[mvapich-discuss] List of all Exit Codes

[Jonathan Perkins]

On Thu, Nov 15, 2012 at 03:27:57PM +0100, hecke wrote:
> Dear Jonathan,
> 
> thanks a lot for your quick reply.
> 
> Actually, I already found the error, which was addressing a vector
> on processes, where it was not initialized.

Glad that you found the cause.

> 
> Nevertheless, I would be glad to have a place to look up error codes.
> You said, that negative ones are just forwarded from the system
> signals, right?
> Is there any documentation on this?

To determine what the error code meant I looked at `man 7 signal'.  As
far as a general error code documentation, that doesn't exist for exit
codes since that should be related to the application and not the MPI
library.  The MPI library will return various return codes through the
MPI functions which is documented by the MPI standard (in this case, MPI
1.2).

> 
> Concerning the version we use, I will have a word with our sysad,
> but usually, he has severe reasons for using certain versions. Even
> if the reason is only: That one works, so I don't change it;-)

That's reasonable.  However if you do find something that you consider a
bug, please do upgrade as newer versions may have resolved the bug as
well performance and functionality enhancements.

> 
> cheers
>     Hecke
> 
> 
> On 11/15/2012 03:18 PM, Jonathan Perkins wrote:
> >On Thu, Nov 15, 2012 at 01:38:25PM +0100, hecke wrote:
> >>Hi,
> >>
> >>even though i crawled the entire web, I was not able to find a list
> >>of all exit codes of mvapich.
> >>
> >>I am using (and therefore particularly interested in) version 1.2.0
> >>on a Mellanox IB network.
> >>
> >>Currently, my program crashes more or less irregularly with exit code -5.
> >>
> >>thanks a lot
> >>      Hecke
> >I believe this is a caught SIGTRAP signal (negative exit code usually
> >means it was an unhandled signal).
> >
> >According to the signal man page this is
> >        SIGTRAP        5        Core    Trace/breakpoint trap
> >
> >I don't believe this is coming from mvapich but from your application.
> >Can you get a backtrace from the core file that should have been
> >generated?
> >
> >Also please note that mvapich-1.2 is old and unsupported.  I strongly
> >suggest that you update to the latest stable mvapich2 (1.8.1) or try our
> >latest release (1.9a2).
> >
> 

> begin:vcard
> fn:Hecke
> n:Degering, geb. Schrobsdorff;Hecke
> org:Max-Planck Institute for Dynamics and Self-Organization;Nonlinear Dynamics
> adr;quoted-printable:;;Am Fassberg 17;G=C3=B6ttingen;;37077;Germany
> email;internet:hecke at nld.ds.mpg.de
> tel;work:+49-551-5176441
> tel;fax:+495176439
> x-mozilla-html:FALSE
> url:http://www.nld.ds.mpg.de/~hecke
> version:2.1
> end:vcard
> 

> _______________________________________________
> mvapich-discuss mailing list
> mvapich-discuss at cse.ohio-state.edu
> http://mail.cse.ohio-state.edu/mailman/listinfo/mvapich-discuss


-- 
Jonathan Perkins
http://www.cse.ohio-state.edu/~perkinjo