Parallel Processing
 

Hi,

I use ICEM on a Linux cluster for creating and meshing a complicated geometry. Since the geometry is huge and composed of different parts, it probably contains more than 15 million mesh elements. Is there any way that I can run ICEM in parallel (such that we run Fluent in parallel on different nodes of the cluster)?

I would appreciate your help.

Although this is possible on windows platform but have no idea abt Linux.

Thank you for your response.

Could you please explain how can I do this in windows?

settings > general > no.of processors. I guess this is only valid for local machines with no. of processors.

I'm attempting to do the same thing with a CFX mesh in Workbench.

My machine meshes in a reasonable amount of time if nodes are under 1M, but any more than that and I'm wasting my day sitting around waiting for the thing to mesh.

The frustrating part is that once I get the setup file (.def), I can submit it to the Linux server and run on 24 cores, or more, giving me the solution in a matter of minutes.

Why isn't there a way to do this with the mesh? Is the problem that a patch-conforming tetra mesh algorithm cannot be partitioned?

Thanks,
-Trues