[AArch64] Add CodeGen tests for vector reduction intrinsics. NFC

[RosieSumpter]

Summary

Tests are added for vector reduce OR, AND and XOR.

[SjoerdMeijer]

Thanks for adding these test. A few nits inlined, but no need for another review if you're happy to make a few changes before committing (but I can always have another look).

llvm/test/CodeGen/AArch64/reduce-or.ll

355

You may want to add <2 x i32> as that is a common type and supported by the vector instructions.

366

I think you can remove this one because the 32 x i8 case above is already testing a type that doesn't fit in a vector register.

The same 2 comments here apply to the other files as well.

llvm/test/CodeGen/AArch64/reduce-xor.ll

4

Nit: test_redor_v1i1 -> test_redxor_v1i1

And a few more below.

[fhahn]

llvm/test/CodeGen/AArch64/reduce-and.ll

5

For boolean reductions, there's already llvm/test/CodeGen/AArch64/vecreduce-bool.ll. Can those tests be unified to reduce duplication?

[RosieSumpter]

llvm/test/CodeGen/AArch64/reduce-and.ll

5

It seems that vecreduce-bool.ll is testing something slightly different as it is looking at the pattern of a compare with a reduction, so generates e.g. uminv and umaxv instructions. I have agreed with @SjoerdMeijer to keep these tests separate for now.

[fhahn]

llvm/test/CodeGen/AArch64/reduce-and.ll

5

ok fair enough.

