
; RV64-NEXT:    vslidedown.vi v16, v8, 8
; RV64-NEXT:    vmul.vv v8, v8, v16
; RV64-NEXT:    vslidedown.vi v16, v8, 4
; RV64-NEXT:    vmul.vv v8, v8, v16
; RV64-NEXT:    vslidedown.vi v16, v8, 2
; RV64-NEXT:    vmul.vv v8, v8, v16
; RV64-NEXT:    vrgather.vi v16, v8, 1
; RV64-NEXT:    vmul.vv v8, v8, v16
; RV64-NEXT:    vmv.x.s a0, v8
; RV64-NEXT:    ret
  %v = load <64 x i64>, ptr %x
  %red = call i64 @llvm.vector.reduce.mul.v64i64(<64 x i64> %v)
  ret i64 %red
}