clang/test/CodeGen/aarch64-sve2p1-intrinsics/acle_sve2p1_stnt1.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 +sve2p1 -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
   4 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2p1 -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
   5 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2p1 -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
   6 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2p1 -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
   7 #include <arm_sve.h>
   8
   9 #ifdef SVE_OVERLOADED_FORMS
  10 // A simple used,unused... macro, long enough to represent any SVE builtin.
  11 #define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
  12 #else
  13 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
  14 #endif
  15
  16
  17 // CHECK-LABEL: @test_svstnt1_u8_x2(
  18 // CHECK-NEXT:  entry:
  19 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[V:%.*]], i64 0)
  20 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[V]], i64 16)
  21 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
  22 // CHECK-NEXT:    ret void
  23 //
  24 // CPP-CHECK-LABEL: @_Z18test_svstnt1_u8_x2u11__SVCount_tPh11svuint8x2_t(
  25 // CPP-CHECK-NEXT:  entry:
  26 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[V:%.*]], i64 0)
  27 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[V]], i64 16)
  28 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
  29 // CPP-CHECK-NEXT:    ret void
  30 //
  31 void test_svstnt1_u8_x2(svcount_t pn, uint8_t *base, svuint8x2_t v)
  32 {
  33   return SVE_ACLE_FUNC(svstnt1,_u8_x2,,)(pn, base, v);
  34 }
  35
  36
  37 // CHECK-LABEL: @test_svstnt1_u16_x2(
  38 // CHECK-NEXT:  entry:
  39 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[V:%.*]], i64 0)
  40 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[V]], i64 8)
  41 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv8i16(<vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
  42 // CHECK-NEXT:    ret void
  43 //
  44 // CPP-CHECK-LABEL: @_Z19test_svstnt1_u16_x2u11__SVCount_tPt12svuint16x2_t(
  45 // CPP-CHECK-NEXT:  entry:
  46 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[V:%.*]], i64 0)
  47 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[V]], i64 8)
  48 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv8i16(<vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
  49 // CPP-CHECK-NEXT:    ret void
  50 //
  51 void test_svstnt1_u16_x2(svcount_t pn, uint16_t *base, svuint16x2_t v)
  52 {
  53   return SVE_ACLE_FUNC(svstnt1,_u16_x2,,)(pn, base, v);
  54 }
  55
  56
  57 // CHECK-LABEL: @test_svstnt1_u32_x2(
  58 // CHECK-NEXT:  entry:
  59 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[V:%.*]], i64 0)
  60 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[V]], i64 4)
  61 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv4i32(<vscale x 4 x i32> [[TMP0]], <vscale x 4 x i32> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
  62 // CHECK-NEXT:    ret void
  63 //
  64 // CPP-CHECK-LABEL: @_Z19test_svstnt1_u32_x2u11__SVCount_tPj12svuint32x2_t(
  65 // CPP-CHECK-NEXT:  entry:
  66 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[V:%.*]], i64 0)
  67 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[V]], i64 4)
  68 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv4i32(<vscale x 4 x i32> [[TMP0]], <vscale x 4 x i32> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
  69 // CPP-CHECK-NEXT:    ret void
  70 //
  71 void test_svstnt1_u32_x2(svcount_t pn, uint32_t *base, svuint32x2_t v)
  72 {
  73   return SVE_ACLE_FUNC(svstnt1,_u32_x2,,)(pn, base, v);
  74 }
  75
  76
  77 // CHECK-LABEL: @test_svstnt1_u64_x2(
  78 // CHECK-NEXT:  entry:
  79 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[V:%.*]], i64 0)
  80 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[V]], i64 2)
  81 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv2i64(<vscale x 2 x i64> [[TMP0]], <vscale x 2 x i64> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
  82 // CHECK-NEXT:    ret void
  83 //
  84 // CPP-CHECK-LABEL: @_Z19test_svstnt1_u64_x2u11__SVCount_tPm12svuint64x2_t(
  85 // CPP-CHECK-NEXT:  entry:
  86 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[V:%.*]], i64 0)
  87 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[V]], i64 2)
  88 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv2i64(<vscale x 2 x i64> [[TMP0]], <vscale x 2 x i64> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
  89 // CPP-CHECK-NEXT:    ret void
  90 //
  91 void test_svstnt1_u64_x2(svcount_t pn, uint64_t *base, svuint64x2_t v)
  92 {
  93   return SVE_ACLE_FUNC(svstnt1,_u64_x2,,)(pn, base, v);
  94 }
  95
  96
  97 // CHECK-LABEL: @test_svstnt1_u8_x4(
  98 // CHECK-NEXT:  entry:
  99 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V:%.*]], i64 0)
 100 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 16)
 101 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 32)
 102 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 48)
 103 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 104 // CHECK-NEXT:    ret void
 105 //
 106 // CPP-CHECK-LABEL: @_Z18test_svstnt1_u8_x4u11__SVCount_tPh11svuint8x4_t(
 107 // CPP-CHECK-NEXT:  entry:
 108 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V:%.*]], i64 0)
 109 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 16)
 110 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 32)
 111 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 48)
 112 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 113 // CPP-CHECK-NEXT:    ret void
 114 //
 115 void test_svstnt1_u8_x4(svcount_t pn, uint8_t *base, svuint8x4_t v)
 116 {
 117   return SVE_ACLE_FUNC(svstnt1,_u8_x4,,)(pn, base, v);
 118 }
 119
 120
 121 // CHECK-LABEL: @test_svstnt1_u16_x4(
 122 // CHECK-NEXT:  entry:
 123 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V:%.*]], i64 0)
 124 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 8)
 125 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 16)
 126 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 24)
 127 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv8i16(<vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 128 // CHECK-NEXT:    ret void
 129 //
 130 // CPP-CHECK-LABEL: @_Z19test_svstnt1_u16_x4u11__SVCount_tPt12svuint16x4_t(
 131 // CPP-CHECK-NEXT:  entry:
 132 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V:%.*]], i64 0)
 133 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 8)
 134 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 16)
 135 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 24)
 136 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv8i16(<vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 137 // CPP-CHECK-NEXT:    ret void
 138 //
