Intel (News - Alert) ‘Many Integrated Core’ Architecture or Intel MIC is a coprocessor computer architecture developed by Intel. It incorporates earlier work on the Larrabee many core architecture, which is a canceled general-purpose computing on graphics processing unit, which Intel was developing.
The Intel MIC also includes the Teraflops Research Chip multicore chip research project and the Intel Single-chip Cloud Computer multicore microprocessor. The second generation product has been codenamed Knights Landing and uses a 14nm process.
Lately, we have been hearing a lot about supercomputers. This is a segment that should be receiving a huge boost from Intel’s latest version of Xeon Phi, or Knights Landing. The chip is a 72-core coprocessor solution manufactured on a 14nm process with 3D Tri-Gate transistors.
Actually, they are coprocessors built around Intel's MIC architecture that combines several cores into a single chip. In turn, this becomes part of a larger PCI (News - Alert)-E add-in card solution designed for supercomputing applications. Knights Landing will be able to deliver double-precision peak performance of more than 3 teraflops and single-precision performance of more than 8 teraflops.
While there has been a slight delay, Knights Landing was originally introduced in 2013, and we can expect to finally see Intel’s fastest processor ever inside supercomputers by early next year. According to Charlie Wuischpard, vice president of the data center group and general manager of the HPC group at Intel, we should see over 50 manufacturers shipping systems that include Knights Landing soon.
Earlier this year in July, Intel announced that it teamed up with HP to build servers around Intel’s suite of HPC technologies, or as Intel calls it, their Scalable System Framework. HP will be releasing a series of systems under the company’s Apollo brand of HPC servers that will integrate Knights Landing.
Intel didn't provide details on the first supercomputers with Knights Landing. The U.S. Department of Energy, however, said that the chip will be used in Cori, a 9,300-core supercomputer that will be deployed in the latter half of 2016 at the National Energy Research Scientific Computing Center in Berkeley, California.
Due to the fact that every company, every agency, pretty much everyone imaginable is collecting more and more data on a daily basis, there is a distinct change in the way supercomputers are being used. We are not looking at application deployment systems, or general processing power; today’s supercomputers are becoming analytical tools.
Machine learning, or the study of pattern recognition and computational learning theory in artificial intelligence, is what is needed today. Big data is useless if it cannot be analyzed and made sense of on a timely basis.
Although there was a delay in getting Knights Landing to market, it does seem that Intel is on track now. We are seeing a supercomputer growth in China, and the U.S. plans to build at least two supercomputers in the next year. The goal of new technologies is to provide more performance out of supercomputers without drawing more power and Intel’s Xeon Phi is there to take up the cause.
Edited by Rory J. Thompson