Page 1 of 1

Big model problem with OpenSeesSP

Posted: Mon Nov 25, 2013 8:27 pm
by nvzitejing
Hi, everyone

I run a big soil model using OpenSeesSP, getting a error message:

terminate called after throwing an instance of 'std::bad_alloc'
what(): std::bad_alloc

=====================================================================================
= BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
= EXIT CODE: 6
= CLEANING UP REMAINING PROCESSES
= YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
=====================================================================================
APPLICATION TERMINATED WITH THE EXIT STRING: Aborted (signal 6)

How to solve it ?

Thanks in advance!

regard


nvzitejing

Re: Big model problem with OpenSeesSP

Posted: Wed Nov 27, 2013 10:39 am
by fmk
any more info? can you build the model and this fails when the analyze command is issued?

Re: Big model problem with OpenSeesSP

Posted: Thu Nov 28, 2013 4:20 am
by nvzitejing
Hi,fmk
The problem have been solved. I used -genBandArpack (default) as solver of eigen analysis. In fact with the model becoming bigger, genBandArpack was not available. Now I use SuperLU, the computing can be run. Which solver is the best using OpenSeesSP? Sometimes Mumps is not available.
Thank you for your reply!

Best wishes!

nvzitejing

Re: Big model problem with OpenSeesSP

Posted: Mon Dec 02, 2013 9:13 am
by fmk
mumps is the fastest, but only works for parallel. though you can run them as sequential applications by using n=1 with mpiexec or just starting them as you would OpenSees (at least that works on my machine, not sure if mpi needs to be installed first)