Windows Azure has been around for a while now and companies are getting used to the idea of migrating data services to “the cloud”. But what about high performance computing?
Scientists working with big data are often working in tools and platforms which don’t necessarily scale to cloud-based services by design, for example MATLAB (an analytical, engineering and computation tool). However, these tools provide a necessary framework for scientific computing which enables scientists to work more easily with their code – something .NET for example isn’t designed to provide.
However, one case study that has recently emerged from Finland has highlighted how one Finnish company, Techila, accelerated a cancer research study from an incredible 15 years of predicted execution time to just 4.5 days on 1,200 Azure instances.
The Techila technology allows MATLAB or R code (also often used in scientific computing) to be accelerated by distributing it across Azure, and an open API to hook in just about any other application you could possibly want.
“The fact that computational resources provided by the cloud service are available, increases prospects to conduct really demanding projects that were not possible few years ago” said Sampsa Hautaniemi of the University of Helsinki, who was involved with the research.