test/CodeGen/AArch64/arm64-neon-aba-abd.ll

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=arm64-none-linux-gnu < %s | FileCheck %s --check-prefixes=CHECK,CHECK-SD
; RUN: llc -mtriple=arm64-none-linux-gnu -global-isel < %s | FileCheck %s --check-prefixes=CHECK,CHECK-GI

declare <8 x i8> @llvm.aarch64.neon.uabd.v8i8(<8 x i8>, <8 x i8>)
declare <8 x i8> @llvm.aarch64.neon.sabd.v8i8(<8 x i8>, <8 x i8>)

define <8 x i8> @test_uabd_v8i8(<8 x i8> %lhs, <8 x i8> %rhs) {
; CHECK-LABEL: test_uabd_v8i8:
; CHECK:       // %bb.0:
; CHECK-NEXT:    uabd v0.8b, v0.8b, v1.8b
; CHECK-NEXT:    ret
  %abd = call <8 x i8> @llvm.aarch64.neon.uabd.v8i8(<8 x i8> %lhs, <8 x i8> %rhs)
  ret <8 x i8> %abd
}

define <8 x i8> @test_uaba_v8i8(<8 x i8> %lhs, <8 x i8> %rhs) {
; CHECK-LABEL: test_uaba_v8i8:
; CHECK:       // %bb.0:
; CHECK-NEXT:    uaba v0.8b, v0.8b, v1.8b
; CHECK-NEXT:    ret
  %abd = call <8 x i8> @llvm.aarch64.neon.uabd.v8i8(<8 x i8> %lhs, <8 x i8> %rhs)
  %aba = add <8 x i8> %lhs, %abd
  ret <8 x i8> %aba
}

define <8 x i8> @test_sabd_v8i8(<8 x i8> %lhs, <8 x i8> %rhs) {
; CHECK-LABEL: test_sabd_v8i8:
; CHECK:       // %bb.0:
; CHECK-NEXT:    sabd v0.8b, v0.8b, v1.8b
; CHECK-NEXT:    ret
  %abd = call <8 x i8> @llvm.aarch64.neon.sabd.v8i8(<8 x i8> %lhs, <8 x i8> %rhs)
  ret <8 x i8> %abd
}

define <8 x i8> @test_saba_v8i8(<8 x i8> %lhs, <8 x i8> %rhs) {
; CHECK-LABEL: test_saba_v8i8:
; CHECK:       // %bb.0:
; CHECK-NEXT:    saba v0.8b, v0.8b, v1.8b
; CHECK-NEXT:    ret
  %abd = call <8 x i8> @llvm.aarch64.neon.sabd.v8i8(<8 x i8> %lhs, <8 x i8> %rhs)
  %aba = add <8 x i8> %lhs, %abd
  ret <8 x i8> %aba
}

declare <16 x i8> @llvm.aarch64.neon.uabd.v16i8(<16 x i8>, <16 x i8>)
declare <16 x i8> @llvm.aarch64.neon.sabd.v16i8(<16 x i8>, <16 x i8>)

define <16 x i8> @test_uabd_v16i8(<16 x i8> %lhs, <16 x i8> %rhs) {
; CHECK-LABEL: test_uabd_v16i8:
; CHECK:       // %bb.0:
; CHECK-NEXT:    uabd v0.16b, v0.16b, v1.16b
; CHECK-NEXT:    ret
  %abd = call <16 x i8> @llvm.aarch64.neon.uabd.v16i8(<16 x i8> %lhs, <16 x i8> %rhs)
  ret <16 x i8> %abd
}

define <16 x i8> @test_uaba_v16i8(<16 x i8> %lhs, <16 x i8> %rhs) {
; CHECK-LABEL: test_uaba_v16i8:
; CHECK:       // %bb.0:
; CHECK-NEXT:    uaba v0.16b, v0.16b, v1.16b
; CHECK-NEXT:    ret
  %abd = call <16 x i8> @llvm.aarch64.neon.uabd.v16i8(<16 x i8> %lhs, <16 x i8> %rhs)
  %aba = add <16 x i8> %lhs, %abd
  ret <16 x i8> %aba
}

define <16 x i8> @test_sabd_v16i8(<16 x i8> %lhs, <16 x i8> %rhs) {
; CHECK-LABEL: test_sabd_v16i8:
; CHECK:       // %bb.0:
; CHECK-NEXT:    sabd v0.16b, v0.16b, v1.16b
; CHECK-NEXT:    ret
  %abd = call <16 x i8> @llvm.aarch64.neon.sabd.v16i8(<16 x i8> %lhs, <16 x i8> %rhs)
  ret <16 x i8> %abd
}

define <16 x i8> @test_saba_v16i8(<16 x i8> %lhs, <16 x i8> %rhs) {
; CHECK-LABEL: test_saba_v16i8:
; CHECK:       // %bb.0:
; CHECK-NEXT:    saba v0.16b, v0.16b, v1.16b
; CHECK-NEXT:    ret
  %abd = call <16 x i8> @llvm.aarch64.neon.sabd.v16i8(<16 x i8> %lhs, <16 x i8> %rhs)
  %aba = add <16 x i8> %lhs, %abd
  ret <16 x i8> %aba
}

