blob: 6c4594eb34fb5bde7e302ec3bbb8430f910e5e99 (
plain)
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
|
package com.vaadin.benchmarks;
import com.vaadin.ui.Label;
/*
* This simple test shows the performance difference between the StringTokenizer implementation and the String.split() implementation in AbstractComponent.
* Your results will vary.
* The real world use case motivating it was a 10k Row table, which generated labels for 10 columns.
* This is 1/10th of what this performance tester demonstrates.
*
* Please run with -server and -Xloggc:/tmp/gclog.vgc -verbose:gc -XX:+PrintCompilation
*
* My results Win 7 64, i7 2760QM 2.4Ghz, Java 7 21.
*
* Proposed Patch with StringTokenizer:
* 13 GC activations, 1.009GB allocated memory over time, total time 948ms
*
* Current String.split implementation:
* 31 GC activations, 2.277 GB allocated memory over time, total time 1557ms
*
*/
public class PerformanceTester8759 {
public static void main(String[] args) throws InterruptedException {
warmup();
long start = System.currentTimeMillis();
runBenchmark(1000000);
long end = System.currentTimeMillis();
System.out.println("took " + (end - start) + " ms");
}
private static void warmup() throws InterruptedException {
runBenchmark(10000);
System.gc();
System.out.println("warmup and gc complete. sleeping 5 seconds.");
Thread.sleep(5000l);
System.out.println("woke up - go.");
}
private static void runBenchmark(int loops) {
Label label = null;
for (int i = 0; i < loops; i++) {
label = new Label();
label.setStyleName("mainStyle");
label.addStyleName("foo bar baz");
label.addStyleName("vaadin");
}
}
}
|
