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
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
|
// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve \
// RUN: -fallow-half-arguments-and-returns -disable-O0-optnone \
// RUN: -emit-llvm -o - %s | opt -S -sroa | FileCheck %s
// REQUIRES: aarch64-registered-target
#include <arm_sve.h>
// CHECK-LABEL: @_Z9cond_boolu10__SVBool_tu10__SVBool_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = icmp ult <vscale x 16 x i1> [[A:%.*]], [[B:%.*]]
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 16 x i1> [[CMP]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 16 x i1> [[VECTOR_COND]], <vscale x 16 x i1> [[A]], <vscale x 16 x i1> [[B]]
// CHECK-NEXT: ret <vscale x 16 x i1> [[VECTOR_SELECT]]
//
svbool_t cond_bool(svbool_t a, svbool_t b) {
return a < b ? a : b;
}
// CHECK-LABEL: @_Z7cond_i8u10__SVInt8_tu10__SVInt8_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = icmp ult <vscale x 16 x i8> [[A:%.*]], [[B:%.*]]
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 16 x i1> [[CMP]] to <vscale x 16 x i8>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 16 x i8> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 16 x i1> [[VECTOR_COND]], <vscale x 16 x i8> [[A]], <vscale x 16 x i8> [[B]]
// CHECK-NEXT: ret <vscale x 16 x i8> [[VECTOR_SELECT]]
//
svint8_t cond_i8(svint8_t a, svint8_t b) {
return a < b ? a : b;
}
// CHECK-LABEL: @_Z7cond_u8u11__SVUint8_tu11__SVUint8_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = icmp ult <vscale x 16 x i8> [[A:%.*]], [[B:%.*]]
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 16 x i1> [[CMP]] to <vscale x 16 x i8>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 16 x i8> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 16 x i1> [[VECTOR_COND]], <vscale x 16 x i8> [[A]], <vscale x 16 x i8> [[B]]
// CHECK-NEXT: ret <vscale x 16 x i8> [[VECTOR_SELECT]]
//
svuint8_t cond_u8(svuint8_t a, svuint8_t b) {
return a < b ? a : b;
}
// CHECK-LABEL: @_Z8cond_i16u11__SVInt16_tu11__SVInt16_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = icmp ult <vscale x 8 x i16> [[A:%.*]], [[B:%.*]]
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 8 x i1> [[CMP]] to <vscale x 8 x i16>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 8 x i16> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 8 x i1> [[VECTOR_COND]], <vscale x 8 x i16> [[A]], <vscale x 8 x i16> [[B]]
// CHECK-NEXT: ret <vscale x 8 x i16> [[VECTOR_SELECT]]
//
svint16_t cond_i16(svint16_t a, svint16_t b) {
return a < b ? a : b;
}
// CHECK-LABEL: @_Z8cond_u16u12__SVUint16_tu12__SVUint16_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = icmp ult <vscale x 8 x i16> [[A:%.*]], [[B:%.*]]
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 8 x i1> [[CMP]] to <vscale x 8 x i16>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 8 x i16> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 8 x i1> [[VECTOR_COND]], <vscale x 8 x i16> [[A]], <vscale x 8 x i16> [[B]]
// CHECK-NEXT: ret <vscale x 8 x i16> [[VECTOR_SELECT]]
//
svuint16_t cond_u16(svuint16_t a, svuint16_t b) {
return a < b ? a : b;
}
// CHECK-LABEL: @_Z8cond_i32u11__SVInt32_tu11__SVInt32_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = icmp ult <vscale x 4 x i32> [[A:%.*]], [[B:%.*]]
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 4 x i1> [[CMP]] to <vscale x 4 x i32>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 4 x i32> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 4 x i1> [[VECTOR_COND]], <vscale x 4 x i32> [[A]], <vscale x 4 x i32> [[B]]
// CHECK-NEXT: ret <vscale x 4 x i32> [[VECTOR_SELECT]]
//
svint32_t cond_i32(svint32_t a, svint32_t b) {
return a < b ? a : b;
}
