blow up

Report or discuss software problems and other woes

Moderators: arango, robertson

Post Reply
Message
Author
rao_sc
Posts: 9
Joined: Tue Aug 04, 2009 3:00 pm
Location: Naval Physical & Oceanographic Laboratory

blow up

#1 Unread post by rao_sc »

Hi all,
I am running a ROMS for Indian ocean. I successfull ran the model on a cluster. But when I want to run the same model configuration without any changes on a multiprocessor (8 processors) HP machine, it getting blow up with same time step (1800sec). What may be problem ?
At the same time, when I want to run the same configuration for 1/6th Indian West coast, it is blowing up with whatever timestep. I am not getting what will be the problem.
Thanks in advance.

Regards,
GNRao.

User avatar
kate
Posts: 4091
Joined: Wed Jul 02, 2003 5:29 pm
Location: CFOS/UAF, USA

Re: blow up

#2 Unread post by kate »

We can't tell from here what the problems are - you will have to do some investigation on your own. If you have a domain running on a cluster but not on a smaller parallel system it could be a memory issue. It could be that the smaller system has a problem in its MPI. It could be a problem in the output if you are writing at that timestep.

Any new domain can also have a multitude of reasons to not run. The first thing to try is a shorter timestep. If you can get it to run for a step or two, then blows up, you can try writing a file and making plots from it. Or learn to use a debugger.

rao_sc
Posts: 9
Joined: Tue Aug 04, 2009 3:00 pm
Location: Naval Physical & Oceanographic Laboratory

Re: blow up

#3 Unread post by rao_sc »

thank u kate
i think last one is right, i.e.,writing problem. Because it was blown up at the time step where it has to write avg, his files. then what will be the problem.

thanks in advance.

GNRao

Post Reply