declare <4 x i16> @llvm.aarch64.neon.uabd.v4i16(<4 x i16>, <4 x i16>)
declare <4 x i16> @llvm.aarch64.neon.sabd.v4i16(<4 x i16>, <4 x i16>)

define <4 x i16> @test_uabd_v4i16(<4 x i16> %lhs, <4 x i16> %rhs) {
; CHECK-LABEL: test_uabd_v4i16:
; CHECK:       // %bb.0:
; CHECK-NEXT:    uabd v0.4h, v0.4h, v1.4h
; CHECK-NEXT:    ret
  %abd = call <4 x i16> @llvm.aarch64.neon.uabd.v4i16(<4 x i16> %lhs, <4 x i16> %rhs)
  ret <4 x i16> %abd
}

define <4 x i16> @test_uaba_v4i16(<4 x i16> %lhs, <4 x i16> %rhs) {
; CHECK-LABEL: test_uaba_v4i16:
; CHECK:       // %bb.0:
; CHECK-NEXT:    uaba v0.4h, v0.4h, v1.4h
; CHECK-NEXT:    ret
  %abd = call <4 x i16> @llvm.aarch64.neon.uabd.v4i16(<4 x i16> %lhs, <4 x i16> %rhs)
  %aba = add <4 x i16> %lhs, %abd
  ret <4 x i16> %aba
}

define <4 x i16> @test_sabd_v4i16(<4 x i16> %lhs, <4 x i16> %rhs) {
; CHECK-LABEL: test_sabd_v4i16:
; CHECK:       // %bb.0:
; CHECK-NEXT:    sabd v0.4h, v0.4h, v1.4h
; CHECK-NEXT:    ret
  %abd = call <4 x i16> @llvm.aarch64.neon.sabd.v4i16(<4 x i16> %lhs, <4 x i16> %rhs)
  ret <4 x i16> %abd
}

define <4 x i16> @test_saba_v4i16(<4 x i16> %lhs, <4 x i16> %rhs) {
; CHECK-LABEL: test_saba_v4i16:
; CHECK:       // %bb.0:
; CHECK-NEXT:    saba v0.4h, v0.4h, v1.4h
; CHECK-NEXT:    ret
  %abd = call <4 x i16> @llvm.aarch64.neon.sabd.v4i16(<4 x i16> %lhs, <4 x i16> %rhs)
  %aba = add <4 x i16> %lhs, %abd
  ret <4 x i16> %aba
}

declare <8 x i16> @llvm.aarch64.neon.uabd.v8i16(<8 x i16>, <8 x i16>)
declare <8 x i16> @llvm.aarch64.neon.sabd.v8i16(<8 x i16>, <8 x i16>)

define <8 x i16> @test_uabd_v8i16(<8 x i16> %lhs, <8 x i16> %rhs) {
; CHECK-LABEL: test_uabd_v8i16:
; CHECK:       // %bb.0:
; CHECK-NEXT:    uabd v0.8h, v0.8h, v1.8h
; CHECK-NEXT:    ret
  %abd = call <8 x i16> @llvm.aarch64.neon.uabd.v8i16(<8 x i16> %lhs, <8 x i16> %rhs)
  ret <8 x i16> %abd
}

define <8 x i16> @test_uaba_v8i16(<8 x i16> %lhs, <8 x i16> %rhs) {
; CHECK-LABEL: test_uaba_v8i16:
; CHECK:       // %bb.0:
; CHECK-NEXT:    uaba v0.8h, v0.8h, v1.8h
; CHECK-NEXT:    ret
  %abd = call <8 x i16> @llvm.aarch64.neon.uabd.v8i16(<8 x i16> %lhs, <8 x i16> %rhs)
  %aba = add <8 x i16> %lhs, %abd
  ret <8 x i16> %aba
}

define <8 x i16> @test_sabd_v8i16(<8 x i16> %lhs, <8 x i16> %rhs) {
; CHECK-LABEL: test_sabd_v8i16:
; CHECK:       // %bb.0:
; CHECK-NEXT:    sabd v0.8h, v0.8h, v1.8h
; CHECK-NEXT:    ret
  %abd = call <8 x i16> @llvm.aarch64.neon.sabd.v8i16(<8 x i16> %lhs, <8 x i16> %rhs)
  ret <8 x i16> %abd
}

define <8 x i16> @test_saba_v8i16(<8 x i16> %lhs, <8 x i16> %rhs) {
; CHECK-LABEL: test_saba_v8i16:
; CHECK:       // %bb.0:
; CHECK-NEXT:    saba v0.8h, v0.8h, v1.8h
; CHECK-NEXT:    ret
  %abd = call <8 x i16> @llvm.aarch64.neon.sabd.v8i16(<8 x i16> %lhs, <8 x i16> %rhs)
  %aba = add <8 x i16> %lhs, %abd
  ret <8 x i16> %aba
}

declare <2 x i32> @llvm.aarch64.neon.uabd.v2i32(<2 x i32>, <2 x i32>)
declare <2 x i32> @llvm.aarch64.neon.sabd.v2i32(<2 x i32>, <2 x i32>)

