1 Java日志框架性能比较
前面几章,笔者分别介绍了log4j,logback,log4j2三大日志实现框架。
接下来,就用具体的数据比较下,哪个日志框架的性能更好!
单线程:外循环100次,内循环100000次;
多线程:开启100个线程,每个线程执行100000次;
1.1 测试代码:
(1)log4j:
public class log4jDemo {
Logger logger = Logger.getLogger(log4jDemo.class);
@Test
public void testThread() throws InterruptedException {
int THREAD_NUM = 100;
final int LOOP_NUM = 100000;
final CountDownLatch countDownLatch = new CountDownLatch(THREAD_NUM);
long start = System.currentTimeMillis();
for(int x= 0;x < THREAD_NUM;x++){
new Thread(new Runnable() {
public void run() {
for (int y = 0; y < LOOP_NUM; y++) {
logger.info("Info Message!");
}
countDownLatch.countDown();
}
}).start();
}
countDownLatch.await();
System.out.println(System.currentTimeMillis() - start);
}
@Test
public void test() throws InterruptedException {
int X_NUM = 100;
int Y_NUM = 100000;
long start = System.currentTimeMillis();
for(int x=0;x < X_NUM;x++) {
for (int y = 0; y < Y_NUM; y++) {
logger.info("Info Message!");
}
}
System.out.print(System.currentTimeMillis() - start);
}
}
(2)logback:
public class logbackDemo {
Logger logger = LoggerFactory.getLogger(logbackDemo.class);
@Test
public void testThread() throws InterruptedException {
int THREAD_NUM = 100;
final int LOOP_NUM = 100000;
final CountDownLatch countDownLatch = new CountDownLatch(THREAD_NUM);
long start = System.currentTimeMillis();
for(int x= 0;x < THREAD_NUM;x++){
new Thread(new Runnable() {
public void run() {
for (int y = 0; y < LOOP_NUM; y++) {
logger.info("Info Message!");
}
countDownLatch.countDown();
}
}).start();
}
countDownLatch.await();
System.out.println(System.currentTimeMillis() - start);
}
@Test
public void test() {
int X_NUM = 100;
int Y_NUM = 100000;
long start = System.currentTimeMillis();
for(int x=0;x
for (int y = 0; y < Y_NUM; y++) {
logger.info("Info Message!");
}
}
System.out.print(System.currentTimeMillis()-start);
}
}
(3)log4j2:
public class log4j2Demo {
private Logger logger = LogManager.getLogger(log4j2Demo.class);
@Test
public void testThread() throws InterruptedException {
int THREAD_NUM = 100;
final int LOOP_NUM = 100000;
final CountDownLatch countDownLatch = new CountDownLatch(THREAD_NUM);
long start = System.currentTimeMillis();
for(int x= 0;x < THREAD_NUM;x++){
new Thread(new Runnable() {
public void run() {
for (int y = 0; y < LOOP_NUM; y++) {
logger.info("Info Message!");
}
countDownLatch.countDown();
}
}).start();
}
countDownLatch.await();
System.out.println(System.currentTimeMillis() - start);
}
@Test
public void test() throws InterruptedException {
int X_NUM = 100;
int Y_NUM = 100000;
long start = System.currentTimeMillis();
for(int x=0;x
for (int y = 0; y < Y_NUM; y++) {
logger.info("Info Message!");
}
}
System.out.print(System.currentTimeMillis() - start);
}
}
1.2 配置文件:
(1)log4j:
(2)logback:
e:/log.out
true
%d{HH:mm:ss.SSS} %p %c - %m%n
e:/log.out
true
false
8192
%d{HH:mm:ss.SSS} %p %c - %m%n
0
128
(3)log4j2:
<
%d{HH:mm:ss.SSS} %p %c - %m%n
immediateFlush="false" bufferedIO="true" bufferSize="8192">
%d{HH:mm:ss.SSS} %p %c - %m%n
1.3 结果比较(毫秒)
笔者从单线程、多线程2种情况下进行的测试!
无论是多线程还是单线程,在启用缓存的情况下,系统性能得到了巨大的提升;
在单线程情况下,相比较来说,启用异步Appender并没有对性能有较大的提升!
值得一提的是,在log4j2中,多线程情况下,相对于同步logger来说,异步logger并没有进一步提高系统的性能,两者不相上下;
但是,对于其他情况而言,异步logger还是有较大的提升!
单线程
(1)单线程,未开启缓存,立即刷出
log4j:29772、29959、30911
logback:25423、24552、26006
log4j2:37927、38240、40164
(2)单线程,开启缓存,不立即刷出
log4j:9858、9677、9665
logback:5561、5604、5611
log4j2:5782、5505、5499
(3)单线程,异步appender,未开启缓存,立即输出
log4j:29683、29929、29385
logback:33102、31779、30516
log4j2:39298、39562、41872
(4)单线程,异步appender,开启缓存,不立即输出
log4j:10110、10068、10177
logback:8753、9112、8922
log4j2:8692、8400、8252
多线程
(1)多线程,未开启缓存,立即刷出
log4j:38541、37791、38366
logback:35644、35463、35442
log4j2:38544、38746、38706
(2)多线程,开启缓存,不立即刷出
log4j:13296、12938、12686
logback:6547、6294、6576
log4j2:5596、5423、5421
(3)多线程,异步appender,未开启缓存,立即输出
log4j:30844、32088、30734
logback:44203、42191、43228
log4j2:46804、46034、46232
(4)多线程,异步appender,开启缓存,不立即输出
log4j:10422、10204、10495
logback:40249、40437、40173
log4j2:7832、8447、8660
(5)多线程,异步logger,未开启缓存,立即输出
log4j2:40555、40245、40325
(6)多线程,异步logger,开启缓存,不立即刷出
log4j2:5319、5407、5305