 139 void test_svstnt1_u16_x4(svcount_t pn, uint16_t *base, svuint16x4_t v)
 140 {
 141   return SVE_ACLE_FUNC(svstnt1,_u16_x4,,)(pn, base, v);
 142 }
 143
 144
 145 // CHECK-LABEL: @test_svstnt1_u32_x4(
 146 // CHECK-NEXT:  entry:
 147 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V:%.*]], i64 0)
 148 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 4)
 149 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 8)
 150 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 12)
 151 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv4i32(<vscale x 4 x i32> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 152 // CHECK-NEXT:    ret void
 153 //
 154 // CPP-CHECK-LABEL: @_Z19test_svstnt1_u32_x4u11__SVCount_tPj12svuint32x4_t(
 155 // CPP-CHECK-NEXT:  entry:
 156 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V:%.*]], i64 0)
 157 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 4)
 158 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 8)
 159 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 12)
 160 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv4i32(<vscale x 4 x i32> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 161 // CPP-CHECK-NEXT:    ret void
 162 //
 163 void test_svstnt1_u32_x4(svcount_t pn, uint32_t *base, svuint32x4_t v)
 164 {
 165   return SVE_ACLE_FUNC(svstnt1,_u32_x4,,)(pn, base, v);
 166 }
 167
 168
 169 // CHECK-LABEL: @test_svstnt1_u64_x4(
 170 // CHECK-NEXT:  entry:
 171 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V:%.*]], i64 0)
 172 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 2)
 173 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 4)
 174 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 6)
 175 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv2i64(<vscale x 2 x i64> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 176 // CHECK-NEXT:    ret void
 177 //
 178 // CPP-CHECK-LABEL: @_Z19test_svstnt1_u64_x4u11__SVCount_tPm12svuint64x4_t(
 179 // CPP-CHECK-NEXT:  entry:
 180 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V:%.*]], i64 0)
 181 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 2)
 182 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 4)
 183 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 6)
 184 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv2i64(<vscale x 2 x i64> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 185 // CPP-CHECK-NEXT:    ret void
 186 //
 187 void test_svstnt1_u64_x4(svcount_t pn, uint64_t *base, svuint64x4_t v)
 188 {
 189   return SVE_ACLE_FUNC(svstnt1,_u64_x4,,)(pn, base, v);
 190 }
 191
 192
 193 // CHECK-LABEL: @test_svstnt1_s8_x2(
 194 // CHECK-NEXT:  entry:
 195 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[V:%.*]], i64 0)
 196 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[V]], i64 16)
 197 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 198 // CHECK-NEXT:    ret void
 199 //
 200 // CPP-CHECK-LABEL: @_Z18test_svstnt1_s8_x2u11__SVCount_tPa10svint8x2_t(
 201 // CPP-CHECK-NEXT:  entry:
 202 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[V:%.*]], i64 0)
 203 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[V]], i64 16)
 204 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 205 // CPP-CHECK-NEXT:    ret void
 206 //
 207 void test_svstnt1_s8_x2(svcount_t pn, int8_t *base, svint8x2_t v)
 208 {
 209   return SVE_ACLE_FUNC(svstnt1,_s8_x2,,)(pn, base, v);
 210 }
 211
 212
 213 // CHECK-LABEL: @test_svstnt1_s16_x2(
 214 // CHECK-NEXT:  entry:
 215 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[V:%.*]], i64 0)
 216 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[V]], i64 8)
 217 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv8i16(<vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 218 // CHECK-NEXT:    ret void
 219 //
 220 // CPP-CHECK-LABEL: @_Z19test_svstnt1_s16_x2u11__SVCount_tPs11svint16x2_t(
 221 // CPP-CHECK-NEXT:  entry:
 222 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[V:%.*]], i64 0)
 223 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[V]], i64 8)
 224 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv8i16(<vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 225 // CPP-CHECK-NEXT:    ret void
 226 //
 227 void test_svstnt1_s16_x2(svcount_t pn, int16_t *base, svint16x2_t v)
 228 {
 229   return SVE_ACLE_FUNC(svstnt1,_s16_x2,,)(pn, base, v);
 230 }
 231
 232
 233 // CHECK-LABEL: @test_svstnt1_s32_x2(
 234 // CHECK-NEXT:  entry:
 235 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[V:%.*]], i64 0)
 236 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[V]], i64 4)
 237 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv4i32(<vscale x 4 x i32> [[TMP0]], <vscale x 4 x i32> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 238 // CHECK-NEXT:    ret void
 239 //
 240 // CPP-CHECK-LABEL: @_Z19test_svstnt1_s32_x2u11__SVCount_tPi11svint32x2_t(
 241 // CPP-CHECK-NEXT:  entry:
 242 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[V:%.*]], i64 0)
 243 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[V]], i64 4)
 244 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv4i32(<vscale x 4 x i32> [[TMP0]], <vscale x 4 x i32> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 245 // CPP-CHECK-NEXT:    ret void
 246 //
 247 void test_svstnt1_s32_x2(svcount_t pn, int32_t *base, svint32x2_t v)
 248 {
 249   return SVE_ACLE_FUNC(svstnt1,_s32_x2,,)(pn, base, v);
 250 }
 251
 252
 253 // CHECK-LABEL: @test_svstnt1_s64_x2(
 254 // CHECK-NEXT:  entry:
 255 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[V:%.*]], i64 0)
 256 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[V]], i64 2)
 257 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv2i64(<vscale x 2 x i64> [[TMP0]], <vscale x 2 x i64> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 258 // CHECK-NEXT:    ret void
 259 //
 260 // CPP-CHECK-LABEL: @_Z19test_svstnt1_s64_x2u11__SVCount_tPl11svint64x2_t(
 261 // CPP-CHECK-NEXT:  entry:
 262 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[V:%.*]], i64 0)
 263 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[V]], i64 2)
 264 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv2i64(<vscale x 2 x i64> [[TMP0]], <vscale x 2 x i64> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 265 // CPP-CHECK-NEXT:    ret void
 266 //
 267 void test_svstnt1_s64_x2(svcount_t pn, int64_t *base, svint64x2_t v)
 268 {
 269   return SVE_ACLE_FUNC(svstnt1,_s64_x2,,)(pn, base, v);
 270 }
 271
 272
 273 // CHECK-LABEL: @test_svstnt1_s8_x4(
 274 // CHECK-NEXT:  entry:
 275 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V:%.*]], i64 0)
 276 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 16)
 277 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 32)
 278 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 48)
