clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq.c

   1 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
   2 // REQUIRES: aarch64-registered-target
   3 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s
   4 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK
   5 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s
   6 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK
   7 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -S -disable-O0-optnone -Werror -o /dev/null %s
   8 #include <arm_sve.h>
   9
  10 #ifdef SVE_OVERLOADED_FORMS
  11 // A simple used,unused... macro, long enough to represent any SVE builtin.
  12 #define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
  13 #else
  14 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
  15 #endif
  16
  17 // CHECK-LABEL: @test_svdupq_lane_s8(
  18 // CHECK-NEXT:  entry:
  19 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], i64 [[INDEX:%.*]])
  20 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
  21 //
  22 // CPP-CHECK-LABEL: @_Z19test_svdupq_lane_s8u10__SVInt8_tm(
  23 // CPP-CHECK-NEXT:  entry:
  24 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], i64 [[INDEX:%.*]])
  25 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
  26 //
  27 svint8_t test_svdupq_lane_s8(svint8_t data, uint64_t index)
  28 {
  29   return SVE_ACLE_FUNC(svdupq_lane,_s8,,)(data, index);
  30 }
  31
  32 // CHECK-LABEL: @test_svdupq_lane_s16(
  33 // CHECK-NEXT:  entry:
  34 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], i64 [[INDEX:%.*]])
  35 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
  36 //
  37 // CPP-CHECK-LABEL: @_Z20test_svdupq_lane_s16u11__SVInt16_tm(
  38 // CPP-CHECK-NEXT:  entry:
  39 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], i64 [[INDEX:%.*]])
  40 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
  41 //
  42 svint16_t test_svdupq_lane_s16(svint16_t data, uint64_t index)
  43 {
  44   return SVE_ACLE_FUNC(svdupq_lane,_s16,,)(data, index);
  45 }
  46
  47 // CHECK-LABEL: @test_svdupq_lane_s32(
  48 // CHECK-NEXT:  entry:
  49 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], i64 [[INDEX:%.*]])
  50 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
  51 //
  52 // CPP-CHECK-LABEL: @_Z20test_svdupq_lane_s32u11__SVInt32_tm(
  53 // CPP-CHECK-NEXT:  entry:
  54 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], i64 [[INDEX:%.*]])
  55 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
  56 //
  57 svint32_t test_svdupq_lane_s32(svint32_t data, uint64_t index)
  58 {
  59   return SVE_ACLE_FUNC(svdupq_lane,_s32,,)(data, index);
  60 }
  61
  62 // CHECK-LABEL: @test_svdupq_lane_s64(
  63 // CHECK-NEXT:  entry:
  64 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], i64 [[INDEX:%.*]])
  65 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
  66 //
  67 // CPP-CHECK-LABEL: @_Z20test_svdupq_lane_s64u11__SVInt64_tm(
  68 // CPP-CHECK-NEXT:  entry:
  69 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], i64 [[INDEX:%.*]])
  70 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
  71 //
  72 svint64_t test_svdupq_lane_s64(svint64_t data, uint64_t index)
  73 {
  74   return SVE_ACLE_FUNC(svdupq_lane,_s64,,)(data, index);
  75 }
  76
  77 // CHECK-LABEL: @test_svdupq_lane_u8(
  78 // CHECK-NEXT:  entry:
  79 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], i64 [[INDEX:%.*]])
  80 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
  81 //
  82 // CPP-CHECK-LABEL: @_Z19test_svdupq_lane_u8u11__SVUint8_tm(
  83 // CPP-CHECK-NEXT:  entry:
  84 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], i64 [[INDEX:%.*]])
  85 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
  86 //
  87 svuint8_t test_svdupq_lane_u8(svuint8_t data, uint64_t index)
  88 {
  89   return SVE_ACLE_FUNC(svdupq_lane,_u8,,)(data, index);
  90 }
  91
  92 // CHECK-LABEL: @test_svdupq_lane_u16(
  93 // CHECK-NEXT:  entry:
  94 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], i64 [[INDEX:%.*]])
  95 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
  96 //
  97 // CPP-CHECK-LABEL: @_Z20test_svdupq_lane_u16u12__SVUint16_tm(
  98 // CPP-CHECK-NEXT:  entry:
  99 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], i64 [[INDEX:%.*]])
 100 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 101 //
 102 svuint16_t test_svdupq_lane_u16(svuint16_t data, uint64_t index)
 103 {
 104   return SVE_ACLE_FUNC(svdupq_lane,_u16,,)(data, index);
 105 }
 106
 107 // CHECK-LABEL: @test_svdupq_lane_u32(
 108 // CHECK-NEXT:  entry:
 109 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], i64 [[INDEX:%.*]])
 110 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 111 //
 112 // CPP-CHECK-LABEL: @_Z20test_svdupq_lane_u32u12__SVUint32_tm(
 113 // CPP-CHECK-NEXT:  entry:
 114 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], i64 [[INDEX:%.*]])
 115 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 116 //
 117 svuint32_t test_svdupq_lane_u32(svuint32_t data, uint64_t index)
 118 {
 119   return SVE_ACLE_FUNC(svdupq_lane,_u32,,)(data, index);
 120 }
 121
 122 // CHECK-LABEL: @test_svdupq_lane_u64(
 123 // CHECK-NEXT:  entry:
 124 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], i64 [[INDEX:%.*]])
 125 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 126 //
 127 // CPP-CHECK-LABEL: @_Z20test_svdupq_lane_u64u12__SVUint64_tm(
 128 // CPP-CHECK-NEXT:  entry:
 129 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], i64 [[INDEX:%.*]])
 130 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 131 //
 132 svuint64_t test_svdupq_lane_u64(svuint64_t data, uint64_t index)
 133 {
 134   return SVE_ACLE_FUNC(svdupq_lane,_u64,,)(data, index);
 135 }
 136
 137 // CHECK-LABEL: @test_svdupq_lane_f16(
 138 // CHECK-NEXT:  entry:
 139 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.dupq.lane.nxv8f16(<vscale x 8 x half> [[DATA:%.*]], i64 [[INDEX:%.*]])
 140 // CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 141 //
 142 // CPP-CHECK-LABEL: @_Z20test_svdupq_lane_f16u13__SVFloat16_tm(
 143 // CPP-CHECK-NEXT:  entry:
 144 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.dupq.lane.nxv8f16(<vscale x 8 x half> [[DATA:%.*]], i64 [[INDEX:%.*]])
 145 // CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 146 //
 147 svfloat16_t test_svdupq_lane_f16(svfloat16_t data, uint64_t index)
 148 {
 149   return SVE_ACLE_FUNC(svdupq_lane,_f16,,)(data, index);
 150 }
 151
 152 // CHECK-LABEL: @test_svdupq_lane_f32(
 153 // CHECK-NEXT:  entry:
 154 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.dupq.lane.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], i64 [[INDEX:%.*]])
 155 // CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 156 //
 157 // CPP-CHECK-LABEL: @_Z20test_svdupq_lane_f32u13__SVFloat32_tm(
 158 // CPP-CHECK-NEXT:  entry:
 159 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.dupq.lane.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], i64 [[INDEX:%.*]])
 160 // CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 161 //
 162 svfloat32_t test_svdupq_lane_f32(svfloat32_t data, uint64_t index)
 163 {
 164   return SVE_ACLE_FUNC(svdupq_lane,_f32,,)(data, index);
 165 }
 166
 167 // CHECK-LABEL: @test_svdupq_lane_f64(
 168 // CHECK-NEXT:  entry:
 169 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.dupq.lane.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], i64 [[INDEX:%.*]])
 170 // CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 171 //
