Intel has confirmed it is doubling-down on artificial intelligence, announcing the launch of the Intel Nervana platform, a partnership with Google, and a $25 million partnership with the Broad Institute to work on improving high-performance computing for the fiend of genomics analysis.
That Intel is bullish on artificial intelligence and deep learning should not come as news, of course: it was only in August that the chip maker announced its acquisition of AI specialist Nervana Systems
in a deal valued at $400 million. It's a field in which Intel sees its Xeon Phi many-core processor family as offering considerable potential, though it directly competes with high-powered general-purpose GPU (GPGPU) products from AMD and Nvidia which have thus far proven difficult to dislodge from their position at the top of the market.
Now, the company is confirming that it believes in artificial intelligence as the future of technology, and - naturally - that Intel products should be at the heart of it. Just three months after acquiring Nervana, the Intel Nervana platform has been put on the roadmap: the first half of 2017 will see Intel producing test-level silicon dubbed Lake Crest, followed by a commercial release as Knights Crest which combines Nervana's technology with Intel's Xeon cores.
'We expect the Intel Nervana platform to produce breakthrough performance and dramatic reductions in the time to train complex neural networks,
' claimed Diane Bryant, executive vice president and general manager of the Data Centre Group at Intel, of the company's plans. 'Before the end of the decade, Intel will deliver a 100-fold increase in performance that will turbocharge the pace of innovation in the emerging deep learning space.
Bryant also teased preliminary performance results for Knights Mill, the next generation of many-core Xeon Phi processor, which will launch next year with a claimed four-fold performance boost for deep learning projects. A partnership with Google on containerised cloud machine learning projects was also announced, alongside the launch of the Nervana AI Academy to boost education.
Finally, Intel's AI arm confirmed a $25 million partnership with the Board Institute as a means of improving the suitability of high-performance computing tools for genomics analysis, promising that 'researchers and software engineers at the Intel-Broad Centre for Genomic Data Engineering will build, optimise and widely share new tools and infrastructure that will help scientists integrate and process genomic data
' over a five-year period.