| 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 -target-feature +sve2 -target-feature +sve2p1 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s |
| 4 |
// RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64 -target-feature +sve -target-feature +sve2 -target-feature +sve2p1 -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK |
| 5 |
// RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64 -target-feature +sme -target-feature +sme2 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s |
| 6 |
// RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64 -target-feature +sme -target-feature +sme2 -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK |
| 7 |
// RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64 -target-feature +sve -target-feature +sme -target-feature +sme2 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s |
| 8 |
// RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64 -target-feature +sme -target-feature +sve -target-feature +sve2 -target-feature +sve2p1 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s |
| 9 |
// RUN: %clang_cc1 -fclang-abi-compat=latest -DSVE_OVERLOADED_FORMS -triple aarch64 -target-feature +sve -target-feature +sve2 -target-feature +sve2p1 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s |
| 10 |
// RUN: %clang_cc1 -fclang-abi-compat=latest -DSVE_OVERLOADED_FORMS -triple aarch64 -target-feature +sve -target-feature +sve2 -target-feature +sve2p1 -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK |
| 11 |
// RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64 -target-feature +sve -target-feature +sve2 -target-feature +sve2p1 -S -disable-O0-optnone -Werror -Wall -o /dev/null %s |
| 12 |
// RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64 -target-feature +sme -target-feature +sme2 -S -disable-O0-optnone -Werror -Wall -o /dev/null %s |
| 13 |
#include <arm_sve.h> |
| 14 |
|
| 15 |
#if defined(__ARM_FEATURE_SME) && defined(__ARM_FEATURE_SVE) |
| 16 |
#define ATTR __arm_streaming_compatible |
| 17 |
#elif defined(__ARM_FEATURE_SME) |
| 18 |
#define ATTR __arm_streaming |
| 19 |
#else |
| 20 |
#define ATTR |
| 21 |
#endif |
| 22 |
|
| 23 |
#ifdef SVE_OVERLOADED_FORMS |
| 24 |
// A simple used,unused... macro, long enough to represent any SVE builtin. |
| 25 |
#define SVE_ACLE_FUNC(A1,A2_UNUSED,A3) A1##A3 |
| 26 |
#else |
| 27 |
#define SVE_ACLE_FUNC(A1,A2,A3) A1##A2##A3 |
| 28 |
#endif |
| 29 |
|
| 30 |
// CHECK-LABEL: @test_svdot_s32_x2( |
| 31 |
// CHECK-NEXT: entry: |
| 32 |
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sdot.x2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]]) |
| 33 |
// CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]] |
| 34 |
// |
| 35 |
// CPP-CHECK-LABEL: @_Z17test_svdot_s32_x2u11__SVInt32_tu11__SVInt16_tS0_( |
| 36 |
// CPP-CHECK-NEXT: entry: |
| 37 |
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sdot.x2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]]) |
| 38 |
// CPP-CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]] |
| 39 |
// |
| 40 |
svint32_t test_svdot_s32_x2(svint32_t op1, svint16_t op2, svint16_t op3) ATTR |
| 41 |
{ |
| 42 |
return SVE_ACLE_FUNC(svdot,_s32_s16,)(op1, op2, op3); |
| 43 |
} |
| 44 |
|
| 45 |
// CHECK-LABEL: @test_svdot_u32_x2( |
| 46 |
// CHECK-NEXT: entry: |
| 47 |
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.udot.x2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]]) |
| 48 |
// CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]] |
| 49 |
// |
| 50 |
// CPP-CHECK-LABEL: @_Z17test_svdot_u32_x2u12__SVUint32_tu12__SVUint16_tS0_( |
| 51 |
// CPP-CHECK-NEXT: entry: |
| 52 |
