SAN DIEGO, Calif., Oct. 19 — Cirrascale Corporation, a premier developer of server and cloud solutions enabling GPU-driven deep learning infrastructure, has announced the future availability of the IBM Power Systems S822LC for HPC in multiple configurations for its GPU-as-a-Service cloud platform. The configurations will support both the 8-core and 10-core POWER8 CPU and up to four Pascal architecture-based NVIDIA Tesla P100 GPU accelerators using NVIDIA NVLink, a high-speed, energy-efficient bidirectional interconnect. NVIDIA NVLink is embedded at the silicon level and tightly coupled with the CPU. This enables data transmission rates up to 80GB/s bidirectional between CPU and GPU, and 115GB/s between CPU and board memory making it 2.5x faster than competing offerings.
The service is differentiated because it offers the industry’s most up-to-date, dedicated GPUs as a bare metal offering. This gives users the full power of modern, dedicated, directly accessible GPUs and associated compute hardware, as well as ultimate flexibility in system configuration. The available configurations ensure that a wide variety of application types can be addressed including deep learning, artificial intelligence, high performance data analytics, and accelerated databases.
“Deep learning and HPC workloads require an advanced level of computing architecture between CPUs and GPUs that must go beyond standard offerings to resolve the PCI-E bottleneck,” said Sumit Gupta, Vice President, High Performance Computing and Data Analytics, IBM. “With the Cirrascale GPU-as-a-Service cloud platform offering the new IBM POWER8 with NVIDIA NVLink-based system, customers can now experience the new potential for GPU-accelerated computing across industries.”
“We have had great success with our GPU-based cloud service recently and continue to look for new ways to expand our offerings and address key market segments,” said PJ Go, president, Cirrascale Corporation. “This latest addition of an NVIDIA NVLink-based system supporting IBM POWER8 processors will provide our customers with some significantly increased performance and scalability, while still being cost effective in its overall rental model.”
The company’s bare-metal cloud service offers the ability for customers to load their very own instances of popular deep learning frameworks, such as Caffe, Torch, Theano and TensorFlow. Because Cirrascale’s offering gives users access to the raw horsepower of a modern multi-GPU system, it is proving attractive to customers with various deep learning and HPC applications.
“Data scientists are looking to extract intelligence from a tsunami of data by leveraging deep learning on GPU-accelerated computing resources,” said Roy Kim, Tesla Product Lead at NVIDIA. “With the Cirrascale cloud platform, powered by Pascal-based NVIDIA Tesla P100 GPU accelerators and NVLink, scientists and HPC users can deploy scalable compute resources on demand that deliver a dramatic boost in throughput for deep learning workloads.”
The Cirrascale cloud platform is available immediately with configurations supporting the latest GPU accelerators from NVIDIA, such as the NVIDIA Tesla P100 GPU accelerators. Customers interested in renting time can visit www.gpuasaservice.com to sign up or contact Cirrascale at (888) 942-3800.
About Cirrascale Corporation
Cirrascale Corporation is a premier developer of hardware and cloud-based solutions enabling GPU-driven deep learning infrastructure. Cirrascale leverages its patented Vertical Cooling Technology and proprietary PCIe switch riser technology to provide the industry’s densest rackmount and blade-based peered multi-GPU platforms. The company sells hardware solutions to large-scale deep learning infrastructure operators, hosting and cloud service providers, and HPC users. Cirrascale also licenses its award winning technology to partners globally. To learn more about Cirrascale and its unique multi-GPU infrastructure solutions, please visit http://www.cirrascale.com or call (888) 942-3800.