With Neural Concept Shape, you can use 3D numerical simulations as input to train your deep learning models. If we take the example of aerodynamic simulations, these CFD simulations results are usually much larger files than images or text (a single result can reach several hundreds of GB). Hence storing a large amount of them can become an issue in the long term, as it would require to regularly scale up the hardware infrastructures accordingly.
The ability to do more faster and better with a GPU solver, can increase your resources by 10x. Imagine going fast and then the turbo kicks in; that’s what the cloud can do.
This white paper discusses how synthetic datasets for training AI can be generated in hours using the OnScale cloud simulation platform. The demonstrated approach of using synthetic datasets to train AI networks can drastically reduce cost, risk, and time for the development of new hardware technologies.
In this June, 2018 post, Digital Engineering's Brian Albright discusses Cloud-based Simulation.
Opin Kerfi‘s HPC resources and UberCloud’s HPC software container for Abaqus provided the compute power and environment needed to reduce the simulation time and allowed a large number of iterations to be run in order to get the optimal design and accurate device placement.
