Sign up for our newsletter and get the latest HPC news and analysis.
Send me information from insideHPC:


Latest Xeon Processors come to SGI UV In-Memory Supercomputers

SGI_logo_platinum_smlToday SGI announced a 30 percent increase in TOP500 systems and new additions to their SGI UV family of large-memory supercomputers.

SGI has increased the number of systems on the TOP500 from 23 systems in November 2014 to 30 systems in July 2015. Building on over 20 years of in-memory computing leadership and overcoming the limitations and complexity of clusters, the SGI UV line of advanced symmetric multiprocessing systems provide unparalleled performance and scale to spur innovation, achieve greater insight and achieve faster time to discovery.

New SGI UV 3000 and UV 30

The SGI UV line of scale-up in-memory supercomputers is designed to help customers solve complex problems at greater scale, reduce overhead costs, and accelerate time to results. Now available with Intel Xeon v3 processors, the new SGI UVTM 3000 and enhanced SGI UVTM 300 offer increased computational capacity and efficiency.

The appetite for in-memory computing has seen a dramatic increase over the years as more organizations need to analyze complex information at massive scale, while also minimizing costs and decreasing the time to discovery,” said Jorge Titinger, president and CEO, at SGI. “Working closely with Intel, we can be sure that whether someone is researching the next scientific breakthrough or analyzing data to gain a competitive edge in their business, our system will provide them with the speed, scale and flexibility they need to get the job done.”

Designed for compute-intensive, fast algorithm workloads such as computer-aided engineering (CAE), computational fluid dynamics (CFD), and scientific simulations, the new SGI UV 3000 , enables customers to solve larger problems, achieve high levels of accuracy, run more applications in less time without complexity, and accelerate time to value. Featuring a high processor to memory ratio utilizing Intel Xeon E5-4600 v3 processors, and coupled with SGI’s ultra-fast, low latency NUMAlink interconnect, the UV 3000 can scale up to 256 CPU sockets and 64 terabytes (TB) of cache-coherent shared memory in a single system.

SGI also introduced the new UV 30, designed for smaller, compute-intensive environments. The 2U, 4- socket server will also feature the Intel Xeon E5-4600 v3 processors and provide up to 3TBs of in- memory computing power.

Enhanced SGI UV 300 and UV 30 EX

Designed for data-intensive, I/O heavy workloads such as data visualization, real-time streaming, and high performance data analytics, the newly enhanced UV 300 delivers high memory to processor ratio for visualization of massive data sets, analyzing Big Data in real-time, and accelerating time to action. By incorporating the latest Intel Xeon E7-8800 v3 processor and extending the 7th generation SGI NUMAlink interconnect topology, SGI more than doubled the capacity of its UV 300 to 64 CPU sockets and 64TBs of cache coherent shared memory.

Many HPC and Big data jobs today require high-speed processing of very large data sets, and find that large in-memory systems provide major performance gains. To create leading edge systems requires strong partnerships between industry leading system vendors and industry leading component suppliers in order to enable enterprises to maximize the performance potential for in-memory computing.” Earl Joseph, vice president of high performance computing, IDC. “By integrating the latest in-memory systems from SGI and the newest processor family from Intel, the companies provide solutions that promise even greater performance for businesses that demand business insight in real time from vast volumes of data.”

Complementing the UV 300, SGI will also extend the Intel Xeon E7-8800 v3 processors to the UV 30EX, a 5U, 4-socket server now providing up to 6TB of in-memory computing power.

The UV 3000, UV 300 and UV 30 EX will be available in the third quarter of Calendar Year 2015. The UV 30 is currently available.

Sign up for our insideHPC Newsletter.

Resource Links: