| author | tpivovarova |
| Tue, 12 Jul 2016 18:24:48 +0300 | |
| changeset 40059 | c2304140ed64 |
| parent 36057 | 867e857daf84 |
| child 41705 | 332239c052cc |
| permissions | -rw-r--r-- |
| 30588 | 1 |
/* |
2 |
* Copyright (c) 2015, Oracle and/or its affiliates. All rights reserved. |
|
3 |
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. |
|
4 |
* |
|
5 |
* This code is free software; you can redistribute it and/or modify it |
|
6 |
* under the terms of the GNU General Public License version 2 only, as |
|
7 |
* published by the Free Software Foundation. |
|
8 |
* |
|
9 |
* This code is distributed in the hope that it will be useful, but WITHOUT |
|
10 |
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or |
|
11 |
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License |
|
12 |
* version 2 for more details (a copy is included in the LICENSE file that |
|
13 |
* accompanied this code). |
|
14 |
* |
|
15 |
* You should have received a copy of the GNU General Public License version |
|
16 |
* 2 along with this work; if not, write to the Free Software Foundation, |
|
17 |
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. |
|
18 |
* |
|
19 |
* Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA |
|
20 |
* or visit www.oracle.com if you need additional information or have any |
|
21 |
* questions. |
|
22 |
* |
|
23 |
*/ |
|
24 |
||
25 |
/** |
|
26 |
* @test |
|
27 |
* @bug 8074981 |
|
28 |
* @summary Add C2 x86 Superword support for scalar product reduction optimizations : int test |
|
|
36057
867e857daf84
8149123: [TESTBUG] compiler/loopopts/superword/SumRed* tests running on non-x86 platforms
thartmann
parents:
30588
diff
changeset
|
29 |
* @requires os.arch=="x86" | os.arch=="i386" | os.arch=="amd64" | os.arch=="x86_64" | os.arch=="aarch64" |
| 30588 | 30 |
* |
| 40059 | 31 |
* @run main/othervm -XX:+IgnoreUnrecognizedVMOptions |
32 |
* -XX:LoopUnrollLimit=250 -XX:CompileThresholdScaling=0.1 |
|
33 |
* -XX:CompileCommand=exclude,compiler.loopopts.superword.ReductionPerf::main |
|
34 |
* -XX:+SuperWordReductions |
|
35 |
* compiler.loopopts.superword.ReductionPerf |
|
36 |
* @run main/othervm -XX:+IgnoreUnrecognizedVMOptions |
|
37 |
* -XX:LoopUnrollLimit=250 -XX:CompileThresholdScaling=0.1 |
|
38 |
* -XX:CompileCommand=exclude,compiler.loopopts.superword.ReductionPerf::main |
|
39 |
* -XX:-SuperWordReductions |
|
40 |
* compiler.loopopts.superword.ReductionPerf |
|
| 30588 | 41 |
*/ |
42 |
||
| 40059 | 43 |
package compiler.loopopts.superword; |
44 |
||
| 30588 | 45 |
public class ReductionPerf {
|
| 40059 | 46 |
public static void main(String[] args) throws Exception {
|
47 |
int[] a1 = new int[8 * 1024]; |
|
48 |
int[] a2 = new int[8 * 1024]; |
|
49 |
int[] a3 = new int[8 * 1024]; |
|
50 |
long[] b1 = new long[8 * 1024]; |
|
51 |
long[] b2 = new long[8 * 1024]; |
|
52 |
long[] b3 = new long[8 * 1024]; |
|
53 |
float[] c1 = new float[8 * 1024]; |
|
54 |
float[] c2 = new float[8 * 1024]; |
|
55 |
float[] c3 = new float[8 * 1024]; |
|
56 |
double[] d1 = new double[8 * 1024]; |
|
57 |
double[] d2 = new double[8 * 1024]; |
|
58 |
double[] d3 = new double[8 * 1024]; |
|
| 30588 | 59 |
|
| 40059 | 60 |
ReductionInit(a1, a2, a3, b1, b2, b3, c1, c2, c3, d1, d2, d3); |
61 |
||
62 |
int sumIv = sumInt(a1, a2, a3); |
|
63 |
long sumLv = sumLong(b1, b2, b3); |
|
64 |
float sumFv = sumFloat(c1, c2, c3); |
|
65 |
double sumDv = sumDouble(d1, d2, d3); |
|
66 |
int mulIv = prodInt(a1, a2, a3); |
|
67 |
long mulLv = prodLong(b1, b2, b3); |
|
68 |
