acle_sve_mulx.c
12 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
// REQUIRES: aarch64-registered-target
// 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
// 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
// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null 2>%t
// RUN: FileCheck --check-prefix=ASM --allow-empty %s <%t
// If this check fails please read test/CodeGen/aarch64-sve-intrinsics/README for instructions on how to resolve it.
// ASM-NOT: warning
#include <arm_sve.h>
#ifdef SVE_OVERLOADED_FORMS
// A simple used,unused... macro, long enough to represent any SVE builtin.
#define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
#else
#define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
#endif
svfloat16_t test_svmulx_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
{
// CHECK-LABEL: test_svmulx_f16_z
// CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
// CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2)
// CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_f16,_z,)(pg, op1, op2);
}
svfloat32_t test_svmulx_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
{
// CHECK-LABEL: test_svmulx_f32_z
// CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
// CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2)
// CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_f32,_z,)(pg, op1, op2);
}
svfloat64_t test_svmulx_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
{
// CHECK-LABEL: test_svmulx_f64_z
// CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
// CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2)
// CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_f64,_z,)(pg, op1, op2);
}
svfloat16_t test_svmulx_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
{
// CHECK-LABEL: test_svmulx_f16_m
// CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
// CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_f16,_m,)(pg, op1, op2);
}
svfloat32_t test_svmulx_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
{
// CHECK-LABEL: test_svmulx_f32_m
// CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
// CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_f32,_m,)(pg, op1, op2);
}
svfloat64_t test_svmulx_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
{
// CHECK-LABEL: test_svmulx_f64_m
// CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
// CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_f64,_m,)(pg, op1, op2);
}
svfloat16_t test_svmulx_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
{
// CHECK-LABEL: test_svmulx_f16_x
// CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
// CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_f16,_x,)(pg, op1, op2);
}
svfloat32_t test_svmulx_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
{
// CHECK-LABEL: test_svmulx_f32_x
// CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
// CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_f32,_x,)(pg, op1, op2);
}
svfloat64_t test_svmulx_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
{
// CHECK-LABEL: test_svmulx_f64_x
// CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
// CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_f64,_x,)(pg, op1, op2);
}
svfloat16_t test_svmulx_n_f16_z(svbool_t pg, svfloat16_t op1, float16_t op2)
{
// CHECK-LABEL: test_svmulx_n_f16_z
// CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
// CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
// CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %[[DUP]])
// CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_n_f16,_z,)(pg, op1, op2);
}
svfloat32_t test_svmulx_n_f32_z(svbool_t pg, svfloat32_t op1, float32_t op2)
{
// CHECK-LABEL: test_svmulx_n_f32_z
// CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
// CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
// CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %[[DUP]])
// CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_n_f32,_z,)(pg, op1, op2);
}
svfloat64_t test_svmulx_n_f64_z(svbool_t pg, svfloat64_t op1, float64_t op2)
{
// CHECK-LABEL: test_svmulx_n_f64_z
// CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
// CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
// CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %[[DUP]])
// CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_n_f64,_z,)(pg, op1, op2);
}
svfloat16_t test_svmulx_n_f16_m(svbool_t pg, svfloat16_t op1, float16_t op2)
{
// CHECK-LABEL: test_svmulx_n_f16_m
// CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
// CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
// CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_n_f16,_m,)(pg, op1, op2);
}
svfloat32_t test_svmulx_n_f32_m(svbool_t pg, svfloat32_t op1, float32_t op2)
{
// CHECK-LABEL: test_svmulx_n_f32_m
// CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
// CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
// CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_n_f32,_m,)(pg, op1, op2);
}
svfloat64_t test_svmulx_n_f64_m(svbool_t pg, svfloat64_t op1, float64_t op2)
{
// CHECK-LABEL: test_svmulx_n_f64_m
// CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
// CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
// CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_n_f64,_m,)(pg, op1, op2);
}
svfloat16_t test_svmulx_n_f16_x(svbool_t pg, svfloat16_t op1, float16_t op2)
{
// CHECK-LABEL: test_svmulx_n_f16_x
// CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
// CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
// CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_n_f16,_x,)(pg, op1, op2);
}
svfloat32_t test_svmulx_n_f32_x(svbool_t pg, svfloat32_t op1, float32_t op2)
{
// CHECK-LABEL: test_svmulx_n_f32_x
// CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
// CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
// CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_n_f32,_x,)(pg, op1, op2);
}
svfloat64_t test_svmulx_n_f64_x(svbool_t pg, svfloat64_t op1, float64_t op2)
{
// CHECK-LABEL: test_svmulx_n_f64_x
// CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
// CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
// CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
// CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
return SVE_ACLE_FUNC(svmulx,_n_f64,_x,)(pg, op1, op2);
}