clang/test/CodeGen/AArch64/sve-intrinsics/acle_sve_acle.c

   1 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
   2 // REQUIRES: aarch64-registered-target
   3 // RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64 -target-feature +sve -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s
   4 // RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64 -target-feature +sve -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 -fclang-abi-compat=latest -DSVE_OVERLOADED_FORMS -triple aarch64 -target-feature +sve -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s
   6 // RUN: %clang_cc1 -fclang-abi-compat=latest -DSVE_OVERLOADED_FORMS -triple aarch64 -target-feature +sve -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 -fclang-abi-compat=latest -triple aarch64 -target-feature +sve -S -disable-O0-optnone -Werror -o /dev/null %s
   8 // RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64 -target-feature +sme -S -disable-O0-optnone -Werror -o /dev/null %s
   9
  10 #include <arm_sve.h>
  11
  12 #if defined __ARM_FEATURE_SME
  13 #define MODE_ATTR __arm_streaming
  14 #else
  15 #define MODE_ATTR
  16 #endif
  17
  18 #ifdef SVE_OVERLOADED_FORMS
  19 // A simple used,unused... macro, long enough to represent any SVE builtin.
  20 #define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
  21 #else
  22 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
  23 #endif
  24
  25 // CHECK-LABEL: @test_svacle_f16(
  26 // CHECK-NEXT:  entry:
  27 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
  28 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.facge.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP1:%.*]])
  29 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
  30 // CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
  31 //
  32 // CPP-CHECK-LABEL: @_Z15test_svacle_f16u10__SVBool_tu13__SVFloat16_tS0_(
  33 // CPP-CHECK-NEXT:  entry:
  34 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
  35 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.facge.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP1:%.*]])
  36 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
  37 // CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
  38 //
  39 svbool_t test_svacle_f16(svbool_t pg, svfloat16_t op1, svfloat16_t op2) MODE_ATTR
  40 {
  41   return SVE_ACLE_FUNC(svacle,_f16,,)(pg, op1, op2);
  42 }
  43
  44 // CHECK-LABEL: @test_svacle_f32(
  45 // CHECK-NEXT:  entry:
  46 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
  47 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.facge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP1:%.*]])
  48 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
  49 // CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
  50 //
  51 // CPP-CHECK-LABEL: @_Z15test_svacle_f32u10__SVBool_tu13__SVFloat32_tS0_(
  52 // CPP-CHECK-NEXT:  entry:
  53 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
  54 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.facge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP1:%.*]])
  55 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
  56 // CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
  57 //
  58 svbool_t test_svacle_f32(svbool_t pg, svfloat32_t op1, svfloat32_t op2) MODE_ATTR
  59 {
  60   return SVE_ACLE_FUNC(svacle,_f32,,)(pg, op1, op2);
  61 }
  62
  63 // CHECK-LABEL: @test_svacle_f64(
  64 // CHECK-NEXT:  entry:
  65 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
  66 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.facge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP1:%.*]])
  67 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
  68 // CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
  69 //
  70 // CPP-CHECK-LABEL: @_Z15test_svacle_f64u10__SVBool_tu13__SVFloat64_tS0_(
  71 // CPP-CHECK-NEXT:  entry:
  72 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
  73 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.facge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP1:%.*]])
  74 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
  75 // CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
  76 //
  77 svbool_t test_svacle_f64(svbool_t pg, svfloat64_t op1, svfloat64_t op2) MODE_ATTR
  78 {
  79   return SVE_ACLE_FUNC(svacle,_f64,,)(pg, op1, op2);
  80 }
  81
  82 // CHECK-LABEL: @test_svacle_n_f32(
  83 // CHECK-NEXT:  entry:
  84 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
  85 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 4 x float> poison, float [[OP2:%.*]], i64 0
  86 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 4 x float> [[DOTSPLATINSERT]], <vscale x 4 x float> poison, <vscale x 4 x i32> zeroinitializer
  87 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.facge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[DOTSPLAT]], <vscale x 4 x float> [[OP1:%.*]])
  88 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
  89 // CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
  90 //
  91 // CPP-CHECK-LABEL: @_Z17test_svacle_n_f32u10__SVBool_tu13__SVFloat32_tf(
  92 // CPP-CHECK-NEXT:  entry:
  93 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
  94 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 4 x float> poison, float [[OP2:%.*]], i64 0
  95 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 4 x float> [[DOTSPLATINSERT]], <vscale x 4 x float> poison, <vscale x 4 x i32> zeroinitializer
  96 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.facge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[DOTSPLAT]], <vscale x 4 x float> [[OP1:%.*]])
  97 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
  98 // CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
  99 //
 100 svbool_t test_svacle_n_f32(svbool_t pg, svfloat32_t op1, float32_t op2) MODE_ATTR
 101 {
 102   return SVE_ACLE_FUNC(svacle,_n_f32,,)(pg, op1, op2);
 103 }
 104
 105 // CHECK-LABEL: @test_svacle_n_f64(
 106 // CHECK-NEXT:  entry:
 107 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 108 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 2 x double> poison, double [[OP2:%.*]], i64 0
 109 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 2 x double> [[DOTSPLATINSERT]], <vscale x 2 x double> poison, <vscale x 2 x i32> zeroinitializer
 110 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.facge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[DOTSPLAT]], <vscale x 2 x double> [[OP1:%.*]])
 111 // CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
 112 // CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
 113 //
 114 // CPP-CHECK-LABEL: @_Z17test_svacle_n_f64u10__SVBool_tu13__SVFloat64_td(
 115 // CPP-CHECK-NEXT:  entry:
 116 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 117 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 2 x double> poison, double [[OP2:%.*]], i64 0
 118 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 2 x double> [[DOTSPLATINSERT]], <vscale x 2 x double> poison, <vscale x 2 x i32> zeroinitializer
 119 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.facge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[DOTSPLAT]], <vscale x 2 x double> [[OP1:%.*]])
 120 // CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
 121 // CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
 122 //
 123 svbool_t test_svacle_n_f64(svbool_t pg, svfloat64_t op1, float64_t op2) MODE_ATTR
 124 {
 125   return SVE_ACLE_FUNC(svacle,_n_f64,,)(pg, op1, op2);
 126 }