Saeed,
This might need some procedural debugging, or be directed to your HPC admin/helpdesk. Generally, signal 9 means that the simulation exceeded its resources. Given that your grid just got four times bigger, you might run into space allocation issues for your global variables/arrays — you have nearly 13 million double variables (if you compiled in double precision) per 2D variable of field size.
On our HPC distributed systems, if this arose, we would submit hybrid jobs for more nodes, but assign fewer cores per node to the job (say half or quarter), this way the shared memory that sits on each node would shared across fewer cores. Hence, each core (or thread) would have more memory at its disposal to accommodate larger arrays.
Matt
Dear FUNWAVE Users,