An Ignition Computing project on accelerating simulations has been awarded a research grant (MIT Viability Study)!
In this project we intend to demonstrate the viability of accelerating simulations by providing pseudo-solutions (preconditioners). These can lead to a greatly improved ‘time to solve’, increasing the productivity of simulation engineers and potentially reducing energy use of simulations.
Some up-front training time is needed to generate these preconditioners. This can be reused across many later simulations, potentially leading to additional energy and cost savings.
We are currently looking for partners who develop or use FEM codes, and who would be interested in exploring the potential of this approach together. If this sounds interesting, please contact us.