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
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
|
/*
* Copyright (C) 2017 The Android Open Source Project
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
/**
* Functional tests for SIMD vectorization.
*/
public class SimdChar {
static char[] a;
//
// Arithmetic operations.
//
/// CHECK-START: void SimdChar.add(int) loop_optimization (before)
/// CHECK-DAG: ArrayGet loop:<<Loop:B\d+>> outer_loop:none
/// CHECK-DAG: ArraySet loop:<<Loop>> outer_loop:none
//
/// CHECK-START-{ARM,ARM64}: void SimdChar.add(int) loop_optimization (after)
/// CHECK-DAG: VecLoad loop:<<Loop:B\d+>> outer_loop:none
/// CHECK-DAG: VecAdd loop:<<Loop>> outer_loop:none
/// CHECK-DAG: VecStore loop:<<Loop>> outer_loop:none
static void add(int x) {
for (int i = 0; i < 128; i++)
a[i] += x;
}
/// CHECK-START: void SimdChar.sub(int) loop_optimization (before)
/// CHECK-DAG: ArrayGet loop:<<Loop:B\d+>> outer_loop:none
/// CHECK-DAG: ArraySet loop:<<Loop>> outer_loop:none
//
/// CHECK-START-{ARM,ARM64}: void SimdChar.sub(int) loop_optimization (after)
/// CHECK-DAG: VecLoad loop:<<Loop:B\d+>> outer_loop:none
/// CHECK-DAG: VecSub loop:<<Loop>> outer_loop:none
/// CHECK-DAG: VecStore loop:<<Loop>> outer_loop:none
static void sub(int x) {
for (int i = 0; i < 128; i++)
a[i] -= x;
}
/// CHECK-START: void SimdChar.mul(int) loop_optimization (before)
/// CHECK-DAG: ArrayGet loop:<<Loop:B\d+>> outer_loop:none
/// CHECK-DAG: ArraySet loop:<<Loop>> outer_loop:none
//
/// CHECK-START-{ARM,ARM64}: void SimdChar.mul(int) loop_optimization (after)
/// CHECK-DAG: VecLoad loop:<<Loop:B\d+>> outer_loop:none
/// CHECK-DAG: VecMul loop:<<Loop>> outer_loop:none
/// CHECK-DAG: VecStore loop:<<Loop>> outer_loop:none
static void mul(int x) {
for (int i = 0; i < 128; i++)
a[i] *= x;
}
/// CHECK-START: void SimdChar.div(int) loop_optimization (before)
/// CHECK-DAG: ArrayGet loop:<<Loop:B\d+>> outer_loop:none
/// CHECK-DAG: ArraySet loop:<<Loop>> outer_loop:none
//
/// CHECK-START: void SimdChar.div(int) loop_optimization (after)
/// CHECK-NOT: VecDiv
//
// Not supported on any architecture.
//
static void div(int x) {
for (int i = 0; i < 128; i++)
a[i] /= x;
}
/// CHECK-START: void SimdChar.neg() loop_optimization (before)
/// CHECK-DAG: ArrayGet loop:<<Loop:B\d+>> outer_loop:none
/// CHECK-DAG: ArraySet loop:<<Loop>> outer_loop:none
//
/// CHECK-START-{ARM,ARM64}: void SimdChar.neg() loop_optimization (after)
/// CHECK-DAG: VecLoad loop:<<Loop:B\d+>> outer_loop:none
/// CHECK-DAG: VecNeg loop:<<Loop>> outer_loop:none
/// CHECK-DAG: VecStore loop:<<Loop>> outer_loop:none
static void neg() {
for (int i = 0; i < 128; i++)
a[i] = (char) -a[i];
}
/// CHECK-START: void SimdChar.not() loop_optimization (before)
/// CHECK-DAG: ArrayGet loop:<<Loop:B\d+>> outer_loop:none
/// CHECK-DAG: ArraySet loop:<<Loop>> outer_loop:none
//
/// CHECK-START-{ARM,ARM64}: void SimdChar.not() loop_optimization (after)
/// CHECK-DAG: VecLoad loop:<<Loop:B\d+>> outer_loop:none
/// CHECK-DAG: VecNot loop:<<Loop>> outer_loop:none
/// CHECK-DAG: VecStore loop:<<Loop>> outer_loop:none
static void not() {
for (int i = 0; i < 128; i++)
a[i] = (char) ~a[i];
}
/// CHECK-START: void SimdChar.shl4() loop_optimization (before)
/// CHECK-DAG: ArrayGet loop:<<Loop:B\d+>> outer_loop:none
/// CHECK-DAG: ArraySet loop:<<Loop>> outer_loop:none
//
/// CHECK-START-{ARM,ARM64}: void SimdChar.shl4() loop_optimization (after)
/// CHECK-DAG: VecLoad loop:<<Loop:B\d+>> outer_loop:none
/// CHECK-DAG: VecShl loop:<<Loop>> outer_loop:none
/// CHECK-DAG: VecStore loop:<<Loop>> outer_loop:none
static void shl4() {
for (int i = 0; i < 128; i++)
a[i] <<= 4;
}
/// CHECK-START: void SimdChar.sar2() loop_optimization (before)
/// CHECK-DAG: ArrayGet loop:<<Loop:B\d+>> outer_loop:none
/// CHECK-DAG: ArraySet loop:<<Loop>> outer_loop:none
//
// TODO: would need signess flip.
