clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3.c

   1 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
   2 // 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
   3 // 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
   4 // 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
   5 // 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
   6
   7 // REQUIRES: aarch64-registered-target
   8
   9 #include <arm_sve.h>
  10
  11 #ifdef SVE_OVERLOADED_FORMS
  12 // A simple used,unused... macro, long enough to represent any SVE builtin.
  13 #define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
  14 #else
  15 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
  16 #endif
  17
  18
  19 // NOTE: For these tests clang converts the struct parameter into
  20 // several parameters, one for each member of the original struct.
  21 // CHECK-LABEL: @test_svset3_s8(
  22 // CHECK-NEXT:  entry:
  23 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TUPLE:%.*]], <vscale x 16 x i8> [[X:%.*]], i64 16)
  24 // CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
  25 //
  26 // CPP-CHECK-LABEL: @_Z14test_svset3_s810svint8x3_tu10__SVInt8_t(
  27 // CPP-CHECK-NEXT:  entry:
  28 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TUPLE:%.*]], <vscale x 16 x i8> [[X:%.*]], i64 16)
  29 // CPP-CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
  30 //
  31 svint8x3_t test_svset3_s8(svint8x3_t tuple, svint8_t x)
  32 {
  33   return SVE_ACLE_FUNC(svset3,_s8,,)(tuple, 1, x);
  34 }
  35
  36 // CHECK-LABEL: @test_svset3_s16(
  37 // CHECK-NEXT:  entry:
  38 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TUPLE:%.*]], <vscale x 8 x i16> [[X:%.*]], i64 16)
  39 // CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
  40 //
  41 // CPP-CHECK-LABEL: @_Z15test_svset3_s1611svint16x3_tu11__SVInt16_t(
  42 // CPP-CHECK-NEXT:  entry:
  43 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TUPLE:%.*]], <vscale x 8 x i16> [[X:%.*]], i64 16)
  44 // CPP-CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
  45 //
  46 svint16x3_t test_svset3_s16(svint16x3_t tuple, svint16_t x)
  47 {
  48   return SVE_ACLE_FUNC(svset3,_s16,,)(tuple, 2, x);
  49 }
  50
  51 // CHECK-LABEL: @test_svset3_s32(
  52 // CHECK-NEXT:  entry:
  53 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TUPLE:%.*]], <vscale x 4 x i32> [[X:%.*]], i64 0)
  54 // CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
  55 //
  56 // CPP-CHECK-LABEL: @_Z15test_svset3_s3211svint32x3_tu11__SVInt32_t(
  57 // CPP-CHECK-NEXT:  entry:
  58 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TUPLE:%.*]], <vscale x 4 x i32> [[X:%.*]], i64 0)
  59 // CPP-CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
  60 //
  61 svint32x3_t test_svset3_s32(svint32x3_t tuple, svint32_t x)
  62 {
  63   return SVE_ACLE_FUNC(svset3,_s32,,)(tuple, 0, x);
  64 }
  65
  66 // CHECK-LABEL: @test_svset3_s64(
  67 // CHECK-NEXT:  entry:
  68 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TUPLE:%.*]], <vscale x 2 x i64> [[X:%.*]], i64 2)
  69 // CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
  70 //
  71 // CPP-CHECK-LABEL: @_Z15test_svset3_s6411svint64x3_tu11__SVInt64_t(
  72 // CPP-CHECK-NEXT:  entry:
  73 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TUPLE:%.*]], <vscale x 2 x i64> [[X:%.*]], i64 2)
  74 // CPP-CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
  75 //
  76 svint64x3_t test_svset3_s64(svint64x3_t tuple, svint64_t x)
  77 {
  78   return SVE_ACLE_FUNC(svset3,_s64,,)(tuple, 1, x);
  79 }
  80
  81 // CHECK-LABEL: @test_svset3_u8(
  82 // CHECK-NEXT:  entry:
  83 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TUPLE:%.*]], <vscale x 16 x i8> [[X:%.*]], i64 32)
  84 // CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
  85 //
  86 // CPP-CHECK-LABEL: @_Z14test_svset3_u811svuint8x3_tu11__SVUint8_t(
  87 // CPP-CHECK-NEXT:  entry:
  88 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 48 x i8> @llvm.vector.insert.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TUPLE:%.*]], <vscale x 16 x i8> [[X:%.*]], i64 32)
  89 // CPP-CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
  90 //
  91 svuint8x3_t test_svset3_u8(svuint8x3_t tuple, svuint8_t x)
  92 {
  93   return SVE_ACLE_FUNC(svset3,_u8,,)(tuple, 2, x);
  94 }
  95
  96 // CHECK-LABEL: @test_svset3_u16(
  97 // CHECK-NEXT:  entry:
  98 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TUPLE:%.*]], <vscale x 8 x i16> [[X:%.*]], i64 0)
  99 // CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