define <2 x i32> @test_uabd_v2i32(<2 x i32> %lhs, <2 x i32> %rhs) {
; CHECK-LABEL: test_uabd_v2i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    uabd v0.2s, v0.2s, v1.2s
; CHECK-NEXT:    ret
  %abd = call <2 x i32> @llvm.aarch64.neon.uabd.v2i32(<2 x i32> %lhs, <2 x i32> %rhs)
  ret <2 x i32> %abd
}

define <2 x i32> @test_uaba_v2i32(<2 x i32> %lhs, <2 x i32> %rhs) {
; CHECK-LABEL: test_uaba_v2i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    uaba v0.2s, v0.2s, v1.2s
; CHECK-NEXT:    ret
  %abd = call <2 x i32> @llvm.aarch64.neon.uabd.v2i32(<2 x i32> %lhs, <2 x i32> %rhs)
  %aba = add <2 x i32> %lhs, %abd
  ret <2 x i32> %aba
}

define <2 x i32> @test_sabd_v2i32(<2 x i32> %lhs, <2 x i32> %rhs) {
; CHECK-LABEL: test_sabd_v2i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    sabd v0.2s, v0.2s, v1.2s
; CHECK-NEXT:    ret
  %abd = call <2 x i32> @llvm.aarch64.neon.sabd.v2i32(<2 x i32> %lhs, <2 x i32> %rhs)
  ret <2 x i32> %abd
}

define <2 x i32> @test_sabd_v2i32_const() {
; CHECK-SD-LABEL: test_sabd_v2i32_const:
; CHECK-SD:       // %bb.0:
; CHECK-SD-NEXT:    adrp x8, .LCPI19_0
; CHECK-SD-NEXT:    ldr d0, [x8, :lo12:.LCPI19_0]
; CHECK-SD-NEXT:    ret
;
; CHECK-GI-LABEL: test_sabd_v2i32_const:
; CHECK-GI:       // %bb.0:
; CHECK-GI-NEXT:    adrp x8, .LCPI19_1
; CHECK-GI-NEXT:    adrp x9, .LCPI19_0
; CHECK-GI-NEXT:    ldr d0, [x8, :lo12:.LCPI19_1]
; CHECK-GI-NEXT:    ldr d1, [x9, :lo12:.LCPI19_0]
; CHECK-GI-NEXT:    sabd v0.2s, v0.2s, v1.2s
; CHECK-GI-NEXT:    ret
  %1 = tail call <2 x i32> @llvm.aarch64.neon.sabd.v2i32(
    <2 x i32> <i32 -2147483648, i32 2147450880>,
    <2 x i32> <i32 -65536, i32 65535>)
  ret <2 x i32> %1
}

define <2 x i32> @test_saba_v2i32(<2 x i32> %lhs, <2 x i32> %rhs) {
; CHECK-LABEL: test_saba_v2i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    saba v0.2s, v0.2s, v1.2s
; CHECK-NEXT:    ret
  %abd = call <2 x i32> @llvm.aarch64.neon.sabd.v2i32(<2 x i32> %lhs, <2 x i32> %rhs)
  %aba = add <2 x i32> %lhs, %abd
  ret <2 x i32> %aba
}

declare <4 x i32> @llvm.aarch64.neon.uabd.v4i32(<4 x i32>, <4 x i32>)
declare <4 x i32> @llvm.aarch64.neon.sabd.v4i32(<4 x i32>, <4 x i32>)

define <4 x i32> @test_uabd_v4i32(<4 x i32> %lhs, <4 x i32> %rhs) {
; CHECK-LABEL: test_uabd_v4i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    uabd v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    ret
  %abd = call <4 x i32> @llvm.aarch64.neon.uabd.v4i32(<4 x i32> %lhs, <4 x i32> %rhs)
  ret <4 x i32> %abd
}

define <4 x i32> @test_uaba_v4i32(<4 x i32> %lhs, <4 x i32> %rhs) {
; CHECK-LABEL: test_uaba_v4i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    uaba v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    ret
  %abd = call <4 x i32> @llvm.aarch64.neon.uabd.v4i32(<4 x i32> %lhs, <4 x i32> %rhs)
  %aba = add <4 x i32> %lhs, %abd
  ret <4 x i32> %aba
}

define <4 x i32> @test_sabd_v4i32(<4 x i32> %lhs, <4 x i32> %rhs) {
; CHECK-LABEL: test_sabd_v4i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    sabd v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    ret
  %abd = call <4 x i32> @llvm.aarch64.neon.sabd.v4i32(<4 x i32> %lhs, <4 x i32> %rhs)
  ret <4 x i32> %abd
}

define <4 x i32> @test_saba_v4i32(<4 x i32> %lhs, <4 x i32> %rhs) {
; CHECK-LABEL: test_saba_v4i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    saba v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    ret
  %abd = call <4 x i32> @llvm.aarch64.neon.sabd.v4i32(<4 x i32> %lhs, <4 x i32> %rhs)
  %aba = add <4 x i32> %lhs, %abd
  ret <4 x i32> %aba
}

declare <2 x float> @llvm.aarch64.neon.fabd.v2f32(<2 x float>, <2 x float>)