 279 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 280 // CHECK-NEXT:    ret void
 281 //
 282 // CPP-CHECK-LABEL: @_Z18test_svstnt1_s8_x4u11__SVCount_tPa10svint8x4_t(
 283 // CPP-CHECK-NEXT:  entry:
 284 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V:%.*]], i64 0)
 285 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 16)
 286 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 32)
 287 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 48)
 288 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 289 // CPP-CHECK-NEXT:    ret void
 290 //
 291 void test_svstnt1_s8_x4(svcount_t pn, int8_t *base, svint8x4_t v)
 292 {
 293   return SVE_ACLE_FUNC(svstnt1,_s8_x4,,)(pn, base, v);
 294 }
 295
 296
 297 // CHECK-LABEL: @test_svstnt1_s16_x4(
 298 // CHECK-NEXT:  entry:
 299 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V:%.*]], i64 0)
 300 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 8)
 301 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 16)
 302 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 24)
 303 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv8i16(<vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 304 // CHECK-NEXT:    ret void
 305 //
 306 // CPP-CHECK-LABEL: @_Z19test_svstnt1_s16_x4u11__SVCount_tPs11svint16x4_t(
 307 // CPP-CHECK-NEXT:  entry:
 308 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V:%.*]], i64 0)
 309 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 8)
 310 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 16)
 311 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 24)
 312 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv8i16(<vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 313 // CPP-CHECK-NEXT:    ret void
 314 //
 315 void test_svstnt1_s16_x4(svcount_t pn, int16_t *base, svint16x4_t v)
 316 {
 317   return SVE_ACLE_FUNC(svstnt1,_s16_x4,,)(pn, base, v);
 318 }
 319
 320
 321 // CHECK-LABEL: @test_svstnt1_s32_x4(
 322 // CHECK-NEXT:  entry:
 323 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V:%.*]], i64 0)
 324 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 4)
 325 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 8)
 326 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 12)
 327 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv4i32(<vscale x 4 x i32> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 328 // CHECK-NEXT:    ret void
 329 //
 330 // CPP-CHECK-LABEL: @_Z19test_svstnt1_s32_x4u11__SVCount_tPi11svint32x4_t(
 331 // CPP-CHECK-NEXT:  entry:
 332 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V:%.*]], i64 0)
 333 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 4)
 334 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 8)
 335 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 12)
 336 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv4i32(<vscale x 4 x i32> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 337 // CPP-CHECK-NEXT:    ret void
 338 //
 339 void test_svstnt1_s32_x4(svcount_t pn, int32_t *base, svint32x4_t v)
 340 {
 341   return SVE_ACLE_FUNC(svstnt1,_s32_x4,,)(pn, base, v);
 342 }
 343
 344
 345 // CHECK-LABEL: @test_svstnt1_s64_x4(
 346 // CHECK-NEXT:  entry:
 347 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V:%.*]], i64 0)
 348 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 2)
 349 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 4)
 350 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 6)
 351 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv2i64(<vscale x 2 x i64> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 352 // CHECK-NEXT:    ret void
 353 //
 354 // CPP-CHECK-LABEL: @_Z19test_svstnt1_s64_x4u11__SVCount_tPl11svint64x4_t(
 355 // CPP-CHECK-NEXT:  entry:
 356 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V:%.*]], i64 0)
 357 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 2)
 358 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 4)
 359 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 6)
 360 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv2i64(<vscale x 2 x i64> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 361 // CPP-CHECK-NEXT:    ret void
 362 //
 363 void test_svstnt1_s64_x4(svcount_t pn, int64_t *base, svint64x4_t v)
 364 {
 365   return SVE_ACLE_FUNC(svstnt1,_s64_x4,,)(pn, base, v);
 366 }
 367
 368
 369 // CHECK-LABEL: @test_svstnt1_f16_x2(
 370 // CHECK-NEXT:  entry:
 371 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[V:%.*]], i64 0)
 372 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[V]], i64 8)
 373 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv8f16(<vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 374 // CHECK-NEXT:    ret void
 375 //
 376 // CPP-CHECK-LABEL: @_Z19test_svstnt1_f16_x2u11__SVCount_tPDh13svfloat16x2_t(
 377 // CPP-CHECK-NEXT:  entry:
 378 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[V:%.*]], i64 0)
 379 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[V]], i64 8)
 380 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv8f16(<vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 381 // CPP-CHECK-NEXT:    ret void
 382 //
 383 void test_svstnt1_f16_x2(svcount_t pn, float16_t *base, svfloat16x2_t v)
 384 {
 385   return SVE_ACLE_FUNC(svstnt1,_f16_x2,,)(pn, base, v);
 386 }
 387
 388
 389 // CHECK-LABEL: @test_svstnt1_f32_x2(
 390 // CHECK-NEXT:  entry:
 391 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[V:%.*]], i64 0)
 392 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[V]], i64 4)
 393 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv4f32(<vscale x 4 x float> [[TMP0]], <vscale x 4 x float> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 394 // CHECK-NEXT:    ret void
 395 //
 396 // CPP-CHECK-LABEL: @_Z19test_svstnt1_f32_x2u11__SVCount_tPf13svfloat32x2_t(
 397 // CPP-CHECK-NEXT:  entry:
 398 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[V:%.*]], i64 0)
 399 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[V]], i64 4)
 400 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv4f32(<vscale x 4 x float> [[TMP0]], <vscale x 4 x float> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 401 // CPP-CHECK-NEXT:    ret void
 402 //
 403 void test_svstnt1_f32_x2(svcount_t pn, float32_t *base, svfloat32x2_t v)
 404 {
 405   return SVE_ACLE_FUNC(svstnt1,_f32_x2,,)(pn, base, v);
 406 }
 407
 408
 409 // CHECK-LABEL: @test_svstnt1_f64_x2(
 410 // CHECK-NEXT:  entry:
 411 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[V:%.*]], i64 0)
 412 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[V]], i64 2)
 413 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv2f64(<vscale x 2 x double> [[TMP0]], <vscale x 2 x double> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 414 // CHECK-NEXT:    ret void
 415 //
 416 // CPP-CHECK-LABEL: @_Z19test_svstnt1_f64_x2u11__SVCount_tPd13svfloat64x2_t(