llvm/test/CodeGen/AArch64/reduce-and.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=aarch64-unknown-linux-gnu | FileCheck %s
+
+define i1 @test_redor_v1i1(<1 x i1> %a) {
+; CHECK-LABEL: test_redor_v1i1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and w0, w0, #0x1
+; CHECK-NEXT:    ret
+  %or_result = call i1 @llvm.vector.reduce.and.v1i1(<1 x i1> %a)
+  ret i1 %or_result
+}
+
+define i1 @test_redor_v2i1(<2 x i1> %a) {
+; CHECK-LABEL: test_redor_v2i1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    mov w8, v0.s[1]
+; CHECK-NEXT:    fmov w9, s0
+; CHECK-NEXT:    and w8, w9, w8
+; CHECK-NEXT:    and w0, w8, #0x1
+; CHECK-NEXT:    ret
+  %or_result = call i1 @llvm.vector.reduce.and.v2i1(<2 x i1> %a)
+  ret i1 %or_result
+}
+
+define i1 @test_redor_v4i1(<4 x i1> %a) {
+; CHECK-LABEL: test_redor_v4i1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w10, v0.h[1]
+; CHECK-NEXT:    umov w11, v0.h[0]
+; CHECK-NEXT:    umov w9, v0.h[2]
+; CHECK-NEXT:    and w10, w11, w10
+; CHECK-NEXT:    umov w8, v0.h[3]
+; CHECK-NEXT:    and w9, w10, w9
+; CHECK-NEXT:    and w8, w9, w8
+; CHECK-NEXT:    and w0, w8, #0x1
+; CHECK-NEXT:    ret
+  %or_result = call i1 @llvm.vector.reduce.and.v4i1(<4 x i1> %a)
+  ret i1 %or_result
+}
+
+define i1 @test_redor_v8i1(<8 x i1> %a) {
+; CHECK-LABEL: test_redor_v8i1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w14, v0.b[1]
+; CHECK-NEXT:    umov w15, v0.b[0]
+; CHECK-NEXT:    umov w13, v0.b[2]
+; CHECK-NEXT:    and w14, w15, w14
+; CHECK-NEXT:    umov w12, v0.b[3]
+; CHECK-NEXT:    and w13, w14, w13
+; CHECK-NEXT:    umov w11, v0.b[4]
+; CHECK-NEXT:    and w12, w13, w12
+; CHECK-NEXT:    umov w10, v0.b[5]
+; CHECK-NEXT:    and w11, w12, w11
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    and w10, w11, w10
+; CHECK-NEXT:    umov w8, v0.b[7]
+; CHECK-NEXT:    and w9, w10, w9
+; CHECK-NEXT:    and w8, w9, w8
+; CHECK-NEXT:    and w0, w8, #0x1
+; CHECK-NEXT:    ret
+  %or_result = call i1 @llvm.vector.reduce.and.v8i1(<8 x i1> %a)
+  ret i1 %or_result
+}
+
+define i1 @test_redor_v16i1(<16 x i1> %a) {
+; CHECK-LABEL: test_redor_v16i1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    and v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.b[1]
+; CHECK-NEXT:    umov w9, v0.b[0]
+; CHECK-NEXT:    and w8, w9, w8
+; CHECK-NEXT:    umov w9, v0.b[2]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[3]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[4]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[5]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[7]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    and w0, w8, #0x1
+; CHECK-NEXT:    ret
+  %or_result = call i1 @llvm.vector.reduce.and.v16i1(<16 x i1> %a)
+  ret i1 %or_result
+}
+
+define i8 @test_redand_v1i8(<1 x i8> %a) {
+; CHECK-LABEL: test_redand_v1i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w0, v0.b[0]
+; CHECK-NEXT:    ret
+  %and_result = call i8 @llvm.vector.reduce.and.v1i8(<1 x i8> %a)
+  ret i8 %and_result
+}
+
+define i8 @test_redand_v3i8(<3 x i8> %a) {
+; CHECK-LABEL: test_redand_v3i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and w8, w0, w1
+; CHECK-NEXT:    and w8, w8, w2
+; CHECK-NEXT:    and w0, w8, #0xff
+; CHECK-NEXT:    ret
+  %and_result = call i8 @llvm.vector.reduce.and.v3i8(<3 x i8> %a)
+  ret i8 %and_result
+}
+
+define i8 @test_redand_v4i8(<4 x i8> %a) {
+; CHECK-LABEL: test_redand_v4i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w10, v0.h[1]
+; CHECK-NEXT:    umov w11, v0.h[0]
+; CHECK-NEXT:    umov w9, v0.h[2]
+; CHECK-NEXT:    and w10, w11, w10
+; CHECK-NEXT:    umov w8, v0.h[3]
+; CHECK-NEXT:    and w9, w10, w9
+; CHECK-NEXT:    and w0, w9, w8
+; CHECK-NEXT:    ret
+  %and_result = call i8 @llvm.vector.reduce.and.v4i8(<4 x i8> %a)
+  ret i8 %and_result
+}
+
+define i8 @test_redand_v8i8(<8 x i8> %a) {
+; CHECK-LABEL: test_redand_v8i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w14, v0.b[1]
+; CHECK-NEXT:    umov w15, v0.b[0]
+; CHECK-NEXT:    umov w13, v0.b[2]
+; CHECK-NEXT:    and w14, w15, w14
+; CHECK-NEXT:    umov w12, v0.b[3]
+; CHECK-NEXT:    and w13, w14, w13
+; CHECK-NEXT:    umov w11, v0.b[4]
+; CHECK-NEXT:    and w12, w13, w12
+; CHECK-NEXT:    umov w10, v0.b[5]
+; CHECK-NEXT:    and w11, w12, w11
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    and w10, w11, w10
+; CHECK-NEXT:    umov w8, v0.b[7]
+; CHECK-NEXT:    and w9, w10, w9
+; CHECK-NEXT:    and w0, w9, w8
+; CHECK-NEXT:    ret
+  %and_result = call i8 @llvm.vector.reduce.and.v8i8(<8 x i8> %a)
+  ret i8 %and_result
+}
+
+define i8 @test_redand_v16i8(<16 x i8> %a) {
+; CHECK-LABEL: test_redand_v16i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    and v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.b[1]
+; CHECK-NEXT:    umov w9, v0.b[0]
+; CHECK-NEXT:    and w8, w9, w8
+; CHECK-NEXT:    umov w9, v0.b[2]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[3]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[4]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[5]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[7]
+; CHECK-NEXT:    and w0, w8, w9
+; CHECK-NEXT:    ret
+  %and_result = call i8 @llvm.vector.reduce.and.v16i8(<16 x i8> %a)
+  ret i8 %and_result
+}
+
+define i8 @test_redand_v32i8(<32 x i8> %a) {
+; CHECK-LABEL: test_redand_v32i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and v0.16b, v0.16b, v1.16b
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    and v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.b[1]
+; CHECK-NEXT:    umov w9, v0.b[0]