define <2 x float> @test_fabd_v2f32(<2 x float> %lhs, <2 x float> %rhs) {
; CHECK-LABEL: test_fabd_v2f32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    fabd v0.2s, v0.2s, v1.2s
; CHECK-NEXT:    ret
  %abd = call <2 x float> @llvm.aarch64.neon.fabd.v2f32(<2 x float> %lhs, <2 x float> %rhs)
  ret <2 x float> %abd
}

declare <4 x float> @llvm.aarch64.neon.fabd.v4f32(<4 x float>, <4 x float>)

define <4 x float> @test_fabd_v4f32(<4 x float> %lhs, <4 x float> %rhs) {
; CHECK-LABEL: test_fabd_v4f32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    fabd v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    ret
  %abd = call <4 x float> @llvm.aarch64.neon.fabd.v4f32(<4 x float> %lhs, <4 x float> %rhs)
  ret <4 x float> %abd
}

declare <2 x double> @llvm.aarch64.neon.fabd.v2f64(<2 x double>, <2 x double>)

define <2 x double> @test_fabd_v2f64(<2 x double> %lhs, <2 x double> %rhs) {
; CHECK-LABEL: test_fabd_v2f64:
; CHECK:       // %bb.0:
; CHECK-NEXT:    fabd v0.2d, v0.2d, v1.2d
; CHECK-NEXT:    ret
  %abd = call <2 x double> @llvm.aarch64.neon.fabd.v2f64(<2 x double> %lhs, <2 x double> %rhs)
  ret <2 x double> %abd
}

define <8 x i16> @test_uabd_knownbits_vec8i16(<8 x i16> %lhs, <8 x i16> %rhs) {
; CHECK-SD-LABEL: test_uabd_knownbits_vec8i16:
; CHECK-SD:       // %bb.0:
; CHECK-SD-NEXT:    movi v2.8h, #15
; CHECK-SD-NEXT:    and v0.16b, v0.16b, v2.16b
; CHECK-SD-NEXT:    and v1.16b, v1.16b, v2.16b
; CHECK-SD-NEXT:    uabd v0.8h, v0.8h, v1.8h
; CHECK-SD-NEXT:    rev64 v0.8h, v0.8h
; CHECK-SD-NEXT:    ext v0.16b, v0.16b, v0.16b, #8
; CHECK-SD-NEXT:    ret
;
; CHECK-GI-LABEL: test_uabd_knownbits_vec8i16:
; CHECK-GI:       // %bb.0:
; CHECK-GI-NEXT:    movi v2.8h, #15
; CHECK-GI-NEXT:    and v0.16b, v0.16b, v2.16b
; CHECK-GI-NEXT:    and v1.16b, v1.16b, v2.16b
; CHECK-GI-NEXT:    uabd v0.8h, v0.8h, v1.8h
; CHECK-GI-NEXT:    rev64 v0.8h, v0.8h
; CHECK-GI-NEXT:    ext v0.16b, v0.16b, v0.16b, #8
; CHECK-GI-NEXT:    and v0.16b, v0.16b, v2.16b
; CHECK-GI-NEXT:    ret
  %and1 = and <8 x i16> %lhs, <i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15>
  %and2 = and <8 x i16> %rhs, <i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15>
  %uabd = call <8 x i16> @llvm.aarch64.neon.uabd.v8i16(<8 x i16> %and1, <8 x i16> %and2)
  %suff = shufflevector <8 x i16> %uabd, <8 x i16> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
  %res = and <8 x i16> %suff, <i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15>
  ret <8 x i16> %res
}

define <4 x i32> @knownbits_uabd_mask_and_shuffle_lshr(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-SD-LABEL: knownbits_uabd_mask_and_shuffle_lshr:
; CHECK-SD:       // %bb.0:
; CHECK-SD-NEXT:    movi v0.2d, #0000000000000000
; CHECK-SD-NEXT:    ushr v0.4s, v0.4s, #17
; CHECK-SD-NEXT:    ret
;
; CHECK-GI-LABEL: knownbits_uabd_mask_and_shuffle_lshr:
; CHECK-GI:       // %bb.0:
; CHECK-GI-NEXT:    movi v2.2d, #0x00ffff0000ffff
; CHECK-GI-NEXT:    and v0.16b, v0.16b, v2.16b
; CHECK-GI-NEXT:    and v1.16b, v1.16b, v2.16b
; CHECK-GI-NEXT:    uabd v0.4s, v0.4s, v1.4s
; CHECK-GI-NEXT:    rev64 v0.4s, v0.4s
; CHECK-GI-NEXT:    ext v0.16b, v0.16b, v0.16b, #8
; CHECK-GI-NEXT:    ushr v0.4s, v0.4s, #17
; CHECK-GI-NEXT:    ret
  %1 = and <4 x i32> %a0, <i32 65535, i32 65535, i32 65535, i32 65535>
  %2 = and <4 x i32> %a1, <i32 65535, i32 65535, i32 65535, i32 65535>
  %3 = call <4 x i32> @llvm.aarch64.neon.uabd.v4i32(<4 x i32> %1, <4 x i32> %2)
  %4 = shufflevector <4 x i32> %3, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
  %5 = lshr <4 x i32> %4, <i32 17, i32 17, i32 17, i32 17>
  ret <4 x i32> %5
}