 172 // CPP-CHECK-LABEL: @_Z20test_svdupq_lane_f64u13__SVFloat64_tm(
 173 // CPP-CHECK-NEXT:  entry:
 174 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.dupq.lane.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], i64 [[INDEX:%.*]])
 175 // CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 176 //
 177 svfloat64_t test_svdupq_lane_f64(svfloat64_t data, uint64_t index)
 178 {
 179   return SVE_ACLE_FUNC(svdupq_lane,_f64,,)(data, index);
 180 }
 181
 182 // CHECK-LABEL: @test_svdupq_n_s8(
 183 // CHECK-NEXT:  entry:
 184 // CHECK-NEXT:    [[TMP0:%.*]] = insertelement <16 x i8> poison, i8 [[X0:%.*]], i64 0
 185 // CHECK-NEXT:    [[TMP1:%.*]] = insertelement <16 x i8> [[TMP0]], i8 [[X1:%.*]], i64 1
 186 // CHECK-NEXT:    [[TMP2:%.*]] = insertelement <16 x i8> [[TMP1]], i8 [[X2:%.*]], i64 2
 187 // CHECK-NEXT:    [[TMP3:%.*]] = insertelement <16 x i8> [[TMP2]], i8 [[X3:%.*]], i64 3
 188 // CHECK-NEXT:    [[TMP4:%.*]] = insertelement <16 x i8> [[TMP3]], i8 [[X4:%.*]], i64 4
 189 // CHECK-NEXT:    [[TMP5:%.*]] = insertelement <16 x i8> [[TMP4]], i8 [[X5:%.*]], i64 5
 190 // CHECK-NEXT:    [[TMP6:%.*]] = insertelement <16 x i8> [[TMP5]], i8 [[X6:%.*]], i64 6
 191 // CHECK-NEXT:    [[TMP7:%.*]] = insertelement <16 x i8> [[TMP6]], i8 [[X7:%.*]], i64 7
 192 // CHECK-NEXT:    [[TMP8:%.*]] = insertelement <16 x i8> [[TMP7]], i8 [[X8:%.*]], i64 8
 193 // CHECK-NEXT:    [[TMP9:%.*]] = insertelement <16 x i8> [[TMP8]], i8 [[X9:%.*]], i64 9
 194 // CHECK-NEXT:    [[TMP10:%.*]] = insertelement <16 x i8> [[TMP9]], i8 [[X10:%.*]], i64 10
 195 // CHECK-NEXT:    [[TMP11:%.*]] = insertelement <16 x i8> [[TMP10]], i8 [[X11:%.*]], i64 11
 196 // CHECK-NEXT:    [[TMP12:%.*]] = insertelement <16 x i8> [[TMP11]], i8 [[X12:%.*]], i64 12
 197 // CHECK-NEXT:    [[TMP13:%.*]] = insertelement <16 x i8> [[TMP12]], i8 [[X13:%.*]], i64 13
 198 // CHECK-NEXT:    [[TMP14:%.*]] = insertelement <16 x i8> [[TMP13]], i8 [[X14:%.*]], i64 14
 199 // CHECK-NEXT:    [[TMP15:%.*]] = insertelement <16 x i8> [[TMP14]], i8 [[X15:%.*]], i64 15
 200 // CHECK-NEXT:    [[TMP16:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.insert.nxv16i8.v16i8(<vscale x 16 x i8> poison, <16 x i8> [[TMP15]], i64 0)
 201 // CHECK-NEXT:    [[TMP17:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[TMP16]], i64 0)
 202 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP17]]
 203 //
 204 // CPP-CHECK-LABEL: @_Z16test_svdupq_n_s8aaaaaaaaaaaaaaaa(
 205 // CPP-CHECK-NEXT:  entry:
 206 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <16 x i8> poison, i8 [[X0:%.*]], i64 0
 207 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <16 x i8> [[TMP0]], i8 [[X1:%.*]], i64 1
 208 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <16 x i8> [[TMP1]], i8 [[X2:%.*]], i64 2
 209 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <16 x i8> [[TMP2]], i8 [[X3:%.*]], i64 3
 210 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = insertelement <16 x i8> [[TMP3]], i8 [[X4:%.*]], i64 4
 211 // CPP-CHECK-NEXT:    [[TMP5:%.*]] = insertelement <16 x i8> [[TMP4]], i8 [[X5:%.*]], i64 5
 212 // CPP-CHECK-NEXT:    [[TMP6:%.*]] = insertelement <16 x i8> [[TMP5]], i8 [[X6:%.*]], i64 6
 213 // CPP-CHECK-NEXT:    [[TMP7:%.*]] = insertelement <16 x i8> [[TMP6]], i8 [[X7:%.*]], i64 7
 214 // CPP-CHECK-NEXT:    [[TMP8:%.*]] = insertelement <16 x i8> [[TMP7]], i8 [[X8:%.*]], i64 8
 215 // CPP-CHECK-NEXT:    [[TMP9:%.*]] = insertelement <16 x i8> [[TMP8]], i8 [[X9:%.*]], i64 9
 216 // CPP-CHECK-NEXT:    [[TMP10:%.*]] = insertelement <16 x i8> [[TMP9]], i8 [[X10:%.*]], i64 10
 217 // CPP-CHECK-NEXT:    [[TMP11:%.*]] = insertelement <16 x i8> [[TMP10]], i8 [[X11:%.*]], i64 11
 218 // CPP-CHECK-NEXT:    [[TMP12:%.*]] = insertelement <16 x i8> [[TMP11]], i8 [[X12:%.*]], i64 12
 219 // CPP-CHECK-NEXT:    [[TMP13:%.*]] = insertelement <16 x i8> [[TMP12]], i8 [[X13:%.*]], i64 13
 220 // CPP-CHECK-NEXT:    [[TMP14:%.*]] = insertelement <16 x i8> [[TMP13]], i8 [[X14:%.*]], i64 14
 221 // CPP-CHECK-NEXT:    [[TMP15:%.*]] = insertelement <16 x i8> [[TMP14]], i8 [[X15:%.*]], i64 15
 222 // CPP-CHECK-NEXT:    [[TMP16:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.insert.nxv16i8.v16i8(<vscale x 16 x i8> poison, <16 x i8> [[TMP15]], i64 0)
 223 // CPP-CHECK-NEXT:    [[TMP17:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[TMP16]], i64 0)
 224 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP17]]
 225 //
 226 svint8_t test_svdupq_n_s8(int8_t x0, int8_t x1, int8_t x2, int8_t x3,
 227                           int8_t x4, int8_t x5, int8_t x6, int8_t x7,
 228                           int8_t x8, int8_t x9, int8_t x10, int8_t x11,
 229                           int8_t x12, int8_t x13, int8_t x14, int8_t x15)
 230 {
 231   // <assume other insertelement>
 232   return SVE_ACLE_FUNC(svdupq,_n,_s8,)(x0, x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, x11, x12, x13, x14, x15);
 233 }
 234
 235 // CHECK-LABEL: @test_svdupq_n_s16(
 236 // CHECK-NEXT:  entry:
 237 // CHECK-NEXT:    [[TMP0:%.*]] = insertelement <8 x i16> poison, i16 [[X0:%.*]], i64 0
 238 // CHECK-NEXT:    [[TMP1:%.*]] = insertelement <8 x i16> [[TMP0]], i16 [[X1:%.*]], i64 1
 239 // CHECK-NEXT:    [[TMP2:%.*]] = insertelement <8 x i16> [[TMP1]], i16 [[X2:%.*]], i64 2
 240 // CHECK-NEXT:    [[TMP3:%.*]] = insertelement <8 x i16> [[TMP2]], i16 [[X3:%.*]], i64 3
 241 // CHECK-NEXT:    [[TMP4:%.*]] = insertelement <8 x i16> [[TMP3]], i16 [[X4:%.*]], i64 4
 242 // CHECK-NEXT:    [[TMP5:%.*]] = insertelement <8 x i16> [[TMP4]], i16 [[X5:%.*]], i64 5
 243 // CHECK-NEXT:    [[TMP6:%.*]] = insertelement <8 x i16> [[TMP5]], i16 [[X6:%.*]], i64 6
 244 // CHECK-NEXT:    [[TMP7:%.*]] = insertelement <8 x i16> [[TMP6]], i16 [[X7:%.*]], i64 7
 245 // CHECK-NEXT:    [[TMP8:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.insert.nxv8i16.v8i16(<vscale x 8 x i16> poison, <8 x i16> [[TMP7]], i64 0)
 246 // CHECK-NEXT:    [[TMP9:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[TMP8]], i64 0)
 247 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP9]]
 248 //
 249 // CPP-CHECK-LABEL: @_Z17test_svdupq_n_s16ssssssss(
 250 // CPP-CHECK-NEXT:  entry:
 251 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <8 x i16> poison, i16 [[X0:%.*]], i64 0
 252 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <8 x i16> [[TMP0]], i16 [[X1:%.*]], i64 1
 253 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <8 x i16> [[TMP1]], i16 [[X2:%.*]], i64 2
 254 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <8 x i16> [[TMP2]], i16 [[X3:%.*]], i64 3
 255 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = insertelement <8 x i16> [[TMP3]], i16 [[X4:%.*]], i64 4
 256 // CPP-CHECK-NEXT:    [[TMP5:%.*]] = insertelement <8 x i16> [[TMP4]], i16 [[X5:%.*]], i64 5
 257 // CPP-CHECK-NEXT:    [[TMP6:%.*]] = insertelement <8 x i16> [[TMP5]], i16 [[X6:%.*]], i64 6
 258 // CPP-CHECK-NEXT:    [[TMP7:%.*]] = insertelement <8 x i16> [[TMP6]], i16 [[X7:%.*]], i64 7
 259 // CPP-CHECK-NEXT:    [[TMP8:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.insert.nxv8i16.v8i16(<vscale x 8 x i16> poison, <8 x i16> [[TMP7]], i64 0)
 260 // CPP-CHECK-NEXT:    [[TMP9:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[TMP8]], i64 0)
 261 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP9]]
 262 //
 263 svint16_t test_svdupq_n_s16(int16_t x0, int16_t x1, int16_t x2, int16_t x3,
 264                             int16_t x4, int16_t x5, int16_t x6, int16_t x7)
 265 {
