Beside their use for improving the graphic card on the common desktop computer, a few years ago graphics processing units (GPUs) were enabled by NVIDIA to be used as powerful computational units. This was possible by the development of the so-called compute unified device architecture (CUDA) parallel computing programming model and its implementation by the GPUs.
So, what does GPU and simulation have in common?
First let me share a video in which the famous Mythbusters compare CPU vs GPU. Check the video below:
As this technology allows a boost in computation speed, having the right software your GPU equipped laptop or desktop can act as a powerful HPC workstation and perform simulation and modeling on a scale never seen before. Back in the 80’s theoreticians were dreaming about this, now it is reality.
Beside the obvious videogames business, drug design, 3d materials simulation, and 3d visualization are among the fields in which this technology is already being used and it is extremely beneficial. Here are some links to more recent applications in materials design, virus structure determinationand unfolding of genome architecture.
There is even a molecular modeling and simulation software written from scratch, specifically designed with the CUDA architecture in mind called Terachem. Originally it was developed at the University of Illinois at Urbana-Champaign, now it is distributed by PetaChem, LLC.
Overall, the use of GPU accelerated materials modeling and simulation could prove to be disruptive for future design a novel materials. The increased speed of computation could provide significant advantage both for research labs and industry, while a significant return on investment for companies performing in-silico research in drug or material design.