[GE users] Open MPI tight integration in HOWTO page

Heywood, Todd heywood at cshl.edu
Thu Feb 1 23:06:31 GMT 2007


Hi Reuti,

No, it is regular IP (gigE).

By "hang", I mean the daemons don't start up, but you have to ^C to get
a prompt back after doing typing "smpd <options> -d 0" (I was trying to
figure out how the integration worked, and was not starting the
daemons).

Todd

-----Original Message-----
From: Reuti [mailto:reuti at staff.uni-marburg.de] 
Sent: Thursday, February 01, 2007 5:46 PM
To: users at gridengine.sunsource.net
Subject: Re: [GE users] Open MPI tight integration in HOWTO page


--cut---


Are you using any special communication lib? Myrinet, Infiniband,... ?
>
>
>
>
> I've also installed MPICH2 (SMPD daemon-based), and found some  
> issues with the .smpd configuration file being corrupted when too  
> many tasks access it at once. Having solved that to run large jobs  
> successfully outside of SGE, I tried Reuti's tight integration, and  
> found that smpd daemons hang when started with the "-d 0" option  
> (hard coded into the start_mpich2 program). But that's another story.
The -d 0 option I got from the MPICH2 developers. It's purpose is to  
avoid the forking of the daemons (and leaving the process tree). The  
forking is instead handled by the start_mpich2 program. So: what do  
you mean with "hang" in detail? After the start of the daemons by the  
PE start_proc_args they should stay there, still bound to the  
shepered and wait for connections.

-- Reuti


---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net




More information about the gridengine-users mailing list