 266   // <assume other insertelement>
 267   return SVE_ACLE_FUNC(svdupq,_n,_s16,)(x0, x1, x2, x3, x4, x5, x6, x7);
 268 }
 269
 270 // CHECK-LABEL: @test_svdupq_n_s32(
 271 // CHECK-NEXT:  entry:
 272 // CHECK-NEXT:    [[TMP0:%.*]] = insertelement <4 x i32> poison, i32 [[X0:%.*]], i64 0
 273 // CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x i32> [[TMP0]], i32 [[X1:%.*]], i64 1
 274 // CHECK-NEXT:    [[TMP2:%.*]] = insertelement <4 x i32> [[TMP1]], i32 [[X2:%.*]], i64 2
 275 // CHECK-NEXT:    [[TMP3:%.*]] = insertelement <4 x i32> [[TMP2]], i32 [[X3:%.*]], i64 3
 276 // CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.insert.nxv4i32.v4i32(<vscale x 4 x i32> poison, <4 x i32> [[TMP3]], i64 0)
 277 // CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[TMP4]], i64 0)
 278 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP5]]
 279 //
 280 // CPP-CHECK-LABEL: @_Z17test_svdupq_n_s32iiii(
 281 // CPP-CHECK-NEXT:  entry:
 282 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <4 x i32> poison, i32 [[X0:%.*]], i64 0
 283 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x i32> [[TMP0]], i32 [[X1:%.*]], i64 1
 284 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <4 x i32> [[TMP1]], i32 [[X2:%.*]], i64 2
 285 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <4 x i32> [[TMP2]], i32 [[X3:%.*]], i64 3
 286 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.insert.nxv4i32.v4i32(<vscale x 4 x i32> poison, <4 x i32> [[TMP3]], i64 0)
 287 // CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[TMP4]], i64 0)
 288 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP5]]
 289 //
 290 svint32_t test_svdupq_n_s32(int32_t x0, int32_t x1, int32_t x2, int32_t x3)
 291 {
 292   // <assume other insertelement>
 293   return SVE_ACLE_FUNC(svdupq,_n,_s32,)(x0, x1, x2, x3);
 294 }
 295
 296 // CHECK-LABEL: @test_svdupq_n_s64(
 297 // CHECK-NEXT:  entry:
 298 // CHECK-NEXT:    [[TMP0:%.*]] = insertelement <2 x i64> poison, i64 [[X0:%.*]], i64 0
 299 // CHECK-NEXT:    [[TMP1:%.*]] = insertelement <2 x i64> [[TMP0]], i64 [[X1:%.*]], i64 1
 300 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.insert.nxv2i64.v2i64(<vscale x 2 x i64> poison, <2 x i64> [[TMP1]], i64 0)
 301 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[TMP2]], i64 0)
 302 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
 303 //
 304 // CPP-CHECK-LABEL: @_Z17test_svdupq_n_s64ll(
 305 // CPP-CHECK-NEXT:  entry:
 306 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <2 x i64> poison, i64 [[X0:%.*]], i64 0
 307 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <2 x i64> [[TMP0]], i64 [[X1:%.*]], i64 1
 308 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.insert.nxv2i64.v2i64(<vscale x 2 x i64> poison, <2 x i64> [[TMP1]], i64 0)
 309 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[TMP2]], i64 0)
 310 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
 311 //
 312 svint64_t test_svdupq_n_s64(int64_t x0, int64_t x1)
 313 {
 314   return SVE_ACLE_FUNC(svdupq,_n,_s64,)(x0, x1);
 315 }
 316
 317 // CHECK-LABEL: @test_svdupq_n_u8(
 318 // CHECK-NEXT:  entry:
 319 // CHECK-NEXT:    [[TMP0:%.*]] = insertelement <16 x i8> poison, i8 [[X0:%.*]], i64 0
 320 // CHECK-NEXT:    [[TMP1:%.*]] = insertelement <16 x i8> [[TMP0]], i8 [[X1:%.*]], i64 1
 321 // CHECK-NEXT:    [[TMP2:%.*]] = insertelement <16 x i8> [[TMP1]], i8 [[X2:%.*]], i64 2
 322 // CHECK-NEXT:    [[TMP3:%.*]] = insertelement <16 x i8> [[TMP2]], i8 [[X3:%.*]], i64 3
 323 // CHECK-NEXT:    [[TMP4:%.*]] = insertelement <16 x i8> [[TMP3]], i8 [[X4:%.*]], i64 4
 324 // CHECK-NEXT:    [[TMP5:%.*]] = insertelement <16 x i8> [[TMP4]], i8 [[X5:%.*]], i64 5
 325 // CHECK-NEXT:    [[TMP6:%.*]] = insertelement <16 x i8> [[TMP5]], i8 [[X6:%.*]], i64 6
 326 // CHECK-NEXT:    [[TMP7:%.*]] = insertelement <16 x i8> [[TMP6]], i8 [[X7:%.*]], i64 7
 327 // CHECK-NEXT:    [[TMP8:%.*]] = insertelement <16 x i8> [[TMP7]], i8 [[X8:%.*]], i64 8
 328 // CHECK-NEXT:    [[TMP9:%.*]] = insertelement <16 x i8> [[TMP8]], i8 [[X9:%.*]], i64 9
 329 // CHECK-NEXT:    [[TMP10:%.*]] = insertelement <16 x i8> [[TMP9]], i8 [[X10:%.*]], i64 10
 330 // CHECK-NEXT:    [[TMP11:%.*]] = insertelement <16 x i8> [[TMP10]], i8 [[X11:%.*]], i64 11
 331 // CHECK-NEXT:    [[TMP12:%.*]] = insertelement <16 x i8> [[TMP11]], i8 [[X12:%.*]], i64 12
 332 // CHECK-NEXT:    [[TMP13:%.*]] = insertelement <16 x i8> [[TMP12]], i8 [[X13:%.*]], i64 13
 333 // CHECK-NEXT:    [[TMP14:%.*]] = insertelement <16 x i8> [[TMP13]], i8 [[X14:%.*]], i64 14
 334 // CHECK-NEXT:    [[TMP15:%.*]] = insertelement <16 x i8> [[TMP14]], i8 [[X15:%.*]], i64 15
 335 // CHECK-NEXT:    [[TMP16:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.insert.nxv16i8.v16i8(<vscale x 16 x i8> poison, <16 x i8> [[TMP15]], i64 0)
 336 // CHECK-NEXT:    [[TMP17:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[TMP16]], i64 0)
 337 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP17]]
 338 //
 339 // CPP-CHECK-LABEL: @_Z16test_svdupq_n_u8hhhhhhhhhhhhhhhh(
 340 // CPP-CHECK-NEXT:  entry:
 341 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <16 x i8> poison, i8 [[X0:%.*]], i64 0
 342 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <16 x i8> [[TMP0]], i8 [[X1:%.*]], i64 1
 343 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <16 x i8> [[TMP1]], i8 [[X2:%.*]], i64 2
 344 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <16 x i8> [[TMP2]], i8 [[X3:%.*]], i64 3
 345 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = insertelement <16 x i8> [[TMP3]], i8 [[X4:%.*]], i64 4
 346 // CPP-CHECK-NEXT:    [[TMP5:%.*]] = insertelement <16 x i8> [[TMP4]], i8 [[X5:%.*]], i64 5
 347 // CPP-CHECK-NEXT:    [[TMP6:%.*]] = insertelement <16 x i8> [[TMP5]], i8 [[X6:%.*]], i64 6
 348 // CPP-CHECK-NEXT:    [[TMP7:%.*]] = insertelement <16 x i8> [[TMP6]], i8 [[X7:%.*]], i64 7
 349 // CPP-CHECK-NEXT:    [[TMP8:%.*]] = insertelement <16 x i8> [[TMP7]], i8 [[X8:%.*]], i64 8
 350 // CPP-CHECK-NEXT:    [[TMP9:%.*]] = insertelement <16 x i8> [[TMP8]], i8 [[X9:%.*]], i64 9
 351 // CPP-CHECK-NEXT:    [[TMP10:%.*]] = insertelement <16 x i8> [[TMP9]], i8 [[X10:%.*]], i64 10
 352 // CPP-CHECK-NEXT:    [[TMP11:%.*]] = insertelement <16 x i8> [[TMP10]], i8 [[X11:%.*]], i64 11
 353 // CPP-CHECK-NEXT:    [[TMP12:%.*]] = insertelement <16 x i8> [[TMP11]], i8 [[X12:%.*]], i64 12
 354 // CPP-CHECK-NEXT:    [[TMP13:%.*]] = insertelement <16 x i8> [[TMP12]], i8 [[X13:%.*]], i64 13
 355 // CPP-CHECK-NEXT:    [[TMP14:%.*]] = insertelement <16 x i8> [[TMP13]], i8 [[X14:%.*]], i64 14
 356 // CPP-CHECK-NEXT:    [[TMP15:%.*]] = insertelement <16 x i8> [[TMP14]], i8 [[X15:%.*]], i64 15
 357 // CPP-CHECK-NEXT:    [[TMP16:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.insert.nxv16i8.v16i8(<vscale x 16 x i8> poison, <16 x i8> [[TMP15]], i64 0)
 358 // CPP-CHECK-NEXT:    [[TMP17:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[TMP16]], i64 0)
 359 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP17]]
 360 //
 361 svuint8_t test_svdupq_n_u8(uint8_t x0, uint8_t x1, uint8_t x2, uint8_t x3,
 362                            uint8_t x4, uint8_t x5, uint8_t x6, uint8_t x7,
 363                            uint8_t x8, uint8_t x9, uint8_t x10, uint8_t x11,
 364                            uint8_t x12, uint8_t x13, uint8_t x14, uint8_t x15)
 365 {
 366   // <assume other insertelement>
 367   return SVE_ACLE_FUNC(svdupq,_n,_u8,)(x0, x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, x11, x12, x13, x14, x15);