+; CHECK-NEXT:    and w8, w9, w8
+; CHECK-NEXT:    umov w9, v0.b[2]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[3]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[4]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[5]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[7]
+; CHECK-NEXT:    and w0, w8, w9
+; CHECK-NEXT:    ret
+  %and_result = call i8 @llvm.vector.reduce.and.v32i8(<32 x i8> %a)
+  ret i8 %and_result
+}
+
+define i8 @test_redand_v64i8(<64 x i8> %a) {
+; CHECK-LABEL: test_redand_v64i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and v1.16b, v1.16b, v3.16b
+; CHECK-NEXT:    and v0.16b, v0.16b, v2.16b
+; CHECK-NEXT:    and v0.16b, v0.16b, v1.16b
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    and v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.b[1]
+; CHECK-NEXT:    umov w9, v0.b[0]
+; CHECK-NEXT:    and w8, w9, w8
+; CHECK-NEXT:    umov w9, v0.b[2]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[3]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[4]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[5]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    and w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[7]
+; CHECK-NEXT:    and w0, w8, w9
+; CHECK-NEXT:    ret
+  %and_result = call i8 @llvm.vector.reduce.and.v64i8(<64 x i8> %a)
+  ret i8 %and_result
+}
+
+define i16 @test_redand_v4i16(<4 x i16> %a) {
+; CHECK-LABEL: test_redand_v4i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w10, v0.h[1]
+; CHECK-NEXT:    umov w11, v0.h[0]
+; CHECK-NEXT:    umov w9, v0.h[2]
+; CHECK-NEXT:    and w10, w11, w10
+; CHECK-NEXT:    umov w8, v0.h[3]
+; CHECK-NEXT:    and w9, w10, w9
+; CHECK-NEXT:    and w0, w9, w8
+; CHECK-NEXT:    ret
+  %and_result = call i16 @llvm.vector.reduce.and.v4i16(<4 x i16> %a)
+  ret i16 %and_result
+}
+
+define i16 @test_redand_v8i16(<8 x i16> %a) {
+; CHECK-LABEL: test_redand_v8i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    and v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.h[1]
+; CHECK-NEXT:    umov w9, v0.h[0]
+; CHECK-NEXT:    umov w10, v0.h[2]
+; CHECK-NEXT:    and w8, w9, w8
+; CHECK-NEXT:    and w8, w8, w10
+; CHECK-NEXT:    umov w9, v0.h[3]
+; CHECK-NEXT:    and w0, w8, w9
+; CHECK-NEXT:    ret
+  %and_result = call i16 @llvm.vector.reduce.and.v8i16(<8 x i16> %a)
+  ret i16 %and_result
+}
+
+define i16 @test_redand_v16i16(<16 x i16> %a) {
+; CHECK-LABEL: test_redand_v16i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and v0.16b, v0.16b, v1.16b
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    and v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.h[1]
+; CHECK-NEXT:    umov w9, v0.h[0]
+; CHECK-NEXT:    umov w10, v0.h[2]
+; CHECK-NEXT:    and w8, w9, w8
+; CHECK-NEXT:    and w8, w8, w10
+; CHECK-NEXT:    umov w9, v0.h[3]
+; CHECK-NEXT:    and w0, w8, w9
+; CHECK-NEXT:    ret
+  %and_result = call i16 @llvm.vector.reduce.and.v16i16(<16 x i16> %a)
+  ret i16 %and_result
+}
+
+define i32 @test_redand_v4i32(<4 x i32> %a) {
+; CHECK-LABEL: test_redand_v4i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    and v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    mov w8, v0.s[1]
+; CHECK-NEXT:    fmov w9, s0
+; CHECK-NEXT:    and w0, w9, w8
+; CHECK-NEXT:    ret
+  %and_result = call i32 @llvm.vector.reduce.and.v4i32(<4 x i32> %a)
+  ret i32 %and_result
+}
+
+define i32 @test_redand_v8i32(<8 x i32> %a) {
+; CHECK-LABEL: test_redand_v8i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and v0.16b, v0.16b, v1.16b
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    and v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    mov w8, v0.s[1]
+; CHECK-NEXT:    fmov w9, s0
+; CHECK-NEXT:    and w0, w9, w8
+; CHECK-NEXT:    ret
+  %and_result = call i32 @llvm.vector.reduce.and.v8i32(<8 x i32> %a)
+  ret i32 %and_result
+}
+
+define i64 @test_redand_v2i64(<2 x i64> %a) {
+; CHECK-LABEL: test_redand_v2i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    and v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    fmov x0, d0
+; CHECK-NEXT:    ret
+  %and_result = call i64 @llvm.vector.reduce.and.v2i64(<2 x i64> %a)
+  ret i64 %and_result
+}
+
+define i64 @test_redand_v4i64(<4 x i64> %a) {
+; CHECK-LABEL: test_redand_v4i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and v0.16b, v0.16b, v1.16b
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    and v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    fmov x0, d0
+; CHECK-NEXT:    ret
+  %and_result = call i64 @llvm.vector.reduce.and.v4i64(<4 x i64> %a)
+  ret i64 %and_result
+}
+
+declare i1 @llvm.vector.reduce.and.v1i1(<1 x i1>)
+declare i1 @llvm.vector.reduce.and.v2i1(<2 x i1>)
+declare i1 @llvm.vector.reduce.and.v4i1(<4 x i1>)
+declare i1 @llvm.vector.reduce.and.v8i1(<8 x i1>)
+declare i1 @llvm.vector.reduce.and.v16i1(<16 x i1>)
+declare i64 @llvm.vector.reduce.and.v2i64(<2 x i64>)
+declare i64 @llvm.vector.reduce.and.v4i64(<4 x i64>)
+declare i32 @llvm.vector.reduce.and.v4i32(<4 x i32>)
+declare i32 @llvm.vector.reduce.and.v8i32(<8 x i32>)
+declare i16 @llvm.vector.reduce.and.v4i16(<4 x i16>)
+declare i16 @llvm.vector.reduce.and.v8i16(<8 x i16>)
+declare i16 @llvm.vector.reduce.and.v16i16(<16 x i16>)
+declare i8 @llvm.vector.reduce.and.v1i8(<1 x i8>)
+declare i8 @llvm.vector.reduce.and.v3i8(<3 x i8>)
+declare i8 @llvm.vector.reduce.and.v4i8(<4 x i8>)
+declare i8 @llvm.vector.reduce.and.v8i8(<8 x i8>)
+declare i8 @llvm.vector.reduce.and.v16i8(<16 x i8>)
+declare i8 @llvm.vector.reduce.and.v32i8(<32 x i8>)
+declare i8 @llvm.vector.reduce.and.v64i8(<64 x i8>)