/// CHECK-START: void SimdChar.sar2() loop_optimization (after)
/// CHECK-NOT: VecShr
static void sar2() {
for (int i = 0; i < 128; i++)
a[i] >>= 2;
}
/// CHECK-START: void SimdChar.shr2() loop_optimization (before)
/// CHECK-DAG: ArrayGet loop:<<Loop:B\d+>> outer_loop:none
/// CHECK-DAG: ArraySet loop:<<Loop>> outer_loop:none
//
/// CHECK-START-{ARM,ARM64}: void SimdChar.shr2() loop_optimization (after)
/// CHECK-DAG: VecLoad loop:<<Loop:B\d+>> outer_loop:none
/// CHECK-DAG: VecUShr loop:<<Loop>> outer_loop:none
/// CHECK-DAG: VecStore loop:<<Loop>> outer_loop:none
static void shr2() {
for (int i = 0; i < 128; i++)
a[i] >>>= 2;
}
//
// Shift checks.
//
static void sar31() {
for (int i = 0; i < 128; i++)
a[i] >>= 31;
}
static void shr31() {
for (int i = 0; i < 128; i++)
a[i] >>>= 31;
}
static void shr32() {
for (int i = 0; i < 128; i++)
a[i] >>>= 32; // 0, since & 31
}
static void shr33() {
for (int i = 0; i < 128; i++)
a[i] >>>= 33; // 1, since & 31
}
//
// Loop bounds.
//
static void bounds() {
for (int i = 1; i < 127; i++)
a[i] += 11;
}
//
// Test Driver.
//
public static void main() {
// Set up.
a = new char[128];
for (int i = 0; i < 128; i++) {
a[i] = (char) i;
}
// Arithmetic operations.
add(2);
for (int i = 0; i < 128; i++) {
expectEquals(i + 2, a[i], "add");
}
sub(2);
for (int i = 0; i < 128; i++) {
expectEquals(i, a[i], "sub");
}
mul(2);
for (int i = 0; i < 128; i++) {
expectEquals(i + i, a[i], "mul");
}
div(2);
for (int i = 0; i < 128; i++) {
expectEquals(i, a[i], "div");
}
neg();
for (int i = 0; i < 128; i++) {
expectEquals((char)-i, a[i], "neg");
}
// Loop bounds.
bounds();
expectEquals(0, a[0], "bounds0");
for (int i = 1; i < 127; i++) {
expectEquals((char)(11 - i), a[i], "bounds");
}
expectEquals((char)-127, a[127], "bounds127");
// Shifts.
for (int i = 0; i < 128; i++) {
a[i] = (char) 0xffff;
}
shl4();
for (int i = 0; i < 128; i++) {
expectEquals((char) 0xfff0, a[i], "shl4");
}
sar2();
for (int i = 0; i < 128; i++) {
expectEquals((char) 0x3ffc, a[i], "sar2");
}
shr2();
for (int i = 0; i < 128; i++) {
expectEquals((char) 0x0fff, a[i], "shr2");
a[i] = (char) 0xffff; // reset
}
sar31();
for (int i = 0; i < 128; i++) {
expectEquals(0, a[i], "sar31");
a[i] = (char) 0xffff; // reset
}
shr31();
for (int i = 0; i < 128; i++) {
expectEquals(0, a[i], "shr31");
a[i] = (char) 0x1200; // reset
}
shr32();
for (int i = 0; i < 128; i++) {
expectEquals((char) 0x1200, a[i], "shr32");
}
shr33();
for (int i = 0; i < 128; i++) {
expectEquals((char) 0x0900, a[i], "shr33");
a[i] = (char) 0xf1f0; // reset
}
not();
for (int i = 0; i < 128; i++) {
expectEquals((char) 0x0e0f, a[i], "not");
}
// Done.
System.out.println("SimdChar passed");
}
private static void expectEquals(int expected, int result, String action) {
if (expected != result) {
throw new Error("Expected: " + expected + ", found: " + result + " for " + action);
}
}
}
|