你来测测看谁说C比Java快？_Software_Java

你来测测看谁说C比Java快？

Posted by: maomao Date: April 25, 2008 09:58PM

MegaEntry - Social networking and discussion site!

我之前在某人的博客中看到一篇文章，他比较了很多语言的运行速度，包括Ruby、Io、 PHP、Python、Lua、Java、Perl、Applescript、TCL、ELispe、Javascript、OCaml、Ghostscript和C等。数据如下：

语言	时间	相对速度
ocaml compiled 3.09.2	0.05 seconds	1.00 x
SBCL 1.0.2	0.13 seconds	2.43 x
C gcc-4.0.1	0.14 seconds	2.67 x
Java 1.4.2	0.39 seconds	7.49 x
Lua 5.1	1.25 seconds	23.81 x
Io 20070410 Vector	1.37 seconds	26.13 x
ocaml bytecode 3.09.2	3.75 seconds	71.48 x
Python 2.5.1	9.99 seconds	190.33 x
Ghostscript 8.51	11.79 seconds	224.51 x
Perl 5.8.6 Optimized	12.37 seconds	235.57 x
TCL 8.4 Optimized	16.00 seconds	304.76 x
Perl 5.8.6	21.75 seconds	414.29 x
PHP 5.1.4	23.10 seconds	440.05 x
Javascript SpiderMonkey v1.6	31.14 seconds	593.10 x
Ruby 1.8.4	33.05 seconds	629.54 x
Emacs Lisp	47.00 seconds	895.24 x
Applescript	71.75 seconds	1366.67 x
Io 20070410	85.44 seconds	1627.47 x

可以看到Java几乎比C慢两倍！

但是接着我发现他用的是Java的老版本而且只实验了一次，这实际上并不能够真实地体现Java的速度。

于是我迅速写了点代码运行100三次，并用我认为 “快”的方式运行（其实还有更快的，只是我比较懒没有设定）。结果如下：

$ java -server -XX:CompileThreshold=1 Mandelbrot 2>/dev/nullJava Elapsed 2.994Java Elapsed 1.926Java Elapsed 1.955
$ gcc -O8 mandelbrot.c$ ./a.out 2>/dev/nullC Elapsed 2.03C Elapsed 2.04C Elapsed 2.05

C仍然赢得了第一轮，但接下来的两次明显Java快。

当然，结果会因为代码和机器的不同而不一样，但有一点可以确定：Java运行相当快。

这次测试用的版本是Java 1.6.0—b105 和gcc 4.1.2，用的电脑是苹果，内核是酷睿双核2.33GHz ，Linux操作系统。我测试用的代码为：

Java测试用的代码

import java.util.*;

class MandelbrotCopyRight owned by the original author.--(www.MegaEntry.com)
{  static int BAILOUT = 16;static int MAX_ITERATIONS = 1000;

     private static int iterate(float x, float y){float cr = y-0.5f;float ci = x;float zi = 0.0f;float zr = 0.0f;MegaEntry - Social networking and discussion site!
int i = 0;while (true) {i  ;float temp = zr * zi;float zr2 = zr * zr;float zi2 = zi * zi;zr = zr2 - zi2   cr;zi = temp   temp   ci;MegaEntry - Social networking and discussion site!
if (zi2   zr2 > BAILOUT)return i;if (i > MAX_ITERATIONS)return 0;}}

     public static void run2(){int x,y;CopyRight owned by the original author.--(www.MegaEntry.com)
for (y = -39; y < 39; y  ) {System.err.print("\n");for (x = -39; x < 39; x  ) {if (iterate(x/40.0f,y/40.0f) == 0) System.err.print("*");elseSystem.err.print(" ");}CopyRight owned by the original author.--(www.MegaEntry.com)
}          }

  public static void run() {Date d1 = new Date();for (int i = 0; i < 100; i  ) run2();Date d2 = new Date();long diff = d2.getTime() - d1.getTime();System.out.println("\nJava Elapsed "   diff/1000.0f);}

            public static void main(String args[]) {MegaEntry - Social networking and discussion site!
run();run();run();}}

C测试用的代码

#include #import #define BAILOUT 16#define MAX_ITERATIONS 1000

MegaEntry - Social networking and discussion site!

int mandelbrot(float x, float y){float cr = y - 0.5;float ci = x;float zi = 0.0;float zr = 0.0;int i = 0;while(1) {

MegaEntry - Social networking and discussion site!

i ;float temp = zr * zi;float zr2 = zr * zr;float zi2 = zi * zi;zr = zr2 - zi2 cr;zi = temp temp ci;if (zi2 zr2 > BAILOUT)return i;

if (i > MAX_ITERATIONS)return 0;} }

void run2() {int x,y;for (y = -39; y < 39; y ) {fputs("\n", stderr);for (x = -39; x < 39; x ) {

MegaEntry - Social networking and discussion site!

int i = mandelbrot(x/40.0, y/40.0);if (i==0)fputs("*", stderr);elsefputs(" ", stderr);} }fputs("\n", stderr);

MegaEntry - Social networking and discussion site!

}

void run() {struct timeval aTv;gettimeofday(&aTv, NULL);long init_time = aTv.tv_sec;long init_usec = aTv.tv_usec;int i;for (i = 0; i < 100; i )

MegaEntry - Social networking and discussion site!

run2();gettimeofday(&aTv,NULL);double query_time = (aTv.tv_sec - init_time) (double)(aTv.tv_usec - init_usec)/1000000.0; printf ("C Elapsed %0.2f\n", query_time);}

int main (int argc, const char * argv[]) {run();run();

MegaEntry - Social networking and discussion site!

run();}

处于好玩，我还在Rhino编辑器上运行了JS的测试：

$ java -cp rhino1_6R5/js.jar -server -XX:CompileThreshold=1 org.mozilla.javascript.tools.shell.Main -O 9 mandelbrot.js 2>/dev/nullJavaScript Elapsed 21.95JavaScript Elapsed 17.039JavaScript Elapsed 17.466JavaScript Elapsed 17.147

在这个测试中，JS比C慢九倍。如果CPU的速度没18个月翻一番，那么2007年JS的运行速度就赶上2002 C 的速度了。

后来在C语言的测试中加了些cpp，-march=pentium4发挥了点作用，但结果还是比Java慢：

$ gcc -O9 -march=pentium4 mandelbrot2.c$ ./a.out 2>/dev/nullCopyRight owned by the original author.--(www.MegaEntry.com)
C Elapsed 1.99C Elapsed 1.99C Elapsed 1.99

Prev:用Java解决国际化问题

Next:Java编译器对于String常量表达式的优化