define <4 x i32> @knownbits_mask_and_shuffle_lshr(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-SD-LABEL: knownbits_mask_and_shuffle_lshr:
; CHECK-SD:       // %bb.0:
; CHECK-SD-NEXT:    movi v0.2d, #0000000000000000
; CHECK-SD-NEXT:    ret
;
; CHECK-GI-LABEL: knownbits_mask_and_shuffle_lshr:
; CHECK-GI:       // %bb.0:
; CHECK-GI-NEXT:    movi v2.4s, #127, msl #8
; CHECK-GI-NEXT:    and v0.16b, v0.16b, v2.16b
; CHECK-GI-NEXT:    and v1.16b, v1.16b, v2.16b
; CHECK-GI-NEXT:    uabd v0.4s, v0.4s, v1.4s
; CHECK-GI-NEXT:    ushr v0.4s, v0.4s, #17
; CHECK-GI-NEXT:    ret
  %1 = and <4 x i32> %a0, <i32 32767, i32 32767, i32 32767, i32 32767>
  %2 = and <4 x i32> %a1, <i32 32767, i32 32767, i32 32767, i32 32767>
  %3 = call <4 x i32> @llvm.aarch64.neon.uabd.v4i32(<4 x i32> %1, <4 x i32> %2)
  %4 = shufflevector <4 x i32> %3, <4 x i32> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %5 = lshr <4 x i32> %4, <i32 17, i32 17, i32 17, i32 17>
  ret <4 x i32> %5
}

define <4 x i32> @test_sabd_knownbits_vec4i32(<4 x i32> %lhs, <4 x i32> %rhs) {
; CHECK-SD-LABEL: test_sabd_knownbits_vec4i32:
; CHECK-SD:       // %bb.0:
; CHECK-SD-NEXT:    adrp x8, .LCPI31_0
; CHECK-SD-NEXT:    adrp x9, .LCPI31_1
; CHECK-SD-NEXT:    ldr q2, [x8, :lo12:.LCPI31_0]
; CHECK-SD-NEXT:    ldr q3, [x9, :lo12:.LCPI31_1]
; CHECK-SD-NEXT:    and v0.16b, v0.16b, v2.16b
; CHECK-SD-NEXT:    and v1.16b, v1.16b, v3.16b
; CHECK-SD-NEXT:    sabd v0.4s, v0.4s, v1.4s
; CHECK-SD-NEXT:    movi v1.2d, #0x0000ff000000ff
; CHECK-SD-NEXT:    mov v0.s[1], v0.s[0]
; CHECK-SD-NEXT:    trn2 v0.4s, v0.4s, v0.4s
; CHECK-SD-NEXT:    and v0.16b, v0.16b, v1.16b
; CHECK-SD-NEXT:    ret
;
; CHECK-GI-LABEL: test_sabd_knownbits_vec4i32:
; CHECK-GI:       // %bb.0:
; CHECK-GI-NEXT:    adrp x8, .LCPI31_2
; CHECK-GI-NEXT:    adrp x9, .LCPI31_1
; CHECK-GI-NEXT:    ldr q2, [x8, :lo12:.LCPI31_2]
; CHECK-GI-NEXT:    ldr q3, [x9, :lo12:.LCPI31_1]
; CHECK-GI-NEXT:    adrp x8, .LCPI31_0
; CHECK-GI-NEXT:    and v0.16b, v0.16b, v2.16b
; CHECK-GI-NEXT:    and v1.16b, v1.16b, v3.16b
; CHECK-GI-NEXT:    ldr q2, [x8, :lo12:.LCPI31_0]
; CHECK-GI-NEXT:    movi v3.2d, #0x0000ff000000ff
; CHECK-GI-NEXT:    sabd v0.4s, v0.4s, v1.4s
; CHECK-GI-NEXT:    tbl v0.16b, { v0.16b, v1.16b }, v2.16b
; CHECK-GI-NEXT:    and v0.16b, v0.16b, v3.16b
; CHECK-GI-NEXT:    ret
  %and1 = and <4 x i32> %lhs, <i32 255, i32 -1, i32 -1, i32 255>
  %and2 = and <4 x i32> %rhs, <i32 255, i32 255, i32 -1, i32 -1>
  %abd = call <4 x i32> @llvm.aarch64.neon.sabd.v4i32(<4 x i32> %and1, <4 x i32> %and2)
  %s = shufflevector <4 x i32> %abd, <4 x i32> undef, <4 x i32> <i32 0, i32 0, i32 3, i32 3>
  %4 = and <4 x i32> %s, <i32 255, i32 255, i32 255, i32 255>
  ret <4 x i32> %4
}

