Named the Xeon Phi 7290, Intel’s top-of-the-line chip will feature a mammoth 72 cores and has clocked in as the company’s faster chipset. In addition to core volume, the 7290 will also feature 16 GB of integrated memory and will be able to support a system that’s loaded with 384 GB of DDR4 memory.
Currently, the plan for the Xeon 7290 will be a roll out to supercomputing centers that’ve already had success using the Xeon Phi series. In fact, the likes of Sandia National Laboratories, Los Alamos, the Barcelona Supercomputer Center and others have thrown their lot in with the Xeon Phi series.
“Our sparse linear algebra library greatly benefits from the massive bandwidth of the Intel Xeon Phi processor. Thanks to MKL’s cross-platform compatibility, we were able to port it in a matter of hours,” stated Mauricio Hanzich of the Barcelona Supercomputing Center’s HPC Software Engineering Group. “Xeon Phi’s vectorial registers along with its massive memory bandwidth are just the perfect combination for finite differences schemes.”
But aside from supercomputing, Intel also has plans to begin outfitting HP Enterprise, Dell and Lenovo with the next generation of Xeon Phi chips for their servers. That could have a big impact on how these chips are used by everyday engineers. Sure it isn’t likely that engineers will be sitting at a workstation equipped with a Xeon Phi 7290 at year’s end, but many of the machines that process the virtual reality data, simulations and IoT edge analytics that engineers tapinto at the office or in the cloud could be using Intel’s next-gen chips to get results faster.
Will Intel’s new CPUs ever surpass GPUs as a better option for high-end computing? My sense is that GPUs will be the dominant driver of high performance and supercomputing power for the foreseeable future. At the moment, GPUs just cost much less and deliver excellent results.
Which brings me to my last point.
According to Intel, the next-generation Xeon Phi chip will be available in September for the staggering price of $6,294.