From: heller@lrz.de (Helmut Heller)
Newsgroups: comp.parallel.mpi
Subject: Re: MPI_ABORT on SGI O2K
Date: 9 Sep 1998 12:54:12 GMT
Organization: [posted via] Leibniz-Rechenzentrum, Muenchen (Germany)
Distribution: world
Message-Id: <6t5tpk$ada$1@sparcserver.lrz-muenchen.de>
References: <6t5353$nic$1@fir.prod.itd.earthlink.net>
Reply-To: heller@lrz.de


In article <6t5353$nic$1@fir.prod.itd.earthlink.net> "Andy M Wissink"  
<ahwissink@earthlink.net> writes:
> My code uses MPI_ABORT to stop all processors in case any hit a
> fatal error.  It works fine on an IBM SP and a SUN HPC machine, but
> on the SGI Origin, I get the error message: "MPI_FINALIZE() not called" and
> a bunch of core dumps from the nodes.

I am trying to use MPI_Abort() on an SP2, but so far the MPI_Abort() call  
does not return! Therefore, all processes are halted, but still hang around.  
Very annoying. How did you manage to use MPI_Abort() on an SP2??

I have to add that I am using UNIX-signals in my program directly: if the  
process receives a SIGTERM, it does a few things and then calls MPI_Abort().  
Could it be that this interferes with the MPI_Abort() mechanism? Does anyone  
know??

Thanks for your help!
Helmut

--
Servus, Helmut  (DH0MAD)    ______________NeXT-mail welcome_________________
FAX: +49-89-280-9460        "Knowledge must be gathered and cannot be given"
heller@lrz.de                                       ZEN, one of BLAKES7
Phone: +49-89-289-28823     ------------------------------------------------
Dr. Helmut Heller                                    
Leibniz-Rechenzentrum (LRZ)