llvm/test/CodeGen/AArch64/reduce-or.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=aarch64-unknown-linux-gnu | FileCheck %s
+
+define i1 @test_redor_v1i1(<1 x i1> %a) {
+; CHECK-LABEL: test_redor_v1i1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and w0, w0, #0x1
+; CHECK-NEXT:    ret
+  %or_result = call i1 @llvm.vector.reduce.or.v1i1(<1 x i1> %a)
+  ret i1 %or_result
+}
+
+define i1 @test_redor_v2i1(<2 x i1> %a) {
+; CHECK-LABEL: test_redor_v2i1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    mov w8, v0.s[1]
+; CHECK-NEXT:    fmov w9, s0
+; CHECK-NEXT:    orr w8, w9, w8
+; CHECK-NEXT:    and w0, w8, #0x1
+; CHECK-NEXT:    ret
+  %or_result = call i1 @llvm.vector.reduce.or.v2i1(<2 x i1> %a)
+  ret i1 %or_result
+}
+
+define i1 @test_redor_v4i1(<4 x i1> %a) {
+; CHECK-LABEL: test_redor_v4i1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w10, v0.h[1]
+; CHECK-NEXT:    umov w11, v0.h[0]
+; CHECK-NEXT:    umov w9, v0.h[2]
+; CHECK-NEXT:    orr w10, w11, w10
+; CHECK-NEXT:    umov w8, v0.h[3]
+; CHECK-NEXT:    orr w9, w10, w9
+; CHECK-NEXT:    orr w8, w9, w8
+; CHECK-NEXT:    and w0, w8, #0x1
+; CHECK-NEXT:    ret
+  %or_result = call i1 @llvm.vector.reduce.or.v4i1(<4 x i1> %a)
+  ret i1 %or_result
+}
+
+define i1 @test_redor_v8i1(<8 x i1> %a) {
+; CHECK-LABEL: test_redor_v8i1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w14, v0.b[1]
+; CHECK-NEXT:    umov w15, v0.b[0]
+; CHECK-NEXT:    umov w13, v0.b[2]
+; CHECK-NEXT:    orr w14, w15, w14
+; CHECK-NEXT:    umov w12, v0.b[3]
+; CHECK-NEXT:    orr w13, w14, w13
+; CHECK-NEXT:    umov w11, v0.b[4]
+; CHECK-NEXT:    orr w12, w13, w12
+; CHECK-NEXT:    umov w10, v0.b[5]
+; CHECK-NEXT:    orr w11, w12, w11
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    orr w10, w11, w10
+; CHECK-NEXT:    umov w8, v0.b[7]
+; CHECK-NEXT:    orr w9, w10, w9
+; CHECK-NEXT:    orr w8, w9, w8
+; CHECK-NEXT:    and w0, w8, #0x1
+; CHECK-NEXT:    ret
+  %or_result = call i1 @llvm.vector.reduce.or.v8i1(<8 x i1> %a)
+  ret i1 %or_result
+}
+
+define i1 @test_redor_v16i1(<16 x i1> %a) {
+; CHECK-LABEL: test_redor_v16i1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    orr v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.b[1]
+; CHECK-NEXT:    umov w9, v0.b[0]
+; CHECK-NEXT:    orr w8, w9, w8
+; CHECK-NEXT:    umov w9, v0.b[2]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[3]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[4]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[5]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[7]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    and w0, w8, #0x1
+; CHECK-NEXT:    ret
+  %or_result = call i1 @llvm.vector.reduce.or.v16i1(<16 x i1> %a)
+  ret i1 %or_result
+}
+
+define i8 @test_redor_v1i8(<1 x i8> %a) {
+;
+; CHECK-LABEL: test_redor_v1i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w0, v0.b[0]
+; CHECK-NEXT:    ret
+  %or_result = call i8 @llvm.vector.reduce.or.v1i8(<1 x i8> %a)
+  ret i8 %or_result
+}
+
+define i8 @test_redor_v3i8(<3 x i8> %a) {
+;
+; CHECK-LABEL: test_redor_v3i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    orr w8, w0, w1
+; CHECK-NEXT:    orr w0, w8, w2
+; CHECK-NEXT:    ret
+  %or_result = call i8 @llvm.vector.reduce.or.v3i8(<3 x i8> %a)
+  ret i8 %or_result
+}
+
+define i8 @test_redor_v4i8(<4 x i8> %a) {
+;
+; CHECK-LABEL: test_redor_v4i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w10, v0.h[1]
+; CHECK-NEXT:    umov w11, v0.h[0]
+; CHECK-NEXT:    umov w9, v0.h[2]
+; CHECK-NEXT:    orr w10, w11, w10
+; CHECK-NEXT:    umov w8, v0.h[3]
+; CHECK-NEXT:    orr w9, w10, w9
+; CHECK-NEXT:    orr w0, w9, w8
+; CHECK-NEXT:    ret
+  %or_result = call i8 @llvm.vector.reduce.or.v4i8(<4 x i8> %a)
+  ret i8 %or_result
+}
+
+define i8 @test_redor_v8i8(<8 x i8> %a) {
+;
+; CHECK-LABEL: test_redor_v8i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w14, v0.b[1]
+; CHECK-NEXT:    umov w15, v0.b[0]
+; CHECK-NEXT:    umov w13, v0.b[2]
+; CHECK-NEXT:    orr w14, w15, w14
+; CHECK-NEXT:    umov w12, v0.b[3]
+; CHECK-NEXT:    orr w13, w14, w13
+; CHECK-NEXT:    umov w11, v0.b[4]
+; CHECK-NEXT:    orr w12, w13, w12
+; CHECK-NEXT:    umov w10, v0.b[5]
+; CHECK-NEXT:    orr w11, w12, w11
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    orr w10, w11, w10
+; CHECK-NEXT:    umov w8, v0.b[7]
+; CHECK-NEXT:    orr w9, w10, w9
+; CHECK-NEXT:    orr w0, w9, w8
+; CHECK-NEXT:    ret
+  %or_result = call i8 @llvm.vector.reduce.or.v8i8(<8 x i8> %a)
+  ret i8 %or_result
+}
+
+define i8 @test_redor_v16i8(<16 x i8> %a) {
+;
+; CHECK-LABEL: test_redor_v16i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    orr v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.b[1]
+; CHECK-NEXT:    umov w9, v0.b[0]
+; CHECK-NEXT:    orr w8, w9, w8
+; CHECK-NEXT:    umov w9, v0.b[2]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[3]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[4]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[5]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[7]
+; CHECK-NEXT:    orr w0, w8, w9
+; CHECK-NEXT:    ret
+  %or_result = call i8 @llvm.vector.reduce.or.v16i8(<16 x i8> %a)
+  ret i8 %or_result
+}
+
+define i8 @test_redor_v32i8(<32 x i8> %a) {
+;
+; CHECK-LABEL: test_redor_v32i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    orr v0.16b, v0.16b, v1.16b
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    orr v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.b[1]
+; CHECK-NEXT:    umov w9, v0.b[0]
+; CHECK-NEXT:    orr w8, w9, w8
