> -------- Original Message --------
> Subject: Re: Using argv[1] in C program
> From: "Jerry Davis" <jfdecd@wi.rr.com>
> Date: Sun, January 16, 2005 9:05 am
> To: plug-discuss@lists.plug.phoenix.az.us

> [snip]

> #include <stdio.h>
>
> void usage(void);
>
> int main(int argc, char** argv) {
>   if ( argc > 1 ) {
>     if (isdigit(argv[1][0])) {
>       usleep((unsigned long) atol(argv[1]));
>       return(0);
>     }
>     if ((strcmp(argv[1],"--usage") == 0)) {
>       usage();
>       return(0);
>     } else {
>       printf("\nusleep: bad argument %s: unknown option\n",argv[1]);
>       usage();
>       return(-1);
>     }
>   }
>   usage();
>   return(-1);
> }
>
> void usage(void) {
>   printf("usage: usleep [--usage] [microseconds]\n");
> }

> [/snip]

Thank you to all that participated in this fun little script, I learned
a couple things :)

I have replaced my system's usleep binary with a compile of the above
code.

How much overhead does adding stdio.h add?  I mean if I want to usleep
1000, in order to be accurate, I would want the usleep binary to be as
small as possible.

So, I mini benchmarked it...

usleep2 is the stripped down version without stdio.h
usleep is the full version stdio.h and useage output

let i=0; time while [ $i -lt 1000000 ]; do let i=$i+100; ./usleep2 1000;
done
real    0m59.565s
user    0m5.230s
sys     0m15.377s

let i=0; time while [ $i -lt 1000000 ]; do let i=$i+100; ./usleep 1000;
done
real    0m59.201s
user    0m5.099s
sys     0m15.429s

Looks like including stdio.h makes it a tad faster!  Just kidding, the
variance is so small it can be disregarded.  Cool.




---------------------------------------------------
PLUG-discuss mailing list - PLUG-discuss@lists.plug.phoenix.az.us
To subscribe, unsubscribe, or to change  you mail settings:
http://lists.PLUG.phoenix.az.us/mailman/listinfo/plug-discuss