Infrastructure-as-a-Service (IaaS) firm SoftLayer has introduced a new cloud-based high performance computing (HPC) service, looking to service traditional HPC needs such as data mining, numerical and seismic analysis as well as graphic-intensive projects including interactive games, applications and 3D content. In partnership with graphics processing unit (GPU) vendor Nvidia (NASDAQ: NVDA), SoftLayer is using Nvidia’s latest GPU cards on dedicated cloud servers to make its HCP offering available to scientists, engineers and software developers.

According to SoftLayer, its HPC service offers performance up to 10 times faster than regular computing models, thanks to the dedicated Intel ED 2600-based servers working with one or two Nvidia Tesla M2090 GPUs. The offering is now generally available with an on-demand billing model.

“We are offering breakthrough performance in a unique delivery model,” said Duke Skarda, CTO of SoftLayer, in a prepared statement. “The beauty of this solution is the way users can break down to the toughest computational problems and the largest blocks of data to work them in parallel over our network driving quicker results. GPU technology is still in its infancy but customer demand continues to soar as more and more HPC applications get migrated over to cloud infrastructures.”

Key features and benefits of the HPC offering, according to SoftLayer, include:

Earlier in April 2012 SoftLayer partnered with IT distributor Synnex to make its cloud services available through the Synnex CloudSolv platform. In February 2012 the company launched  SoftLayer Flex Images, a unified server provisioning system.

Talkin’ Cloud readers can click here to read about SoftLayer’s partner program.