 368 }
 369
 370 // CHECK-LABEL: @test_svdupq_n_u16(
 371 // CHECK-NEXT:  entry:
 372 // CHECK-NEXT:    [[TMP0:%.*]] = insertelement <8 x i16> poison, i16 [[X0:%.*]], i64 0
 373 // CHECK-NEXT:    [[TMP1:%.*]] = insertelement <8 x i16> [[TMP0]], i16 [[X1:%.*]], i64 1
 374 // CHECK-NEXT:    [[TMP2:%.*]] = insertelement <8 x i16> [[TMP1]], i16 [[X2:%.*]], i64 2
 375 // CHECK-NEXT:    [[TMP3:%.*]] = insertelement <8 x i16> [[TMP2]], i16 [[X3:%.*]], i64 3
 376 // CHECK-NEXT:    [[TMP4:%.*]] = insertelement <8 x i16> [[TMP3]], i16 [[X4:%.*]], i64 4
 377 // CHECK-NEXT:    [[TMP5:%.*]] = insertelement <8 x i16> [[TMP4]], i16 [[X5:%.*]], i64 5
 378 // CHECK-NEXT:    [[TMP6:%.*]] = insertelement <8 x i16> [[TMP5]], i16 [[X6:%.*]], i64 6
 379 // CHECK-NEXT:    [[TMP7:%.*]] = insertelement <8 x i16> [[TMP6]], i16 [[X7:%.*]], i64 7
 380 // CHECK-NEXT:    [[TMP8:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.insert.nxv8i16.v8i16(<vscale x 8 x i16> poison, <8 x i16> [[TMP7]], i64 0)
 381 // CHECK-NEXT:    [[TMP9:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[TMP8]], i64 0)
 382 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP9]]
 383 //
 384 // CPP-CHECK-LABEL: @_Z17test_svdupq_n_u16tttttttt(
 385 // CPP-CHECK-NEXT:  entry:
 386 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <8 x i16> poison, i16 [[X0:%.*]], i64 0
 387 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <8 x i16> [[TMP0]], i16 [[X1:%.*]], i64 1
 388 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <8 x i16> [[TMP1]], i16 [[X2:%.*]], i64 2
 389 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <8 x i16> [[TMP2]], i16 [[X3:%.*]], i64 3
 390 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = insertelement <8 x i16> [[TMP3]], i16 [[X4:%.*]], i64 4
 391 // CPP-CHECK-NEXT:    [[TMP5:%.*]] = insertelement <8 x i16> [[TMP4]], i16 [[X5:%.*]], i64 5
 392 // CPP-CHECK-NEXT:    [[TMP6:%.*]] = insertelement <8 x i16> [[TMP5]], i16 [[X6:%.*]], i64 6
 393 // CPP-CHECK-NEXT:    [[TMP7:%.*]] = insertelement <8 x i16> [[TMP6]], i16 [[X7:%.*]], i64 7
 394 // CPP-CHECK-NEXT:    [[TMP8:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.insert.nxv8i16.v8i16(<vscale x 8 x i16> poison, <8 x i16> [[TMP7]], i64 0)
 395 // CPP-CHECK-NEXT:    [[TMP9:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[TMP8]], i64 0)
 396 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP9]]
 397 //
 398 svuint16_t test_svdupq_n_u16(uint16_t x0, uint16_t x1, uint16_t x2, uint16_t x3,
 399                              uint16_t x4, uint16_t x5, uint16_t x6, uint16_t x7)
 400 {
 401   // <assume other insertelement>
 402   return SVE_ACLE_FUNC(svdupq,_n,_u16,)(x0, x1, x2, x3, x4, x5, x6, x7);
 403 }
 404
 405 // CHECK-LABEL: @test_svdupq_n_u32(
 406 // CHECK-NEXT:  entry:
 407 // CHECK-NEXT:    [[TMP0:%.*]] = insertelement <4 x i32> poison, i32 [[X0:%.*]], i64 0
 408 // CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x i32> [[TMP0]], i32 [[X1:%.*]], i64 1
 409 // CHECK-NEXT:    [[TMP2:%.*]] = insertelement <4 x i32> [[TMP1]], i32 [[X2:%.*]], i64 2
 410 // CHECK-NEXT:    [[TMP3:%.*]] = insertelement <4 x i32> [[TMP2]], i32 [[X3:%.*]], i64 3
 411 // CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.insert.nxv4i32.v4i32(<vscale x 4 x i32> poison, <4 x i32> [[TMP3]], i64 0)
 412 // CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[TMP4]], i64 0)
 413 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP5]]
 414 //
 415 // CPP-CHECK-LABEL: @_Z17test_svdupq_n_u32jjjj(
 416 // CPP-CHECK-NEXT:  entry:
 417 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <4 x i32> poison, i32 [[X0:%.*]], i64 0
 418 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x i32> [[TMP0]], i32 [[X1:%.*]], i64 1
 419 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <4 x i32> [[TMP1]], i32 [[X2:%.*]], i64 2
 420 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <4 x i32> [[TMP2]], i32 [[X3:%.*]], i64 3
 421 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.insert.nxv4i32.v4i32(<vscale x 4 x i32> poison, <4 x i32> [[TMP3]], i64 0)
 422 // CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[TMP4]], i64 0)
 423 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP5]]
 424 //
 425 svuint32_t test_svdupq_n_u32(uint32_t x0, uint32_t x1, uint32_t x2, uint32_t x3)
 426 {
 427   // <assume other insertelement>
 428   return SVE_ACLE_FUNC(svdupq,_n,_u32,)(x0, x1, x2, x3);
 429 }
 430
 431 // CHECK-LABEL: @test_svdupq_n_u64(
 432 // CHECK-NEXT:  entry:
 433 // CHECK-NEXT:    [[TMP0:%.*]] = insertelement <2 x i64> poison, i64 [[X0:%.*]], i64 0
 434 // CHECK-NEXT:    [[TMP1:%.*]] = insertelement <2 x i64> [[TMP0]], i64 [[X1:%.*]], i64 1
 435 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.insert.nxv2i64.v2i64(<vscale x 2 x i64> poison, <2 x i64> [[TMP1]], i64 0)
 436 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[TMP2]], i64 0)
 437 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
 438 //
 439 // CPP-CHECK-LABEL: @_Z17test_svdupq_n_u64mm(
 440 // CPP-CHECK-NEXT:  entry:
 441 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <2 x i64> poison, i64 [[X0:%.*]], i64 0
 442 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <2 x i64> [[TMP0]], i64 [[X1:%.*]], i64 1
 443 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.insert.nxv2i64.v2i64(<vscale x 2 x i64> poison, <2 x i64> [[TMP1]], i64 0)
 444 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[TMP2]], i64 0)
 445 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
 446 //
 447 svuint64_t test_svdupq_n_u64(uint64_t x0, uint64_t x1)
 448 {
 449   return SVE_ACLE_FUNC(svdupq,_n,_u64,)(x0, x1);
 450 }
 451
 452 // CHECK-LABEL: @test_svdupq_n_f16(
 453 // CHECK-NEXT:  entry:
 454 // CHECK-NEXT:    [[TMP0:%.*]] = insertelement <8 x half> poison, half [[X0:%.*]], i64 0
 455 // CHECK-NEXT:    [[TMP1:%.*]] = insertelement <8 x half> [[TMP0]], half [[X1:%.*]], i64 1
 456 // CHECK-NEXT:    [[TMP2:%.*]] = insertelement <8 x half> [[TMP1]], half [[X2:%.*]], i64 2
 457 // CHECK-NEXT:    [[TMP3:%.*]] = insertelement <8 x half> [[TMP2]], half [[X3:%.*]], i64 3
 458 // CHECK-NEXT:    [[TMP4:%.*]] = insertelement <8 x half> [[TMP3]], half [[X4:%.*]], i64 4
 459 // CHECK-NEXT:    [[TMP5:%.*]] = insertelement <8 x half> [[TMP4]], half [[X5:%.*]], i64 5
 460 // CHECK-NEXT:    [[TMP6:%.*]] = insertelement <8 x half> [[TMP5]], half [[X6:%.*]], i64 6
 461 // CHECK-NEXT:    [[TMP7:%.*]] = insertelement <8 x half> [[TMP6]], half [[X7:%.*]], i64 7
 462 // CHECK-NEXT:    [[TMP8:%.*]] = tail call <vscale x 8 x half> @llvm.vector.insert.nxv8f16.v8f16(<vscale x 8 x half> poison, <8 x half> [[TMP7]], i64 0)
 463 // CHECK-NEXT:    [[TMP9:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.dupq.lane.nxv8f16(<vscale x 8 x half> [[TMP8]], i64 0)
 464 // CHECK-NEXT:    ret <vscale x 8 x half> [[TMP9]]
 465 //
 466 // CPP-CHECK-LABEL: @_Z17test_svdupq_n_f16DhDhDhDhDhDhDhDh(
 467 // CPP-CHECK-NEXT:  entry:
 468 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <8 x half> poison, half [[X0:%.*]], i64 0
 469 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <8 x half> [[TMP0]], half [[X1:%.*]], i64 1
 470 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <8 x half> [[TMP1]], half [[X2:%.*]], i64 2
 471 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <8 x half> [[TMP2]], half [[X3:%.*]], i64 3
 472 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = insertelement <8 x half> [[TMP3]], half [[X4:%.*]], i64 4