// CHECK-LABEL: @_Z8cond_u32u12__SVUint32_tu12__SVUint32_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = icmp ult <vscale x 4 x i32> [[A:%.*]], [[B:%.*]]
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 4 x i1> [[CMP]] to <vscale x 4 x i32>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 4 x i32> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 4 x i1> [[VECTOR_COND]], <vscale x 4 x i32> [[A]], <vscale x 4 x i32> [[B]]
// CHECK-NEXT: ret <vscale x 4 x i32> [[VECTOR_SELECT]]
//
svuint32_t cond_u32(svuint32_t a, svuint32_t b) {
return a < b ? a : b;
}
// CHECK-LABEL: @_Z8cond_i64u11__SVInt64_tu11__SVInt64_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = icmp ult <vscale x 2 x i64> [[A:%.*]], [[B:%.*]]
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 2 x i1> [[CMP]] to <vscale x 2 x i64>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 2 x i64> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 2 x i1> [[VECTOR_COND]], <vscale x 2 x i64> [[A]], <vscale x 2 x i64> [[B]]
// CHECK-NEXT: ret <vscale x 2 x i64> [[VECTOR_SELECT]]
//
svint64_t cond_i64(svint64_t a, svint64_t b) {
return a < b ? a : b;
}
// CHECK-LABEL: @_Z8cond_u64u12__SVUint64_tu12__SVUint64_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = icmp ult <vscale x 2 x i64> [[A:%.*]], [[B:%.*]]
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 2 x i1> [[CMP]] to <vscale x 2 x i64>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 2 x i64> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 2 x i1> [[VECTOR_COND]], <vscale x 2 x i64> [[A]], <vscale x 2 x i64> [[B]]
// CHECK-NEXT: ret <vscale x 2 x i64> [[VECTOR_SELECT]]
//
svuint64_t cond_u64(svuint64_t a, svuint64_t b) {
return a < b ? a : b;
}
// CHECK-LABEL: @_Z8cond_f16u13__SVFloat16_tu13__SVFloat16_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = fcmp olt <vscale x 8 x half> [[A:%.*]], [[B:%.*]]
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 8 x i1> [[CMP]] to <vscale x 8 x i16>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 8 x i16> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 8 x i1> [[VECTOR_COND]], <vscale x 8 x half> [[A]], <vscale x 8 x half> [[B]]
// CHECK-NEXT: ret <vscale x 8 x half> [[VECTOR_SELECT]]
//
svfloat16_t cond_f16(svfloat16_t a, svfloat16_t b) {
return a < b ? a : b;
}
// CHECK-LABEL: @_Z8cond_f32u13__SVFloat32_tu13__SVFloat32_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = fcmp olt <vscale x 4 x float> [[A:%.*]], [[B:%.*]]
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 4 x i1> [[CMP]] to <vscale x 4 x i32>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 4 x i32> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 4 x i1> [[VECTOR_COND]], <vscale x 4 x float> [[A]], <vscale x 4 x float> [[B]]
// CHECK-NEXT: ret <vscale x 4 x float> [[VECTOR_SELECT]]
//
svfloat32_t cond_f32(svfloat32_t a, svfloat32_t b) {
return a < b ? a : b;
}
// CHECK-LABEL: @_Z8cond_f64u13__SVFloat64_tu13__SVFloat64_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = fcmp olt <vscale x 2 x double> [[A:%.*]], [[B:%.*]]
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 2 x i1> [[CMP]] to <vscale x 2 x i64>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 2 x i64> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 2 x i1> [[VECTOR_COND]], <vscale x 2 x double> [[A]], <vscale x 2 x double> [[B]]
// CHECK-NEXT: ret <vscale x 2 x double> [[VECTOR_SELECT]]
//
svfloat64_t cond_f64(svfloat64_t a, svfloat64_t b) {
return a < b ? a : b;
}
// CHECK-LABEL: @_Z14cond_i32_splatu11__SVInt32_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = icmp ult <vscale x 4 x i32> [[A:%.*]], zeroinitializer
