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

declare i8 @llvm.vp.reduce.add.nxv1i8(i8, <vscale x 1 x i8>, <vscale x 1 x i1>, i32)

define signext i8 @vpreduce_add_nxv1i8(i8 signext %s, <vscale x 1 x i8> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_add_nxv1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf8, ta, ma
; CHECK-NEXT:    vredsum.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.add.nxv1i8(i8 %s, <vscale x 1 x i8> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.umax.nxv1i8(i8, <vscale x 1 x i8>, <vscale x 1 x i1>, i32)

define signext i8 @vpreduce_umax_nxv1i8(i8 signext %s, <vscale x 1 x i8> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_umax_nxv1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    andi a0, a0, 255
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf8, ta, ma
; CHECK-NEXT:    vredmaxu.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.umax.nxv1i8(i8 %s, <vscale x 1 x i8> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.smax.nxv1i8(i8, <vscale x 1 x i8>, <vscale x 1 x i1>, i32)

define signext i8 @vpreduce_smax_nxv1i8(i8 signext %s, <vscale x 1 x i8> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smax_nxv1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf8, ta, ma
; CHECK-NEXT:    vredmax.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.smax.nxv1i8(i8 %s, <vscale x 1 x i8> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.umin.nxv1i8(i8, <vscale x 1 x i8>, <vscale x 1 x i1>, i32)

define signext i8 @vpreduce_umin_nxv1i8(i8 signext %s, <vscale x 1 x i8> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_umin_nxv1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    andi a0, a0, 255
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf8, ta, ma
; CHECK-NEXT:    vredminu.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.umin.nxv1i8(i8 %s, <vscale x 1 x i8> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.smin.nxv1i8(i8, <vscale x 1 x i8>, <vscale x 1 x i1>, i32)

define signext i8 @vpreduce_smin_nxv1i8(i8 signext %s, <vscale x 1 x i8> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smin_nxv1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf8, ta, ma
; CHECK-NEXT:    vredmin.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.smin.nxv1i8(i8 %s, <vscale x 1 x i8> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.and.nxv1i8(i8, <vscale x 1 x i8>, <vscale x 1 x i1>, i32)

define signext i8 @vpreduce_and_nxv1i8(i8 signext %s, <vscale x 1 x i8> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_and_nxv1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf8, ta, ma
; CHECK-NEXT:    vredand.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.and.nxv1i8(i8 %s, <vscale x 1 x i8> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.or.nxv1i8(i8, <vscale x 1 x i8>, <vscale x 1 x i1>, i32)

define signext i8 @vpreduce_or_nxv1i8(i8 signext %s, <vscale x 1 x i8> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_or_nxv1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf8, ta, ma
; CHECK-NEXT:    vredor.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.or.nxv1i8(i8 %s, <vscale x 1 x i8> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.xor.nxv1i8(i8, <vscale x 1 x i8>, <vscale x 1 x i1>, i32)

define signext i8 @vpreduce_xor_nxv1i8(i8 signext %s, <vscale x 1 x i8> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_xor_nxv1i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf8, ta, ma
; CHECK-NEXT:    vredxor.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.xor.nxv1i8(i8 %s, <vscale x 1 x i8> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.add.nxv2i8(i8, <vscale x 2 x i8>, <vscale x 2 x i1>, i32)

define signext i8 @vpreduce_add_nxv2i8(i8 signext %s, <vscale x 2 x i8> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_add_nxv2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf4, ta, ma
; CHECK-NEXT:    vredsum.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.add.nxv2i8(i8 %s, <vscale x 2 x i8> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.umax.nxv2i8(i8, <vscale x 2 x i8>, <vscale x 2 x i1>, i32)

define signext i8 @vpreduce_umax_nxv2i8(i8 signext %s, <vscale x 2 x i8> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_umax_nxv2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    andi a0, a0, 255
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf4, ta, ma
; CHECK-NEXT:    vredmaxu.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.umax.nxv2i8(i8 %s, <vscale x 2 x i8> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.smax.nxv2i8(i8, <vscale x 2 x i8>, <vscale x 2 x i1>, i32)

define signext i8 @vpreduce_smax_nxv2i8(i8 signext %s, <vscale x 2 x i8> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smax_nxv2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf4, ta, ma
; CHECK-NEXT:    vredmax.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.smax.nxv2i8(i8 %s, <vscale x 2 x i8> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.umin.nxv2i8(i8, <vscale x 2 x i8>, <vscale x 2 x i1>, i32)

define signext i8 @vpreduce_umin_nxv2i8(i8 signext %s, <vscale x 2 x i8> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_umin_nxv2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    andi a0, a0, 255
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf4, ta, ma
; CHECK-NEXT:    vredminu.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.umin.nxv2i8(i8 %s, <vscale x 2 x i8> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.smin.nxv2i8(i8, <vscale x 2 x i8>, <vscale x 2 x i1>, i32)

define signext i8 @vpreduce_smin_nxv2i8(i8 signext %s, <vscale x 2 x i8> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smin_nxv2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf4, ta, ma
; CHECK-NEXT:    vredmin.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.smin.nxv2i8(i8 %s, <vscale x 2 x i8> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.and.nxv2i8(i8, <vscale x 2 x i8>, <vscale x 2 x i1>, i32)

define signext i8 @vpreduce_and_nxv2i8(i8 signext %s, <vscale x 2 x i8> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_and_nxv2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf4, ta, ma
; CHECK-NEXT:    vredand.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.and.nxv2i8(i8 %s, <vscale x 2 x i8> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.or.nxv2i8(i8, <vscale x 2 x i8>, <vscale x 2 x i1>, i32)

define signext i8 @vpreduce_or_nxv2i8(i8 signext %s, <vscale x 2 x i8> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_or_nxv2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf4, ta, ma
; CHECK-NEXT:    vredor.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.or.nxv2i8(i8 %s, <vscale x 2 x i8> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.xor.nxv2i8(i8, <vscale x 2 x i8>, <vscale x 2 x i1>, i32)

define signext i8 @vpreduce_xor_nxv2i8(i8 signext %s, <vscale x 2 x i8> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_xor_nxv2i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf4, ta, ma
; CHECK-NEXT:    vredxor.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.xor.nxv2i8(i8 %s, <vscale x 2 x i8> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.smax.nxv3i8(i8, <vscale x 3 x i8>, <vscale x 3 x i1>, i32)

define signext i8 @vpreduce_smax_nxv3i8(i8 signext %s, <vscale x 3 x i8> %v, <vscale x 3 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smax_nxv3i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf2, ta, ma
; CHECK-NEXT:    vredmax.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.smax.nxv3i8(i8 %s, <vscale x 3 x i8> %v, <vscale x 3 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.add.nxv4i8(i8, <vscale x 4 x i8>, <vscale x 4 x i1>, i32)

define signext i8 @vpreduce_add_nxv4i8(i8 signext %s, <vscale x 4 x i8> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_add_nxv4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf2, ta, ma
; CHECK-NEXT:    vredsum.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.add.nxv4i8(i8 %s, <vscale x 4 x i8> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.umax.nxv4i8(i8, <vscale x 4 x i8>, <vscale x 4 x i1>, i32)