 473 // CPP-CHECK-NEXT:    [[TMP5:%.*]] = insertelement <8 x half> [[TMP4]], half [[X5:%.*]], i64 5
 474 // CPP-CHECK-NEXT:    [[TMP6:%.*]] = insertelement <8 x half> [[TMP5]], half [[X6:%.*]], i64 6
 475 // CPP-CHECK-NEXT:    [[TMP7:%.*]] = insertelement <8 x half> [[TMP6]], half [[X7:%.*]], i64 7
 476 // CPP-CHECK-NEXT:    [[TMP8:%.*]] = tail call <vscale x 8 x half> @llvm.vector.insert.nxv8f16.v8f16(<vscale x 8 x half> poison, <8 x half> [[TMP7]], i64 0)
 477 // CPP-CHECK-NEXT:    [[TMP9:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.dupq.lane.nxv8f16(<vscale x 8 x half> [[TMP8]], i64 0)
 478 // CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP9]]
 479 //
 480 svfloat16_t test_svdupq_n_f16(float16_t x0, float16_t x1, float16_t x2, float16_t x3,
 481                               float16_t x4, float16_t x5, float16_t x6, float16_t x7)
 482 {
 483   // <assume other insertelement>
 484   return SVE_ACLE_FUNC(svdupq,_n,_f16,)(x0, x1, x2, x3, x4, x5, x6, x7);
 485 }
 486
 487 // CHECK-LABEL: @test_svdupq_n_f32(
 488 // CHECK-NEXT:  entry:
 489 // CHECK-NEXT:    [[TMP0:%.*]] = insertelement <4 x float> poison, float [[X0:%.*]], i64 0
 490 // CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x float> [[TMP0]], float [[X1:%.*]], i64 1
 491 // CHECK-NEXT:    [[TMP2:%.*]] = insertelement <4 x float> [[TMP1]], float [[X2:%.*]], i64 2
 492 // CHECK-NEXT:    [[TMP3:%.*]] = insertelement <4 x float> [[TMP2]], float [[X3:%.*]], i64 3
 493 // CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.vector.insert.nxv4f32.v4f32(<vscale x 4 x float> poison, <4 x float> [[TMP3]], i64 0)
 494 // CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.dupq.lane.nxv4f32(<vscale x 4 x float> [[TMP4]], i64 0)
 495 // CHECK-NEXT:    ret <vscale x 4 x float> [[TMP5]]
 496 //
 497 // CPP-CHECK-LABEL: @_Z17test_svdupq_n_f32ffff(
 498 // CPP-CHECK-NEXT:  entry:
 499 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <4 x float> poison, float [[X0:%.*]], i64 0
 500 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x float> [[TMP0]], float [[X1:%.*]], i64 1
 501 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <4 x float> [[TMP1]], float [[X2:%.*]], i64 2
 502 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <4 x float> [[TMP2]], float [[X3:%.*]], i64 3
 503 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.vector.insert.nxv4f32.v4f32(<vscale x 4 x float> poison, <4 x float> [[TMP3]], i64 0)
 504 // CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.dupq.lane.nxv4f32(<vscale x 4 x float> [[TMP4]], i64 0)
 505 // CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP5]]
 506 //
 507 svfloat32_t test_svdupq_n_f32(float32_t x0, float32_t x1, float32_t x2, float32_t x3)
 508 {
 509   // <assume other insertelement>
 510   return SVE_ACLE_FUNC(svdupq,_n,_f32,)(x0, x1, x2, x3);
 511 }
 512
 513 // CHECK-LABEL: @test_svdupq_n_f64(
 514 // CHECK-NEXT:  entry:
 515 // CHECK-NEXT:    [[TMP0:%.*]] = insertelement <2 x double> poison, double [[X0:%.*]], i64 0
 516 // CHECK-NEXT:    [[TMP1:%.*]] = insertelement <2 x double> [[TMP0]], double [[X1:%.*]], i64 1
 517 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.vector.insert.nxv2f64.v2f64(<vscale x 2 x double> poison, <2 x double> [[TMP1]], i64 0)
 518 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.dupq.lane.nxv2f64(<vscale x 2 x double> [[TMP2]], i64 0)
 519 // CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
 520 //
 521 // CPP-CHECK-LABEL: @_Z17test_svdupq_n_f64dd(
 522 // CPP-CHECK-NEXT:  entry:
 523 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <2 x double> poison, double [[X0:%.*]], i64 0
 524 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <2 x double> [[TMP0]], double [[X1:%.*]], i64 1
 525 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.vector.insert.nxv2f64.v2f64(<vscale x 2 x double> poison, <2 x double> [[TMP1]], i64 0)
 526 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.dupq.lane.nxv2f64(<vscale x 2 x double> [[TMP2]], i64 0)
 527 // CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
 528 //
 529 svfloat64_t test_svdupq_n_f64(float64_t x0, float64_t x1)
 530 {
 531   return SVE_ACLE_FUNC(svdupq,_n,_f64,)(x0, x1);
 532 }
 533
 534 // CHECK-LABEL: @test_svdupq_n_b8(
 535 // CHECK-NEXT:  entry:
 536 // CHECK-NEXT:    [[FROMBOOL:%.*]] = zext i1 [[X0:%.*]] to i8
 537 // CHECK-NEXT:    [[FROMBOOL1:%.*]] = zext i1 [[X1:%.*]] to i8
 538 // CHECK-NEXT:    [[FROMBOOL2:%.*]] = zext i1 [[X2:%.*]] to i8
 539 // CHECK-NEXT:    [[FROMBOOL3:%.*]] = zext i1 [[X3:%.*]] to i8
 540 // CHECK-NEXT:    [[FROMBOOL4:%.*]] = zext i1 [[X4:%.*]] to i8
 541 // CHECK-NEXT:    [[FROMBOOL5:%.*]] = zext i1 [[X5:%.*]] to i8
 542 // CHECK-NEXT:    [[FROMBOOL6:%.*]] = zext i1 [[X6:%.*]] to i8
 543 // CHECK-NEXT:    [[FROMBOOL7:%.*]] = zext i1 [[X7:%.*]] to i8
 544 // CHECK-NEXT:    [[FROMBOOL8:%.*]] = zext i1 [[X8:%.*]] to i8
 545 // CHECK-NEXT:    [[FROMBOOL9:%.*]] = zext i1 [[X9:%.*]] to i8
 546 // CHECK-NEXT:    [[FROMBOOL10:%.*]] = zext i1 [[X10:%.*]] to i8
 547 // CHECK-NEXT:    [[FROMBOOL11:%.*]] = zext i1 [[X11:%.*]] to i8
 548 // CHECK-NEXT:    [[FROMBOOL12:%.*]] = zext i1 [[X12:%.*]] to i8
 549 // CHECK-NEXT:    [[FROMBOOL13:%.*]] = zext i1 [[X13:%.*]] to i8
 550 // CHECK-NEXT:    [[FROMBOOL14:%.*]] = zext i1 [[X14:%.*]] to i8
 551 // CHECK-NEXT:    [[FROMBOOL15:%.*]] = zext i1 [[X15:%.*]] to i8
 552 // CHECK-NEXT:    [[TOBOOL:%.*]] = trunc i8 [[FROMBOOL]] to i1
 553 // CHECK-NEXT:    [[TOBOOL16:%.*]] = trunc i8 [[FROMBOOL1]] to i1
 554 // CHECK-NEXT:    [[TOBOOL17:%.*]] = trunc i8 [[FROMBOOL2]] to i1
 555 // CHECK-NEXT:    [[TOBOOL18:%.*]] = trunc i8 [[FROMBOOL3]] to i1
 556 // CHECK-NEXT:    [[TOBOOL19:%.*]] = trunc i8 [[FROMBOOL4]] to i1
 557 // CHECK-NEXT:    [[TOBOOL20:%.*]] = trunc i8 [[FROMBOOL5]] to i1
 558 // CHECK-NEXT:    [[TOBOOL21:%.*]] = trunc i8 [[FROMBOOL6]] to i1
 559 // CHECK-NEXT:    [[TOBOOL22:%.*]] = trunc i8 [[FROMBOOL7]] to i1
 560 // CHECK-NEXT:    [[TOBOOL23:%.*]] = trunc i8 [[FROMBOOL8]] to i1
 561 // CHECK-NEXT:    [[TOBOOL24:%.*]] = trunc i8 [[FROMBOOL9]] to i1
 562 // CHECK-NEXT:    [[TOBOOL25:%.*]] = trunc i8 [[FROMBOOL10]] to i1
 563 // CHECK-NEXT:    [[TOBOOL26:%.*]] = trunc i8 [[FROMBOOL11]] to i1
 564 // CHECK-NEXT:    [[TOBOOL27:%.*]] = trunc i8 [[FROMBOOL12]] to i1
 565 // CHECK-NEXT:    [[TOBOOL28:%.*]] = trunc i8 [[FROMBOOL13]] to i1
 566 // CHECK-NEXT:    [[TOBOOL29:%.*]] = trunc i8 [[FROMBOOL14]] to i1
 567 // CHECK-NEXT:    [[TOBOOL30:%.*]] = trunc i8 [[FROMBOOL15]] to i1
 568 // CHECK-NEXT:    [[TMP0:%.*]] = zext i1 [[TOBOOL]] to i8
 569 // CHECK-NEXT:    [[TMP1:%.*]] = zext i1 [[TOBOOL16]] to i8
 570 // CHECK-NEXT:    [[TMP2:%.*]] = zext i1 [[TOBOOL17]] to i8
 571 // CHECK-NEXT:    [[TMP3:%.*]] = zext i1 [[TOBOOL18]] to i8
 572 // CHECK-NEXT:    [[TMP4:%.*]] = zext i1 [[TOBOOL19]] to i8
 573 // CHECK-NEXT:    [[TMP5:%.*]] = zext i1 [[TOBOOL20]] to i8
 574 // CHECK-NEXT:    [[TMP6:%.*]] = zext i1 [[TOBOOL21]] to i8
 575 // CHECK-NEXT:    [[TMP7:%.*]] = zext i1 [[TOBOOL22]] to i8
 576 // CHECK-NEXT:    [[TMP8:%.*]] = zext i1 [[TOBOOL23]] to i8
 577 // CHECK-NEXT:    [[TMP9:%.*]] = zext i1 [[TOBOOL24]] to i8
 578 // CHECK-NEXT:    [[TMP10:%.*]] = zext i1 [[TOBOOL25]] to i8
 579 // CHECK-NEXT:    [[TMP11:%.*]] = zext i1 [[TOBOOL26]] to i8
 580 // CHECK-NEXT:    [[TMP12:%.*]] = zext i1 [[TOBOOL27]] to i8
 581 // CHECK-NEXT:    [[TMP13:%.*]] = zext i1 [[TOBOOL28]] to i8
 582 // CHECK-NEXT:    [[TMP14:%.*]] = zext i1 [[TOBOOL29]] to i8
