Speeding up Science

URL

http://www.scientificcomputing.com/articles-HPC-Speeding-up-Science-111510.aspx

Description

Scientific simulations can tackle larger problems and produce more accurate answers, as computing grows increasingly parallel. This need for parallel processing proves particularly crucial for some questions, including how to make better drugs and how elements formed after the Big Bang. Such simulations may require millions of iterations, with results feeding back into more calculations. In the past, the parallelism came from using more CPUs — either from a cluster or from a multicore chip — but graphics processing units (GPUs) now offer an alternative. As a result, today’s scientific calculations run faster than ever.