define signext i8 @vpreduce_umax_nxv4i8(i8 signext %s, <vscale x 4 x i8> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_umax_nxv4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    andi a0, a0, 255
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf2, ta, ma
; CHECK-NEXT:    vredmaxu.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.umax.nxv4i8(i8 %s, <vscale x 4 x i8> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.smax.nxv4i8(i8, <vscale x 4 x i8>, <vscale x 4 x i1>, i32)

define signext i8 @vpreduce_smax_nxv4i8(i8 signext %s, <vscale x 4 x i8> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smax_nxv4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf2, ta, ma
; CHECK-NEXT:    vredmax.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.smax.nxv4i8(i8 %s, <vscale x 4 x i8> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.umin.nxv4i8(i8, <vscale x 4 x i8>, <vscale x 4 x i1>, i32)

define signext i8 @vpreduce_umin_nxv4i8(i8 signext %s, <vscale x 4 x i8> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_umin_nxv4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    andi a0, a0, 255
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf2, ta, ma
; CHECK-NEXT:    vredminu.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.umin.nxv4i8(i8 %s, <vscale x 4 x i8> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.smin.nxv4i8(i8, <vscale x 4 x i8>, <vscale x 4 x i1>, i32)

define signext i8 @vpreduce_smin_nxv4i8(i8 signext %s, <vscale x 4 x i8> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smin_nxv4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf2, ta, ma
; CHECK-NEXT:    vredmin.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.smin.nxv4i8(i8 %s, <vscale x 4 x i8> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.and.nxv4i8(i8, <vscale x 4 x i8>, <vscale x 4 x i1>, i32)

define signext i8 @vpreduce_and_nxv4i8(i8 signext %s, <vscale x 4 x i8> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_and_nxv4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf2, ta, ma
; CHECK-NEXT:    vredand.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.and.nxv4i8(i8 %s, <vscale x 4 x i8> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.or.nxv4i8(i8, <vscale x 4 x i8>, <vscale x 4 x i1>, i32)

define signext i8 @vpreduce_or_nxv4i8(i8 signext %s, <vscale x 4 x i8> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_or_nxv4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf2, ta, ma
; CHECK-NEXT:    vredor.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.or.nxv4i8(i8 %s, <vscale x 4 x i8> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i8 @llvm.vp.reduce.xor.nxv4i8(i8, <vscale x 4 x i8>, <vscale x 4 x i1>, i32)

define signext i8 @vpreduce_xor_nxv4i8(i8 signext %s, <vscale x 4 x i8> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_xor_nxv4i8:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e8, mf2, ta, ma
; CHECK-NEXT:    vredxor.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i8 @llvm.vp.reduce.xor.nxv4i8(i8 %s, <vscale x 4 x i8> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i8 %r
}

declare i16 @llvm.vp.reduce.add.nxv1i16(i16, <vscale x 1 x i16>, <vscale x 1 x i1>, i32)

define signext i16 @vpreduce_add_nxv1i16(i16 signext %s, <vscale x 1 x i16> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_add_nxv1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, mf4, ta, ma
; CHECK-NEXT:    vredsum.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.add.nxv1i16(i16 %s, <vscale x 1 x i16> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.umax.nxv1i16(i16, <vscale x 1 x i16>, <vscale x 1 x i1>, i32)

define signext i16 @vpreduce_umax_nxv1i16(i16 signext %s, <vscale x 1 x i16> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umax_nxv1i16:
; RV32:       # %bb.0:
; RV32-NEXT:    slli a0, a0, 16
; RV32-NEXT:    srli a0, a0, 16
; RV32-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; RV32-NEXT:    vmv.s.x v9, a0
; RV32-NEXT:    vsetvli zero, a1, e16, mf4, ta, ma
; RV32-NEXT:    vredmaxu.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umax_nxv1i16:
; RV64:       # %bb.0:
; RV64-NEXT:    slli a0, a0, 48
; RV64-NEXT:    srli a0, a0, 48
; RV64-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e16, mf4, ta, ma
; RV64-NEXT:    vredmaxu.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.umax.nxv1i16(i16 %s, <vscale x 1 x i16> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.smax.nxv1i16(i16, <vscale x 1 x i16>, <vscale x 1 x i1>, i32)

define signext i16 @vpreduce_smax_nxv1i16(i16 signext %s, <vscale x 1 x i16> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smax_nxv1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, mf4, ta, ma
; CHECK-NEXT:    vredmax.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.smax.nxv1i16(i16 %s, <vscale x 1 x i16> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.umin.nxv1i16(i16, <vscale x 1 x i16>, <vscale x 1 x i1>, i32)

define signext i16 @vpreduce_umin_nxv1i16(i16 signext %s, <vscale x 1 x i16> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umin_nxv1i16:
; RV32:       # %bb.0:
; RV32-NEXT:    slli a0, a0, 16
; RV32-NEXT:    srli a0, a0, 16
; RV32-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; RV32-NEXT:    vmv.s.x v9, a0
; RV32-NEXT:    vsetvli zero, a1, e16, mf4, ta, ma
; RV32-NEXT:    vredminu.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umin_nxv1i16:
; RV64:       # %bb.0:
; RV64-NEXT:    slli a0, a0, 48
; RV64-NEXT:    srli a0, a0, 48
; RV64-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e16, mf4, ta, ma
; RV64-NEXT:    vredminu.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.umin.nxv1i16(i16 %s, <vscale x 1 x i16> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.smin.nxv1i16(i16, <vscale x 1 x i16>, <vscale x 1 x i1>, i32)

define signext i16 @vpreduce_smin_nxv1i16(i16 signext %s, <vscale x 1 x i16> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smin_nxv1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, mf4, ta, ma
; CHECK-NEXT:    vredmin.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.smin.nxv1i16(i16 %s, <vscale x 1 x i16> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.and.nxv1i16(i16, <vscale x 1 x i16>, <vscale x 1 x i1>, i32)

define signext i16 @vpreduce_and_nxv1i16(i16 signext %s, <vscale x 1 x i16> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_and_nxv1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, mf4, ta, ma
; CHECK-NEXT:    vredand.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.and.nxv1i16(i16 %s, <vscale x 1 x i16> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.or.nxv1i16(i16, <vscale x 1 x i16>, <vscale x 1 x i1>, i32)

define signext i16 @vpreduce_or_nxv1i16(i16 signext %s, <vscale x 1 x i16> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_or_nxv1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, mf4, ta, ma
; CHECK-NEXT:    vredor.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.or.nxv1i16(i16 %s, <vscale x 1 x i16> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.xor.nxv1i16(i16, <vscale x 1 x i16>, <vscale x 1 x i1>, i32)

define signext i16 @vpreduce_xor_nxv1i16(i16 signext %s, <vscale x 1 x i16> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_xor_nxv1i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, mf4, ta, ma
; CHECK-NEXT:    vredxor.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.xor.nxv1i16(i16 %s, <vscale x 1 x i16> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.add.nxv2i16(i16, <vscale x 2 x i16>, <vscale x 2 x i1>, i32)

