clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4.c

   1 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
   2 // REQUIRES: aarch64-registered-target
   3 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s
   4 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK
   5 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s
   6 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK
   7 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -S -disable-O0-optnone -Werror -Wall -o /dev/null %s
   8 #include <arm_sve.h>
   9
  10 #ifdef SVE_OVERLOADED_FORMS
  11 // A simple used,unused... macro, long enough to represent any SVE builtin.
  12 #define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
  13 #else
  14 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
  15 #endif
  16
  17 // NOTE: For these tests clang converts the struct parameter into
  18 // several parameters, one for each member of the original struct.
  19 // CHECK-LABEL: @test_svget4_s8(
  20 // CHECK-NEXT:  entry:
  21 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[TUPLE:%.*]], i64 0)
  22 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
  23 //
  24 // CPP-CHECK-LABEL: @_Z14test_svget4_s810svint8x4_t(
  25 // CPP-CHECK-NEXT:  entry:
  26 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[TUPLE:%.*]], i64 0)
  27 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
  28 //
  29 svint8_t test_svget4_s8(svint8x4_t tuple)
  30 {
  31   return SVE_ACLE_FUNC(svget4,_s8,,)(tuple, 0);
  32 }
  33
  34 // CHECK-LABEL: @test_svget4_s16(
  35 // CHECK-NEXT:  entry:
  36 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[TUPLE:%.*]], i64 16)
  37 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
  38 //
  39 // CPP-CHECK-LABEL: @_Z15test_svget4_s1611svint16x4_t(
  40 // CPP-CHECK-NEXT:  entry:
  41 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[TUPLE:%.*]], i64 16)
  42 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
  43 //
  44 svint16_t test_svget4_s16(svint16x4_t tuple)
  45 {
  46   return SVE_ACLE_FUNC(svget4,_s16,,)(tuple, 2);
  47 }
  48
  49 // CHECK-LABEL: @test_svget4_s32(
  50 // CHECK-NEXT:  entry:
  51 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[TUPLE:%.*]], i64 8)
  52 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
  53 //
  54 // CPP-CHECK-LABEL: @_Z15test_svget4_s3211svint32x4_t(
  55 // CPP-CHECK-NEXT:  entry:
  56 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[TUPLE:%.*]], i64 8)
  57 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
  58 //
  59 svint32_t test_svget4_s32(svint32x4_t tuple)
  60 {
  61   return SVE_ACLE_FUNC(svget4,_s32,,)(tuple, 2);
  62 }
  63
  64 // CHECK-LABEL: @test_svget4_s64(
  65 // CHECK-NEXT:  entry:
  66 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[TUPLE:%.*]], i64 6)
  67 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
  68 //
  69 // CPP-CHECK-LABEL: @_Z15test_svget4_s6411svint64x4_t(
  70 // CPP-CHECK-NEXT:  entry:
  71 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[TUPLE:%.*]], i64 6)
  72 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
  73 //
  74 svint64_t test_svget4_s64(svint64x4_t tuple)
  75 {
  76   return SVE_ACLE_FUNC(svget4,_s64,,)(tuple, 3);
  77 }
  78
  79 // CHECK-LABEL: @test_svget4_u8(
  80 // CHECK-NEXT:  entry:
  81 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[TUPLE:%.*]], i64 32)
  82 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
  83 //
  84 // CPP-CHECK-LABEL: @_Z14test_svget4_u811svuint8x4_t(
  85 // CPP-CHECK-NEXT:  entry:
  86 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv64i8(<vscale x 64 x i8> [[TUPLE:%.*]], i64 32)
  87 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
  88 //
