The entire body of the ttime function is:
-----
double ttime(void)
{
struct timeval now;
double sec;
if (gettimeofday(&now, (struct timezone *) 0)) {
return(0);
}
sec = (double) now.tv_sec + (((double) now.tv_usec) / 1000000.0);
return(sec);
}
-----
So if something is seg faulting in this routine, it is likely that there
is some other kind of memory corruption in the application.
Have you tried running it through a memory checking debugger, such as
purify, bcheck, or valgrind?
On Fri, 16 Jan 2004, Dev Anand wrote:
> I was trying to run the irs benchmark,
> but i get the following. I think ttime () is
> a lam function.
> Dev
>
> (gdb) r -k seq zrad.0001
> Starting program: /home/dev/irs-1.4/decks/irs0116 -k seq zrad.0001
> [New Thread 1077286592 (LWP 10186)]
> (c) Copyright 1996-2001 Regents of University of California
> MPI ID:PID MAPPING 0: 10186
> MPI ID:PID MAPPING 0: 10186
>
> irs version 01/16/04
>
> Program received signal SIGSEGV, Segmentation fault.
> [Switching to Thread 1077286592 (LWP 10186)]
> 0x0832a89c in ttime ()
> (gdb)
>
>
>
>
> _______________________________________________
> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>
--
{+} Jeff Squyres
{+} jsquyres_at_[hidden]
{+} http://www.lam-mpi.org/
|