define signext i16 @vpreduce_add_nxv2i16(i16 signext %s, <vscale x 2 x i16> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_add_nxv2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, mf2, ta, ma
; CHECK-NEXT:    vredsum.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.add.nxv2i16(i16 %s, <vscale x 2 x i16> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.umax.nxv2i16(i16, <vscale x 2 x i16>, <vscale x 2 x i1>, i32)

define signext i16 @vpreduce_umax_nxv2i16(i16 signext %s, <vscale x 2 x i16> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umax_nxv2i16:
; RV32:       # %bb.0:
; RV32-NEXT:    slli a0, a0, 16
; RV32-NEXT:    srli a0, a0, 16
; RV32-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; RV32-NEXT:    vmv.s.x v9, a0
; RV32-NEXT:    vsetvli zero, a1, e16, mf2, ta, ma
; RV32-NEXT:    vredmaxu.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umax_nxv2i16:
; RV64:       # %bb.0:
; RV64-NEXT:    slli a0, a0, 48
; RV64-NEXT:    srli a0, a0, 48
; RV64-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e16, mf2, ta, ma
; RV64-NEXT:    vredmaxu.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.umax.nxv2i16(i16 %s, <vscale x 2 x i16> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.smax.nxv2i16(i16, <vscale x 2 x i16>, <vscale x 2 x i1>, i32)

define signext i16 @vpreduce_smax_nxv2i16(i16 signext %s, <vscale x 2 x i16> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smax_nxv2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, mf2, ta, ma
; CHECK-NEXT:    vredmax.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.smax.nxv2i16(i16 %s, <vscale x 2 x i16> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.umin.nxv2i16(i16, <vscale x 2 x i16>, <vscale x 2 x i1>, i32)

define signext i16 @vpreduce_umin_nxv2i16(i16 signext %s, <vscale x 2 x i16> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umin_nxv2i16:
; RV32:       # %bb.0:
; RV32-NEXT:    slli a0, a0, 16
; RV32-NEXT:    srli a0, a0, 16
; RV32-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; RV32-NEXT:    vmv.s.x v9, a0
; RV32-NEXT:    vsetvli zero, a1, e16, mf2, ta, ma
; RV32-NEXT:    vredminu.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umin_nxv2i16:
; RV64:       # %bb.0:
; RV64-NEXT:    slli a0, a0, 48
; RV64-NEXT:    srli a0, a0, 48
; RV64-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e16, mf2, ta, ma
; RV64-NEXT:    vredminu.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.umin.nxv2i16(i16 %s, <vscale x 2 x i16> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.smin.nxv2i16(i16, <vscale x 2 x i16>, <vscale x 2 x i1>, i32)

define signext i16 @vpreduce_smin_nxv2i16(i16 signext %s, <vscale x 2 x i16> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smin_nxv2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, mf2, ta, ma
; CHECK-NEXT:    vredmin.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.smin.nxv2i16(i16 %s, <vscale x 2 x i16> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.and.nxv2i16(i16, <vscale x 2 x i16>, <vscale x 2 x i1>, i32)

define signext i16 @vpreduce_and_nxv2i16(i16 signext %s, <vscale x 2 x i16> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_and_nxv2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, mf2, ta, ma
; CHECK-NEXT:    vredand.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.and.nxv2i16(i16 %s, <vscale x 2 x i16> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.or.nxv2i16(i16, <vscale x 2 x i16>, <vscale x 2 x i1>, i32)

define signext i16 @vpreduce_or_nxv2i16(i16 signext %s, <vscale x 2 x i16> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_or_nxv2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, mf2, ta, ma
; CHECK-NEXT:    vredor.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.or.nxv2i16(i16 %s, <vscale x 2 x i16> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.xor.nxv2i16(i16, <vscale x 2 x i16>, <vscale x 2 x i1>, i32)

define signext i16 @vpreduce_xor_nxv2i16(i16 signext %s, <vscale x 2 x i16> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_xor_nxv2i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, mf2, ta, ma
; CHECK-NEXT:    vredxor.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.xor.nxv2i16(i16 %s, <vscale x 2 x i16> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.add.nxv4i16(i16, <vscale x 4 x i16>, <vscale x 4 x i1>, i32)

define signext i16 @vpreduce_add_nxv4i16(i16 signext %s, <vscale x 4 x i16> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_add_nxv4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, m1, ta, ma
; CHECK-NEXT:    vredsum.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.add.nxv4i16(i16 %s, <vscale x 4 x i16> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.umax.nxv4i16(i16, <vscale x 4 x i16>, <vscale x 4 x i1>, i32)

define signext i16 @vpreduce_umax_nxv4i16(i16 signext %s, <vscale x 4 x i16> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umax_nxv4i16:
; RV32:       # %bb.0:
; RV32-NEXT:    slli a0, a0, 16
; RV32-NEXT:    srli a0, a0, 16
; RV32-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; RV32-NEXT:    vmv.s.x v9, a0
; RV32-NEXT:    vsetvli zero, a1, e16, m1, ta, ma
; RV32-NEXT:    vredmaxu.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umax_nxv4i16:
; RV64:       # %bb.0:
; RV64-NEXT:    slli a0, a0, 48
; RV64-NEXT:    srli a0, a0, 48
; RV64-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e16, m1, ta, ma
; RV64-NEXT:    vredmaxu.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.umax.nxv4i16(i16 %s, <vscale x 4 x i16> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.smax.nxv4i16(i16, <vscale x 4 x i16>, <vscale x 4 x i1>, i32)

define signext i16 @vpreduce_smax_nxv4i16(i16 signext %s, <vscale x 4 x i16> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smax_nxv4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, m1, ta, ma
; CHECK-NEXT:    vredmax.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.smax.nxv4i16(i16 %s, <vscale x 4 x i16> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.umin.nxv4i16(i16, <vscale x 4 x i16>, <vscale x 4 x i1>, i32)

define signext i16 @vpreduce_umin_nxv4i16(i16 signext %s, <vscale x 4 x i16> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umin_nxv4i16:
; RV32:       # %bb.0:
; RV32-NEXT:    slli a0, a0, 16
; RV32-NEXT:    srli a0, a0, 16
; RV32-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; RV32-NEXT:    vmv.s.x v9, a0
; RV32-NEXT:    vsetvli zero, a1, e16, m1, ta, ma
; RV32-NEXT:    vredminu.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umin_nxv4i16:
; RV64:       # %bb.0:
; RV64-NEXT:    slli a0, a0, 48
; RV64-NEXT:    srli a0, a0, 48
; RV64-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e16, m1, ta, ma
; RV64-NEXT:    vredminu.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.umin.nxv4i16(i16 %s, <vscale x 4 x i16> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.smin.nxv4i16(i16, <vscale x 4 x i16>, <vscale x 4 x i1>, i32)

define signext i16 @vpreduce_smin_nxv4i16(i16 signext %s, <vscale x 4 x i16> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smin_nxv4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, m1, ta, ma
; CHECK-NEXT:    vredmin.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.smin.nxv4i16(i16 %s, <vscale x 4 x i16> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.and.nxv4i16(i16, <vscale x 4 x i16>, <vscale x 4 x i1>, i32)

