On Tue, Jan 04, 2005 at 06:41:24PM -0500, Jacob Vos wrote:
>With usysv: always...
>With sysv: not yet...
ok, that's weird.
I can't see a LAM + usysv problem with
2.4.25 kernel + redhat 7.3 + LAM 7.1a1cvs20031031
or with
2.6.10 kernel + fedora core 3 + LAM 7.2b1svn-2004-11-23
maybe OS X is buggy, or Linux is masking a LAM bug?
I changed a few lines in the code to be straight C instead of C++, but I
doubt if that changes anything. code attached.
cheers,
robin
>Hardware: Dual 2.5 GHz PowerPC G5, 4 GB DDR SDRAM
>OS: Mac OS X Version 10.3.6, using ssh from a Windows XP Box
>LAM: 7.1.1 "localhost cpu=8"
>
>./configure --with-rsh=ssh --with-memory-manager=none
>--prefix=/user/local/lam
>
>and $FC=xlf
>
>
>-----Original Message-----
>From: Robin Humble [mailto:rjh_at_[hidden]]
>Sent: Tuesday, January 04, 2005 6:04 PM
>To: General LAM/MPI mailing list
>Subject: Re: LAM: data corruption
>
>
>How often does this error occur?
>I just ran 30 of these with sysv and 30 with usysv and didn't see any
>problems.
>
>config is dual Xeon, linux 2.4.25 kernel, nodes file 'localhost cpu=8',
>LAM is 7.1a1cvs20031031, with
> mpirun -ssi rpi usysv -c 8 program
>or
> mpirun -ssi rpi sysv -c 8 program
>
>cheers,
>robin
>_______________________________________________
>This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>
>
>
>_______________________________________________
>This list is archived at http://www.lam-mpi.org/MailArchives/lam/
|