define <4 x i32> @knownbits_sabd_and_mask(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-SD-LABEL: knownbits_sabd_and_mask:
; CHECK-SD:       // %bb.0:
; CHECK-SD-NEXT:    adrp x8, .LCPI32_0
; CHECK-SD-NEXT:    ldr q2, [x8, :lo12:.LCPI32_0]
; CHECK-SD-NEXT:    and v0.16b, v0.16b, v2.16b
; CHECK-SD-NEXT:    and v1.16b, v1.16b, v2.16b
; CHECK-SD-NEXT:    sabd v0.4s, v0.4s, v1.4s
; CHECK-SD-NEXT:    zip2 v0.4s, v0.4s, v0.4s
; CHECK-SD-NEXT:    ret
;
; CHECK-GI-LABEL: knownbits_sabd_and_mask:
; CHECK-GI:       // %bb.0:
; CHECK-GI-NEXT:    adrp x8, .LCPI32_1
; CHECK-GI-NEXT:    ldr q2, [x8, :lo12:.LCPI32_1]
; CHECK-GI-NEXT:    adrp x8, .LCPI32_0
; CHECK-GI-NEXT:    and v0.16b, v0.16b, v2.16b
; CHECK-GI-NEXT:    and v1.16b, v1.16b, v2.16b
; CHECK-GI-NEXT:    ldr q2, [x8, :lo12:.LCPI32_0]
; CHECK-GI-NEXT:    sabd v0.4s, v0.4s, v1.4s
; CHECK-GI-NEXT:    tbl v0.16b, { v0.16b, v1.16b }, v2.16b
; CHECK-GI-NEXT:    ret
  %1 = and <4 x i32> %a0, <i32 -1, i32 -1, i32 255, i32 4085>
  %2 = and <4 x i32> %a1, <i32 -1, i32 -1, i32 255, i32 4085>
  %3 = call <4 x i32> @llvm.aarch64.neon.sabd.v4i32(<4 x i32> %1, <4 x i32> %2)
  %4 = shufflevector <4 x i32> %3, <4 x i32> undef, <4 x i32> <i32 2, i32 2, i32 3, i32 3>
  ret <4 x i32> %4
}

define <4 x i32> @knownbits_sabd_and_or_mask(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-SD-LABEL: knownbits_sabd_and_or_mask:
; CHECK-SD:       // %bb.0:
; CHECK-SD-NEXT:    movi v0.2d, #0000000000000000
; CHECK-SD-NEXT:    ret
;
; CHECK-GI-LABEL: knownbits_sabd_and_or_mask:
; CHECK-GI:       // %bb.0:
; CHECK-GI-NEXT:    adrp x8, .LCPI33_1
; CHECK-GI-NEXT:    movi v3.2d, #0x00ffff0000ffff
; CHECK-GI-NEXT:    ldr q2, [x8, :lo12:.LCPI33_1]
; CHECK-GI-NEXT:    adrp x8, .LCPI33_0
; CHECK-GI-NEXT:    and v0.16b, v0.16b, v2.16b
; CHECK-GI-NEXT:    and v1.16b, v1.16b, v2.16b
; CHECK-GI-NEXT:    ldr q2, [x8, :lo12:.LCPI33_0]
; CHECK-GI-NEXT:    orr v0.16b, v0.16b, v3.16b
; CHECK-GI-NEXT:    orr v1.16b, v1.16b, v3.16b
; CHECK-GI-NEXT:    uabd v0.4s, v0.4s, v1.4s
; CHECK-GI-NEXT:    tbl v0.16b, { v0.16b, v1.16b }, v2.16b
; CHECK-GI-NEXT:    ret
  %1 = and <4 x i32> %a0, <i32 -1, i32 -1, i32 255, i32 4085>
  %2 = or <4 x i32> %1, <i32 65535, i32 65535, i32 65535, i32 65535>
  %3 = and <4 x i32> %a1, <i32 -1, i32 -1, i32 255, i32 4085>
  %4 = or <4 x i32> %3, <i32 65535, i32 65535, i32 65535, i32 65535>
  %5 = call <4 x i32> @llvm.aarch64.neon.uabd.v4i32(<4 x i32> %2, <4 x i32> %4)
  %6 = shufflevector <4 x i32> %5, <4 x i32> undef, <4 x i32> <i32 2, i32 2, i32 3, i32 3>
  ret <4 x i32> %6
}