+; CHECK-NEXT:    umov w9, v0.b[2]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[3]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[4]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[5]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[7]
+; CHECK-NEXT:    orr w0, w8, w9
+; CHECK-NEXT:    ret
+  %or_result = call i8 @llvm.vector.reduce.or.v32i8(<32 x i8> %a)
+  ret i8 %or_result
+}
+
+define i8 @test_redor_v64i8(<64 x i8> %a) {
+;
+; CHECK-LABEL: test_redor_v64i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    orr v1.16b, v1.16b, v3.16b
+; CHECK-NEXT:    orr v0.16b, v0.16b, v2.16b
+; CHECK-NEXT:    orr v0.16b, v0.16b, v1.16b
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    orr v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.b[1]
+; CHECK-NEXT:    umov w9, v0.b[0]
+; CHECK-NEXT:    orr w8, w9, w8
+; CHECK-NEXT:    umov w9, v0.b[2]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[3]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[4]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[5]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    orr w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[7]
+; CHECK-NEXT:    orr w0, w8, w9
+; CHECK-NEXT:    ret
+  %or_result = call i8 @llvm.vector.reduce.or.v64i8(<64 x i8> %a)
+  ret i8 %or_result
+}
+
+define i16 @test_redor_v4i16(<4 x i16> %a) {
+;
+; CHECK-LABEL: test_redor_v4i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w10, v0.h[1]
+; CHECK-NEXT:    umov w11, v0.h[0]
+; CHECK-NEXT:    umov w9, v0.h[2]
+; CHECK-NEXT:    orr w10, w11, w10
+; CHECK-NEXT:    umov w8, v0.h[3]
+; CHECK-NEXT:    orr w9, w10, w9
+; CHECK-NEXT:    orr w0, w9, w8
+; CHECK-NEXT:    ret
+  %or_result = call i16 @llvm.vector.reduce.or.v4i16(<4 x i16> %a)
+  ret i16 %or_result
+}
+
+define i16 @test_redor_v8i16(<8 x i16> %a) {
+;
+; CHECK-LABEL: test_redor_v8i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    orr v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.h[1]
+; CHECK-NEXT:    umov w9, v0.h[0]
+; CHECK-NEXT:    umov w10, v0.h[2]
+; CHECK-NEXT:    orr w8, w9, w8
+; CHECK-NEXT:    orr w8, w8, w10
+; CHECK-NEXT:    umov w9, v0.h[3]
+; CHECK-NEXT:    orr w0, w8, w9
+; CHECK-NEXT:    ret
+  %or_result = call i16 @llvm.vector.reduce.or.v8i16(<8 x i16> %a)
+  ret i16 %or_result
+}
+
+define i16 @test_redor_v16i16(<16 x i16> %a) {
+;
+; CHECK-LABEL: test_redor_v16i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    orr v0.16b, v0.16b, v1.16b
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    orr v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.h[1]
+; CHECK-NEXT:    umov w9, v0.h[0]
+; CHECK-NEXT:    umov w10, v0.h[2]
+; CHECK-NEXT:    orr w8, w9, w8
+; CHECK-NEXT:    orr w8, w8, w10
+; CHECK-NEXT:    umov w9, v0.h[3]
+; CHECK-NEXT:    orr w0, w8, w9
+; CHECK-NEXT:    ret
+  %or_result = call i16 @llvm.vector.reduce.or.v16i16(<16 x i16> %a)
+  ret i16 %or_result
+}
+
+define i32 @test_redor_v4i32(<4 x i32> %a) {
+;
+; CHECK-LABEL: test_redor_v4i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    orr v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    mov w8, v0.s[1]
+; CHECK-NEXT:    fmov w9, s0
+; CHECK-NEXT:    orr w0, w9, w8
+; CHECK-NEXT:    ret
+  %or_result = call i32 @llvm.vector.reduce.or.v4i32(<4 x i32> %a)
+  ret i32 %or_result
+}
+
+define i32 @test_redor_v8i32(<8 x i32> %a) {
+;
+; CHECK-LABEL: test_redor_v8i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    orr v0.16b, v0.16b, v1.16b
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    orr v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    mov w8, v0.s[1]
+; CHECK-NEXT:    fmov w9, s0
+; CHECK-NEXT:    orr w0, w9, w8
+; CHECK-NEXT:    ret
+  %or_result = call i32 @llvm.vector.reduce.or.v8i32(<8 x i32> %a)
+  ret i32 %or_result
+}
+
+define i64 @test_redor_v2i64(<2 x i64> %a) {
+;
+; CHECK-LABEL: test_redor_v2i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    orr v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    fmov x0, d0
+; CHECK-NEXT:    ret
+  %or_result = call i64 @llvm.vector.reduce.or.v2i64(<2 x i64> %a)
+  ret i64 %or_result
+}
+
+define i64 @test_redor_v4i64(<4 x i64> %a) {
+;
+; CHECK-LABEL: test_redor_v4i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    orr v0.16b, v0.16b, v1.16b
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    orr v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    fmov x0, d0
+; CHECK-NEXT:    ret
+  %or_result = call i64 @llvm.vector.reduce.or.v4i64(<4 x i64> %a)
+  ret i64 %or_result
+}
+
+declare i1 @llvm.vector.reduce.or.v1i1(<1 x i1>)
+declare i1 @llvm.vector.reduce.or.v2i1(<2 x i1>)
+declare i1 @llvm.vector.reduce.or.v4i1(<4 x i1>)
+declare i1 @llvm.vector.reduce.or.v8i1(<8 x i1>)
+declare i1 @llvm.vector.reduce.or.v16i1(<16 x i1>)
+declare i64 @llvm.vector.reduce.or.v2i64(<2 x i64>)
+declare i64 @llvm.vector.reduce.or.v4i64(<4 x i64>)
+declare i32 @llvm.vector.reduce.or.v4i32(<4 x i32>)
+declare i32 @llvm.vector.reduce.or.v8i32(<8 x i32>)
+declare i16 @llvm.vector.reduce.or.v4i16(<4 x i16>)
+declare i16 @llvm.vector.reduce.or.v8i16(<8 x i16>)
+declare i16 @llvm.vector.reduce.or.v16i16(<16 x i16>)
+declare i8 @llvm.vector.reduce.or.v1i8(<1 x i8>)
+declare i8 @llvm.vector.reduce.or.v3i8(<3 x i8>)
+declare i8 @llvm.vector.reduce.or.v4i8(<4 x i8>)
+declare i8 @llvm.vector.reduce.or.v8i8(<8 x i8>)
+declare i8 @llvm.vector.reduce.or.v16i8(<16 x i8>)
+declare i8 @llvm.vector.reduce.or.v32i8(<32 x i8>)
+declare i8 @llvm.vector.reduce.or.v64i8(<64 x i8>)