float mulFv = prodFloat(c1, c2, c3); |
|
69 |
double mulDv = prodDouble(d1, d2, d3); |
|
| 30588 | 70 |
|
| 40059 | 71 |
int sumI = 0; |
72 |
long sumL = 0; |
|
73 |
float sumF = 0.f; |
|
74 |
double sumD = 0.; |
|
75 |
int mulI = 0; |
|
76 |
long mulL = 0; |
|
77 |
float mulF = 0.f; |
|
78 |
double mulD = 0.; |
|
79 |
||
80 |
System.out.println("Warmup ...");
|
|
81 |
long start = System.currentTimeMillis(); |
|
| 30588 | 82 |
|
| 40059 | 83 |
for (int j = 0; j < 2000; j++) {
|
84 |
sumI = sumInt(a1, a2, a3); |
|
85 |
sumL = sumLong(b1, b2, b3); |
|
86 |
sumF = sumFloat(c1, c2, c3); |
|
87 |
sumD = sumDouble(d1, d2, d3); |
|
88 |
mulI = prodInt(a1, a2, a3); |
|
89 |
mulL = prodLong(b1, b2, b3); |
|
90 |
mulF = prodFloat(c1, c2, c3); |
|
91 |
mulD = prodDouble(d1, d2, d3); |
|
92 |
} |
|
| 30588 | 93 |
|
| 40059 | 94 |
long stop = System.currentTimeMillis(); |
95 |
System.out.println(" Warmup is done in " + (stop - start) + " msec");
|
|
| 30588 | 96 |
|
| 40059 | 97 |
if (sumIv != sumI) {
|
98 |
System.out.println("sum int: " + sumIv + " != " + sumI);
|
|
99 |
} |
|
100 |
if (sumLv != sumL) {
|
|
101 |
System.out.println("sum long: " + sumLv + " != " + sumL);
|
|
102 |
} |
|
103 |
if (sumFv != sumF) {
|
|
104 |
System.out.println("sum float: " + sumFv + " != " + sumF);
|
|
105 |
} |
|
106 |
if (sumDv != sumD) {
|
|
107 |
System.out.println("sum double: " + sumDv + " != " + sumD);
|
|
108 |
} |
|
109 |
if (mulIv != mulI) {
|
|
110 |
System.out.println("prod int: " + mulIv + " != " + mulI);
|
|
111 |
} |
|
112 |
if (mulLv != mulL) {
|
|
113 |
System.out.println("prod long: " + mulLv + " != " + mulL);
|
|
114 |
} |
|
115 |
if (mulFv != mulF) {
|
|
116 |
System.out.println("prod float: " + mulFv + " != " + mulF);
|
|
117 |
} |
|
118 |
if (mulDv != mulD) {
|
|
119 |
System.out.println("prod double: " + mulDv + " != " + mulD);
|
|
120 |
} |
|
| 30588 | 121 |
|
| 40059 | 122 |
start = System.currentTimeMillis(); |
123 |
for (int j = 0; j < 5000; j++) {
|
|
124 |
sumI = sumInt(a1, a2, a3); |
|
125 |
} |
|
126 |
stop = System.currentTimeMillis(); |
|
127 |
System.out.println("sum int: " + (stop - start));
|
|
128 |
||
129 |
start = System.currentTimeMillis(); |
|
130 |
for (int j = 0; j < 5000; j++) {
|
|
131 |
sumL = sumLong(b1, b2, b3); |
|
132 |
} |
|
133 |
stop = System.currentTimeMillis(); |
|
134 |
System.out.println("sum long: " + (stop - start));
|
|
| 30588 | 135 |
|
| 40059 | 136 |
start = System.currentTimeMillis(); |
137 |
for (int j = 0; j < 5000; j++) {
|
|
138 |
sumF = sumFloat(c1, c2, c3); |
|
139 |
} |
|
140 |
stop = System.currentTimeMillis(); |
|
141 |
System.out.println("sum float: " + (stop - start));
|
|
142 |
||
143 |
start = System.currentTimeMillis(); |
|
144 |
for (int j = 0; j < 5000; j++) {
|
|
145 |
sumD = sumDouble(d1, d2, d3); |
|
146 |
} |
|
147 |
stop = System.currentTimeMillis(); |
|
148 |
System.out.println("sum double: " + (stop - start));
|
|
149 |
||
150 |
start = System.currentTimeMillis(); |
|
151 |
for (int j = 0; j < 5000; j++) {
|
|
152 |
mulI = prodInt(a1, a2, a3); |
|
153 |
} |
|
154 |
stop = System.currentTimeMillis(); |
|
155 |
System.out.println("prod int: " + (stop - start));
|
|
156 |
||
157 |
start = System.currentTimeMillis(); |
|
158 |
for (int j = 0; j < 5000; j++) {
|
|
159 |
mulL = prodLong(b1, b2, b3); |
|
160 |
} |
|
161 |
stop = System.currentTimeMillis(); |
|
162 |
System.out.println("prod long: " + (stop - start));
|
|
163 |
||
164 |
start = System.currentTimeMillis(); |
|
165 |
for (int j = 0; j < 5000; j++) {