 417 // CPP-CHECK-NEXT:  entry:
 418 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[V:%.*]], i64 0)
 419 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[V]], i64 2)
 420 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv2f64(<vscale x 2 x double> [[TMP0]], <vscale x 2 x double> [[TMP1]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 421 // CPP-CHECK-NEXT:    ret void
 422 //
 423 void test_svstnt1_f64_x2(svcount_t pn, float64_t *base, svfloat64x2_t v)
 424 {
 425   return SVE_ACLE_FUNC(svstnt1,_f64_x2,,)(pn, base, v);
 426 }
 427
 428
 429 // CHECK-LABEL: @test_svstnt1_f16_x4(
 430 // CHECK-NEXT:  entry:
 431 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[V:%.*]], i64 0)
 432 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[V]], i64 8)
 433 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[V]], i64 16)
 434 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[V]], i64 24)
 435 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv8f16(<vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 436 // CHECK-NEXT:    ret void
 437 //
 438 // CPP-CHECK-LABEL: @_Z19test_svstnt1_f16_x4u11__SVCount_tPDh13svfloat16x4_t(
 439 // CPP-CHECK-NEXT:  entry:
 440 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[V:%.*]], i64 0)
 441 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[V]], i64 8)
 442 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[V]], i64 16)
 443 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[V]], i64 24)
 444 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv8f16(<vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 445 // CPP-CHECK-NEXT:    ret void
 446 //
 447 void test_svstnt1_f16_x4(svcount_t pn, float16_t *base, svfloat16x4_t v)
 448 {
 449   return SVE_ACLE_FUNC(svstnt1,_f16_x4,,)(pn, base, v);
 450 }
 451
 452
 453 // CHECK-LABEL: @test_svstnt1_f32_x4(
 454 // CHECK-NEXT:  entry:
 455 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[V:%.*]], i64 0)
 456 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[V]], i64 4)
 457 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[V]], i64 8)
 458 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[V]], i64 12)
 459 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv4f32(<vscale x 4 x float> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 460 // CHECK-NEXT:    ret void
 461 //
 462 // CPP-CHECK-LABEL: @_Z19test_svstnt1_f32_x4u11__SVCount_tPf13svfloat32x4_t(
 463 // CPP-CHECK-NEXT:  entry:
 464 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[V:%.*]], i64 0)
 465 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[V]], i64 4)
 466 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[V]], i64 8)
 467 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[V]], i64 12)
 468 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv4f32(<vscale x 4 x float> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 469 // CPP-CHECK-NEXT:    ret void
 470 //
 471 void test_svstnt1_f32_x4(svcount_t pn, float32_t *base, svfloat32x4_t v)
 472 {
 473   return SVE_ACLE_FUNC(svstnt1,_f32_x4,,)(pn, base, v);
 474 }
 475
 476
 477 // CHECK-LABEL: @test_svstnt1_f64_x4(
 478 // CHECK-NEXT:  entry:
 479 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[V:%.*]], i64 0)
 480 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[V]], i64 2)
 481 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[V]], i64 4)
 482 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[V]], i64 6)
 483 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv2f64(<vscale x 2 x double> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 484 // CHECK-NEXT:    ret void
 485 //
 486 // CPP-CHECK-LABEL: @_Z19test_svstnt1_f64_x4u11__SVCount_tPd13svfloat64x4_t(
 487 // CPP-CHECK-NEXT:  entry:
 488 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[V:%.*]], i64 0)
 489 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[V]], i64 2)
 490 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[V]], i64 4)
 491 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[V]], i64 6)
 492 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv2f64(<vscale x 2 x double> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP3]], target("aarch64.svcount") [[PN:%.*]], ptr [[BASE:%.*]])
 493 // CPP-CHECK-NEXT:    ret void
 494 //
 495 void test_svstnt1_f64_x4(svcount_t pn, float64_t *base, svfloat64x4_t v)
 496 {
 497   return SVE_ACLE_FUNC(svstnt1,_f64_x4,,)(pn, base, v);
 498 }
 499
 500
 501 // == VNUM variants ==
 502
 503
 504
 505 // CHECK-LABEL: @test_svstnt1_vnum_u8_x2(
 506 // CHECK-NEXT:  entry:
 507 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 16 x i8>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 508 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[V:%.*]], i64 0)
 509 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[V]], i64 16)
 510 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv16i8(<vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 511 // CHECK-NEXT:    ret void
 512 //
 513 // CPP-CHECK-LABEL: @_Z23test_svstnt1_vnum_u8_x2u11__SVCount_tPhl11svuint8x2_t(
 514 // CPP-CHECK-NEXT:  entry:
 515 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 16 x i8>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 516 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[V:%.*]], i64 0)
 517 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[V]], i64 16)
 518 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv16i8(<vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 519 // CPP-CHECK-NEXT:    ret void
 520 //
 521 void test_svstnt1_vnum_u8_x2(svcount_t pn, uint8_t *base, int64_t vnum, svuint8x2_t v)
 522 {
 523   return SVE_ACLE_FUNC(svstnt1_vnum,_u8_x2,,)(pn, base, vnum, v);
 524 }
 525
 526
 527 // CHECK-LABEL: @test_svstnt1_vnum_u16_x2(
 528 // CHECK-NEXT:  entry:
 529 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 8 x i16>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 530 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[V:%.*]], i64 0)
 531 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[V]], i64 8)
 532 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 533 // CHECK-NEXT:    ret void
 534 //
 535 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_u16_x2u11__SVCount_tPtl12svuint16x2_t(
 536 // CPP-CHECK-NEXT:  entry:
 537 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 8 x i16>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 538 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[V:%.*]], i64 0)
 539 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[V]], i64 8)
 540 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 541 // CPP-CHECK-NEXT:    ret void
 542 //
 543 void test_svstnt1_vnum_u16_x2(svcount_t pn, uint16_t *base, int64_t vnum, svuint16x2_t v)
 544 {
 545   return SVE_ACLE_FUNC(svstnt1_vnum,_u16_x2,,)(pn, base, vnum, v);