define signext i16 @vpreduce_and_nxv4i16(i16 signext %s, <vscale x 4 x i16> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_and_nxv4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, m1, ta, ma
; CHECK-NEXT:    vredand.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.and.nxv4i16(i16 %s, <vscale x 4 x i16> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.or.nxv4i16(i16, <vscale x 4 x i16>, <vscale x 4 x i1>, i32)

define signext i16 @vpreduce_or_nxv4i16(i16 signext %s, <vscale x 4 x i16> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_or_nxv4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, m1, ta, ma
; CHECK-NEXT:    vredor.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.or.nxv4i16(i16 %s, <vscale x 4 x i16> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i16 @llvm.vp.reduce.xor.nxv4i16(i16, <vscale x 4 x i16>, <vscale x 4 x i1>, i32)

define signext i16 @vpreduce_xor_nxv4i16(i16 signext %s, <vscale x 4 x i16> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_xor_nxv4i16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e16, m1, ta, ma
; CHECK-NEXT:    vredxor.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i16 @llvm.vp.reduce.xor.nxv4i16(i16 %s, <vscale x 4 x i16> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i16 %r
}

declare i32 @llvm.vp.reduce.add.nxv1i32(i32, <vscale x 1 x i32>, <vscale x 1 x i1>, i32)

define signext i32 @vpreduce_add_nxv1i32(i32 signext %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_add_nxv1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, mf2, ta, ma
; CHECK-NEXT:    vredsum.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.add.nxv1i32(i32 %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.umax.nxv1i32(i32, <vscale x 1 x i32>, <vscale x 1 x i1>, i32)

define signext i32 @vpreduce_umax_nxv1i32(i32 signext %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umax_nxv1i32:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; RV32-NEXT:    vmv.s.x v9, a0
; RV32-NEXT:    vsetvli zero, a1, e32, mf2, ta, ma
; RV32-NEXT:    vredmaxu.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umax_nxv1i32:
; RV64:       # %bb.0:
; RV64-NEXT:    andi a0, a0, -1
; RV64-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e32, mf2, ta, ma
; RV64-NEXT:    vredmaxu.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.umax.nxv1i32(i32 %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.smax.nxv1i32(i32, <vscale x 1 x i32>, <vscale x 1 x i1>, i32)

define signext i32 @vpreduce_smax_nxv1i32(i32 signext %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smax_nxv1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, mf2, ta, ma
; CHECK-NEXT:    vredmax.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.smax.nxv1i32(i32 %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.umin.nxv1i32(i32, <vscale x 1 x i32>, <vscale x 1 x i1>, i32)

define signext i32 @vpreduce_umin_nxv1i32(i32 signext %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umin_nxv1i32:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; RV32-NEXT:    vmv.s.x v9, a0
; RV32-NEXT:    vsetvli zero, a1, e32, mf2, ta, ma
; RV32-NEXT:    vredminu.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umin_nxv1i32:
; RV64:       # %bb.0:
; RV64-NEXT:    andi a0, a0, -1
; RV64-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e32, mf2, ta, ma
; RV64-NEXT:    vredminu.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.umin.nxv1i32(i32 %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.smin.nxv1i32(i32, <vscale x 1 x i32>, <vscale x 1 x i1>, i32)

define signext i32 @vpreduce_smin_nxv1i32(i32 signext %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smin_nxv1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, mf2, ta, ma
; CHECK-NEXT:    vredmin.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.smin.nxv1i32(i32 %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.and.nxv1i32(i32, <vscale x 1 x i32>, <vscale x 1 x i1>, i32)

define signext i32 @vpreduce_and_nxv1i32(i32 signext %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_and_nxv1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, mf2, ta, ma
; CHECK-NEXT:    vredand.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.and.nxv1i32(i32 %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.or.nxv1i32(i32, <vscale x 1 x i32>, <vscale x 1 x i1>, i32)

define signext i32 @vpreduce_or_nxv1i32(i32 signext %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_or_nxv1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, mf2, ta, ma
; CHECK-NEXT:    vredor.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.or.nxv1i32(i32 %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.xor.nxv1i32(i32, <vscale x 1 x i32>, <vscale x 1 x i1>, i32)

define signext i32 @vpreduce_xor_nxv1i32(i32 signext %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_xor_nxv1i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, mf2, ta, ma
; CHECK-NEXT:    vredxor.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.xor.nxv1i32(i32 %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.add.nxv2i32(i32, <vscale x 2 x i32>, <vscale x 2 x i1>, i32)

define signext i32 @vpreduce_add_nxv2i32(i32 signext %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_add_nxv2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
; CHECK-NEXT:    vredsum.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.add.nxv2i32(i32 %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.umax.nxv2i32(i32, <vscale x 2 x i32>, <vscale x 2 x i1>, i32)

define signext i32 @vpreduce_umax_nxv2i32(i32 signext %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umax_nxv2i32:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; RV32-NEXT:    vmv.s.x v9, a0
; RV32-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
; RV32-NEXT:    vredmaxu.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umax_nxv2i32:
; RV64:       # %bb.0:
; RV64-NEXT:    andi a0, a0, -1
; RV64-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
; RV64-NEXT:    vredmaxu.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.umax.nxv2i32(i32 %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.smax.nxv2i32(i32, <vscale x 2 x i32>, <vscale x 2 x i1>, i32)

define signext i32 @vpreduce_smax_nxv2i32(i32 signext %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smax_nxv2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
; CHECK-NEXT:    vredmax.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.smax.nxv2i32(i32 %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.umin.nxv2i32(i32, <vscale x 2 x i32>, <vscale x 2 x i1>, i32)

define signext i32 @vpreduce_umin_nxv2i32(i32 signext %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umin_nxv2i32:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; RV32-NEXT:    vmv.s.x v9, a0
; RV32-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
; RV32-NEXT:    vredminu.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umin_nxv2i32:
; RV64:       # %bb.0:
; RV64-NEXT:    andi a0, a0, -1
; RV64-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
; RV64-NEXT:    vredminu.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.umin.nxv2i32(i32 %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.smin.nxv2i32(i32, <vscale x 2 x i32>, <vscale x 2 x i1>, i32)

define signext i32 @vpreduce_smin_nxv2i32(i32 signext %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smin_nxv2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
; CHECK-NEXT:    vredmin.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.smin.nxv2i32(i32 %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.and.nxv2i32(i32, <vscale x 2 x i32>, <vscale x 2 x i1>, i32)

define signext i32 @vpreduce_and_nxv2i32(i32 signext %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_and_nxv2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
; CHECK-NEXT:    vredand.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.and.nxv2i32(i32 %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.or.nxv2i32(i32, <vscale x 2 x i32>, <vscale x 2 x i1>, i32)

define signext i32 @vpreduce_or_nxv2i32(i32 signext %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_or_nxv2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
; CHECK-NEXT:    vredor.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.or.nxv2i32(i32 %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.xor.nxv2i32(i32, <vscale x 2 x i32>, <vscale x 2 x i1>, i32)

define signext i32 @vpreduce_xor_nxv2i32(i32 signext %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_xor_nxv2i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v9, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
; CHECK-NEXT:    vredxor.vs v9, v8, v9, v0.t
; CHECK-NEXT:    vmv.x.s a0, v9
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.xor.nxv2i32(i32 %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.add.nxv4i32(i32, <vscale x 4 x i32>, <vscale x 4 x i1>, i32)