|
|
166 |
mulF = prodFloat(c1, c2, c3); |
|
167 |
} |
|
168 |
stop = System.currentTimeMillis(); |
|
169 |
System.out.println("prod float: " + (stop - start));
|
|
170 |
||
171 |
start = System.currentTimeMillis(); |
|
172 |
for (int j = 0; j < 5000; j++) {
|
|
173 |
mulD = prodDouble(d1, d2, d3); |
|
174 |
} |
|
175 |
stop = System.currentTimeMillis(); |
|
176 |
System.out.println("prod double: " + (stop - start));
|
|
177 |
||
| 30588 | 178 |
} |
179 |
||
| 40059 | 180 |
public static void ReductionInit(int[] a1, int[] a2, int[] a3, |
181 |
long[] b1, long[] b2, long[] b3, |
|
182 |
float[] c1, float[] c2, float[] c3, |
|
183 |
double[] d1, double[] d2, double[] d3) {
|
|
184 |
for(int i = 0; i < a1.length; i++) {
|
|
185 |
a1[i] = (i + 0); |
|
186 |
a2[i] = (i + 1); |
|
187 |
a3[i] = (i + 2); |
|
188 |
b1[i] = (long) (i + 0); |
|
189 |
b2[i] = (long) (i + 1); |
|
190 |
b3[i] = (long) (i + 2); |
|
191 |
c1[i] = (float) (i + 0); |
|
192 |
c2[i] = (float) (i + 1); |
|
193 |
c3[i] = (float) (i + 2); |
|
194 |
d1[i] = (double) (i + 0); |
|
195 |
d2[i] = (double) (i + 1); |
|
196 |
d3[i] = (double) (i + 2); |
|
197 |
} |
|
| 30588 | 198 |
} |
199 |
||
| 40059 | 200 |
public static int sumInt(int[] a1, int[] a2, int[] a3) {
|
201 |
int total = 0; |
|
202 |
for (int i = 0; i < a1.length; i++) {
|
|
203 |
total += (a1[i] * a2[i]) + (a1[i] * a3[i]) + (a2[i] * a3[i]); |
|
204 |
} |
|
205 |
return total; |
|
| 30588 | 206 |
} |
207 |
||
| 40059 | 208 |
public static long sumLong(long[] b1, long[] b2, long[] b3) {
|
209 |
long total = 0; |
|
210 |
for (int i = 0; i < b1.length; i++) {
|
|
211 |
total += (b1[i] * b2[i]) + (b1[i] * b3[i]) + (b2[i] * b3[i]); |
|
212 |
} |
|
213 |
return total; |
|
| 30588 | 214 |
} |
215 |
||
| 40059 | 216 |
public static float sumFloat(float[] c1, float[] c2, float[] c3) {
|
217 |
float total = 0; |
|
218 |
for (int i = 0; i < c1.length; i++) {
|
|
219 |
total += (c1[i] * c2[i]) + (c1[i] * c3[i]) + (c2[i] * c3[i]); |
|
220 |
} |
|
221 |
return total; |
|
| 30588 | 222 |
} |
223 |
||
| 40059 | 224 |
public static double sumDouble(double[] d1, double[] d2, double[] d3) {
|
225 |
double total = 0; |
|
226 |
for (int i = 0; i < d1.length; i++) {
|
|
227 |
total += (d1[i] * d2[i]) + (d1[i] * d3[i]) + (d2[i] * d3[i]); |
|
228 |
} |
|
229 |
return total; |
|
| 30588 | 230 |
} |
231 |
||
| 40059 | 232 |
public static int prodInt(int[] a1, int[] a2, int[] a3) {
|
233 |
int total = 1; |
|
234 |
for (int i = 0; i < a1.length; i++) {
|
|
235 |
total *= (a1[i] * a2[i]) + (a1[i] * a3[i]) + (a2[i] * a3[i]); |
|
236 |
} |
|
237 |
return total; |
|
| 30588 | 238 |
} |
239 |
||
| 40059 | 240 |
public static long prodLong(long[] b1, long[] b2, long[] b3) {
|
241 |
long total = 1; |
|
242 |
for (int i = 0; i < b1.length; i++) {
|
|
243 |
total *= (b1[i] * b2[i]) + (b1[i] * b3[i]) + (b2[i] * b3[i]); |
|
244 |
} |
|
245 |
return total; |
|
| 30588 | 246 |
} |
247 |
||
| 40059 | 248 |
public static float prodFloat(float[] c1, float[] c2, float[] c3) {
|
249 |
float total = 1; |
|
250 |
for (int i = 0; i < c1.length; i++) {
|
|
251 |
total *= (c1[i] * c2[i]) + (c1[i] * c3[i]) + (c2[i] * c3[i]); |
|
252 |
} |
|
253 |
return total; |
|
| 30588 | 254 |
} |
255 |
||
| 40059 | 256 |
public static double prodDouble(double[] d1, double[] d2, double[] d3) {
|
257 |
double total = 1; |
|
258 |
for (int i = 0; i < d1.length; i++) {
|
|
259 |
total *= (d1[i] * d2[i]) + (d1[i] * d3[i]) + (d2[i] * d3[i]); |
|
260 |
} |
|
261 |
return total; |
|
| 30588 | 262 |
} |
263 |
} |