1 // REQUIRES: aarch64-registered-target
2 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
3 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
4 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
5 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
6 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
7 #include <arm_sve.h>
8
9 #ifdef SVE_OVERLOADED_FORMS
10 // A simple used,unused... macro, long enough to represent any SVE builtin.
11 #define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
12 #else
13 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
14 #endif
15
test_svzip2_s8(svint8_t op1,svint8_t op2)16 svint8_t test_svzip2_s8(svint8_t op1, svint8_t op2)
17 {
18 // CHECK-LABEL: test_svzip2_s8
19 // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip2.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
20 // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
21 return SVE_ACLE_FUNC(svzip2,_s8,,)(op1, op2);
22 }
23
test_svzip2_s16(svint16_t op1,svint16_t op2)24 svint16_t test_svzip2_s16(svint16_t op1, svint16_t op2)
25 {
26 // CHECK-LABEL: test_svzip2_s16
27 // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip2.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
28 // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
29 return SVE_ACLE_FUNC(svzip2,_s16,,)(op1, op2);
30 }
31
test_svzip2_s32(svint32_t op1,svint32_t op2)32 svint32_t test_svzip2_s32(svint32_t op1, svint32_t op2)
33 {
34 // CHECK-LABEL: test_svzip2_s32
35 // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip2.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
36 // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
37 return SVE_ACLE_FUNC(svzip2,_s32,,)(op1, op2);
38 }
39
test_svzip2_s64(svint64_t op1,svint64_t op2)40 svint64_t test_svzip2_s64(svint64_t op1, svint64_t op2)
41 {
42 // CHECK-LABEL: test_svzip2_s64
43 // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip2.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
44 // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
45 return SVE_ACLE_FUNC(svzip2,_s64,,)(op1, op2);
46 }
47
test_svzip2_u8(svuint8_t op1,svuint8_t op2)48 svuint8_t test_svzip2_u8(svuint8_t op1, svuint8_t op2)
49 {
50 // CHECK-LABEL: test_svzip2_u8
51 // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip2.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
52 // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
53 return SVE_ACLE_FUNC(svzip2,_u8,,)(op1, op2);
54 }
55
test_svzip2_u16(svuint16_t op1,svuint16_t op2)56 svuint16_t test_svzip2_u16(svuint16_t op1, svuint16_t op2)
57 {
58 // CHECK-LABEL: test_svzip2_u16
59 // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip2.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
60 // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
61 return SVE_ACLE_FUNC(svzip2,_u16,,)(op1, op2);
62 }
63
test_svzip2_u32(svuint32_t op1,svuint32_t op2)64 svuint32_t test_svzip2_u32(svuint32_t op1, svuint32_t op2)
65 {
66 // CHECK-LABEL: test_svzip2_u32
67 // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip2.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
68 // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
69 return SVE_ACLE_FUNC(svzip2,_u32,,)(op1, op2);
70 }
71
test_svzip2_u64(svuint64_t op1,svuint64_t op2)72 svuint64_t test_svzip2_u64(svuint64_t op1, svuint64_t op2)
73 {
74 // CHECK-LABEL: test_svzip2_u64
75 // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip2.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
76 // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
77 return SVE_ACLE_FUNC(svzip2,_u64,,)(op1, op2);
78 }
79
test_svzip2_f16(svfloat16_t op1,svfloat16_t op2)80 svfloat16_t test_svzip2_f16(svfloat16_t op1, svfloat16_t op2)
81 {
82 // CHECK-LABEL: test_svzip2_f16
83 // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.zip2.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
84 // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
85 return SVE_ACLE_FUNC(svzip2,_f16,,)(op1, op2);
86 }
87
test_svzip2_f32(svfloat32_t op1,svfloat32_t op2)88 svfloat32_t test_svzip2_f32(svfloat32_t op1, svfloat32_t op2)
89 {
90 // CHECK-LABEL: test_svzip2_f32
91 // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.zip2.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
92 // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
93 return SVE_ACLE_FUNC(svzip2,_f32,,)(op1, op2);
94 }
95
test_svzip2_f64(svfloat64_t op1,svfloat64_t op2)96 svfloat64_t test_svzip2_f64(svfloat64_t op1, svfloat64_t op2)
97 {
98 // CHECK-LABEL: test_svzip2_f64
99 // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.zip2.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
100 // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
101 return SVE_ACLE_FUNC(svzip2,_f64,,)(op1, op2);
102 }
103
test_svzip2_b8(svbool_t op1,svbool_t op2)104 svbool_t test_svzip2_b8(svbool_t op1, svbool_t op2)
105 {
106 // CHECK-LABEL: test_svzip2_b8
107 // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.zip2.nxv16i1(<vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
108 // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
109 return svzip2_b8(op1, op2);
110 }
111
test_svzip2_b16(svbool_t op1,svbool_t op2)112 svbool_t test_svzip2_b16(svbool_t op1, svbool_t op2)
113 {
114 // CHECK-LABEL: test_svzip2_b16
115 // CHECK-DAG: %[[OP1:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op1)
116 // CHECK-DAG: %[[OP2:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op2)
117 // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.zip2.nxv8i1(<vscale x 8 x i1> %[[OP1]], <vscale x 8 x i1> %[[OP2]])
118 // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
119 // CHECK: ret <vscale x 16 x i1> %[[CAST]]
120 return svzip2_b16(op1, op2);
121 }
122
test_svzip2_b32(svbool_t op1,svbool_t op2)123 svbool_t test_svzip2_b32(svbool_t op1, svbool_t op2)
124 {
125 // CHECK-LABEL: test_svzip2_b32
126 // CHECK-DAG: %[[OP1:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op1)
127 // CHECK-DAG: %[[OP2:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op2)
128 // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.zip2.nxv4i1(<vscale x 4 x i1> %[[OP1]], <vscale x 4 x i1> %[[OP2]])
129 // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
130 // CHECK: ret <vscale x 16 x i1> %[[CAST]]
131 return svzip2_b32(op1, op2);
132 }
133
test_svzip2_b64(svbool_t op1,svbool_t op2)134 svbool_t test_svzip2_b64(svbool_t op1, svbool_t op2)
135 {
136 // CHECK-LABEL: test_svzip2_b64
137 // CHECK-DAG: %[[OP1:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op1)
138 // CHECK-DAG: %[[OP2:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op2)
139 // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.zip2.nxv2i1(<vscale x 2 x i1> %[[OP1]], <vscale x 2 x i1> %[[OP2]])
140 // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
141 // CHECK: ret <vscale x 16 x i1> %[[CAST]]
142 return svzip2_b64(op1, op2);
143 }
144