 546 }
 547
 548
 549 // CHECK-LABEL: @test_svstnt1_vnum_u32_x2(
 550 // CHECK-NEXT:  entry:
 551 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 4 x i32>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 552 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[V:%.*]], i64 0)
 553 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[V]], i64 4)
 554 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 555 // CHECK-NEXT:    ret void
 556 //
 557 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_u32_x2u11__SVCount_tPjl12svuint32x2_t(
 558 // CPP-CHECK-NEXT:  entry:
 559 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 4 x i32>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 560 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[V:%.*]], i64 0)
 561 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[V]], i64 4)
 562 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 563 // CPP-CHECK-NEXT:    ret void
 564 //
 565 void test_svstnt1_vnum_u32_x2(svcount_t pn, uint32_t *base, int64_t vnum, svuint32x2_t v)
 566 {
 567   return SVE_ACLE_FUNC(svstnt1_vnum,_u32_x2,,)(pn, base, vnum, v);
 568 }
 569
 570
 571 // CHECK-LABEL: @test_svstnt1_vnum_u64_x2(
 572 // CHECK-NEXT:  entry:
 573 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 2 x i64>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 574 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[V:%.*]], i64 0)
 575 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[V]], i64 2)
 576 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 577 // CHECK-NEXT:    ret void
 578 //
 579 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_u64_x2u11__SVCount_tPml12svuint64x2_t(
 580 // CPP-CHECK-NEXT:  entry:
 581 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 2 x i64>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 582 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[V:%.*]], i64 0)
 583 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[V]], i64 2)
 584 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 585 // CPP-CHECK-NEXT:    ret void
 586 //
 587 void test_svstnt1_vnum_u64_x2(svcount_t pn, uint64_t *base, int64_t vnum, svuint64x2_t v)
 588 {
 589   return SVE_ACLE_FUNC(svstnt1_vnum,_u64_x2,,)(pn, base, vnum, v);
 590 }
 591
 592
 593 // CHECK-LABEL: @test_svstnt1_vnum_u8_x4(
 594 // CHECK-NEXT:  entry:
 595 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 16 x i8>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 596 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V:%.*]], i64 0)
 597 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 16)
 598 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 32)
 599 // CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 48)
 600 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv16i8(<vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 601 // CHECK-NEXT:    ret void
 602 //
 603 // CPP-CHECK-LABEL: @_Z23test_svstnt1_vnum_u8_x4u11__SVCount_tPhl11svuint8x4_t(
 604 // CPP-CHECK-NEXT:  entry:
 605 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 16 x i8>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 606 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V:%.*]], i64 0)
 607 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 16)
 608 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 32)
 609 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 48)
 610 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv16i8(<vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 611 // CPP-CHECK-NEXT:    ret void
 612 //
 613 void test_svstnt1_vnum_u8_x4(svcount_t pn, uint8_t *base, int64_t vnum, svuint8x4_t v)
 614 {
 615   return SVE_ACLE_FUNC(svstnt1_vnum,_u8_x4,,)(pn, base, vnum, v);
 616 }
 617
 618
 619 // CHECK-LABEL: @test_svstnt1_vnum_u16_x4(
 620 // CHECK-NEXT:  entry:
 621 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 8 x i16>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 622 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V:%.*]], i64 0)
 623 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 8)
 624 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 16)
 625 // CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 24)
 626 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 627 // CHECK-NEXT:    ret void
 628 //
 629 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_u16_x4u11__SVCount_tPtl12svuint16x4_t(
 630 // CPP-CHECK-NEXT:  entry:
 631 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 8 x i16>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 632 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V:%.*]], i64 0)
 633 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 8)
 634 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 16)
 635 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 24)
 636 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 637 // CPP-CHECK-NEXT:    ret void
 638 //
 639 void test_svstnt1_vnum_u16_x4(svcount_t pn, uint16_t *base, int64_t vnum, svuint16x4_t v)
 640 {
 641   return SVE_ACLE_FUNC(svstnt1_vnum,_u16_x4,,)(pn, base, vnum, v);
 642 }
 643
 644
 645 // CHECK-LABEL: @test_svstnt1_vnum_u32_x4(
 646 // CHECK-NEXT:  entry:
 647 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 4 x i32>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 648 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V:%.*]], i64 0)
 649 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 4)
 650 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 8)
 651 // CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 12)
 652 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 653 // CHECK-NEXT:    ret void
 654 //
 655 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_u32_x4u11__SVCount_tPjl12svuint32x4_t(
 656 // CPP-CHECK-NEXT:  entry:
 657 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 4 x i32>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 658 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V:%.*]], i64 0)
 659 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 4)
 660 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 8)
 661 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 12)
 662 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 663 // CPP-CHECK-NEXT:    ret void
 664 //
 665 void test_svstnt1_vnum_u32_x4(svcount_t pn, uint32_t *base, int64_t vnum, svuint32x4_t v)
 666 {
 667   return SVE_ACLE_FUNC(svstnt1_vnum,_u32_x4,,)(pn, base, vnum, v);
 668 }
 669
 670
 671 // CHECK-LABEL: @test_svstnt1_vnum_u64_x4(
 672 // CHECK-NEXT:  entry:
 673 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 2 x i64>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 674 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V:%.*]], i64 0)
 675 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 2)
