Re: Convergence Problem to Simulate Natural Convection (Buoyancy Driven Flow) Simulation
Posted: Thu Jun 10, 2010 8:13 pm
Hello,
I do not know why the job was killed, but I forgot to mention one thing: if you have an error or several error_n* files, check their content. This is important especially if you are running in parallel, as if the code crashes on one processor, it will kill instances on other processors. This means that the interesting message may not be the one in "error", but the one in error_n0014 if the cause of the problem originated on processor 14 for example.
Also, if the code stops at 30000 iterations, you may want to run 29000 iterations, check your results, then run a follow-up (restarted) calculation.
Best regards,
Yvan
I do not know why the job was killed, but I forgot to mention one thing: if you have an error or several error_n* files, check their content. This is important especially if you are running in parallel, as if the code crashes on one processor, it will kill instances on other processors. This means that the interesting message may not be the one in "error", but the one in error_n0014 if the cause of the problem originated on processor 14 for example.
Also, if the code stops at 30000 iterations, you may want to run 29000 iterations, check your results, then run a follow-up (restarted) calculation.
Best regards,
Yvan