Mesh refinement in parallel
I have done simulations of a propagating shock with dynamic mesh refinement (using rhoCentralFoam).
This all works. However, when running parallel I get issues with load balancing. If the shock happens to be in a domain belonging to a few of the processor these processors get to do all the extra work while the remaining processors sit idle for a large fraction of time - which is obviously ineffective.
Is there a way to handle this?
Any updates on this?
|All times are GMT -4. The time now is 18:01.|