define signext i32 @vpreduce_add_nxv4i32(i32 signext %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_add_nxv4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v10, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
; CHECK-NEXT:    vredsum.vs v10, v8, v10, v0.t
; CHECK-NEXT:    vmv.x.s a0, v10
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.add.nxv4i32(i32 %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.umax.nxv4i32(i32, <vscale x 4 x i32>, <vscale x 4 x i1>, i32)

define signext i32 @vpreduce_umax_nxv4i32(i32 signext %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umax_nxv4i32:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; RV32-NEXT:    vmv.s.x v10, a0
; RV32-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
; RV32-NEXT:    vredmaxu.vs v10, v8, v10, v0.t
; RV32-NEXT:    vmv.x.s a0, v10
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umax_nxv4i32:
; RV64:       # %bb.0:
; RV64-NEXT:    andi a0, a0, -1
; RV64-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; RV64-NEXT:    vmv.s.x v10, a0
; RV64-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
; RV64-NEXT:    vredmaxu.vs v10, v8, v10, v0.t
; RV64-NEXT:    vmv.x.s a0, v10
; RV64-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.umax.nxv4i32(i32 %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.umax.nxv32i32(i32, <vscale x 32 x i32>, <vscale x 32 x i1>, i32)

define signext i32 @vpreduce_umax_nxv32i32(i32 signext %s, <vscale x 32 x i32> %v, <vscale x 32 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umax_nxv32i32:
; RV32:       # %bb.0:
; RV32-NEXT:    csrr a3, vlenb
; RV32-NEXT:    srli a2, a3, 2
; RV32-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
; RV32-NEXT:    vslidedown.vx v24, v0, a2
; RV32-NEXT:    slli a3, a3, 1
; RV32-NEXT:    sub a2, a1, a3
; RV32-NEXT:    sltu a4, a1, a2
; RV32-NEXT:    addi a4, a4, -1
; RV32-NEXT:    and a2, a4, a2
; RV32-NEXT:    bltu a1, a3, .LBB67_2
; RV32-NEXT:  # %bb.1:
; RV32-NEXT:    mv a1, a3
; RV32-NEXT:  .LBB67_2:
; RV32-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
; RV32-NEXT:    vmv.s.x v25, a0
; RV32-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; RV32-NEXT:    vredmaxu.vs v25, v8, v25, v0.t
; RV32-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
; RV32-NEXT:    vmv1r.v v0, v24
; RV32-NEXT:    vredmaxu.vs v25, v16, v25, v0.t
; RV32-NEXT:    vmv.x.s a0, v25
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umax_nxv32i32:
; RV64:       # %bb.0:
; RV64-NEXT:    csrr a3, vlenb
; RV64-NEXT:    srli a2, a3, 2
; RV64-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
; RV64-NEXT:    vslidedown.vx v24, v0, a2
; RV64-NEXT:    andi a2, a0, -1
; RV64-NEXT:    slli a3, a3, 1
; RV64-NEXT:    sub a0, a1, a3
; RV64-NEXT:    sltu a4, a1, a0
; RV64-NEXT:    addi a4, a4, -1
; RV64-NEXT:    and a0, a4, a0
; RV64-NEXT:    bltu a1, a3, .LBB67_2
; RV64-NEXT:  # %bb.1:
; RV64-NEXT:    mv a1, a3
; RV64-NEXT:  .LBB67_2:
; RV64-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
; RV64-NEXT:    vmv.s.x v25, a2
; RV64-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
; RV64-NEXT:    vredmaxu.vs v25, v8, v25, v0.t
; RV64-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
; RV64-NEXT:    vmv1r.v v0, v24
; RV64-NEXT:    vredmaxu.vs v25, v16, v25, v0.t
; RV64-NEXT:    vmv.x.s a0, v25
; RV64-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.umax.nxv32i32(i32 %s, <vscale x 32 x i32> %v, <vscale x 32 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.smax.nxv4i32(i32, <vscale x 4 x i32>, <vscale x 4 x i1>, i32)

define signext i32 @vpreduce_smax_nxv4i32(i32 signext %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smax_nxv4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v10, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
; CHECK-NEXT:    vredmax.vs v10, v8, v10, v0.t
; CHECK-NEXT:    vmv.x.s a0, v10
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.smax.nxv4i32(i32 %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.umin.nxv4i32(i32, <vscale x 4 x i32>, <vscale x 4 x i1>, i32)

define signext i32 @vpreduce_umin_nxv4i32(i32 signext %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umin_nxv4i32:
; RV32:       # %bb.0:
; RV32-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; RV32-NEXT:    vmv.s.x v10, a0
; RV32-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
; RV32-NEXT:    vredminu.vs v10, v8, v10, v0.t
; RV32-NEXT:    vmv.x.s a0, v10
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umin_nxv4i32:
; RV64:       # %bb.0:
; RV64-NEXT:    andi a0, a0, -1
; RV64-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; RV64-NEXT:    vmv.s.x v10, a0
; RV64-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
; RV64-NEXT:    vredminu.vs v10, v8, v10, v0.t
; RV64-NEXT:    vmv.x.s a0, v10
; RV64-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.umin.nxv4i32(i32 %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.smin.nxv4i32(i32, <vscale x 4 x i32>, <vscale x 4 x i1>, i32)

define signext i32 @vpreduce_smin_nxv4i32(i32 signext %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_smin_nxv4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v10, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
; CHECK-NEXT:    vredmin.vs v10, v8, v10, v0.t
; CHECK-NEXT:    vmv.x.s a0, v10
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.smin.nxv4i32(i32 %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.and.nxv4i32(i32, <vscale x 4 x i32>, <vscale x 4 x i1>, i32)

define signext i32 @vpreduce_and_nxv4i32(i32 signext %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_and_nxv4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v10, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
; CHECK-NEXT:    vredand.vs v10, v8, v10, v0.t
; CHECK-NEXT:    vmv.x.s a0, v10
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.and.nxv4i32(i32 %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.or.nxv4i32(i32, <vscale x 4 x i32>, <vscale x 4 x i1>, i32)

define signext i32 @vpreduce_or_nxv4i32(i32 signext %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_or_nxv4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v10, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
; CHECK-NEXT:    vredor.vs v10, v8, v10, v0.t
; CHECK-NEXT:    vmv.x.s a0, v10
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.or.nxv4i32(i32 %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i32 @llvm.vp.reduce.xor.nxv4i32(i32, <vscale x 4 x i32>, <vscale x 4 x i1>, i32)

define signext i32 @vpreduce_xor_nxv4i32(i32 signext %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vpreduce_xor_nxv4i32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetivli zero, 1, e32, m1, ta, ma
; CHECK-NEXT:    vmv.s.x v10, a0
; CHECK-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
; CHECK-NEXT:    vredxor.vs v10, v8, v10, v0.t
; CHECK-NEXT:    vmv.x.s a0, v10
; CHECK-NEXT:    ret
  %r = call i32 @llvm.vp.reduce.xor.nxv4i32(i32 %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i32 %r
}

