The supercomputer manufacturer, Cray, has developed two new systems – the CS-Storm 500GT and 500NX, both of which aim to meet customer demand for AI-capable infrastructure. Both systems are designed to work with Nvidia's Tesla GPU, but offer different feature-sets and capabilities.
The RAM capacity of both devices is staggering. The CS-Storm 500GT supports up to 2TB DDR4 (16 x 128GB DIMMs), while the 500NX supports up to 3TB DDR4 (24 x 128GB DIMMs). As for their processors, the 500GT has two Intel Xeon Skylake family processors and the 500NX has Intel Xeon E5-2600 v4 family processors.
Fred Kohout, Cray's senior vice president of products and chief marketing offer, said:
“Customer demand for AI-capable infrastructure is growing quickly, and the introduction of our new CS-Storm systems will give our customers a powerful solution for tackling a broad range of deep learning and machine leaning workloads at scale with the power of a Cray supercomputer. The exponential growth of data sizes, coupled with the need for faster time-to-solutions in AI, dictates the need for a highly-scalable and tuned infrastructure.”
While scoring well in the TOP500 supercomputer rankings, Cray is underperforming to the likes of Chinese supercomputers, which are coming on leaps and bounds along with the rest of China's science and technologies. With so much interest in AI now, Cray has a new market to exploit by catering for those needs.