define <4 x i32> @knownbits_sabd_and_xor_mask(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-SD-LABEL: knownbits_sabd_and_xor_mask:
; CHECK-SD:       // %bb.0:
; CHECK-SD-NEXT:    adrp x8, .LCPI34_0
; CHECK-SD-NEXT:    movi v3.2d, #0x00ffff0000ffff
; CHECK-SD-NEXT:    ldr q2, [x8, :lo12:.LCPI34_0]
; CHECK-SD-NEXT:    and v0.16b, v0.16b, v2.16b
; CHECK-SD-NEXT:    and v1.16b, v1.16b, v2.16b
; CHECK-SD-NEXT:    eor v0.16b, v0.16b, v3.16b
; CHECK-SD-NEXT:    eor v1.16b, v1.16b, v3.16b
; CHECK-SD-NEXT:    sabd v0.4s, v0.4s, v1.4s
; CHECK-SD-NEXT:    zip2 v0.4s, v0.4s, v0.4s
; CHECK-SD-NEXT:    ret
;
; CHECK-GI-LABEL: knownbits_sabd_and_xor_mask:
; CHECK-GI:       // %bb.0:
; CHECK-GI-NEXT:    adrp x8, .LCPI34_1
; CHECK-GI-NEXT:    movi v3.2d, #0x00ffff0000ffff
; CHECK-GI-NEXT:    ldr q2, [x8, :lo12:.LCPI34_1]
; CHECK-GI-NEXT:    adrp x8, .LCPI34_0
; CHECK-GI-NEXT:    and v0.16b, v0.16b, v2.16b
; CHECK-GI-NEXT:    and v1.16b, v1.16b, v2.16b
; CHECK-GI-NEXT:    ldr q2, [x8, :lo12:.LCPI34_0]
; CHECK-GI-NEXT:    eor v0.16b, v0.16b, v3.16b
; CHECK-GI-NEXT:    eor v1.16b, v1.16b, v3.16b
; CHECK-GI-NEXT:    sabd v0.4s, v0.4s, v1.4s
; CHECK-GI-NEXT:    tbl v0.16b, { v0.16b, v1.16b }, v2.16b
; CHECK-GI-NEXT:    ret
  %1 = and <4 x i32> %a0, <i32 -1, i32 -1, i32 255, i32 4085>
  %2 = xor <4 x i32> %1, <i32 65535, i32 65535, i32 65535, i32 65535>
  %3 = and <4 x i32> %a1, <i32 -1, i32 -1, i32 255, i32 4085>
  %4 = xor <4 x i32> %3, <i32 65535, i32 65535, i32 65535, i32 65535>
  %5 = call <4 x i32> @llvm.aarch64.neon.sabd.v4i32(<4 x i32> %2, <4 x i32> %4)
  %6 = shufflevector <4 x i32> %5, <4 x i32> undef, <4 x i32> <i32 2, i32 2, i32 3, i32 3>
  ret <4 x i32> %6
}

define <4 x i32> @knownbits_sabd_and_shl_mask(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-SD-LABEL: knownbits_sabd_and_shl_mask:
; CHECK-SD:       // %bb.0:
; CHECK-SD-NEXT:    movi v0.2d, #0000000000000000
; CHECK-SD-NEXT:    ret
;
; CHECK-GI-LABEL: knownbits_sabd_and_shl_mask:
; CHECK-GI:       // %bb.0:
; CHECK-GI-NEXT:    adrp x8, .LCPI35_1
; CHECK-GI-NEXT:    ldr q2, [x8, :lo12:.LCPI35_1]
; CHECK-GI-NEXT:    adrp x8, .LCPI35_0
; CHECK-GI-NEXT:    and v0.16b, v0.16b, v2.16b
; CHECK-GI-NEXT:    and v1.16b, v1.16b, v2.16b
; CHECK-GI-NEXT:    ldr q2, [x8, :lo12:.LCPI35_0]
; CHECK-GI-NEXT:    shl v0.4s, v0.4s, #17
; CHECK-GI-NEXT:    shl v1.4s, v1.4s, #17
; CHECK-GI-NEXT:    sabd v0.4s, v0.4s, v1.4s
; CHECK-GI-NEXT:    tbl v0.16b, { v0.16b, v1.16b }, v2.16b
; CHECK-GI-NEXT:    ret
  %1 = and <4 x i32> %a0, <i32 -65536, i32 -7, i32 -7, i32 -65536>
  %2 = shl <4 x i32> %1, <i32 17, i32 17, i32 17, i32 17>
  %3 = and <4 x i32> %a1, <i32 -65536, i32 -7, i32 -7, i32 -65536>
  %4 = shl <4 x i32> %3, <i32 17, i32 17, i32 17, i32 17>
  %5 = call <4 x i32> @llvm.aarch64.neon.sabd.v4i32(<4 x i32> %2, <4 x i32> %4)
  %6 = shufflevector <4 x i32> %5, <4 x i32> undef, <4 x i32> <i32 0, i32 0, i32 3, i32 3>
  ret <4 x i32> %6
}

define <4 x i32> @knownbits_sabd_and_mul_mask(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-SD-LABEL: knownbits_sabd_and_mul_mask:
; CHECK-SD:       // %bb.0:
; CHECK-SD-NEXT:    adrp x8, .LCPI36_0
; CHECK-SD-NEXT:    ldr q2, [x8, :lo12:.LCPI36_0]
; CHECK-SD-NEXT:    and v3.16b, v0.16b, v2.16b
; CHECK-SD-NEXT:    and v2.16b, v1.16b, v2.16b
; CHECK-SD-NEXT:    mul v0.4s, v0.4s, v3.4s
; CHECK-SD-NEXT:    mul v1.4s, v1.4s, v2.4s
; CHECK-SD-NEXT:    sabd v0.4s, v0.4s, v1.4s
; CHECK-SD-NEXT:    mov v0.s[1], v0.s[0]
; CHECK-SD-NEXT:    trn2 v0.4s, v0.4s, v0.4s
; CHECK-SD-NEXT:    ret
;
; CHECK-GI-LABEL: knownbits_sabd_and_mul_mask:
; CHECK-GI:       // %bb.0:
; CHECK-GI-NEXT:    adrp x8, .LCPI36_1
; CHECK-GI-NEXT:    ldr q2, [x8, :lo12:.LCPI36_1]
; CHECK-GI-NEXT:    adrp x8, .LCPI36_0
; CHECK-GI-NEXT:    and v3.16b, v0.16b, v2.16b
; CHECK-GI-NEXT:    and v2.16b, v1.16b, v2.16b
; CHECK-GI-NEXT:    mul v0.4s, v0.4s, v3.4s
; CHECK-GI-NEXT:    mul v1.4s, v1.4s, v2.4s
; CHECK-GI-NEXT:    ldr q2, [x8, :lo12:.LCPI36_0]
; CHECK-GI-NEXT:    sabd v0.4s, v0.4s, v1.4s
; CHECK-GI-NEXT:    tbl v0.16b, { v0.16b, v1.16b }, v2.16b
; CHECK-GI-NEXT:    ret
  %1 = and <4 x i32> %a0, <i32 -65536, i32 -7, i32 -7, i32 -65536>
  %2 = mul <4 x i32> %a0, %1
  %3 = and <4 x i32> %a1, <i32 -65536, i32 -7, i32 -7, i32 -65536>
  %4 = mul <4 x i32> %a1, %3
  %5 = call <4 x i32> @llvm.aarch64.neon.sabd.v4i32(<4 x i32> %2, <4 x i32> %4)
  %6 = shufflevector <4 x i32> %5, <4 x i32> undef, <4 x i32> <i32 0, i32 0, i32 3, i32 3>
  ret <4 x i32> %6
}