llvm/test/CodeGen/AArch64/reduce-xor.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=aarch64-unknown-linux-gnu | FileCheck %s
+
+define i1 @test_redor_v1i1(<1 x i1> %a) {
+; CHECK-LABEL: test_redor_v1i1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    and w0, w0, #0x1
+; CHECK-NEXT:    ret
+  %or_result = call i1 @llvm.vector.reduce.xor.v1i1(<1 x i1> %a)
+  ret i1 %or_result
+}
+
+define i1 @test_redor_v2i1(<2 x i1> %a) {
+; CHECK-LABEL: test_redor_v2i1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    mov w8, v0.s[1]
+; CHECK-NEXT:    fmov w9, s0
+; CHECK-NEXT:    eor w8, w9, w8
+; CHECK-NEXT:    and w0, w8, #0x1
+; CHECK-NEXT:    ret
+  %or_result = call i1 @llvm.vector.reduce.xor.v2i1(<2 x i1> %a)
+  ret i1 %or_result
+}
+
+define i1 @test_redor_v4i1(<4 x i1> %a) {
+; CHECK-LABEL: test_redor_v4i1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w10, v0.h[1]
+; CHECK-NEXT:    umov w11, v0.h[0]
+; CHECK-NEXT:    umov w9, v0.h[2]
+; CHECK-NEXT:    eor w10, w11, w10
+; CHECK-NEXT:    umov w8, v0.h[3]
+; CHECK-NEXT:    eor w9, w10, w9
+; CHECK-NEXT:    eor w8, w9, w8
+; CHECK-NEXT:    and w0, w8, #0x1
+; CHECK-NEXT:    ret
+  %or_result = call i1 @llvm.vector.reduce.xor.v4i1(<4 x i1> %a)
+  ret i1 %or_result
+}
+
+define i1 @test_redor_v8i1(<8 x i1> %a) {
+; CHECK-LABEL: test_redor_v8i1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w14, v0.b[1]
+; CHECK-NEXT:    umov w15, v0.b[0]
+; CHECK-NEXT:    umov w13, v0.b[2]
+; CHECK-NEXT:    eor w14, w15, w14
+; CHECK-NEXT:    umov w12, v0.b[3]
+; CHECK-NEXT:    eor w13, w14, w13
+; CHECK-NEXT:    umov w11, v0.b[4]
+; CHECK-NEXT:    eor w12, w13, w12
+; CHECK-NEXT:    umov w10, v0.b[5]
+; CHECK-NEXT:    eor w11, w12, w11
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    eor w10, w11, w10
+; CHECK-NEXT:    umov w8, v0.b[7]
+; CHECK-NEXT:    eor w9, w10, w9
+; CHECK-NEXT:    eor w8, w9, w8
+; CHECK-NEXT:    and w0, w8, #0x1
+; CHECK-NEXT:    ret
+  %or_result = call i1 @llvm.vector.reduce.xor.v8i1(<8 x i1> %a)
+  ret i1 %or_result
+}
+
+define i1 @test_redor_v16i1(<16 x i1> %a) {
+; CHECK-LABEL: test_redor_v16i1:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    eor v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.b[1]
+; CHECK-NEXT:    umov w9, v0.b[0]
+; CHECK-NEXT:    eor w8, w9, w8
+; CHECK-NEXT:    umov w9, v0.b[2]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[3]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[4]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[5]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[7]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    and w0, w8, #0x1
+; CHECK-NEXT:    ret
+  %or_result = call i1 @llvm.vector.reduce.xor.v16i1(<16 x i1> %a)
+  ret i1 %or_result
+}
+
+define i8 @test_redxor_v1i8(<1 x i8> %a) {
+; CHECK-LABEL: test_redxor_v1i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w0, v0.b[0]
+; CHECK-NEXT:    ret
+  %xor_result = call i8 @llvm.vector.reduce.xor.v1i8(<1 x i8> %a)
+  ret i8 %xor_result
+}
+
+define i8 @test_redxor_v3i8(<3 x i8> %a) {
+; CHECK-LABEL: test_redxor_v3i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor w8, w0, w1
+; CHECK-NEXT:    eor w0, w8, w2
+; CHECK-NEXT:    ret
+  %xor_result = call i8 @llvm.vector.reduce.xor.v3i8(<3 x i8> %a)
+  ret i8 %xor_result
+}
+
+define i8 @test_redxor_v4i8(<4 x i8> %a) {
+; CHECK-LABEL: test_redxor_v4i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w10, v0.h[1]
+; CHECK-NEXT:    umov w11, v0.h[0]
+; CHECK-NEXT:    umov w9, v0.h[2]
+; CHECK-NEXT:    eor w10, w11, w10
+; CHECK-NEXT:    umov w8, v0.h[3]
+; CHECK-NEXT:    eor w9, w10, w9
+; CHECK-NEXT:    eor w0, w9, w8
+; CHECK-NEXT:    ret