 676 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 4)
 677 // CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 6)
 678 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 679 // CHECK-NEXT:    ret void
 680 //
 681 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_u64_x4u11__SVCount_tPml12svuint64x4_t(
 682 // CPP-CHECK-NEXT:  entry:
 683 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 2 x i64>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 684 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V:%.*]], i64 0)
 685 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 2)
 686 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 4)
 687 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 6)
 688 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 689 // CPP-CHECK-NEXT:    ret void
 690 //
 691 void test_svstnt1_vnum_u64_x4(svcount_t pn, uint64_t *base, int64_t vnum, svuint64x4_t v)
 692 {
 693   return SVE_ACLE_FUNC(svstnt1_vnum,_u64_x4,,)(pn, base, vnum, v);
 694 }
 695
 696
 697 // CHECK-LABEL: @test_svstnt1_vnum_s8_x2(
 698 // CHECK-NEXT:  entry:
 699 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 16 x i8>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 700 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[V:%.*]], i64 0)
 701 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[V]], i64 16)
 702 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv16i8(<vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 703 // CHECK-NEXT:    ret void
 704 //
 705 // CPP-CHECK-LABEL: @_Z23test_svstnt1_vnum_s8_x2u11__SVCount_tPal10svint8x2_t(
 706 // CPP-CHECK-NEXT:  entry:
 707 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 16 x i8>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 708 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[V:%.*]], i64 0)
 709 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv32i8(<vscale x 32 x i8> [[V]], i64 16)
 710 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv16i8(<vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 711 // CPP-CHECK-NEXT:    ret void
 712 //
 713 void test_svstnt1_vnum_s8_x2(svcount_t pn, int8_t *base, int64_t vnum, svint8x2_t v)
 714 {
 715   return SVE_ACLE_FUNC(svstnt1_vnum,_s8_x2,,)(pn, base, vnum, v);
 716 }
 717
 718
 719 // CHECK-LABEL: @test_svstnt1_vnum_s16_x2(
 720 // CHECK-NEXT:  entry:
 721 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 8 x i16>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 722 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[V:%.*]], i64 0)
 723 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[V]], i64 8)
 724 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 725 // CHECK-NEXT:    ret void
 726 //
 727 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_s16_x2u11__SVCount_tPsl11svint16x2_t(
 728 // CPP-CHECK-NEXT:  entry:
 729 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 8 x i16>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 730 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[V:%.*]], i64 0)
 731 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[V]], i64 8)
 732 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 733 // CPP-CHECK-NEXT:    ret void
 734 //
 735 void test_svstnt1_vnum_s16_x2(svcount_t pn, int16_t *base, int64_t vnum, svint16x2_t v)
 736 {
 737   return SVE_ACLE_FUNC(svstnt1_vnum,_s16_x2,,)(pn, base, vnum, v);
 738 }
 739
 740
 741 // CHECK-LABEL: @test_svstnt1_vnum_s32_x2(
 742 // CHECK-NEXT:  entry:
 743 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 4 x i32>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 744 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[V:%.*]], i64 0)
 745 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[V]], i64 4)
 746 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 747 // CHECK-NEXT:    ret void
 748 //
 749 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_s32_x2u11__SVCount_tPil11svint32x2_t(
 750 // CPP-CHECK-NEXT:  entry:
 751 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 4 x i32>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 752 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[V:%.*]], i64 0)
 753 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv8i32(<vscale x 8 x i32> [[V]], i64 4)
 754 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 755 // CPP-CHECK-NEXT:    ret void
 756 //
 757 void test_svstnt1_vnum_s32_x2(svcount_t pn, int32_t *base, int64_t vnum, svint32x2_t v)
 758 {
 759   return SVE_ACLE_FUNC(svstnt1_vnum,_s32_x2,,)(pn, base, vnum, v);
 760 }
 761
 762
 763 // CHECK-LABEL: @test_svstnt1_vnum_s64_x2(
 764 // CHECK-NEXT:  entry:
 765 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 2 x i64>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 766 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[V:%.*]], i64 0)
 767 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[V]], i64 2)
 768 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 769 // CHECK-NEXT:    ret void
 770 //
 771 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_s64_x2u11__SVCount_tPll11svint64x2_t(
 772 // CPP-CHECK-NEXT:  entry:
 773 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 2 x i64>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 774 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[V:%.*]], i64 0)
 775 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv4i64(<vscale x 4 x i64> [[V]], i64 2)
 776 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 777 // CPP-CHECK-NEXT:    ret void
 778 //
 779 void test_svstnt1_vnum_s64_x2(svcount_t pn, int64_t *base, int64_t vnum, svint64x2_t v)
 780 {
 781   return SVE_ACLE_FUNC(svstnt1_vnum,_s64_x2,,)(pn, base, vnum, v);
 782 }
 783
 784
 785 // CHECK-LABEL: @test_svstnt1_vnum_s8_x4(
 786 // CHECK-NEXT:  entry:
 787 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 16 x i8>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 788 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V:%.*]], i64 0)
 789 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 16)
 790 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 32)
 791 // CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 48)
 792 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv16i8(<vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 793 // CHECK-NEXT:    ret void
 794 //
 795 // CPP-CHECK-LABEL: @_Z23test_svstnt1_vnum_s8_x4u11__SVCount_tPal10svint8x4_t(
 796 // CPP-CHECK-NEXT:  entry:
 797 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 16 x i8>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 798 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V:%.*]], i64 0)
 799 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 16)
 800 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 32)
 801 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[V]], i64 48)
