On November 8, 2012, Cray announced the launch of a new generation of supercomputers, the Cray XC 30. Previously, the project was known as Cascade.

“After several years of incredibly hard work focused on completing the most ambitious project in the history of our company, today's launch of the Cray XC 30 supercomputer is exciting for Cray employees, as well as for company customers who are looking forward to the release of a new product,” said Peter Ungaro, President and CEO of the company. “In addition to our most successful products, the new Cray XC 30 will provide researchers, scientists and engineers with a system that they can adapt to fulfill their most demanding tasks. We have a great start and already have contracts for more than $ 100 million for this product. We believe that the Cray XC 30 will allow a wider range of users to use the most advanced technology of supercomputers. ”
')
Already, several leading HPC centers (High-performance computing) have signed contracts with Cray for the purchase of new supercomputers, among them:
- Swiss National Supercomputer Center (CSCS)
- Finnish Supercomputer Center (CSC)
- National Computing Center for Energy Research under the control of the National Laboratory. Lawrence at Berkeley (NERSC)
- Academic Computing Center (ACCMS) of Kyoto University in Japan
- Supercomputer Center of the University of Stuttgart (HLRS) in Germany
The first customer to sign the contract for the XC 30 was HLRS from Stuttgart. This happened back in 2010. “The system will be a valuable resource for our researchers and scientists, as well as for our partners in the automotive and aerospace industries,” said Professor Michael Resch, director of HLRS. “We have been working closely with Cray for many years to provide our users with innovative and advanced developments in the field of supercomputer technologies. We hope that our cooperation will continue to develop fruitfully. ”
The Cray XC30 configuration includes an Aries internal interconnect system and Intel Xeon processors. In addition, the manufacturer notes the presence of proprietary integrated software environment and innovative power and cooling systems. Allegedly, due to scaling, the performance of the Cray XC30 can exceed 100 petaflops.

Each XC30 rack will provide a performance of 66 teraflop / s, and the maximum performance limit can be about 100 petaflop / s. XC30 consists of the so-called "cabinets". One “cabinet” is three chassis for blade servers. Each chassis has 16 blades servers with 4 cores each. Initially, they are used as Xeon E5-2600 processors (more than a million processor cores can be included in the system). But the system can work with Xeon Phi co-processors, as well as Nvidia Tesla graphics chips based on the Tesla architecture. This, by the way, will allow optimizing the system configuration for certain tasks. The operating system used here is Cray Linux Enviroment, the proprietary version of Linux.
A wide release of computers is scheduled for early 2013.
The cost of this device in the budget option starts from $ 500,000.
More complete technical data presented on the
official website of the company.