ARM Research Summit: Research Roadmap Keynote Live Blogby Ian Cutress on September 15, 2016 7:59 AM EST
12:33PM EDT - We're still at the ARM Research Summit, and the keynote talk on the ARM roadmap is about to start!
12:33PM EDT - Eric Hennenhoefer, VP Research for ARM, on stage
12:34PM EDT - Has a background in verification, moved to ARM when the company was acquired
12:35PM EDT - ARM Licencing model: IP Provider, licence fees and royalties
12:35PM EDT - 'Everyone who participates, wins'
12:35PM EDT - 'Design things once, and reuse it across multiple partners'
12:35PM EDT - 1250 licences, increases 100+ per year
12:36PM EDT - 350+ partners for potential royalties
12:36PM EDT - 86 billion ARM chips (not cores) shipped to date, 25% compound annual growth rate over last 5 years
12:37PM EDT - Ongoing royalties are typically based on a percentage of chip price
12:37PM EDT - Upfront licence fees cover development cost
12:38PM EDT - 'There's no point making something, signing up partners, and they don't ship anything'
12:40PM EDT - ARM Artisan IP is popular apparently
12:40PM EDT - Acquired Apical in the last year
12:41PM EDT - Price ranges from 50c / chip to over $25 (depends on partners, ARM doesn't make the chips remember)
12:42PM EDT - ARM has silicon partners, software/training/consortia partners, design support partners
12:43PM EDT - ARM Research as a unit are 3-7 years ahead of the product teams. That's important
12:43PM EDT - ARM Research are based from n+2
12:43PM EDT - A combination of targeted and blue-sky research
12:44PM EDT - Combination of internal research teams and academic/research partnerships
12:44PM EDT - 'We work on disruptive technology roadmaps'
12:45PM EDT - 'One of the jobs is to look at potential disruptive technologies, what they mean to ARM, and what it takes to make them real'
12:45PM EDT - This includes attending conferences and knowing lots of people with potential
12:47PM EDT - As the number of devices increases, research like energy harvesting becomes important
12:47PM EDT - especially for 'edge' devices (ones at the end of the communication chain)
12:48PM EDT - 'The compute is there, you can see the future coming'
12:49PM EDT - ARM Research is split into skill areas
12:49PM EDT - Some technologies die before it reaches ARM, but ARM still needs to keep abreast of partner development
12:50PM EDT - Memory/Interconnect team deals with 3D Memory, non-volatile memory etc
12:51PM EDT - The architecture group works with the product teams, but is working on long term projects
12:51PM EDT - e.g. Scaleable Vector Extensions was in the works for 7 years
12:51PM EDT - Also, 'More without Moore' is discussed in this group
12:52PM EDT - Such as arch/uArch, or new methodology, or new materials
12:53PM EDT - 'Let's not leave security to software, too many unsecure video teddy bears are being hacked - we need to bake it into hardware'
12:53PM EDT - 'We also need to do it in cost effective ways'
12:53PM EDT - 'The more protocols, the more ways to screw it up'
12:55PM EDT - The Applied Silicon group goes through integration, sensor nodes, or things like printed electronics
12:55PM EDT - Also future silicon technology, what's after MOS
12:56PM EDT - 'Aggressively looking for new and interesting things'
12:56PM EDT - 'Please invent something. Disruptive. In this area.'
12:58PM EDT - Research into verification, HPC, efficiency, compute near memory
12:59PM EDT - Machine learning, Graphics, Computer vision and mobile systems have strong research teams
01:00PM EDT - Special projects such as motors and low power radio
01:01PM EDT - ARM has doubled in since in a few short years
01:01PM EDT - Didn't used to be interesting to Academia, was considered the keyboard chip people
01:02PM EDT - ARM sees value in creating a Research Ecosystem, with discussions and events like this
01:03PM EDT - Academia and ARM can work together - it depends on what you want to do
01:04PM EDT - Cooperations involve publishing, open source, non-exclusive IP rights, pooled funding
01:04PM EDT - Most academics are on the pre-competitive side
01:04PM EDT - Some work in the Competitive side, which has more rules
01:05PM EDT - NDAs and Patents/IP are issues to cover on a case-by-case basis
01:06PM EDT - ARM encourages university enablement via free materials
01:06PM EDT - starter kits, tools, some IP are free
01:06PM EDT - More access can be had, have to ask. There may be free toys which can help depending on what we have
01:08PM EDT - Accelerators are a common topic of discussion
01:09PM EDT - IP enablement is a win-win when it accelerates research, especially when it gets disruptive
01:09PM EDT - ARM Research has strong IP and legal teams to help smooth IP sponsorship with academia as well
01:10PM EDT - ARM developing the Research ecosystem. The current Summit is an example
01:11PM EDT - Participation in many EU academic collaborative projects
01:11PM EDT - Lots of positions available at ARM, doubled in size in four years
01:12PM EDT - Graduate positions, internships as well
01:12PM EDT - Shanghai site is being ramped up as well
01:14PM EDT - That's a wrap!