 802 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv16i8(<vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 803 // CPP-CHECK-NEXT:    ret void
 804 //
 805 void test_svstnt1_vnum_s8_x4(svcount_t pn, int8_t *base, int64_t vnum, svint8x4_t v)
 806 {
 807   return SVE_ACLE_FUNC(svstnt1_vnum,_s8_x4,,)(pn, base, vnum, v);
 808 }
 809
 810
 811 // CHECK-LABEL: @test_svstnt1_vnum_s16_x4(
 812 // CHECK-NEXT:  entry:
 813 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 8 x i16>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 814 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V:%.*]], i64 0)
 815 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 8)
 816 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 16)
 817 // CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 24)
 818 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 819 // CHECK-NEXT:    ret void
 820 //
 821 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_s16_x4u11__SVCount_tPsl11svint16x4_t(
 822 // CPP-CHECK-NEXT:  entry:
 823 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 8 x i16>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 824 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V:%.*]], i64 0)
 825 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 8)
 826 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 16)
 827 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[V]], i64 24)
 828 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 829 // CPP-CHECK-NEXT:    ret void
 830 //
 831 void test_svstnt1_vnum_s16_x4(svcount_t pn, int16_t *base, int64_t vnum, svint16x4_t v)
 832 {
 833   return SVE_ACLE_FUNC(svstnt1_vnum,_s16_x4,,)(pn, base, vnum, v);
 834 }
 835
 836
 837 // CHECK-LABEL: @test_svstnt1_vnum_s32_x4(
 838 // CHECK-NEXT:  entry:
 839 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 4 x i32>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 840 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V:%.*]], i64 0)
 841 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 4)
 842 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 8)
 843 // CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 12)
 844 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 845 // CHECK-NEXT:    ret void
 846 //
 847 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_s32_x4u11__SVCount_tPil11svint32x4_t(
 848 // CPP-CHECK-NEXT:  entry:
 849 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 4 x i32>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 850 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V:%.*]], i64 0)
 851 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 4)
 852 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 8)
 853 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[V]], i64 12)
 854 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 855 // CPP-CHECK-NEXT:    ret void
 856 //
 857 void test_svstnt1_vnum_s32_x4(svcount_t pn, int32_t *base, int64_t vnum, svint32x4_t v)
 858 {
 859   return SVE_ACLE_FUNC(svstnt1_vnum,_s32_x4,,)(pn, base, vnum, v);
 860 }
 861
 862
 863 // CHECK-LABEL: @test_svstnt1_vnum_s64_x4(
 864 // CHECK-NEXT:  entry:
 865 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 2 x i64>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 866 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V:%.*]], i64 0)
 867 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 2)
 868 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 4)
 869 // CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 6)
 870 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 871 // CHECK-NEXT:    ret void
 872 //
 873 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_s64_x4u11__SVCount_tPll11svint64x4_t(
 874 // CPP-CHECK-NEXT:  entry:
 875 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 2 x i64>, ptr [[BASE:%.*]], i64 [[VNUM:%.*]]
 876 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V:%.*]], i64 0)
 877 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 2)
 878 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 4)
 879 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[V]], i64 6)
 880 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 881 // CPP-CHECK-NEXT:    ret void
 882 //
 883 void test_svstnt1_vnum_s64_x4(svcount_t pn, int64_t *base, int64_t vnum, svint64x4_t v)
 884 {
 885   return SVE_ACLE_FUNC(svstnt1_vnum,_s64_x4,,)(pn, base, vnum, v);
 886 }
 887
 888
 889 // CHECK-LABEL: @test_svstnt1_vnum_f16_x2(
 890 // CHECK-NEXT:  entry:
 891 // CHECK-NEXT:    [[CONV:%.*]] = fptosi double [[VNUM:%.*]] to i64
 892 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 8 x half>, ptr [[BASE:%.*]], i64 [[CONV]]
 893 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[V:%.*]], i64 0)
 894 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[V]], i64 8)
 895 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv8f16(<vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 896 // CHECK-NEXT:    ret void
 897 //
 898 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_f16_x2u11__SVCount_tPDhd13svfloat16x2_t(
 899 // CPP-CHECK-NEXT:  entry:
 900 // CPP-CHECK-NEXT:    [[CONV:%.*]] = fptosi double [[VNUM:%.*]] to i64
 901 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 8 x half>, ptr [[BASE:%.*]], i64 [[CONV]]
 902 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[V:%.*]], i64 0)
 903 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[V]], i64 8)
 904 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv8f16(<vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 905 // CPP-CHECK-NEXT:    ret void
 906 //
 907 void test_svstnt1_vnum_f16_x2(svcount_t pn, float16_t *base, float64_t vnum, svfloat16x2_t v)
 908 {
 909   return SVE_ACLE_FUNC(svstnt1_vnum,_f16_x2,,)(pn, base, vnum, v);
 910 }
 911
 912
 913 // CHECK-LABEL: @test_svstnt1_vnum_f32_x2(
 914 // CHECK-NEXT:  entry:
 915 // CHECK-NEXT:    [[CONV:%.*]] = fptosi double [[VNUM:%.*]] to i64
 916 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 4 x float>, ptr [[BASE:%.*]], i64 [[CONV]]
 917 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[V:%.*]], i64 0)
 918 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[V]], i64 4)
 919 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv4f32(<vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 920 // CHECK-NEXT:    ret void
 921 //
 922 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_f32_x2u11__SVCount_tPfd13svfloat32x2_t(
 923 // CPP-CHECK-NEXT:  entry:
 924 // CPP-CHECK-NEXT:    [[CONV:%.*]] = fptosi double [[VNUM:%.*]] to i64
 925 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 4 x float>, ptr [[BASE:%.*]], i64 [[CONV]]
 926 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[V:%.*]], i64 0)
 927 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> [[V]], i64 4)
