This site may earn chapter commissions from the links on this page. Terms of use.

KNL-Feature

Information technology'southward been about three years since Intel outset detailed its upcoming Xeon Phi hardware, codenamed Knights Landing, and the company is now shipping silicon to developers who tin pony upwardly the cash.

There are two Knights Landing-based Xeon Phi systems available. 1 is a liquid-cooled organization with half dozen DIMM slots, dual gigabit Ethernet ports, and the Developer Edition of the Xeon Phi — 16GB of MCDRAM, AVX-512, and 72 CPU cores. The other is a college-end 2U rack with 4 hot-swappable nodes. Base price on the equipment is $5,000 and $20,000 — and while that's extremely high by consumer standards, the $five,000 system isn't necessarily a bad bargain if yous piece of work in the field and are looking to experiment with high performance computing optimizations.

What Knights Landing brings to the tabular array

Knights Landing is the tertiary manycore x86 architecture that Intel has developed. The previous iteration of the hardware, Knights Corner, used a modified Intel Pentium core (P54C). The new Knights Landing, in dissimilarity, is based on a heavily modified Airmont 14nm Atom core — and while Intel hasn't shared all the details of those modifications, what they accept shared point to pregnant changes.

KnightsLanding

Knights Landing is made up of 36 tiles. Each tile contains ii CPU cores and two VPUs (vector processing units) per cadre (total of iv per tile). Dissimilar previous Xeon Phi cards, which functioned solely every bit co-processors, the new Knights Landing is designed to self-boot and can control an operating system as the native processor.

KL-2

Each CPU tin can handle up to four threads and the chip's out of order resource have been essentially improved. Intel is claiming a 3x per-thread performance improvement over older Knights Corner products with an estimated 3 TFLOPS of performance bachelor per chip. The core likewise uses 16GB of MCDRAM and can be configured for multiple retention configurations that tap both the onboard retention and the available DDR4 memory channels depending on which model is most advantageous for a particular piece of software.

On the one hand, Knights Landing is a straightforward development of Knights Corner, meant to proceed pace with Pascal and keep Intel competitive in the lucrative high-performance computing (HPC) marketplace. On the other, notwithstanding, information technology's meant to help developers write more parallel software. With single-thread performance about stalled out in modern CPUs, the manufacture needs to foster approaches that increase parallelism in common applications. Such efforts frequently begin at the supercomputing level and hopefully lead to all-time practices that become commonplace a few years later.

Nvidia's Pascal will become toe-to-toe with Knights Landing when it debuts, but Intel's got a substantial development pb on this 1, and information technology's obviously eager to put the architecture into developer hands.