declare i64 @llvm.vp.reduce.add.nxv1i64(i64, <vscale x 1 x i64>, <vscale x 1 x i1>, i32)

define signext i64 @vpreduce_add_nxv1i64(i64 signext %s, <vscale x 1 x i64> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_add_nxv1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v9, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m1, ta, ma
; RV32-NEXT:    vredsum.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v9, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_add_nxv1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m1, ta, ma
; RV64-NEXT:    vredsum.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.add.nxv1i64(i64 %s, <vscale x 1 x i64> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i64 %r
}

define signext i64 @vpwreduce_add_nxv1i32(i64 signext %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpwreduce_add_nxv1i32:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v9, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e32, mf2, ta, ma
; RV32-NEXT:    vwredsum.vs v9, v8, v9, v0.t
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsrl.vx v8, v9, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpwreduce_add_nxv1i32:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e32, mf2, ta, ma
; RV64-NEXT:    vwredsum.vs v9, v8, v9, v0.t
; RV64-NEXT:    vsetvli zero, zero, e64, m1, ta, ma
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %e = sext <vscale x 1 x i32> %v to <vscale x 1 x i64>
  %r = call i64 @llvm.vp.reduce.add.nxv1i64(i64 %s, <vscale x 1 x i64> %e, <vscale x 1 x i1> %m, i32 %evl)
  ret i64 %r
}

define signext i64 @vpwreduce_uadd_nxv1i32(i64 signext %s, <vscale x 1 x i32> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpwreduce_uadd_nxv1i32:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v9, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e32, mf2, ta, ma
; RV32-NEXT:    vwredsum.vs v9, v8, v9, v0.t
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsrl.vx v8, v9, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpwreduce_uadd_nxv1i32:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e32, mf2, ta, ma
; RV64-NEXT:    vwredsum.vs v9, v8, v9, v0.t
; RV64-NEXT:    vsetvli zero, zero, e64, m1, ta, ma
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %e = sext <vscale x 1 x i32> %v to <vscale x 1 x i64>
  %r = call i64 @llvm.vp.reduce.add.nxv1i64(i64 %s, <vscale x 1 x i64> %e, <vscale x 1 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.umax.nxv1i64(i64, <vscale x 1 x i64>, <vscale x 1 x i1>, i32)

define signext i64 @vpreduce_umax_nxv1i64(i64 signext %s, <vscale x 1 x i64> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umax_nxv1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v9, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m1, ta, ma
; RV32-NEXT:    vredmaxu.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v9, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umax_nxv1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m1, ta, ma
; RV64-NEXT:    vredmaxu.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.umax.nxv1i64(i64 %s, <vscale x 1 x i64> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.smax.nxv1i64(i64, <vscale x 1 x i64>, <vscale x 1 x i1>, i32)

define signext i64 @vpreduce_smax_nxv1i64(i64 signext %s, <vscale x 1 x i64> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_smax_nxv1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v9, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m1, ta, ma
; RV32-NEXT:    vredmax.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v9, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_smax_nxv1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m1, ta, ma
; RV64-NEXT:    vredmax.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.smax.nxv1i64(i64 %s, <vscale x 1 x i64> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.umin.nxv1i64(i64, <vscale x 1 x i64>, <vscale x 1 x i1>, i32)

define signext i64 @vpreduce_umin_nxv1i64(i64 signext %s, <vscale x 1 x i64> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umin_nxv1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v9, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m1, ta, ma
; RV32-NEXT:    vredminu.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v9, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umin_nxv1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m1, ta, ma
; RV64-NEXT:    vredminu.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.umin.nxv1i64(i64 %s, <vscale x 1 x i64> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.smin.nxv1i64(i64, <vscale x 1 x i64>, <vscale x 1 x i1>, i32)

define signext i64 @vpreduce_smin_nxv1i64(i64 signext %s, <vscale x 1 x i64> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_smin_nxv1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v9, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m1, ta, ma
; RV32-NEXT:    vredmin.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v9, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_smin_nxv1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m1, ta, ma
; RV64-NEXT:    vredmin.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.smin.nxv1i64(i64 %s, <vscale x 1 x i64> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.and.nxv1i64(i64, <vscale x 1 x i64>, <vscale x 1 x i1>, i32)

define signext i64 @vpreduce_and_nxv1i64(i64 signext %s, <vscale x 1 x i64> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_and_nxv1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v9, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m1, ta, ma
; RV32-NEXT:    vredand.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v9, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_and_nxv1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m1, ta, ma
; RV64-NEXT:    vredand.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.and.nxv1i64(i64 %s, <vscale x 1 x i64> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.or.nxv1i64(i64, <vscale x 1 x i64>, <vscale x 1 x i1>, i32)

define signext i64 @vpreduce_or_nxv1i64(i64 signext %s, <vscale x 1 x i64> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_or_nxv1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v9, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m1, ta, ma
; RV32-NEXT:    vredor.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v9, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_or_nxv1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m1, ta, ma
; RV64-NEXT:    vredor.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.or.nxv1i64(i64 %s, <vscale x 1 x i64> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.xor.nxv1i64(i64, <vscale x 1 x i64>, <vscale x 1 x i1>, i32)

define signext i64 @vpreduce_xor_nxv1i64(i64 signext %s, <vscale x 1 x i64> %v, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_xor_nxv1i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v9, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m1, ta, ma
; RV32-NEXT:    vredxor.vs v9, v8, v9, v0.t
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v9, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_xor_nxv1i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m1, ta, ma
; RV64-NEXT:    vredxor.vs v9, v8, v9, v0.t
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.xor.nxv1i64(i64 %s, <vscale x 1 x i64> %v, <vscale x 1 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.add.nxv2i64(i64, <vscale x 2 x i64>, <vscale x 2 x i1>, i32)

define signext i64 @vpreduce_add_nxv2i64(i64 signext %s, <vscale x 2 x i64> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_add_nxv2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v10, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m2, ta, ma
; RV32-NEXT:    vredsum.vs v10, v8, v10, v0.t
; RV32-NEXT:    vmv.x.s a0, v10
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v10, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_add_nxv2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v10, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m2, ta, ma
; RV64-NEXT:    vredsum.vs v10, v8, v10, v0.t
; RV64-NEXT:    vmv.x.s a0, v10
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.add.nxv2i64(i64 %s, <vscale x 2 x i64> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i64 %r
}

define signext i64 @vwpreduce_add_nxv2i32(i64 signext %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vwpreduce_add_nxv2i32:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v9, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e32, m1, ta, ma
; RV32-NEXT:    vwredsum.vs v9, v8, v9, v0.t
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsrl.vx v8, v9, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vwpreduce_add_nxv2i32:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
; RV64-NEXT:    vwredsum.vs v9, v8, v9, v0.t
; RV64-NEXT:    vsetvli zero, zero, e64, m2, ta, ma
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %e = sext <vscale x 2 x i32> %v to <vscale x 2 x i64>
  %r = call i64 @llvm.vp.reduce.add.nxv2i64(i64 %s, <vscale x 2 x i64> %e, <vscale x 2 x i1> %m, i32 %evl)
  ret i64 %r
}