define <4 x i16> @trunc_abdu_foldable(<4 x i16> %a, <4 x i16> %b) {
; CHECK-SD-LABEL: trunc_abdu_foldable:
; CHECK-SD:       // %bb.0:
; CHECK-SD-NEXT:    uabd v0.4h, v0.4h, v1.4h
; CHECK-SD-NEXT:    ret
;
; CHECK-GI-LABEL: trunc_abdu_foldable:
; CHECK-GI:       // %bb.0:
; CHECK-GI-NEXT:    ushll v0.4s, v0.4h, #0
; CHECK-GI-NEXT:    ushll v1.4s, v1.4h, #0
; CHECK-GI-NEXT:    uabd v0.4s, v0.4s, v1.4s
; CHECK-GI-NEXT:    xtn v0.4h, v0.4s
; CHECK-GI-NEXT:    ret
  %ext_a = zext <4 x i16> %a to <4 x i32>
  %ext_b = zext <4 x i16> %b to <4 x i32>
  %abd = call <4 x i32> @llvm.aarch64.neon.uabd.v4i32(<4 x i32> %ext_a, <4 x i32> %ext_b)
  %trunc = trunc <4 x i32> %abd to <4 x i16>
  ret <4 x i16> %trunc
}

define <4 x i16> @trunc_abds_foldable(<4 x i16> %a, <4 x i16> %b) {
; CHECK-SD-LABEL: trunc_abds_foldable:
; CHECK-SD:       // %bb.0:
; CHECK-SD-NEXT:    sabd v0.4h, v0.4h, v1.4h
; CHECK-SD-NEXT:    ret
;
; CHECK-GI-LABEL: trunc_abds_foldable:
; CHECK-GI:       // %bb.0:
; CHECK-GI-NEXT:    sshll v0.4s, v0.4h, #0
; CHECK-GI-NEXT:    sshll v1.4s, v1.4h, #0
; CHECK-GI-NEXT:    sabd v0.4s, v0.4s, v1.4s
; CHECK-GI-NEXT:    xtn v0.4h, v0.4s
; CHECK-GI-NEXT:    ret
  %a32 = sext <4 x i16> %a to <4 x i32>
  %b32 = sext <4 x i16> %b to <4 x i32>
  %abd32 = call <4 x i32> @llvm.aarch64.neon.sabd.v4i32(<4 x i32> %a32, <4 x i32> %b32)
  %res16 = trunc <4 x i32> %abd32 to <4 x i16>
  ret <4 x i16> %res16
}

define <4 x i16> @trunc_abdu_not_foldable(<4 x i16> %a, <4 x i32> %b) {
; CHECK-LABEL: trunc_abdu_not_foldable:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ushll v0.4s, v0.4h, #0
; CHECK-NEXT:    uabd v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    xtn v0.4h, v0.4s
; CHECK-NEXT:    ret
  %ext_a = zext <4 x i16> %a to <4 x i32>
  %abd = call <4 x i32> @llvm.aarch64.neon.uabd.v4i32(<4 x i32> %ext_a, <4 x i32> %b)
  %trunc = trunc <4 x i32> %abd to <4 x i16>
  ret <4 x i16> %trunc
}

define <4 x i16> @truncate_abds_testcase1(<4 x i16> %a, <4 x i32> %b) {
; CHECK-LABEL: truncate_abds_testcase1:
; CHECK:       // %bb.0:
; CHECK-NEXT:    sshll v0.4s, v0.4h, #0
; CHECK-NEXT:    sabd v0.4s, v0.4s, v1.4s
; CHECK-NEXT:    xtn v0.4h, v0.4s
; CHECK-NEXT:    ret
  %a32 = sext <4 x i16> %a to <4 x i32>
  %abd32 = call <4 x i32> @llvm.aarch64.neon.sabd.v4i32(<4 x i32> %a32, <4 x i32> %b)
  %res16 = trunc <4 x i32> %abd32 to <4 x i16>
  ret <4 x i16> %res16
}