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.udot.x2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]]) |
| 53 |
// CPP-CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]] |
| 54 |
// |
| 55 |
svuint32_t test_svdot_u32_x2(svuint32_t op1, svuint16_t op2, svuint16_t op3) ATTR |
| 56 |
{ |
| 57 |
return SVE_ACLE_FUNC(svdot,_u32_u16,)(op1, op2, op3); |
| 58 |
} |
| 59 |
|
| 60 |
// CHECK-LABEL: @test_svdot_f32_x2( |
| 61 |
// CHECK-NEXT: entry: |
| 62 |
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.fdot.x2.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]]) |
| 63 |
// CHECK-NEXT: ret <vscale x 4 x float> [[TMP0]] |
| 64 |
// |
| 65 |
// CPP-CHECK-LABEL: @_Z17test_svdot_f32_x2u13__SVFloat32_tu13__SVFloat16_tS0_( |
| 66 |
// CPP-CHECK-NEXT: entry: |
| 67 |
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.fdot.x2.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]]) |
| 68 |
// CPP-CHECK-NEXT: ret <vscale x 4 x float> [[TMP0]] |
| 69 |
// |
| 70 |
svfloat32_t test_svdot_f32_x2(svfloat32_t op1, svfloat16_t op2, svfloat16_t op3) ATTR |
| 71 |
{ |
| 72 |
return SVE_ACLE_FUNC(svdot,_f32_f16,)(op1, op2, op3); |
| 73 |
} |
| 74 |
|
| 75 |
|
| 76 |
|
| 77 |
// CHECK-LABEL: @test_svdot_lane_s32_x2( |
| 78 |
// CHECK-NEXT: entry: |
| 79 |
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sdot.lane.x2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 3) |
| 80 |
// CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]] |
| 81 |
// |
| 82 |
// CPP-CHECK-LABEL: @_Z22test_svdot_lane_s32_x2u11__SVInt32_tu11__SVInt16_tS0_( |
| 83 |
// CPP-CHECK-NEXT: entry: |
| 84 |
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sdot.lane.x2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 3) |
| 85 |
// CPP-CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]] |
| 86 |
// |
| 87 |
svint32_t test_svdot_lane_s32_x2(svint32_t op1, svint16_t op2, svint16_t op3) ATTR |
| 88 |
{ |
| 89 |
return SVE_ACLE_FUNC(svdot_lane,_s32_s16,)(op1, op2, op3, 3); |
| 90 |
} |
| 91 |
|
| 92 |
// CHECK-LABEL: @test_svdot_lane_u32_x2( |
| 93 |
// CHECK-NEXT: entry: |
| 94 |
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.udot.lane.x2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 3) |
| 95 |
// CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]] |
| 96 |
// |
| 97 |
// CPP-CHECK-LABEL: @_Z22test_svdot_lane_u32_x2u12__SVUint32_tu12__SVUint16_tS0_( |
| 98 |
// CPP-CHECK-NEXT: entry: |
| 99 |
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.udot.lane.x2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 3) |
| 100 |
// CPP-CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]] |
| 101 |
// |
| 102 |
svuint32_t test_svdot_lane_u32_x2(svuint32_t op1, svuint16_t op2, svuint16_t op3) ATTR |
| 103 |
{ |
| 104 |
return SVE_ACLE_FUNC(svdot_lane,_u32_u16,)(op1, op2, op3, 3); |
| 105 |
} |
| 106 |
|
| 107 |
// CHECK-LABEL: @test_svdot_lane_f32_x2( |
| 108 |
// CHECK-NEXT: entry: |
| 109 |
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.fdot.lane.x2.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 3) |
| 110 |
// CHECK-NEXT: ret <vscale x 4 x float> [[TMP0]] |
| 111 |
// |
| 112 |
// CPP-CHECK-LABEL: @_Z22test_svdot_lane_f32_x2u13__SVFloat32_tu13__SVFloat16_tS0_( |
| 113 |
// CPP-CHECK-NEXT: entry: |
| 114 |
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.fdot.lane.x2.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 3) |
| 115 |
// CPP-CHECK-NEXT: ret <vscale x 4 x float> [[TMP0]] |
| 116 |
// |
| 117 |
svfloat32_t test_svdot_lane_f32_x2(svfloat32_t op1, svfloat16_t op2, svfloat16_t op3) ATTR |
| 118 |
{ |
| 119 |
return SVE_ACLE_FUNC(svdot_lane,_f32_f16,)(op1, op2, op3, 3); |
| 120 |
} |