Multi-GPU deep learning cloud solutions provider, Cirrascale Cloud Services, has further expanded operations into its US-East region adding a new data center location in North Carolina. The extended expansion of this region would provide users with low latency and high bandwidth options at a reduced cost, as well as give clients the ability to access data stored on eastern storage hubs of AWS Cloud, Microsoft Azure or Google Cloud Platform with “very low latency.”

Cirrascale Cloud Services offers a dedicated, bare-metal cloud service with the ability for customers to load their very own instances of popular deep learning frameworks, such as TensorFlow, PyTorch, Caffe 2, and others.

Cirrascale Cloud Services’ cloud servers would use the latest GPU accelerators, which can draw nearly 3000W per server requiring specialized cabinet technology for increased density.

Cirrascale’s east coast facilities are outfitted with Dynamic Density Control (‘DDC’), a cabinet technology delivered by ScaleMatrix. DDC is a fully enclosed, closed-loop platform with integrated air handling, fire suppression, air filtration, access control, and both cabinet and facility-level monitoring. The proprietary cabinet technology would deliver high-density capabilities, and high power efficiency, enabling Cirrascale to deliver the latest deep learning and HPC servers.

“With some of the most advanced, scale-out deep learning hardware available in the cloud, Cirrascale Cloud Services is a perfect fit for our technology,” said Chris Orlando, CEO and co-founder, ScaleMatrix. “Our market-leading DDC enclosure platform enables us to stay ahead of today’s growing power and density needs, and provides organizations like Cirrascale with an even greater competitive advantage.”

Rapid US Growth

Multi-GPU cloud servers on the Cirrascale Cloud Services platform are already in operation at the new North Carolina US-East facility and are available for use immediately.

“Our rapid growth into the US-East market has been fueled by the increase of customers looking for providing reduced latency as well as improved bandwidth for their high-speed, scale-out application needs,” said Mike LaPan, vice president, Cirrascale Cloud Services. “We’re finding more and more companies with a desperate need for tailored deep learning solutions that provide the ability to feed GPUs at a rapid rate. Adding additional data centers allows us to ensure that we are doing whatever we can to support these needs and remain a leader in private and public deep learning cloud services.”