define signext i64 @vwpreduce_uadd_nxv2i32(i64 signext %s, <vscale x 2 x i32> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vwpreduce_uadd_nxv2i32:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v9, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e32, m1, ta, ma
; RV32-NEXT:    vwredsum.vs v9, v8, v9, v0.t
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vmv.x.s a0, v9
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsrl.vx v8, v9, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vwpreduce_uadd_nxv2i32:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v9, a0
; RV64-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
; RV64-NEXT:    vwredsum.vs v9, v8, v9, v0.t
; RV64-NEXT:    vsetvli zero, zero, e64, m2, ta, ma
; RV64-NEXT:    vmv.x.s a0, v9
; RV64-NEXT:    ret
  %e = sext <vscale x 2 x i32> %v to <vscale x 2 x i64>
  %r = call i64 @llvm.vp.reduce.add.nxv2i64(i64 %s, <vscale x 2 x i64> %e, <vscale x 2 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.umax.nxv2i64(i64, <vscale x 2 x i64>, <vscale x 2 x i1>, i32)

define signext i64 @vpreduce_umax_nxv2i64(i64 signext %s, <vscale x 2 x i64> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umax_nxv2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v10, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m2, ta, ma
; RV32-NEXT:    vredmaxu.vs v10, v8, v10, v0.t
; RV32-NEXT:    vmv.x.s a0, v10
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v10, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umax_nxv2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v10, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m2, ta, ma
; RV64-NEXT:    vredmaxu.vs v10, v8, v10, v0.t
; RV64-NEXT:    vmv.x.s a0, v10
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.umax.nxv2i64(i64 %s, <vscale x 2 x i64> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.smax.nxv2i64(i64, <vscale x 2 x i64>, <vscale x 2 x i1>, i32)

define signext i64 @vpreduce_smax_nxv2i64(i64 signext %s, <vscale x 2 x i64> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_smax_nxv2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v10, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m2, ta, ma
; RV32-NEXT:    vredmax.vs v10, v8, v10, v0.t
; RV32-NEXT:    vmv.x.s a0, v10
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v10, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_smax_nxv2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v10, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m2, ta, ma
; RV64-NEXT:    vredmax.vs v10, v8, v10, v0.t
; RV64-NEXT:    vmv.x.s a0, v10
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.smax.nxv2i64(i64 %s, <vscale x 2 x i64> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.umin.nxv2i64(i64, <vscale x 2 x i64>, <vscale x 2 x i1>, i32)

define signext i64 @vpreduce_umin_nxv2i64(i64 signext %s, <vscale x 2 x i64> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umin_nxv2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v10, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m2, ta, ma
; RV32-NEXT:    vredminu.vs v10, v8, v10, v0.t
; RV32-NEXT:    vmv.x.s a0, v10
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v10, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umin_nxv2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v10, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m2, ta, ma
; RV64-NEXT:    vredminu.vs v10, v8, v10, v0.t
; RV64-NEXT:    vmv.x.s a0, v10
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.umin.nxv2i64(i64 %s, <vscale x 2 x i64> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.smin.nxv2i64(i64, <vscale x 2 x i64>, <vscale x 2 x i1>, i32)

define signext i64 @vpreduce_smin_nxv2i64(i64 signext %s, <vscale x 2 x i64> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_smin_nxv2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v10, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m2, ta, ma
; RV32-NEXT:    vredmin.vs v10, v8, v10, v0.t
; RV32-NEXT:    vmv.x.s a0, v10
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v10, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_smin_nxv2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v10, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m2, ta, ma
; RV64-NEXT:    vredmin.vs v10, v8, v10, v0.t
; RV64-NEXT:    vmv.x.s a0, v10
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.smin.nxv2i64(i64 %s, <vscale x 2 x i64> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.and.nxv2i64(i64, <vscale x 2 x i64>, <vscale x 2 x i1>, i32)

define signext i64 @vpreduce_and_nxv2i64(i64 signext %s, <vscale x 2 x i64> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_and_nxv2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v10, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m2, ta, ma
; RV32-NEXT:    vredand.vs v10, v8, v10, v0.t
; RV32-NEXT:    vmv.x.s a0, v10
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v10, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_and_nxv2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v10, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m2, ta, ma
; RV64-NEXT:    vredand.vs v10, v8, v10, v0.t
; RV64-NEXT:    vmv.x.s a0, v10
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.and.nxv2i64(i64 %s, <vscale x 2 x i64> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.or.nxv2i64(i64, <vscale x 2 x i64>, <vscale x 2 x i1>, i32)

define signext i64 @vpreduce_or_nxv2i64(i64 signext %s, <vscale x 2 x i64> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_or_nxv2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v10, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m2, ta, ma
; RV32-NEXT:    vredor.vs v10, v8, v10, v0.t
; RV32-NEXT:    vmv.x.s a0, v10
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v10, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_or_nxv2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v10, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m2, ta, ma
; RV64-NEXT:    vredor.vs v10, v8, v10, v0.t
; RV64-NEXT:    vmv.x.s a0, v10
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.or.nxv2i64(i64 %s, <vscale x 2 x i64> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.xor.nxv2i64(i64, <vscale x 2 x i64>, <vscale x 2 x i1>, i32)

define signext i64 @vpreduce_xor_nxv2i64(i64 signext %s, <vscale x 2 x i64> %v, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_xor_nxv2i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v10, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m2, ta, ma
; RV32-NEXT:    vredxor.vs v10, v8, v10, v0.t
; RV32-NEXT:    vmv.x.s a0, v10
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v10, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_xor_nxv2i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v10, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m2, ta, ma
; RV64-NEXT:    vredxor.vs v10, v8, v10, v0.t
; RV64-NEXT:    vmv.x.s a0, v10
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.xor.nxv2i64(i64 %s, <vscale x 2 x i64> %v, <vscale x 2 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.add.nxv4i64(i64, <vscale x 4 x i64>, <vscale x 4 x i1>, i32)

define signext i64 @vpreduce_add_nxv4i64(i64 signext %s, <vscale x 4 x i64> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_add_nxv4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v12, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m4, ta, ma
; RV32-NEXT:    vredsum.vs v12, v8, v12, v0.t
; RV32-NEXT:    vmv.x.s a0, v12
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v12, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_add_nxv4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v12, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m4, ta, ma
; RV64-NEXT:    vredsum.vs v12, v8, v12, v0.t
; RV64-NEXT:    vmv.x.s a0, v12
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.add.nxv4i64(i64 %s, <vscale x 4 x i64> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i64 %r
}

define signext i64 @vpwreduce_add_nxv4i32(i64 signext %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpwreduce_add_nxv4i32:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v10, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e32, m2, ta, ma
; RV32-NEXT:    vwredsum.vs v10, v8, v10, v0.t
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vmv.x.s a0, v10
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsrl.vx v8, v10, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpwreduce_add_nxv4i32:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v10, a0
; RV64-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
; RV64-NEXT:    vwredsum.vs v10, v8, v10, v0.t
; RV64-NEXT:    vsetvli zero, zero, e64, m4, ta, ma
; RV64-NEXT:    vmv.x.s a0, v10
; RV64-NEXT:    ret
  %e = sext <vscale x 4 x i32> %v to <vscale x 4 x i64>
  %r = call i64 @llvm.vp.reduce.add.nxv4i64(i64 %s, <vscale x 4 x i64> %e, <vscale x 4 x i1> %m, i32 %evl)
  ret i64 %r
}

