Intel Says It's Now Shipping Xeon Phi 'Knights Landing' Processors
Intel has started shipping its long-anticipated second-generation Xeon Phi processors, which aim to expand high-performance computing capabilities, the company said this week at the International Supercomputing Conference in Frankfurt, Germany.
The 14-nanometer co-processors, dubbed Knights Landing, aim to squeeze more performance out of supercomputers. Outside of supercomputers, the processors can also be used for deep-learning systems and artificial intelligence in data centers.
"With Intel supported software defined visualization libraries, our flagship visualization products VTK and ParaView can now scale to the next-generation supercomputers on the latest Intel platform, including systems powered by Intel Xeon Phi Processors," said Berk Geveci, senior director of scientific computing at Kitware, a Clifton, N.Y.-based Intel partner, in a statement. "We are already collaborating with [U.S. Department of Energy] labs to deliver large-scale ... visualization to Trinity and Cori supercomputers.’
Knights Landing, which has 72 individual cores manufactured on the company's 14-nanometer (14nm) Tri-Gate Transistor process, succeeds Intel's current 22nm 61-core Xeon Phi version, dubbed Knights Corner.
While Knights Corner was used with Xeon E5 server processors, Intel's new co-processor will also work as a host processor, capable of running an operating system and applications on its own.
Knights Landing features 16GB of on-package MCDRAM memory, which means that modules are stacked through a wire, allowing for up to five times more bandwidth than DDR4 memory. According to Intel, the coprocessor can deliver over 3 teraflops of peak performance, and is designed for highly paralleled computing.
The co-processors are also the first Xeon Phi family products to divert from the original 32-bit Pentium microarchitecture, and are offered in both a socketed version and a PCIe version.
In terms of performance, Intel says its processors have up to five times more performance than Nvidia’s GPU accelerator high-performance computing (HPC) products.
"With the Omnipath first-generation host fabric interface integrated into the Xeon Phi processor … you get the processor and eliminate the need for something like the PCIe card and the slot and the platform," said Rajeez Hazra, vice president of the Intel Data Center Group. "[It has] tremendous performance and density advantages."
Knights Landing is part of the company's System Scalable Framework (SSF) ecosystem, an advanced architectural approach to enable more scalable, flexible and balanced HPC systems.
Intel said the new scalable system framework -- which arrives as the HPC industry faces increasing changes and challenges, such as divergent infrastructures among visualization, big data, machine learning and HPC -- will advance the accessibility of HPC to more industries and workloads such as data-driven analytics, visualization and machine learning.