How to use Distributed Analysis in SIwave

If you are running SIwave on multiple machines, but the job is running on a single machine, you'll need to configure your job so that it will distribute throughout all the machines within your environment. 

The setup below is a walkthrough of a 256 core job and how to configure SIwave to utilize all cores.

First you need to select Pack instead of pool. To do this you'll need to open the Compute SYZ-parameters and click on "Other Solver Options"

Then click on the Multiprocessing tab and select HPC Pack instead of Pool.  This will eliminate any license error you may be receiving.

Then go into the Distributed Analysis (HPC) tab in the SYZ menu and click "Distributed simulation" and use the "Intel" MPI version. Once you do that, click on "Import Machine List"

Browse to home > nimbix > MACHINES.txt

Once you select this file and click open, it will pull these machines into the main window.  Make sure all are checked and enabled.

Click save settings and Launch.

Was this article helpful?
0 out of 0 found this helpful