SoftLayer Targets GPUs in the Cloud
April 17, 2012
Graphics processing units (GPUs) are taking center stage in SoftLayer’s latest high-performance computing (HPC) strategy in the cloud, as the company now offers HPC servers with NVIDIA GPUs starting at US$879 per month for an entry-level configuration of one GPU, 16GB of RAM, and 500GB of storage.
Dallas-based SoftLayer is targeting traditional scientific environments – think oil & gas and other seismic applications – along with other numerical analyses, data mining, as well as "advertising agencies and web-design shops looking to develop interactive games, applications, and 3D content," according to Nathan Day, SoftLayer’s Chief Scientist…
"We’ve seen demand from our cusomers doing video rendering and biotech, for example. We’re responding to that demand by putting a product out there they can use (immediately). Now they don’t have to put out the capital expenditure, but can just get what they need through a consumable model," he adds.
Although SoftLayer counts very large companies among its customers, Day points out that small companies can really take advantage of the company’s new cloud services. "Many small companies are the risk takers who take advantage of innovation, in the oil and field services businesses (among others)," he says. "Our new service offers a great way to leverage that innovation and apply it to data processing and video rendering."
Day also notes that SoftLayer offers "GPUs in a dedicated manner. Our customers are used to having a physical server they can reach out and touch. Now they can have a dedicated (graphics) server provisioned in under two hours, and don’t have to pay the hypervisor (provider)."
Philbert Shih, managing director at SoftLayer customer Structure Research Ltd. said, in an official statement, "Giving people the ability to scale GPU resources adds incredible flexibility in the way research, engineering and development projects are conducted. By putting such raw computational power into a user’s hands, in an easy pay as you go model, the focus is on research results not acquiring or maintaining the latest technology."
SoftLayer’s new HPC hosting service integrates dual-processor Intel E5-2600 (Sandy Bridge) servers supporting one or two NVIDIA Tesla M2090 GPUs. GPUs work in conjunction with the server’s CPU to accelerate application and processing performance, the company says. The GPU will offload compute-intensive portions of the application from the CPU to accelerate overall performance of the machine, the company says, "dramatically boosting overall application performance."
SoftLayer now has a global cloud infrastructure that spans 13 datacenters, including locations in The Netherlands and Singapore. The company has 100,000 servers under management, and says it is the largest privately held IaaS provider in the world.