I’m interested in setting up a visualization cluster for ParaView, and am currently considering which hardware to buy. I have two questions:
Are there any problems with running ParaView on a cluster with IBM POWER9 servers–for example, the AC922, with either 4 or 6 GPUs per server?
Are there any general guidelines for determining how many servers or GPUs would be required for the cluster based on the properties of the model to be visualized, such as the memory required for the model and its number of degrees of freedom?
Thank you in advance for your help.