+  %xor_result = call i8 @llvm.vector.reduce.xor.v4i8(<4 x i8> %a)
+  ret i8 %xor_result
+}
+
+define i8 @test_redxor_v8i8(<8 x i8> %a) {
+; CHECK-LABEL: test_redxor_v8i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w14, v0.b[1]
+; CHECK-NEXT:    umov w15, v0.b[0]
+; CHECK-NEXT:    umov w13, v0.b[2]
+; CHECK-NEXT:    eor w14, w15, w14
+; CHECK-NEXT:    umov w12, v0.b[3]
+; CHECK-NEXT:    eor w13, w14, w13
+; CHECK-NEXT:    umov w11, v0.b[4]
+; CHECK-NEXT:    eor w12, w13, w12
+; CHECK-NEXT:    umov w10, v0.b[5]
+; CHECK-NEXT:    eor w11, w12, w11
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    eor w10, w11, w10
+; CHECK-NEXT:    umov w8, v0.b[7]
+; CHECK-NEXT:    eor w9, w10, w9
+; CHECK-NEXT:    eor w0, w9, w8
+; CHECK-NEXT:    ret
+  %xor_result = call i8 @llvm.vector.reduce.xor.v8i8(<8 x i8> %a)
+  ret i8 %xor_result
+}
+
+define i8 @test_redxor_v16i8(<16 x i8> %a) {
+; CHECK-LABEL: test_redxor_v16i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    eor v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.b[1]
+; CHECK-NEXT:    umov w9, v0.b[0]
+; CHECK-NEXT:    eor w8, w9, w8
+; CHECK-NEXT:    umov w9, v0.b[2]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[3]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[4]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[5]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[7]
+; CHECK-NEXT:    eor w0, w8, w9
+; CHECK-NEXT:    ret
+  %xor_result = call i8 @llvm.vector.reduce.xor.v16i8(<16 x i8> %a)
+  ret i8 %xor_result
+}
+
+define i8 @test_redxor_v32i8(<32 x i8> %a) {
+; CHECK-LABEL: test_redxor_v32i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor v0.16b, v0.16b, v1.16b
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    eor v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.b[1]
+; CHECK-NEXT:    umov w9, v0.b[0]
+; CHECK-NEXT:    eor w8, w9, w8
+; CHECK-NEXT:    umov w9, v0.b[2]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[3]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[4]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[5]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[7]
+; CHECK-NEXT:    eor w0, w8, w9
+; CHECK-NEXT:    ret
+  %xor_result = call i8 @llvm.vector.reduce.xor.v32i8(<32 x i8> %a)
+  ret i8 %xor_result
+}
+
+define i8 @test_redxor_v64i8(<64 x i8> %a) {
+; CHECK-LABEL: test_redxor_v64i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor v1.16b, v1.16b, v3.16b
+; CHECK-NEXT:    eor v0.16b, v0.16b, v2.16b
+; CHECK-NEXT:    eor v0.16b, v0.16b, v1.16b
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    eor v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.b[1]
+; CHECK-NEXT:    umov w9, v0.b[0]
+; CHECK-NEXT:    eor w8, w9, w8
+; CHECK-NEXT:    umov w9, v0.b[2]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[3]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[4]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[5]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[6]
+; CHECK-NEXT:    eor w8, w8, w9
+; CHECK-NEXT:    umov w9, v0.b[7]
+; CHECK-NEXT:    eor w0, w8, w9
+; CHECK-NEXT:    ret
+  %xor_result = call i8 @llvm.vector.reduce.xor.v64i8(<64 x i8> %a)
+  ret i8 %xor_result
+}
+
+define i16 @test_redxor_v4i16(<4 x i16> %a) {
+; CHECK-LABEL: test_redxor_v4i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $q0
+; CHECK-NEXT:    umov w10, v0.h[1]
+; CHECK-NEXT:    umov w11, v0.h[0]
+; CHECK-NEXT:    umov w9, v0.h[2]
+; CHECK-NEXT:    eor w10, w11, w10
+; CHECK-NEXT:    umov w8, v0.h[3]
+; CHECK-NEXT:    eor w9, w10, w9
+; CHECK-NEXT:    eor w0, w9, w8
+; CHECK-NEXT:    ret
+  %xor_result = call i16 @llvm.vector.reduce.xor.v4i16(<4 x i16> %a)
+  ret i16 %xor_result
+}
+
+define i16 @test_redxor_v8i16(<8 x i16> %a) {
