I have a very compact code example (10 lines) for which running java interpreter
on a class file takes 4.5s and executing optimized binary created from the class
file with gcj takes 5.5s. Something is funny here. Can anyone explain what's
happening?
% uname -a
Linux localdomain 2.6.12-1.1372_FC3 #1 Fri Jul 15 00:59:10 EDT 2005 i686 i686
i386 GNU/Linux
>>>>>>>>>>> The java code (X.java):
import java.lang.*;
import java.util.*;
import java.io.*;
public class X {
public static void main(String[] args) {
Random R = new Random();
for (int i=0; i < 10000000; i++)
R.nextDouble();
}
}
>>>>>>>>>>The Makefile
all:
@javac X.java
@echo "*** Interpreted:"
@time java -classpath . X
@gcj -O3 --main=X -o x X.java
@echo "*** Compiled:"
@time ./x
>>>>>>>>>>Here is the output:
*** Interpreted:
4.30user 0.02system 0:04.51elapsed 95%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (1major+2734minor)pagefaults 0swaps
*** Compiled:
5.55user 0.01system 0:05.58elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2069minor)pagefaults 0swaps
--
Summary: Java interpreter significantly faster than gcc?!
Product: gcc
Version: 3.4.4
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: java
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: netzberg at gmail dot com
CC: gcc-bugs at gcc dot gnu dot org,java-prs at gcc dot gnu
dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23283
