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 +bf16 -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 +bf16 -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 -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
8 // 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
9 // 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
10 // RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64 -target-feature +bf16 -target-feature +sme -target-feature +sme2 -S -disable-O0-optnone -Werror -Wall -o /dev/null %s
13 #ifdef __ARM_FEATURE_SME
14 #define ATTR __arm_streaming
19 #ifdef SVE_OVERLOADED_FORMS
20 // A simple used,unused... macro, long enough to represent any SVE builtin.
21 #define SVE_ACLE_FUNC(A1,A2_UNUSED,A3) A1##A3
23 #define SVE_ACLE_FUNC(A1,A2,A3) A1##A2##A3
26 // CHECK-LABEL: @test_svdot_s32_x2(
28 // 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:%.*]])
29 // CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]]
31 // CPP-CHECK-LABEL: @_Z17test_svdot_s32_x2u11__SVInt32_tu11__SVInt16_tS0_(
32 // CPP-CHECK-NEXT: entry:
33 // 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:%.*]])
34 // CPP-CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]]
36 svint32_t
test_svdot_s32_x2(svint32_t op1
, svint16_t op2
, svint16_t op3
) ATTR
38 return SVE_ACLE_FUNC(svdot
,_s32_s16
,)(op1
, op2
, op3
);
41 // CHECK-LABEL: @test_svdot_u32_x2(
43 // 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:%.*]])
44 // CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]]
46 // CPP-CHECK-LABEL: @_Z17test_svdot_u32_x2u12__SVUint32_tu12__SVUint16_tS0_(
47 // CPP-CHECK-NEXT: entry:
48 // 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:%.*]])
49 // CPP-CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]]
51 svuint32_t
test_svdot_u32_x2(svuint32_t op1
, svuint16_t op2
, svuint16_t op3
) ATTR
53 return SVE_ACLE_FUNC(svdot
,_u32_u16
,)(op1
, op2
, op3
);
56 // CHECK-LABEL: @test_svdot_f32_x2(
58 // 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:%.*]])
59 // CHECK-NEXT: ret <vscale x 4 x float> [[TMP0]]
61 // CPP-CHECK-LABEL: @_Z17test_svdot_f32_x2u13__SVFloat32_tu13__SVFloat16_tS0_(
62 // CPP-CHECK-NEXT: entry:
63 // 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:%.*]])
64 // CPP-CHECK-NEXT: ret <vscale x 4 x float> [[TMP0]]
66 svfloat32_t
test_svdot_f32_x2(svfloat32_t op1
, svfloat16_t op2
, svfloat16_t op3
) ATTR
68 return SVE_ACLE_FUNC(svdot
,_f32_f16
,)(op1
, op2
, op3
);
73 // CHECK-LABEL: @test_svdot_lane_s32_x2(
75 // 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)
76 // CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]]
78 // CPP-CHECK-LABEL: @_Z22test_svdot_lane_s32_x2u11__SVInt32_tu11__SVInt16_tS0_(
79 // CPP-CHECK-NEXT: entry:
80 // 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)
81 // CPP-CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]]
83 svint32_t
test_svdot_lane_s32_x2(svint32_t op1
, svint16_t op2
, svint16_t op3
) ATTR
85 return SVE_ACLE_FUNC(svdot_lane
,_s32_s16
,)(op1
, op2
, op3
, 3);
88 // CHECK-LABEL: @test_svdot_lane_u32_x2(
90 // 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)
91 // CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]]
93 // CPP-CHECK-LABEL: @_Z22test_svdot_lane_u32_x2u12__SVUint32_tu12__SVUint16_tS0_(
94 // CPP-CHECK-NEXT: entry:
95 // 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)
96 // CPP-CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]]
98 svuint32_t
test_svdot_lane_u32_x2(svuint32_t op1
, svuint16_t op2
, svuint16_t op3
) ATTR
100 return SVE_ACLE_FUNC(svdot_lane
,_u32_u16
,)(op1
, op2
, op3
, 3);
103 // CHECK-LABEL: @test_svdot_lane_f32_x2(
104 // CHECK-NEXT: entry:
105 // 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)
106 // CHECK-NEXT: ret <vscale x 4 x float> [[TMP0]]
108 // CPP-CHECK-LABEL: @_Z22test_svdot_lane_f32_x2u13__SVFloat32_tu13__SVFloat16_tS0_(
109 // CPP-CHECK-NEXT: entry:
110 // 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)
111 // CPP-CHECK-NEXT: ret <vscale x 4 x float> [[TMP0]]
113 svfloat32_t
test_svdot_lane_f32_x2(svfloat32_t op1
, svfloat16_t op2
, svfloat16_t op3
) ATTR
115 return SVE_ACLE_FUNC(svdot_lane
,_f32_f16
,)(op1
, op2
, op3
, 3);