 583 // CHECK-NEXT:    [[TMP15:%.*]] = zext i1 [[TOBOOL30]] to i8
 584 // CHECK-NEXT:    [[TMP16:%.*]] = insertelement <16 x i8> poison, i8 [[TMP0]], i64 0
 585 // CHECK-NEXT:    [[TMP17:%.*]] = insertelement <16 x i8> [[TMP16]], i8 [[TMP1]], i64 1
 586 // CHECK-NEXT:    [[TMP18:%.*]] = insertelement <16 x i8> [[TMP17]], i8 [[TMP2]], i64 2
 587 // CHECK-NEXT:    [[TMP19:%.*]] = insertelement <16 x i8> [[TMP18]], i8 [[TMP3]], i64 3
 588 // CHECK-NEXT:    [[TMP20:%.*]] = insertelement <16 x i8> [[TMP19]], i8 [[TMP4]], i64 4
 589 // CHECK-NEXT:    [[TMP21:%.*]] = insertelement <16 x i8> [[TMP20]], i8 [[TMP5]], i64 5
 590 // CHECK-NEXT:    [[TMP22:%.*]] = insertelement <16 x i8> [[TMP21]], i8 [[TMP6]], i64 6
 591 // CHECK-NEXT:    [[TMP23:%.*]] = insertelement <16 x i8> [[TMP22]], i8 [[TMP7]], i64 7
 592 // CHECK-NEXT:    [[TMP24:%.*]] = insertelement <16 x i8> [[TMP23]], i8 [[TMP8]], i64 8
 593 // CHECK-NEXT:    [[TMP25:%.*]] = insertelement <16 x i8> [[TMP24]], i8 [[TMP9]], i64 9
 594 // CHECK-NEXT:    [[TMP26:%.*]] = insertelement <16 x i8> [[TMP25]], i8 [[TMP10]], i64 10
 595 // CHECK-NEXT:    [[TMP27:%.*]] = insertelement <16 x i8> [[TMP26]], i8 [[TMP11]], i64 11
 596 // CHECK-NEXT:    [[TMP28:%.*]] = insertelement <16 x i8> [[TMP27]], i8 [[TMP12]], i64 12
 597 // CHECK-NEXT:    [[TMP29:%.*]] = insertelement <16 x i8> [[TMP28]], i8 [[TMP13]], i64 13
 598 // CHECK-NEXT:    [[TMP30:%.*]] = insertelement <16 x i8> [[TMP29]], i8 [[TMP14]], i64 14
 599 // CHECK-NEXT:    [[TMP31:%.*]] = insertelement <16 x i8> [[TMP30]], i8 [[TMP15]], i64 15
 600 // CHECK-NEXT:    [[TMP32:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.insert.nxv16i8.v16i8(<vscale x 16 x i8> poison, <16 x i8> [[TMP31]], i64 0)
 601 // CHECK-NEXT:    [[TMP33:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[TMP32]], i64 0)
 602 // CHECK-NEXT:    [[TMP34:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 31)
 603 // CHECK-NEXT:    [[TMP35:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.wide.nxv16i8(<vscale x 16 x i1> [[TMP34]], <vscale x 16 x i8> [[TMP33]], <vscale x 2 x i64> zeroinitializer)
 604 // CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP35]]
 605 //
 606 // CPP-CHECK-LABEL: @_Z16test_svdupq_n_b8bbbbbbbbbbbbbbbb(
 607 // CPP-CHECK-NEXT:  entry:
 608 // CPP-CHECK-NEXT:    [[FROMBOOL:%.*]] = zext i1 [[X0:%.*]] to i8
 609 // CPP-CHECK-NEXT:    [[FROMBOOL1:%.*]] = zext i1 [[X1:%.*]] to i8
 610 // CPP-CHECK-NEXT:    [[FROMBOOL2:%.*]] = zext i1 [[X2:%.*]] to i8
 611 // CPP-CHECK-NEXT:    [[FROMBOOL3:%.*]] = zext i1 [[X3:%.*]] to i8
 612 // CPP-CHECK-NEXT:    [[FROMBOOL4:%.*]] = zext i1 [[X4:%.*]] to i8
 613 // CPP-CHECK-NEXT:    [[FROMBOOL5:%.*]] = zext i1 [[X5:%.*]] to i8
 614 // CPP-CHECK-NEXT:    [[FROMBOOL6:%.*]] = zext i1 [[X6:%.*]] to i8
 615 // CPP-CHECK-NEXT:    [[FROMBOOL7:%.*]] = zext i1 [[X7:%.*]] to i8
 616 // CPP-CHECK-NEXT:    [[FROMBOOL8:%.*]] = zext i1 [[X8:%.*]] to i8
 617 // CPP-CHECK-NEXT:    [[FROMBOOL9:%.*]] = zext i1 [[X9:%.*]] to i8
 618 // CPP-CHECK-NEXT:    [[FROMBOOL10:%.*]] = zext i1 [[X10:%.*]] to i8
 619 // CPP-CHECK-NEXT:    [[FROMBOOL11:%.*]] = zext i1 [[X11:%.*]] to i8
 620 // CPP-CHECK-NEXT:    [[FROMBOOL12:%.*]] = zext i1 [[X12:%.*]] to i8
 621 // CPP-CHECK-NEXT:    [[FROMBOOL13:%.*]] = zext i1 [[X13:%.*]] to i8
 622 // CPP-CHECK-NEXT:    [[FROMBOOL14:%.*]] = zext i1 [[X14:%.*]] to i8
 623 // CPP-CHECK-NEXT:    [[FROMBOOL15:%.*]] = zext i1 [[X15:%.*]] to i8
 624 // CPP-CHECK-NEXT:    [[TOBOOL:%.*]] = trunc i8 [[FROMBOOL]] to i1
 625 // CPP-CHECK-NEXT:    [[TOBOOL16:%.*]] = trunc i8 [[FROMBOOL1]] to i1
 626 // CPP-CHECK-NEXT:    [[TOBOOL17:%.*]] = trunc i8 [[FROMBOOL2]] to i1
 627 // CPP-CHECK-NEXT:    [[TOBOOL18:%.*]] = trunc i8 [[FROMBOOL3]] to i1
 628 // CPP-CHECK-NEXT:    [[TOBOOL19:%.*]] = trunc i8 [[FROMBOOL4]] to i1
 629 // CPP-CHECK-NEXT:    [[TOBOOL20:%.*]] = trunc i8 [[FROMBOOL5]] to i1
 630 // CPP-CHECK-NEXT:    [[TOBOOL21:%.*]] = trunc i8 [[FROMBOOL6]] to i1
 631 // CPP-CHECK-NEXT:    [[TOBOOL22:%.*]] = trunc i8 [[FROMBOOL7]] to i1
 632 // CPP-CHECK-NEXT:    [[TOBOOL23:%.*]] = trunc i8 [[FROMBOOL8]] to i1
 633 // CPP-CHECK-NEXT:    [[TOBOOL24:%.*]] = trunc i8 [[FROMBOOL9]] to i1
 634 // CPP-CHECK-NEXT:    [[TOBOOL25:%.*]] = trunc i8 [[FROMBOOL10]] to i1
 635 // CPP-CHECK-NEXT:    [[TOBOOL26:%.*]] = trunc i8 [[FROMBOOL11]] to i1
 636 // CPP-CHECK-NEXT:    [[TOBOOL27:%.*]] = trunc i8 [[FROMBOOL12]] to i1
 637 // CPP-CHECK-NEXT:    [[TOBOOL28:%.*]] = trunc i8 [[FROMBOOL13]] to i1
 638 // CPP-CHECK-NEXT:    [[TOBOOL29:%.*]] = trunc i8 [[FROMBOOL14]] to i1
 639 // CPP-CHECK-NEXT:    [[TOBOOL30:%.*]] = trunc i8 [[FROMBOOL15]] to i1
 640 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = zext i1 [[TOBOOL]] to i8
 641 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = zext i1 [[TOBOOL16]] to i8
 642 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext i1 [[TOBOOL17]] to i8
 643 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext i1 [[TOBOOL18]] to i8
 644 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext i1 [[TOBOOL19]] to i8
 645 // CPP-CHECK-NEXT:    [[TMP5:%.*]] = zext i1 [[TOBOOL20]] to i8
 646 // CPP-CHECK-NEXT:    [[TMP6:%.*]] = zext i1 [[TOBOOL21]] to i8
 647 // CPP-CHECK-NEXT:    [[TMP7:%.*]] = zext i1 [[TOBOOL22]] to i8
 648 // CPP-CHECK-NEXT:    [[TMP8:%.*]] = zext i1 [[TOBOOL23]] to i8
 649 // CPP-CHECK-NEXT:    [[TMP9:%.*]] = zext i1 [[TOBOOL24]] to i8
 650 // CPP-CHECK-NEXT:    [[TMP10:%.*]] = zext i1 [[TOBOOL25]] to i8
 651 // CPP-CHECK-NEXT:    [[TMP11:%.*]] = zext i1 [[TOBOOL26]] to i8
 652 // CPP-CHECK-NEXT:    [[TMP12:%.*]] = zext i1 [[TOBOOL27]] to i8
 653 // CPP-CHECK-NEXT:    [[TMP13:%.*]] = zext i1 [[TOBOOL28]] to i8
 654 // CPP-CHECK-NEXT:    [[TMP14:%.*]] = zext i1 [[TOBOOL29]] to i8
 655 // CPP-CHECK-NEXT:    [[TMP15:%.*]] = zext i1 [[TOBOOL30]] to i8
 656 // CPP-CHECK-NEXT:    [[TMP16:%.*]] = insertelement <16 x i8> poison, i8 [[TMP0]], i64 0
 657 // CPP-CHECK-NEXT:    [[TMP17:%.*]] = insertelement <16 x i8> [[TMP16]], i8 [[TMP1]], i64 1
 658 // CPP-CHECK-NEXT:    [[TMP18:%.*]] = insertelement <16 x i8> [[TMP17]], i8 [[TMP2]], i64 2
 659 // CPP-CHECK-NEXT:    [[TMP19:%.*]] = insertelement <16 x i8> [[TMP18]], i8 [[TMP3]], i64 3
 660 // CPP-CHECK-NEXT:    [[TMP20:%.*]] = insertelement <16 x i8> [[TMP19]], i8 [[TMP4]], i64 4
 661 // CPP-CHECK-NEXT:    [[TMP21:%.*]] = insertelement <16 x i8> [[TMP20]], i8 [[TMP5]], i64 5
 662 // CPP-CHECK-NEXT:    [[TMP22:%.*]] = insertelement <16 x i8> [[TMP21]], i8 [[TMP6]], i64 6
 663 // CPP-CHECK-NEXT:    [[TMP23:%.*]] = insertelement <16 x i8> [[TMP22]], i8 [[TMP7]], i64 7
 664 // CPP-CHECK-NEXT:    [[TMP24:%.*]] = insertelement <16 x i8> [[TMP23]], i8 [[TMP8]], i64 8
 665 // CPP-CHECK-NEXT:    [[TMP25:%.*]] = insertelement <16 x i8> [[TMP24]], i8 [[TMP9]], i64 9
 666 // CPP-CHECK-NEXT:    [[TMP26:%.*]] = insertelement <16 x i8> [[TMP25]], i8 [[TMP10]], i64 10
 667 // CPP-CHECK-NEXT:    [[TMP27:%.*]] = insertelement <16 x i8> [[TMP26]], i8 [[TMP11]], i64 11
 668 // CPP-CHECK-NEXT:    [[TMP28:%.*]] = insertelement <16 x i8> [[TMP27]], i8 [[TMP12]], i64 12
 669 // CPP-CHECK-NEXT:    [[TMP29:%.*]] = insertelement <16 x i8> [[TMP28]], i8 [[TMP13]], i64 13
 670 // CPP-CHECK-NEXT:    [[TMP30:%.*]] = insertelement <16 x i8> [[TMP29]], i8 [[TMP14]], i64 14
 671 // CPP-CHECK-NEXT:    [[TMP31:%.*]] = insertelement <16 x i8> [[TMP30]], i8 [[TMP15]], i64 15
