clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mla.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 +sve2 -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
   3 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
   4 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
   5 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | 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 // CHECK-LABEL: @test_svmla_lane_s16(
  19 // CHECK-NEXT:  entry:
  20 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
  21 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
  22 //
  23 // CPP-CHECK-LABEL: @_Z19test_svmla_lane_s16u11__SVInt16_tS_S_(
  24 // CPP-CHECK-NEXT:  entry:
  25 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
  26 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
  27 //
  28 svint16_t test_svmla_lane_s16(svint16_t op1, svint16_t op2, svint16_t op3)
  29 {
  30   return SVE_ACLE_FUNC(svmla_lane,_s16,,)(op1, op2, op3, 0);
  31 }
  32
  33 // CHECK-LABEL: @test_svmla_lane_s16_1(
  34 // CHECK-NEXT:  entry:
  35 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
  36 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
  37 //
  38 // CPP-CHECK-LABEL: @_Z21test_svmla_lane_s16_1u11__SVInt16_tS_S_(
  39 // CPP-CHECK-NEXT:  entry:
  40 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
  41 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
  42 //
  43 svint16_t test_svmla_lane_s16_1(svint16_t op1, svint16_t op2, svint16_t op3)
  44 {
  45   return SVE_ACLE_FUNC(svmla_lane,_s16,,)(op1, op2, op3, 7);
  46 }
  47
  48 // CHECK-LABEL: @test_svmla_lane_s32(
  49 // CHECK-NEXT:  entry:
  50 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.mla.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
  51 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
  52 //
  53 // CPP-CHECK-LABEL: @_Z19test_svmla_lane_s32u11__SVInt32_tS_S_(
  54 // CPP-CHECK-NEXT:  entry:
  55 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.mla.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
  56 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
  57 //
  58 svint32_t test_svmla_lane_s32(svint32_t op1, svint32_t op2, svint32_t op3)
  59 {
  60   return SVE_ACLE_FUNC(svmla_lane,_s32,,)(op1, op2, op3, 0);
  61 }
  62
  63 // CHECK-LABEL: @test_svmla_lane_s32_1(
  64 // CHECK-NEXT:  entry:
  65 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.mla.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
  66 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
  67 //
  68 // CPP-CHECK-LABEL: @_Z21test_svmla_lane_s32_1u11__SVInt32_tS_S_(
  69 // CPP-CHECK-NEXT:  entry:
  70 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.mla.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
  71 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
  72 //
  73 svint32_t test_svmla_lane_s32_1(svint32_t op1, svint32_t op2, svint32_t op3)
  74 {
  75   return SVE_ACLE_FUNC(svmla_lane,_s32,,)(op1, op2, op3, 3);
  76 }
  77
  78 // CHECK-LABEL: @test_svmla_lane_s64(
  79 // CHECK-NEXT:  entry:
  80 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
  81 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
  82 //
  83 // CPP-CHECK-LABEL: @_Z19test_svmla_lane_s64u11__SVInt64_tS_S_(
  84 // CPP-CHECK-NEXT:  entry:
  85 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
  86 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
  87 //
  88 svint64_t test_svmla_lane_s64(svint64_t op1, svint64_t op2, svint64_t op3)
  89 {
  90   return SVE_ACLE_FUNC(svmla_lane,_s64,,)(op1, op2, op3, 0);
  91 }
  92
  93 // CHECK-LABEL: @test_svmla_lane_s64_1(
  94 // CHECK-NEXT:  entry:
  95 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 1)
  96 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
  97 //
  98 // CPP-CHECK-LABEL: @_Z21test_svmla_lane_s64_1u11__SVInt64_tS_S_(
  99 // CPP-CHECK-NEXT:  entry:
 100 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 1)
 101 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 102 //
 103 svint64_t test_svmla_lane_s64_1(svint64_t op1, svint64_t op2, svint64_t op3)
 104 {
 105   return SVE_ACLE_FUNC(svmla_lane,_s64,,)(op1, op2, op3, 1);
 106 }
 107
 108 // CHECK-LABEL: @test_svmla_lane_u16(
 109 // CHECK-NEXT:  entry:
 110 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
 111 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 112 //
 113 // CPP-CHECK-LABEL: @_Z19test_svmla_lane_u16u12__SVUint16_tS_S_(
 114 // CPP-CHECK-NEXT:  entry:
 115 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
 116 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 117 //
 118 svuint16_t test_svmla_lane_u16(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 119 {
 120   return SVE_ACLE_FUNC(svmla_lane,_u16,,)(op1, op2, op3, 0);
 121 }
 122
 123 // CHECK-LABEL: @test_svmla_lane_u16_1(
 124 // CHECK-NEXT:  entry:
 125 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
 126 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 127 //
 128 // CPP-CHECK-LABEL: @_Z21test_svmla_lane_u16_1u12__SVUint16_tS_S_(
 129 // CPP-CHECK-NEXT:  entry:
 130 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
 131 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 132 //
 133 svuint16_t test_svmla_lane_u16_1(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 134 {
 135   return SVE_ACLE_FUNC(svmla_lane,_u16,,)(op1, op2, op3, 7);
 136 }
 137
 138 // CHECK-LABEL: @test_svmla_lane_u32_1(
 139 // CHECK-NEXT:  entry:
 140 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.mla.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
 141 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 142 //
 143 // CPP-CHECK-LABEL: @_Z21test_svmla_lane_u32_1u12__SVUint32_tS_S_(
 144 // CPP-CHECK-NEXT:  entry:
 145 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.mla.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
 146 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 147 //
 148 svuint32_t test_svmla_lane_u32_1(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 149 {
 150   return SVE_ACLE_FUNC(svmla_lane,_u32,,)(op1, op2, op3, 3);
 151 }
 152
 153 // CHECK-LABEL: @test_svmla_lane_u64(
 154 // CHECK-NEXT:  entry:
 155 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
 156 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 157 //
 158 // CPP-CHECK-LABEL: @_Z19test_svmla_lane_u64u12__SVUint64_tS_S_(
 159 // CPP-CHECK-NEXT:  entry:
 160 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
 161 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 162 //
 163 svuint64_t test_svmla_lane_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 164 {
 165   return SVE_ACLE_FUNC(svmla_lane,_u64,,)(op1, op2, op3, 0);
 166 }
 167
 168 // CHECK-LABEL: @test_svmla_lane_u64_1(
 169 // CHECK-NEXT:  entry:
 170 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 1)
 171 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 172 //
 173 // CPP-CHECK-LABEL: @_Z21test_svmla_lane_u64_1u12__SVUint64_tS_S_(
 174 // CPP-CHECK-NEXT:  entry:
 175 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 1)
 176 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 177 //
 178 svuint64_t test_svmla_lane_u64_1(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 179 {
 180   return SVE_ACLE_FUNC(svmla_lane,_u64,,)(op1, op2, op3, 1);
 181 }