 928 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv4f32(<vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 929 // CPP-CHECK-NEXT:    ret void
 930 //
 931 void test_svstnt1_vnum_f32_x2(svcount_t pn, float32_t *base, float64_t vnum, svfloat32x2_t v)
 932 {
 933   return SVE_ACLE_FUNC(svstnt1_vnum,_f32_x2,,)(pn, base, vnum, v);
 934 }
 935
 936
 937 // CHECK-LABEL: @test_svstnt1_vnum_f64_x2(
 938 // CHECK-NEXT:  entry:
 939 // CHECK-NEXT:    [[CONV:%.*]] = fptosi double [[VNUM:%.*]] to i64
 940 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 2 x double>, ptr [[BASE:%.*]], i64 [[CONV]]
 941 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[V:%.*]], i64 0)
 942 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[V]], i64 2)
 943 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv2f64(<vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 944 // CHECK-NEXT:    ret void
 945 //
 946 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_f64_x2u11__SVCount_tPdd13svfloat64x2_t(
 947 // CPP-CHECK-NEXT:  entry:
 948 // CPP-CHECK-NEXT:    [[CONV:%.*]] = fptosi double [[VNUM:%.*]] to i64
 949 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 2 x double>, ptr [[BASE:%.*]], i64 [[CONV]]
 950 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[V:%.*]], i64 0)
 951 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv4f64(<vscale x 4 x double> [[V]], i64 2)
 952 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x2.nxv2f64(<vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 953 // CPP-CHECK-NEXT:    ret void
 954 //
 955 void test_svstnt1_vnum_f64_x2(svcount_t pn, float64_t *base, float64_t vnum, svfloat64x2_t v)
 956 {
 957   return SVE_ACLE_FUNC(svstnt1_vnum,_f64_x2,,)(pn, base, vnum, v);
 958 }
 959
 960
 961 // CHECK-LABEL: @test_svstnt1_vnum_f16_x4(
 962 // CHECK-NEXT:  entry:
 963 // CHECK-NEXT:    [[CONV:%.*]] = fptosi double [[VNUM:%.*]] to i64
 964 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 8 x half>, ptr [[BASE:%.*]], i64 [[CONV]]
 965 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[V:%.*]], i64 0)
 966 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[V]], i64 8)
 967 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[V]], i64 16)
 968 // CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[V]], i64 24)
 969 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv8f16(<vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 970 // CHECK-NEXT:    ret void
 971 //
 972 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_f16_x4u11__SVCount_tPDhd13svfloat16x4_t(
 973 // CPP-CHECK-NEXT:  entry:
 974 // CPP-CHECK-NEXT:    [[CONV:%.*]] = fptosi double [[VNUM:%.*]] to i64
 975 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 8 x half>, ptr [[BASE:%.*]], i64 [[CONV]]
 976 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[V:%.*]], i64 0)
 977 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[V]], i64 8)
 978 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[V]], i64 16)
 979 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[V]], i64 24)
 980 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv8f16(<vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 981 // CPP-CHECK-NEXT:    ret void
 982 //
 983 void test_svstnt1_vnum_f16_x4(svcount_t pn, float16_t *base, float64_t vnum, svfloat16x4_t v)
 984 {
 985   return SVE_ACLE_FUNC(svstnt1_vnum,_f16_x4,,)(pn, base, vnum, v);
 986 }
 987
 988
 989 // CHECK-LABEL: @test_svstnt1_vnum_f32_x4(
 990 // CHECK-NEXT:  entry:
 991 // CHECK-NEXT:    [[CONV:%.*]] = fptosi double [[VNUM:%.*]] to i64
 992 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 4 x float>, ptr [[BASE:%.*]], i64 [[CONV]]
 993 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[V:%.*]], i64 0)
 994 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[V]], i64 4)
 995 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[V]], i64 8)
 996 // CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[V]], i64 12)
 997 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv4f32(<vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
 998 // CHECK-NEXT:    ret void
 999 //
1000 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_f32_x4u11__SVCount_tPfd13svfloat32x4_t(
1001 // CPP-CHECK-NEXT:  entry:
1002 // CPP-CHECK-NEXT:    [[CONV:%.*]] = fptosi double [[VNUM:%.*]] to i64
1003 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 4 x float>, ptr [[BASE:%.*]], i64 [[CONV]]
1004 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[V:%.*]], i64 0)
1005 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[V]], i64 4)
1006 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[V]], i64 8)
1007 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[V]], i64 12)
1008 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv4f32(<vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
1009 // CPP-CHECK-NEXT:    ret void
1010 //
1011 void test_svstnt1_vnum_f32_x4(svcount_t pn, float32_t *base, float64_t vnum, svfloat32x4_t v)
1012 {
1013   return SVE_ACLE_FUNC(svstnt1_vnum,_f32_x4,,)(pn, base, vnum, v);
1014 }
1015
1016
1017 // CHECK-LABEL: @test_svstnt1_vnum_f64_x4(
1018 // CHECK-NEXT:  entry:
1019 // CHECK-NEXT:    [[CONV:%.*]] = fptosi double [[VNUM:%.*]] to i64
1020 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 2 x double>, ptr [[BASE:%.*]], i64 [[CONV]]
1021 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[V:%.*]], i64 0)
1022 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[V]], i64 2)
1023 // CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[V]], i64 4)
1024 // CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[V]], i64 6)
1025 // CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv2f64(<vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
1026 // CHECK-NEXT:    ret void
1027 //
1028 // CPP-CHECK-LABEL: @_Z24test_svstnt1_vnum_f64_x4u11__SVCount_tPdd13svfloat64x4_t(
1029 // CPP-CHECK-NEXT:  entry:
1030 // CPP-CHECK-NEXT:    [[CONV:%.*]] = fptosi double [[VNUM:%.*]] to i64
1031 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = getelementptr <vscale x 2 x double>, ptr [[BASE:%.*]], i64 [[CONV]]
1032 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[V:%.*]], i64 0)
1033 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[V]], i64 2)
1034 // CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[V]], i64 4)
1035 // CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[V]], i64 6)
1036 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sve.stnt1.pn.x4.nxv2f64(<vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], target("aarch64.svcount") [[PN:%.*]], ptr [[TMP0]])
1037 // CPP-CHECK-NEXT:    ret void
1038 //
1039 void test_svstnt1_vnum_f64_x4(svcount_t pn, float64_t *base, float64_t vnum, svfloat64x4_t v)
1040 {
1041   return SVE_ACLE_FUNC(svstnt1_vnum,_f64_x4,,)(pn, base, vnum, v);
1042 }