 672 // CPP-CHECK-NEXT:    [[TMP32:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.insert.nxv16i8.v16i8(<vscale x 16 x i8> poison, <16 x i8> [[TMP31]], i64 0)
 673 // CPP-CHECK-NEXT:    [[TMP33:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[TMP32]], i64 0)
 674 // CPP-CHECK-NEXT:    [[TMP34:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 31)
 675 // CPP-CHECK-NEXT:    [[TMP35:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.wide.nxv16i8(<vscale x 16 x i1> [[TMP34]], <vscale x 16 x i8> [[TMP33]], <vscale x 2 x i64> zeroinitializer)
 676 // CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP35]]
 677 //
 678 svbool_t test_svdupq_n_b8(bool x0, bool x1, bool x2, bool x3,
 679                           bool x4, bool x5, bool x6, bool x7,
 680                           bool x8, bool x9, bool x10, bool x11,
 681                           bool x12, bool x13, bool x14, bool x15)
 682 {
 683   // <assume other insertelement>
 684   return SVE_ACLE_FUNC(svdupq,_n,_b8,)(x0, x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, x11, x12, x13, x14, x15);
 685 }
 686
 687 // CHECK-LABEL: @test_svdupq_n_b16(
 688 // CHECK-NEXT:  entry:
 689 // CHECK-NEXT:    [[FROMBOOL:%.*]] = zext i1 [[X0:%.*]] to i8
 690 // CHECK-NEXT:    [[FROMBOOL1:%.*]] = zext i1 [[X1:%.*]] to i8
 691 // CHECK-NEXT:    [[FROMBOOL2:%.*]] = zext i1 [[X2:%.*]] to i8
 692 // CHECK-NEXT:    [[FROMBOOL3:%.*]] = zext i1 [[X3:%.*]] to i8
 693 // CHECK-NEXT:    [[FROMBOOL4:%.*]] = zext i1 [[X4:%.*]] to i8
 694 // CHECK-NEXT:    [[FROMBOOL5:%.*]] = zext i1 [[X5:%.*]] to i8
 695 // CHECK-NEXT:    [[FROMBOOL6:%.*]] = zext i1 [[X6:%.*]] to i8
 696 // CHECK-NEXT:    [[FROMBOOL7:%.*]] = zext i1 [[X7:%.*]] to i8
 697 // CHECK-NEXT:    [[TOBOOL:%.*]] = trunc i8 [[FROMBOOL]] to i1
 698 // CHECK-NEXT:    [[TOBOOL8:%.*]] = trunc i8 [[FROMBOOL1]] to i1
 699 // CHECK-NEXT:    [[TOBOOL9:%.*]] = trunc i8 [[FROMBOOL2]] to i1
 700 // CHECK-NEXT:    [[TOBOOL10:%.*]] = trunc i8 [[FROMBOOL3]] to i1
 701 // CHECK-NEXT:    [[TOBOOL11:%.*]] = trunc i8 [[FROMBOOL4]] to i1
 702 // CHECK-NEXT:    [[TOBOOL12:%.*]] = trunc i8 [[FROMBOOL5]] to i1
 703 // CHECK-NEXT:    [[TOBOOL13:%.*]] = trunc i8 [[FROMBOOL6]] to i1
 704 // CHECK-NEXT:    [[TOBOOL14:%.*]] = trunc i8 [[FROMBOOL7]] to i1
 705 // CHECK-NEXT:    [[TMP0:%.*]] = zext i1 [[TOBOOL]] to i16
 706 // CHECK-NEXT:    [[TMP1:%.*]] = zext i1 [[TOBOOL8]] to i16
 707 // CHECK-NEXT:    [[TMP2:%.*]] = zext i1 [[TOBOOL9]] to i16
 708 // CHECK-NEXT:    [[TMP3:%.*]] = zext i1 [[TOBOOL10]] to i16
 709 // CHECK-NEXT:    [[TMP4:%.*]] = zext i1 [[TOBOOL11]] to i16
 710 // CHECK-NEXT:    [[TMP5:%.*]] = zext i1 [[TOBOOL12]] to i16
 711 // CHECK-NEXT:    [[TMP6:%.*]] = zext i1 [[TOBOOL13]] to i16
 712 // CHECK-NEXT:    [[TMP7:%.*]] = zext i1 [[TOBOOL14]] to i16
 713 // CHECK-NEXT:    [[TMP8:%.*]] = insertelement <8 x i16> poison, i16 [[TMP0]], i64 0
 714 // CHECK-NEXT:    [[TMP9:%.*]] = insertelement <8 x i16> [[TMP8]], i16 [[TMP1]], i64 1
 715 // CHECK-NEXT:    [[TMP10:%.*]] = insertelement <8 x i16> [[TMP9]], i16 [[TMP2]], i64 2
 716 // CHECK-NEXT:    [[TMP11:%.*]] = insertelement <8 x i16> [[TMP10]], i16 [[TMP3]], i64 3
 717 // CHECK-NEXT:    [[TMP12:%.*]] = insertelement <8 x i16> [[TMP11]], i16 [[TMP4]], i64 4
 718 // CHECK-NEXT:    [[TMP13:%.*]] = insertelement <8 x i16> [[TMP12]], i16 [[TMP5]], i64 5
 719 // CHECK-NEXT:    [[TMP14:%.*]] = insertelement <8 x i16> [[TMP13]], i16 [[TMP6]], i64 6
 720 // CHECK-NEXT:    [[TMP15:%.*]] = insertelement <8 x i16> [[TMP14]], i16 [[TMP7]], i64 7
 721 // CHECK-NEXT:    [[TMP16:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.insert.nxv8i16.v8i16(<vscale x 8 x i16> poison, <8 x i16> [[TMP15]], i64 0)
 722 // CHECK-NEXT:    [[TMP17:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[TMP16]], i64 0)
 723 // CHECK-NEXT:    [[TMP18:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 31)
 724 // CHECK-NEXT:    [[TMP19:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.wide.nxv8i16(<vscale x 8 x i1> [[TMP18]], <vscale x 8 x i16> [[TMP17]], <vscale x 2 x i64> zeroinitializer)
 725 // CHECK-NEXT:    [[TMP20:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP19]])
 726 // CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP20]]
 727 //
 728 // CPP-CHECK-LABEL: @_Z17test_svdupq_n_b16bbbbbbbb(
 729 // CPP-CHECK-NEXT:  entry:
 730 // CPP-CHECK-NEXT:    [[FROMBOOL:%.*]] = zext i1 [[X0:%.*]] to i8
 731 // CPP-CHECK-NEXT:    [[FROMBOOL1:%.*]] = zext i1 [[X1:%.*]] to i8
 732 // CPP-CHECK-NEXT:    [[FROMBOOL2:%.*]] = zext i1 [[X2:%.*]] to i8
 733 // CPP-CHECK-NEXT:    [[FROMBOOL3:%.*]] = zext i1 [[X3:%.*]] to i8
 734 // CPP-CHECK-NEXT:    [[FROMBOOL4:%.*]] = zext i1 [[X4:%.*]] to i8
 735 // CPP-CHECK-NEXT:    [[FROMBOOL5:%.*]] = zext i1 [[X5:%.*]] to i8
 736 // CPP-CHECK-NEXT:    [[FROMBOOL6:%.*]] = zext i1 [[X6:%.*]] to i8
 737 // CPP-CHECK-NEXT:    [[FROMBOOL7:%.*]] = zext i1 [[X7:%.*]] to i8
 738 // CPP-CHECK-NEXT:    [[TOBOOL:%.*]] = trunc i8 [[FROMBOOL]] to i1
 739 // CPP-CHECK-NEXT:    [[TOBOOL8:%.*]] = trunc i8 [[FROMBOOL1]] to i1
 740 // CPP-CHECK-NEXT:    [[TOBOOL9:%.*]] = trunc i8 [[FROMBOOL2]] to i1
 741 // CPP-CHECK-NEXT:    [[TOBOOL10:%.*]] = trunc i8 [[FROMBOOL3]] to i1
 742 // CPP-CHECK-NEXT:    [[TOBOOL11:%.*]] = trunc i8 [[FROMBOOL4]] to i1
 743 // CPP-CHECK-NEXT:    [[TOBOOL12:%.*]] = trunc i8 [[FROMBOOL5]] to i1
 744 // CPP-CHECK-NEXT:    [[TOBOOL13:%.*]] = trunc i8 [[FROMBOOL6]] to i1
 745 // CPP-CHECK-NEXT:    [[TOBOOL14:%.*]] = trunc i8 [[FROMBOOL7]] to i1
 746 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = zext i1 [[TOBOOL]] to i16
 747 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = zext i1 [[TOBOOL8]] to i16
 748 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext i1 [[TOBOOL9]] to i16
 749 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext i1 [[TOBOOL10]] to i16
 750 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext i1 [[TOBOOL11]] to i16
 751 // CPP-CHECK-NEXT:    [[TMP5:%.*]] = zext i1 [[TOBOOL12]] to i16
 752 // CPP-CHECK-NEXT:    [[TMP6:%.*]] = zext i1 [[TOBOOL13]] to i16
 753 // CPP-CHECK-NEXT:    [[TMP7:%.*]] = zext i1 [[TOBOOL14]] to i16
 754 // CPP-CHECK-NEXT:    [[TMP8:%.*]] = insertelement <8 x i16> poison, i16 [[TMP0]], i64 0
 755 // CPP-CHECK-NEXT:    [[TMP9:%.*]] = insertelement <8 x i16> [[TMP8]], i16 [[TMP1]], i64 1
 756 // CPP-CHECK-NEXT:    [[TMP10:%.*]] = insertelement <8 x i16> [[TMP9]], i16 [[TMP2]], i64 2
 757 // CPP-CHECK-NEXT:    [[TMP11:%.*]] = insertelement <8 x i16> [[TMP10]], i16 [[TMP3]], i64 3
 758 // CPP-CHECK-NEXT:    [[TMP12:%.*]] = insertelement <8 x i16> [[TMP11]], i16 [[TMP4]], i64 4
 759 // CPP-CHECK-NEXT:    [[TMP13:%.*]] = insertelement <8 x i16> [[TMP12]], i16 [[TMP5]], i64 5
 760 // CPP-CHECK-NEXT:    [[TMP14:%.*]] = insertelement <8 x i16> [[TMP13]], i16 [[TMP6]], i64 6
 761 // CPP-CHECK-NEXT:    [[TMP15:%.*]] = insertelement <8 x i16> [[TMP14]], i16 [[TMP7]], i64 7
 762 // CPP-CHECK-NEXT:    [[TMP16:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.insert.nxv8i16.v8i16(<vscale x 8 x i16> poison, <8 x i16> [[TMP15]], i64 0)
 763 // CPP-CHECK-NEXT:    [[TMP17:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[TMP16]], i64 0)
 764 // CPP-CHECK-NEXT:    [[TMP18:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 31)
 765 // CPP-CHECK-NEXT:    [[TMP19:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.wide.nxv8i16(<vscale x 8 x i1> [[TMP18]], <vscale x 8 x i16> [[TMP17]], <vscale x 2 x i64> zeroinitializer)
 766 // CPP-CHECK-NEXT:    [[TMP20:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP19]])