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 4 x i1> [[CMP]] to <vscale x 4 x i32>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 4 x i32> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 4 x i1> [[VECTOR_COND]], <vscale x 4 x i32> [[A]], <vscale x 4 x i32> zeroinitializer
// CHECK-NEXT: ret <vscale x 4 x i32> [[VECTOR_SELECT]]
//
svint32_t cond_i32_splat(svint32_t a) {
return a < 0 ? a : 0;
}
// CHECK-LABEL: @_Z14cond_u32_splatu12__SVUint32_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = icmp ult <vscale x 4 x i32> [[A:%.*]], shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 1, i32 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer)
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 4 x i1> [[CMP]] to <vscale x 4 x i32>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 4 x i32> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 4 x i1> [[VECTOR_COND]], <vscale x 4 x i32> [[A]], <vscale x 4 x i32> shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 1, i32 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer)
// CHECK-NEXT: ret <vscale x 4 x i32> [[VECTOR_SELECT]]
//
svuint32_t cond_u32_splat(svuint32_t a) {
return a < 1u ? a : 1u;
}
// CHECK-LABEL: @_Z14cond_i64_splatu11__SVInt64_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = icmp ult <vscale x 2 x i64> [[A:%.*]], zeroinitializer
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 2 x i1> [[CMP]] to <vscale x 2 x i64>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 2 x i64> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 2 x i1> [[VECTOR_COND]], <vscale x 2 x i64> [[A]], <vscale x 2 x i64> zeroinitializer
// CHECK-NEXT: ret <vscale x 2 x i64> [[VECTOR_SELECT]]
//
svint64_t cond_i64_splat(svint64_t a) {
return a < 0l ? a : 0l;
}
// CHECK-LABEL: @_Z14cond_u64_splatu12__SVUint64_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = icmp ult <vscale x 2 x i64> [[A:%.*]], shufflevector (<vscale x 2 x i64> insertelement (<vscale x 2 x i64> poison, i64 1, i32 0), <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer)
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 2 x i1> [[CMP]] to <vscale x 2 x i64>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 2 x i64> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 2 x i1> [[VECTOR_COND]], <vscale x 2 x i64> [[A]], <vscale x 2 x i64> shufflevector (<vscale x 2 x i64> insertelement (<vscale x 2 x i64> poison, i64 1, i32 0), <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer)
// CHECK-NEXT: ret <vscale x 2 x i64> [[VECTOR_SELECT]]
//
svuint64_t cond_u64_splat(svuint64_t a) {
return a < 1ul ? a : 1ul;
}
// CHECK-LABEL: @_Z14cond_f32_splatu13__SVFloat32_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = fcmp olt <vscale x 4 x float> [[A:%.*]], zeroinitializer
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 4 x i1> [[CMP]] to <vscale x 4 x i32>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 4 x i32> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 4 x i1> [[VECTOR_COND]], <vscale x 4 x float> [[A]], <vscale x 4 x float> zeroinitializer
// CHECK-NEXT: ret <vscale x 4 x float> [[VECTOR_SELECT]]
//
svfloat32_t cond_f32_splat(svfloat32_t a) {
return a < 0.f ? a : 0.f;
}
// CHECK-LABEL: @_Z14cond_f64_splatu13__SVFloat64_t(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CMP:%.*]] = fcmp olt <vscale x 2 x double> [[A:%.*]], zeroinitializer
// CHECK-NEXT: [[CONV:%.*]] = zext <vscale x 2 x i1> [[CMP]] to <vscale x 2 x i64>
// CHECK-NEXT: [[VECTOR_COND:%.*]] = icmp ne <vscale x 2 x i64> [[CONV]], zeroinitializer
// CHECK-NEXT: [[VECTOR_SELECT:%.*]] = select <vscale x 2 x i1> [[VECTOR_COND]], <vscale x 2 x double> [[A]], <vscale x 2 x double> zeroinitializer
// CHECK-NEXT: ret <vscale x 2 x double> [[VECTOR_SELECT]]
//
svfloat64_t cond_f64_splat(svfloat64_t a) {
return a < 0. ? a : 0.;
}
|
