Massive simultaneous cloud computing (MSCC) for multiphysics-simulation applications
Gernot Kurt Boiger, Matthew Everitt, Darren Sharman, Marlon Boldrini · Zürcher Hochschule für Angewandte Wissenschaften digital collection (Zurich University of Applied Sciences) · 2020
The Massive Simultaneous Cloud Computing concept allows appliers and developers of Multiphysics simulation software to utilize any number of cloud-based computers simultaneously. Thus novel workflows in simulation based research and development can be introduced, focusing on simultaneous rather than on sequential computation of individual cases. On the basis of this capability the simulation-researcher/-engineer can (i) dramatically increase the parameter space of any computational parameter study, (ii) devise novel concepts of conducting optimization procedures and (iii) can ultimately even proceed to produce sufficient simulation data in order to train e.g. artificial neuronal networks. The cloud-based software platform Kaleidosim has been devised to effectively enable the handling of MSCC simulation run series of up to 500 simultaneous cloud runs. Case studies will be presented where a speed up factor of 100 has been achieved in terms of comparing MSCC based parameter studies to the standard workflow on in-house hardware.