 767 // CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP20]]
 768 //
 769 svbool_t test_svdupq_n_b16(bool x0, bool x1, bool x2, bool x3,
 770                            bool x4, bool x5, bool x6, bool x7)
 771 {
 772   // <assume other insertelement>
 773   return SVE_ACLE_FUNC(svdupq,_n,_b16,)(x0, x1, x2, x3, x4, x5, x6, x7);
 774 }
 775
 776 // CHECK-LABEL: @test_svdupq_n_b32(
 777 // CHECK-NEXT:  entry:
 778 // CHECK-NEXT:    [[FROMBOOL:%.*]] = zext i1 [[X0:%.*]] to i8
 779 // CHECK-NEXT:    [[FROMBOOL1:%.*]] = zext i1 [[X1:%.*]] to i8
 780 // CHECK-NEXT:    [[FROMBOOL2:%.*]] = zext i1 [[X2:%.*]] to i8
 781 // CHECK-NEXT:    [[FROMBOOL3:%.*]] = zext i1 [[X3:%.*]] to i8
 782 // CHECK-NEXT:    [[TOBOOL:%.*]] = trunc i8 [[FROMBOOL]] to i1
 783 // CHECK-NEXT:    [[TOBOOL4:%.*]] = trunc i8 [[FROMBOOL1]] to i1
 784 // CHECK-NEXT:    [[TOBOOL5:%.*]] = trunc i8 [[FROMBOOL2]] to i1
 785 // CHECK-NEXT:    [[TOBOOL6:%.*]] = trunc i8 [[FROMBOOL3]] to i1
 786 // CHECK-NEXT:    [[TMP0:%.*]] = zext i1 [[TOBOOL]] to i32
 787 // CHECK-NEXT:    [[TMP1:%.*]] = zext i1 [[TOBOOL4]] to i32
 788 // CHECK-NEXT:    [[TMP2:%.*]] = zext i1 [[TOBOOL5]] to i32
 789 // CHECK-NEXT:    [[TMP3:%.*]] = zext i1 [[TOBOOL6]] to i32
 790 // CHECK-NEXT:    [[TMP4:%.*]] = insertelement <4 x i32> poison, i32 [[TMP0]], i64 0
 791 // CHECK-NEXT:    [[TMP5:%.*]] = insertelement <4 x i32> [[TMP4]], i32 [[TMP1]], i64 1
 792 // CHECK-NEXT:    [[TMP6:%.*]] = insertelement <4 x i32> [[TMP5]], i32 [[TMP2]], i64 2
 793 // CHECK-NEXT:    [[TMP7:%.*]] = insertelement <4 x i32> [[TMP6]], i32 [[TMP3]], i64 3
 794 // CHECK-NEXT:    [[TMP8:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.insert.nxv4i32.v4i32(<vscale x 4 x i32> poison, <4 x i32> [[TMP7]], i64 0)
 795 // CHECK-NEXT:    [[TMP9:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[TMP8]], i64 0)
 796 // CHECK-NEXT:    [[TMP10:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.ptrue.nxv4i1(i32 31)
 797 // CHECK-NEXT:    [[TMP11:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.wide.nxv4i32(<vscale x 4 x i1> [[TMP10]], <vscale x 4 x i32> [[TMP9]], <vscale x 2 x i64> zeroinitializer)
 798 // CHECK-NEXT:    [[TMP12:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP11]])
 799 // CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP12]]
 800 //
 801 // CPP-CHECK-LABEL: @_Z17test_svdupq_n_b32bbbb(
 802 // CPP-CHECK-NEXT:  entry:
 803 // CPP-CHECK-NEXT:    [[FROMBOOL:%.*]] = zext i1 [[X0:%.*]] to i8
 804 // CPP-CHECK-NEXT:    [[FROMBOOL1:%.*]] = zext i1 [[X1:%.*]] to i8
 805 // CPP-CHECK-NEXT:    [[FROMBOOL2:%.*]] = zext i1 [[X2:%.*]] to i8
 806 // CPP-CHECK-NEXT:    [[FROMBOOL3:%.*]] = zext i1 [[X3:%.*]] to i8
 807 // CPP-CHECK-NEXT:    [[TOBOOL:%.*]] = trunc i8 [[FROMBOOL]] to i1
 808 // CPP-CHECK-NEXT:    [[TOBOOL4:%.*]] = trunc i8 [[FROMBOOL1]] to i1
 809 // CPP-CHECK-NEXT:    [[TOBOOL5:%.*]] = trunc i8 [[FROMBOOL2]] to i1
 810 // CPP-CHECK-NEXT:    [[TOBOOL6:%.*]] = trunc i8 [[FROMBOOL3]] to i1
 811 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = zext i1 [[TOBOOL]] to i32
 812 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = zext i1 [[TOBOOL4]] to i32
 813 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext i1 [[TOBOOL5]] to i32
 814 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext i1 [[TOBOOL6]] to i32
 815 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = insertelement <4 x i32> poison, i32 [[TMP0]], i64 0
 816 // CPP-CHECK-NEXT:    [[TMP5:%.*]] = insertelement <4 x i32> [[TMP4]], i32 [[TMP1]], i64 1
 817 // CPP-CHECK-NEXT:    [[TMP6:%.*]] = insertelement <4 x i32> [[TMP5]], i32 [[TMP2]], i64 2
 818 // CPP-CHECK-NEXT:    [[TMP7:%.*]] = insertelement <4 x i32> [[TMP6]], i32 [[TMP3]], i64 3
 819 // CPP-CHECK-NEXT:    [[TMP8:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.insert.nxv4i32.v4i32(<vscale x 4 x i32> poison, <4 x i32> [[TMP7]], i64 0)
 820 // CPP-CHECK-NEXT:    [[TMP9:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[TMP8]], i64 0)
 821 // CPP-CHECK-NEXT:    [[TMP10:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.ptrue.nxv4i1(i32 31)
 822 // CPP-CHECK-NEXT:    [[TMP11:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.wide.nxv4i32(<vscale x 4 x i1> [[TMP10]], <vscale x 4 x i32> [[TMP9]], <vscale x 2 x i64> zeroinitializer)
 823 // CPP-CHECK-NEXT:    [[TMP12:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP11]])
 824 // CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP12]]
 825 //
 826 svbool_t test_svdupq_n_b32(bool x0, bool x1, bool x2, bool x3)
 827 {
 828   // <assume other insertelement>
 829   return SVE_ACLE_FUNC(svdupq,_n,_b32,)(x0, x1, x2, x3);
 830 }
 831
 832 // CHECK-LABEL: @test_svdupq_n_b64(
 833 // CHECK-NEXT:  entry:
 834 // CHECK-NEXT:    [[FROMBOOL:%.*]] = zext i1 [[X0:%.*]] to i8
 835 // CHECK-NEXT:    [[FROMBOOL1:%.*]] = zext i1 [[X1:%.*]] to i8
 836 // CHECK-NEXT:    [[TOBOOL:%.*]] = trunc i8 [[FROMBOOL]] to i1
 837 // CHECK-NEXT:    [[TOBOOL2:%.*]] = trunc i8 [[FROMBOOL1]] to i1
 838 // CHECK-NEXT:    [[TMP0:%.*]] = zext i1 [[TOBOOL]] to i64
 839 // CHECK-NEXT:    [[TMP1:%.*]] = zext i1 [[TOBOOL2]] to i64
 840 // CHECK-NEXT:    [[TMP2:%.*]] = insertelement <2 x i64> poison, i64 [[TMP0]], i64 0
 841 // CHECK-NEXT:    [[TMP3:%.*]] = insertelement <2 x i64> [[TMP2]], i64 [[TMP1]], i64 1
 842 // CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.insert.nxv2i64.v2i64(<vscale x 2 x i64> poison, <2 x i64> [[TMP3]], i64 0)
 843 // CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[TMP4]], i64 0)
 844 // CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.ptrue.nxv2i1(i32 31)
 845 // CHECK-NEXT:    [[TMP7:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> [[TMP6]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i64> zeroinitializer)
 846 // CHECK-NEXT:    [[TMP8:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP7]])
 847 // CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP8]]
 848 //
 849 // CPP-CHECK-LABEL: @_Z17test_svdupq_n_b64bb(
 850 // CPP-CHECK-NEXT:  entry:
 851 // CPP-CHECK-NEXT:    [[FROMBOOL:%.*]] = zext i1 [[X0:%.*]] to i8
 852 // CPP-CHECK-NEXT:    [[FROMBOOL1:%.*]] = zext i1 [[X1:%.*]] to i8
 853 // CPP-CHECK-NEXT:    [[TOBOOL:%.*]] = trunc i8 [[FROMBOOL]] to i1
 854 // CPP-CHECK-NEXT:    [[TOBOOL2:%.*]] = trunc i8 [[FROMBOOL1]] to i1
 855 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = zext i1 [[TOBOOL]] to i64
 856 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = zext i1 [[TOBOOL2]] to i64
 857 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <2 x i64> poison, i64 [[TMP0]], i64 0
 858 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <2 x i64> [[TMP2]], i64 [[TMP1]], i64 1
 859 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.insert.nxv2i64.v2i64(<vscale x 2 x i64> poison, <2 x i64> [[TMP3]], i64 0)
 860 // CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[TMP4]], i64 0)
 861 // CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.ptrue.nxv2i1(i32 31)
 862 // CPP-CHECK-NEXT:    [[TMP7:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> [[TMP6]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i64> zeroinitializer)
 863 // CPP-CHECK-NEXT:    [[TMP8:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP7]])
 864 // CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP8]]
 865 //
 866 svbool_t test_svdupq_n_b64(bool x0, bool x1)
 867 {
 868   return SVE_ACLE_FUNC(svdupq,_n,_b64,)(x0, x1);
 869 }