define signext i64 @vpwreduce_uadd_nxv4i32(i64 signext %s, <vscale x 4 x i32> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpwreduce_uadd_nxv4i32:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v10, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e32, m2, ta, ma
; RV32-NEXT:    vwredsumu.vs v10, v8, v10, v0.t
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vmv.x.s a0, v10
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsrl.vx v8, v10, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpwreduce_uadd_nxv4i32:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v10, a0
; RV64-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
; RV64-NEXT:    vwredsumu.vs v10, v8, v10, v0.t
; RV64-NEXT:    vsetvli zero, zero, e64, m4, ta, ma
; RV64-NEXT:    vmv.x.s a0, v10
; RV64-NEXT:    ret
  %e = zext <vscale x 4 x i32> %v to <vscale x 4 x i64>
  %r = call i64 @llvm.vp.reduce.add.nxv4i64(i64 %s, <vscale x 4 x i64> %e, <vscale x 4 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.umax.nxv4i64(i64, <vscale x 4 x i64>, <vscale x 4 x i1>, i32)

define signext i64 @vpreduce_umax_nxv4i64(i64 signext %s, <vscale x 4 x i64> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umax_nxv4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v12, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m4, ta, ma
; RV32-NEXT:    vredmaxu.vs v12, v8, v12, v0.t
; RV32-NEXT:    vmv.x.s a0, v12
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v12, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umax_nxv4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v12, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m4, ta, ma
; RV64-NEXT:    vredmaxu.vs v12, v8, v12, v0.t
; RV64-NEXT:    vmv.x.s a0, v12
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.umax.nxv4i64(i64 %s, <vscale x 4 x i64> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.smax.nxv4i64(i64, <vscale x 4 x i64>, <vscale x 4 x i1>, i32)

define signext i64 @vpreduce_smax_nxv4i64(i64 signext %s, <vscale x 4 x i64> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_smax_nxv4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v12, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m4, ta, ma
; RV32-NEXT:    vredmax.vs v12, v8, v12, v0.t
; RV32-NEXT:    vmv.x.s a0, v12
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v12, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_smax_nxv4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v12, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m4, ta, ma
; RV64-NEXT:    vredmax.vs v12, v8, v12, v0.t
; RV64-NEXT:    vmv.x.s a0, v12
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.smax.nxv4i64(i64 %s, <vscale x 4 x i64> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.umin.nxv4i64(i64, <vscale x 4 x i64>, <vscale x 4 x i1>, i32)

define signext i64 @vpreduce_umin_nxv4i64(i64 signext %s, <vscale x 4 x i64> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_umin_nxv4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v12, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m4, ta, ma
; RV32-NEXT:    vredminu.vs v12, v8, v12, v0.t
; RV32-NEXT:    vmv.x.s a0, v12
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v12, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_umin_nxv4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v12, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m4, ta, ma
; RV64-NEXT:    vredminu.vs v12, v8, v12, v0.t
; RV64-NEXT:    vmv.x.s a0, v12
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.umin.nxv4i64(i64 %s, <vscale x 4 x i64> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.smin.nxv4i64(i64, <vscale x 4 x i64>, <vscale x 4 x i1>, i32)

define signext i64 @vpreduce_smin_nxv4i64(i64 signext %s, <vscale x 4 x i64> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_smin_nxv4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v12, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m4, ta, ma
; RV32-NEXT:    vredmin.vs v12, v8, v12, v0.t
; RV32-NEXT:    vmv.x.s a0, v12
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v12, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_smin_nxv4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v12, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m4, ta, ma
; RV64-NEXT:    vredmin.vs v12, v8, v12, v0.t
; RV64-NEXT:    vmv.x.s a0, v12
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.smin.nxv4i64(i64 %s, <vscale x 4 x i64> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.and.nxv4i64(i64, <vscale x 4 x i64>, <vscale x 4 x i1>, i32)

define signext i64 @vpreduce_and_nxv4i64(i64 signext %s, <vscale x 4 x i64> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_and_nxv4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v12, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m4, ta, ma
; RV32-NEXT:    vredand.vs v12, v8, v12, v0.t
; RV32-NEXT:    vmv.x.s a0, v12
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v12, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_and_nxv4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v12, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m4, ta, ma
; RV64-NEXT:    vredand.vs v12, v8, v12, v0.t
; RV64-NEXT:    vmv.x.s a0, v12
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.and.nxv4i64(i64 %s, <vscale x 4 x i64> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.or.nxv4i64(i64, <vscale x 4 x i64>, <vscale x 4 x i1>, i32)

define signext i64 @vpreduce_or_nxv4i64(i64 signext %s, <vscale x 4 x i64> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_or_nxv4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v12, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m4, ta, ma
; RV32-NEXT:    vredor.vs v12, v8, v12, v0.t
; RV32-NEXT:    vmv.x.s a0, v12
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v12, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_or_nxv4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v12, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m4, ta, ma
; RV64-NEXT:    vredor.vs v12, v8, v12, v0.t
; RV64-NEXT:    vmv.x.s a0, v12
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.or.nxv4i64(i64 %s, <vscale x 4 x i64> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i64 %r
}

declare i64 @llvm.vp.reduce.xor.nxv4i64(i64, <vscale x 4 x i64>, <vscale x 4 x i1>, i32)

define signext i64 @vpreduce_xor_nxv4i64(i64 signext %s, <vscale x 4 x i64> %v, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; RV32-LABEL: vpreduce_xor_nxv4i64:
; RV32:       # %bb.0:
; RV32-NEXT:    addi sp, sp, -16
; RV32-NEXT:    .cfi_def_cfa_offset 16
; RV32-NEXT:    sw a1, 12(sp)
; RV32-NEXT:    sw a0, 8(sp)
; RV32-NEXT:    addi a0, sp, 8
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vlse64.v v12, (a0), zero
; RV32-NEXT:    vsetvli zero, a2, e64, m4, ta, ma
; RV32-NEXT:    vredxor.vs v12, v8, v12, v0.t
; RV32-NEXT:    vmv.x.s a0, v12
; RV32-NEXT:    li a1, 32
; RV32-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV32-NEXT:    vsrl.vx v8, v12, a1
; RV32-NEXT:    vmv.x.s a1, v8
; RV32-NEXT:    addi sp, sp, 16
; RV32-NEXT:    ret
;
; RV64-LABEL: vpreduce_xor_nxv4i64:
; RV64:       # %bb.0:
; RV64-NEXT:    vsetivli zero, 1, e64, m1, ta, ma
; RV64-NEXT:    vmv.s.x v12, a0
; RV64-NEXT:    vsetvli zero, a1, e64, m4, ta, ma
; RV64-NEXT:    vredxor.vs v12, v8, v12, v0.t
; RV64-NEXT:    vmv.x.s a0, v12
; RV64-NEXT:    ret
  %r = call i64 @llvm.vp.reduce.xor.nxv4i64(i64 %s, <vscale x 4 x i64> %v, <vscale x 4 x i1> %m, i32 %evl)
  ret i64 %r
}