  89 svuint8_t test_svget4_u8(svuint8x4_t tuple)
  90 {
  91   return SVE_ACLE_FUNC(svget4,_u8,,)(tuple, 2);
  92 }
  93
  94 // CHECK-LABEL: @test_svget4_u16(
  95 // CHECK-NEXT:  entry:
  96 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[TUPLE:%.*]], i64 24)
  97 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
  98 //
  99 // CPP-CHECK-LABEL: @_Z15test_svget4_u1612svuint16x4_t(
 100 // CPP-CHECK-NEXT:  entry:
 101 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[TUPLE:%.*]], i64 24)
 102 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 103 //
 104 svuint16_t test_svget4_u16(svuint16x4_t tuple)
 105 {
 106   return SVE_ACLE_FUNC(svget4,_u16,,)(tuple, 3);
 107 }
 108
 109 // CHECK-LABEL: @test_svget4_u32(
 110 // CHECK-NEXT:  entry:
 111 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[TUPLE:%.*]], i64 0)
 112 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 113 //
 114 // CPP-CHECK-LABEL: @_Z15test_svget4_u3212svuint32x4_t(
 115 // CPP-CHECK-NEXT:  entry:
 116 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> [[TUPLE:%.*]], i64 0)
 117 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 118 //
 119 svuint32_t test_svget4_u32(svuint32x4_t tuple)
 120 {
 121   return SVE_ACLE_FUNC(svget4,_u32,,)(tuple, 0);
 122 }
 123
 124 // CHECK-LABEL: @test_svget4_u64(
 125 // CHECK-NEXT:  entry:
 126 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[TUPLE:%.*]], i64 6)
 127 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 128 //
 129 // CPP-CHECK-LABEL: @_Z15test_svget4_u6412svuint64x4_t(
 130 // CPP-CHECK-NEXT:  entry:
 131 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.vector.extract.nxv2i64.nxv8i64(<vscale x 8 x i64> [[TUPLE:%.*]], i64 6)
 132 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 133 //
 134 svuint64_t test_svget4_u64(svuint64x4_t tuple)
 135 {
 136   return SVE_ACLE_FUNC(svget4,_u64,,)(tuple, 3);
 137 }
 138
 139 // CHECK-LABEL: @test_svget4_f16(
 140 // CHECK-NEXT:  entry:
 141 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[TUPLE:%.*]], i64 16)
 142 // CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 143 //
 144 // CPP-CHECK-LABEL: @_Z15test_svget4_f1613svfloat16x4_t(
 145 // CPP-CHECK-NEXT:  entry:
 146 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[TUPLE:%.*]], i64 16)
 147 // CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 148 //
 149 svfloat16_t test_svget4_f16(svfloat16x4_t tuple)
 150 {
 151   return SVE_ACLE_FUNC(svget4,_f16,,)(tuple, 2);
 152 }
 153
 154 // CHECK-LABEL: @test_svget4_f32(
 155 // CHECK-NEXT:  entry:
 156 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[TUPLE:%.*]], i64 0)
 157 // CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 158 //
 159 // CPP-CHECK-LABEL: @_Z15test_svget4_f3213svfloat32x4_t(
 160 // CPP-CHECK-NEXT:  entry:
 161 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv16f32(<vscale x 16 x float> [[TUPLE:%.*]], i64 0)
 162 // CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 163 //
 164 svfloat32_t test_svget4_f32(svfloat32x4_t tuple)
 165 {
 166   return SVE_ACLE_FUNC(svget4,_f32,,)(tuple, 0);
 167 }
 168
 169 // CHECK-LABEL: @test_svget4_f64(
 170 // CHECK-NEXT:  entry:
 171 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[TUPLE:%.*]], i64 4)
 172 // CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 173 //
 174 // CPP-CHECK-LABEL: @_Z15test_svget4_f6413svfloat64x4_t(
 175 // CPP-CHECK-NEXT:  entry:
 176 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.vector.extract.nxv2f64.nxv8f64(<vscale x 8 x double> [[TUPLE:%.*]], i64 4)
 177 // CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 178 //
 179 svfloat64_t test_svget4_f64(svfloat64x4_t tuple)
 180 {
 181   return SVE_ACLE_FUNC(svget4,_f64,,)(tuple, 2);
 182 }