Nvidia is using its Tesla 20-series GPUs to power a new scientific tool, Amber 11, that promises to make life easier for scientists by harnessing supercomputer-like speeds on desktop PCs.
Dr. Ross Walker, a research professor at San Diego Supercomputer Centre and AMBER contributor, reckons that the Tesla GPU provides as much power as a high performance cluster of 512 CPU cores when simlulating one nanosecond in the lifetime of a 25,000 atom implicit solvent nucleosome. You bloody what?
What it means is that Nvidia’s gear is fast as heck and is basically needed to run Amber really effectively. Compared to a traditional CPU server, Nvidia’s GPUs boost the tool’s speed about 100-fold, the company claims.
While looking into bio-molecular sims, Dr. Walker noted a dozen reports of speedups over 30 times thanks to Amber 11 boosts. With the new GPUs, in place scientists can now get on with their work significantly quicker, while also being able to simulate many different scenarios at once.
It’s a game-changer for the science community, says Walker, because scientists can now do work straight from the desktop without having to lurk around menacing-looking supercomputers.