Cray will feature these processors in their systems to enable the market to accelerate the addition of AI methodologies to their existing HPC simulation and modelling capabilities. Technologies in Intel Cascade Lake processors, like Intel Deep Learning Boost, will offer customers significant improvements in inference acceleration in a variety of image analysis and feature detection use cases. Cray customers will benefit from XC50 and CS500 systems with large system memory to analyze highly complex images.
The use cases that these types of images are typically found include cryo-electron microscopy (Cryo-EM). For Cryo-EM, researchers are taking advantage of new techniques to capture detailed images of the architecture of cells, proteins and viruses for medical research. In Cryo-EM research, performing image inference for feature detection on full fidelity images is a key requirement for successful analysis. This type of analysis is ideally aligned with the capabilities of Cray supercomputers powered by 2nd gen Intel Xeon Scalable processors.
"The market is entering a new era of exascale-class computing where customers will increasingly use a combination of HPC and AI methodologies to aid in their work and research. To power this new era, customers will have an ever-increasing need for the latest in processing capabilities optimized for this convergence", stated Fred Kohout, senior vice president and CMO at Cray. "The added support for the new 2nd gen Intel Xeon Scalable processors in our XC50 and CS500 supercomputers provides customers with the performance, scalability and flexibility to address their rapidly evolving analytics needs to achieve the fastest time to insight."
"Our 2nd gen Intel Xeon Scalable processors are designed to deliver leadership performance for the broadest range of workloads - including the convergence of traditional HPC and AI", stated Trish Damkroger, vice president and general manager in Intel's Datacenter Group. "With Intel technology and innovators like Cray, we will enable scientists to explore and address the worlds biggest computing challenges."
Cray XC50 systems are optimized for sustained performance, scaling and reliability - essential capabilities in the most demanding supercomputing environments. High-performance communication libraries layered on the Cray-designed Aries network enable the compute nodes to work together efficiently, yielding faster and better simulation and analytics results at scale.
The Cray CS500 cluster supercomputers are highly customizable and expressly designed to handle a range of simulation, research, analytics, machine learning, and deep learning problems. They provide flexible node configurations featuring the latest processor and interconnect technologies, giving customers the ability to tailor a system to specific needs. The systems' liquid-cooled design offers more energy savings and is built to efficiently house highly dense and higher-wattage solutions.