 100 //
 101 // CPP-CHECK-LABEL: @_Z15test_svset3_u1612svuint16x3_tu12__SVUint16_t(
 102 // CPP-CHECK-NEXT:  entry:
 103 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x i16> @llvm.vector.insert.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TUPLE:%.*]], <vscale x 8 x i16> [[X:%.*]], i64 0)
 104 // CPP-CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
 105 //
 106 svuint16x3_t test_svset3_u16(svuint16x3_t tuple, svuint16_t x)
 107 {
 108   return SVE_ACLE_FUNC(svset3,_u16,,)(tuple, 0, x);
 109 }
 110
 111 // CHECK-LABEL: @test_svset3_u32(
 112 // CHECK-NEXT:  entry:
 113 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TUPLE:%.*]], <vscale x 4 x i32> [[X:%.*]], i64 4)
 114 // CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
 115 //
 116 // CPP-CHECK-LABEL: @_Z15test_svset3_u3212svuint32x3_tu12__SVUint32_t(
 117 // CPP-CHECK-NEXT:  entry:
 118 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x i32> @llvm.vector.insert.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TUPLE:%.*]], <vscale x 4 x i32> [[X:%.*]], i64 4)
 119 // CPP-CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
 120 //
 121 svuint32x3_t test_svset3_u32(svuint32x3_t tuple, svuint32_t x)
 122 {
 123   return SVE_ACLE_FUNC(svset3,_u32,,)(tuple, 1, x);
 124 }
 125
 126 // CHECK-LABEL: @test_svset3_u64(
 127 // CHECK-NEXT:  entry:
 128 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TUPLE:%.*]], <vscale x 2 x i64> [[X:%.*]], i64 4)
 129 // CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
 130 //
 131 // CPP-CHECK-LABEL: @_Z15test_svset3_u6412svuint64x3_tu12__SVUint64_t(
 132 // CPP-CHECK-NEXT:  entry:
 133 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x i64> @llvm.vector.insert.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TUPLE:%.*]], <vscale x 2 x i64> [[X:%.*]], i64 4)
 134 // CPP-CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
 135 //
 136 svuint64x3_t test_svset3_u64(svuint64x3_t tuple, svuint64_t x)
 137 {
 138   return SVE_ACLE_FUNC(svset3,_u64,,)(tuple, 2, x);
 139 }
 140
 141 // CHECK-LABEL: @test_svset3_f16(
 142 // CHECK-NEXT:  entry:
 143 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x half> @llvm.vector.insert.nxv24f16.nxv8f16(<vscale x 24 x half> [[TUPLE:%.*]], <vscale x 8 x half> [[X:%.*]], i64 0)
 144 // CHECK-NEXT:    ret <vscale x 24 x half> [[TMP0]]
 145 //
 146 // CPP-CHECK-LABEL: @_Z15test_svset3_f1613svfloat16x3_tu13__SVFloat16_t(
 147 // CPP-CHECK-NEXT:  entry:
 148 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 24 x half> @llvm.vector.insert.nxv24f16.nxv8f16(<vscale x 24 x half> [[TUPLE:%.*]], <vscale x 8 x half> [[X:%.*]], i64 0)
 149 // CPP-CHECK-NEXT:    ret <vscale x 24 x half> [[TMP0]]
 150 //
 151 svfloat16x3_t test_svset3_f16(svfloat16x3_t tuple, svfloat16_t x)
 152 {
 153   return SVE_ACLE_FUNC(svset3,_f16,,)(tuple, 0, x);
 154 }
 155
 156 // CHECK-LABEL: @test_svset3_f32(
 157 // CHECK-NEXT:  entry:
 158 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x float> @llvm.vector.insert.nxv12f32.nxv4f32(<vscale x 12 x float> [[TUPLE:%.*]], <vscale x 4 x float> [[X:%.*]], i64 4)
 159 // CHECK-NEXT:    ret <vscale x 12 x float> [[TMP0]]
 160 //
 161 // CPP-CHECK-LABEL: @_Z15test_svset3_f3213svfloat32x3_tu13__SVFloat32_t(
 162 // CPP-CHECK-NEXT:  entry:
 163 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 12 x float> @llvm.vector.insert.nxv12f32.nxv4f32(<vscale x 12 x float> [[TUPLE:%.*]], <vscale x 4 x float> [[X:%.*]], i64 4)
 164 // CPP-CHECK-NEXT:    ret <vscale x 12 x float> [[TMP0]]
 165 //
 166 svfloat32x3_t test_svset3_f32(svfloat32x3_t tuple, svfloat32_t x)
 167 {
 168   return SVE_ACLE_FUNC(svset3,_f32,,)(tuple, 1, x);
 169 }
 170
 171 // CHECK-LABEL: @test_svset3_f64(
 172 // CHECK-NEXT:  entry:
 173 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x double> @llvm.vector.insert.nxv6f64.nxv2f64(<vscale x 6 x double> [[TUPLE:%.*]], <vscale x 2 x double> [[X:%.*]], i64 4)
 174 // CHECK-NEXT:    ret <vscale x 6 x double> [[TMP0]]
 175 //
 176 // CPP-CHECK-LABEL: @_Z15test_svset3_f6413svfloat64x3_tu13__SVFloat64_t(
 177 // CPP-CHECK-NEXT:  entry:
 178 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 6 x double> @llvm.vector.insert.nxv6f64.nxv2f64(<vscale x 6 x double> [[TUPLE:%.*]], <vscale x 2 x double> [[X:%.*]], i64 4)
 179 // CPP-CHECK-NEXT:    ret <vscale x 6 x double> [[TMP0]]
 180 //
 181 svfloat64x3_t test_svset3_f64(svfloat64x3_t tuple, svfloat64_t x)
 182 {
 183   return SVE_ACLE_FUNC(svset3,_f64,,)(tuple, 2, x);
 184 }