Machine learning is catching up, catching up for mass usage. Artificial intelligence has now advanced to a new height. App Developers and IT solutions companies now clearly have a major opportunity with regards to AI.
One of the aspects of Machine learning is the need for the algorithms to broadly depend on the computing powers. The magnitude of these capabilities have grown exponentially over a matter of time.
Intel’s formal launch of the Xeon Phi processor at ISC High performance 2016 conference has ensured its dedication to machine learning within its family of processors. The idea is to support good performance for the applications which involve high computations for its machine learning algorithms.
The new Xeon Phi family will be a set of co-processors for the general purpose Intel Xeon processors.These introduces an Intel Scalable System framework (Intel SSF), thus there will be less dependency on PCIe or dedicated graphics processing units per say. Per Intel , the Phi processors uses 16GB of high bandwidth memory to deliver upto 500GB/s of memory bandwidth. To improve the performance of parallel application, they have a dual-port Omni-Path Architecture.