+; CHECK-LABEL: test_redxor_v8i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    eor v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.h[1]
+; CHECK-NEXT:    umov w9, v0.h[0]
+; CHECK-NEXT:    umov w10, v0.h[2]
+; CHECK-NEXT:    eor w8, w9, w8
+; CHECK-NEXT:    eor w8, w8, w10
+; CHECK-NEXT:    umov w9, v0.h[3]
+; CHECK-NEXT:    eor w0, w8, w9
+; CHECK-NEXT:    ret
+  %xor_result = call i16 @llvm.vector.reduce.xor.v8i16(<8 x i16> %a)
+  ret i16 %xor_result
+}
+
+define i16 @test_redxor_v16i16(<16 x i16> %a) {
+; CHECK-LABEL: test_redxor_v16i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor v0.16b, v0.16b, v1.16b
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    eor v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    umov w8, v0.h[1]
+; CHECK-NEXT:    umov w9, v0.h[0]
+; CHECK-NEXT:    umov w10, v0.h[2]
+; CHECK-NEXT:    eor w8, w9, w8
+; CHECK-NEXT:    eor w8, w8, w10
+; CHECK-NEXT:    umov w9, v0.h[3]
+; CHECK-NEXT:    eor w0, w8, w9
+; CHECK-NEXT:    ret
+  %xor_result = call i16 @llvm.vector.reduce.xor.v16i16(<16 x i16> %a)
+  ret i16 %xor_result
+}
+
+define i32 @test_redxor_v4i32(<4 x i32> %a) {
+; CHECK-LABEL: test_redxor_v4i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    eor v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    mov w8, v0.s[1]
+; CHECK-NEXT:    fmov w9, s0
+; CHECK-NEXT:    eor w0, w9, w8
+; CHECK-NEXT:    ret
+  %xor_result = call i32 @llvm.vector.reduce.xor.v4i32(<4 x i32> %a)
+  ret i32 %xor_result
+}
+
+define i32 @test_redxor_v8i32(<8 x i32> %a) {
+; CHECK-LABEL: test_redxor_v8i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor v0.16b, v0.16b, v1.16b
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    eor v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    mov w8, v0.s[1]
+; CHECK-NEXT:    fmov w9, s0
+; CHECK-NEXT:    eor w0, w9, w8
+; CHECK-NEXT:    ret
+  %xor_result = call i32 @llvm.vector.reduce.xor.v8i32(<8 x i32> %a)
+  ret i32 %xor_result
+}
+
+define i64 @test_redxor_v2i64(<2 x i64> %a) {
+; CHECK-LABEL: test_redxor_v2i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    eor v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    fmov x0, d0
+; CHECK-NEXT:    ret
+  %xor_result = call i64 @llvm.vector.reduce.xor.v2i64(<2 x i64> %a)
+  ret i64 %xor_result
+}
+
+define i64 @test_redxor_v4i64(<4 x i64> %a) {
+; CHECK-LABEL: test_redxor_v4i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    eor v0.16b, v0.16b, v1.16b
+; CHECK-NEXT:    ext v1.16b, v0.16b, v0.16b, #8
+; CHECK-NEXT:    eor v0.8b, v0.8b, v1.8b
+; CHECK-NEXT:    fmov x0, d0
+; CHECK-NEXT:    ret
+  %xor_result = call i64 @llvm.vector.reduce.xor.v4i64(<4 x i64> %a)
+  ret i64 %xor_result
+}
+
+declare i1 @llvm.vector.reduce.xor.v1i1(<1 x i1>)
+declare i1 @llvm.vector.reduce.xor.v2i1(<2 x i1>)
+declare i1 @llvm.vector.reduce.xor.v4i1(<4 x i1>)
+declare i1 @llvm.vector.reduce.xor.v8i1(<8 x i1>)
+declare i1 @llvm.vector.reduce.xor.v16i1(<16 x i1>)
+declare i64 @llvm.vector.reduce.xor.v2i64(<2 x i64>)
+declare i64 @llvm.vector.reduce.xor.v4i64(<4 x i64>)
+declare i32 @llvm.vector.reduce.xor.v4i32(<4 x i32>)
+declare i32 @llvm.vector.reduce.xor.v8i32(<8 x i32>)
+declare i16 @llvm.vector.reduce.xor.v4i16(<4 x i16>)
+declare i16 @llvm.vector.reduce.xor.v8i16(<8 x i16>)
+declare i16 @llvm.vector.reduce.xor.v16i16(<16 x i16>)
+declare i8 @llvm.vector.reduce.xor.v1i8(<1 x i8>)
+declare i8 @llvm.vector.reduce.xor.v3i8(<3 x i8>)
+declare i8 @llvm.vector.reduce.xor.v4i8(<4 x i8>)
+declare i8 @llvm.vector.reduce.xor.v8i8(<8 x i8>)
+declare i8 @llvm.vector.reduce.xor.v16i8(<16 x i8>)
+declare i8 @llvm.vector.reduce.xor.v32i8(<32 x i8>)
+declare i8 @llvm.vector.reduce.xor.v64i8(<64 x i8>)