Why system gcc that is 4.2.1 produces different code than gcc-4.2.1 compiled from sources?

2010-04-25 Thread Yuri 

I ran this simple example:

#include stdlib.h
#include stdio.h

int fib(int AnArg) {
 if (AnArg = 2) return (1);
 return (fib(AnArg-1)+fib(AnArg-2));
}

int main(int argc, char* argv[]) {
 int n = atoi(argv[1]);
 printf(fib(%i)=%i\n, n, fib(n));
}

through system gcc and gcc built from sources. Both are 4.2.1. Options: -O3.

Average runtime with the argument 45 is quite different: system gcc is 
3.650s and gcc-4.2.1 from sources is 3.740.

CPU: Intel(R) Core(TM) i7 CPU   @ 9200 @ 2.67GHz

Yuri
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org

Re: Why system gcc that is 4.2.1 produces different code than gcc-4.2.1 compiled from sources?

2010-04-25 Thread Leonidas Tsampros 
Yuri y...@rawbw.com writes:

 I ran this simple example:

 #include stdlib.h
 #include stdio.h

 int fib(int AnArg) {
  if (AnArg = 2) return (1);
  return (fib(AnArg-1)+fib(AnArg-2));
 }

 int main(int argc, char* argv[]) {
  int n = atoi(argv[1]);
  printf(fib(%i)=%i\n, n, fib(n));
 }

 through system gcc and gcc built from sources. Both are 4.2.1. Options: -O3.

 Average runtime with the argument 45 is quite different: system gcc is
 3.650s and gcc-4.2.1 from sources is 3.740.
 CPU: Intel(R) Core(TM) i7 CPU   @ 9200 @ 2.67GHz


Hi,

I'm pretty sure that a small difference in execution time does not mean
that the produced code is different. In order to check the code gcc
produced, you need to do something like this:

objdump -d executable

or even use gdb's disassemble command.

Regards,
Leonidas

 Yuri

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org

Re: Why system gcc that is 4.2.1 produces different code than gcc-4.2.1 compiled from sources?

2010-04-25 Thread Michaël Grünewald 

Leonidas Tsampros wrote:

I'm pretty sure that a small difference in execution time does not mean
that the produced code is different.


Actually, execution time of a process is very sensitive to the 
environment of this process. See for instance:


http://www-plan.cs.colorado.edu/diwan/asplos09.pdf
  We see that something external and orthogonal to the program,
   i.e., changing the size (in bytes) of an unused environment variable,
   can dramatically (frequently by about 33% and once by almost 300%)
   change the performance of our program.

(The quotation is taken out of the article.)

I learned about this in a message from Xavier Leroy to the OCaml mailing 
list:


http://caml.inria.fr/pub/ml-archives/caml-list/2009/12/e261eeb95bec6c5a2791335c84234a05.en.html

Cheers,
Michaël
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org