Earlier this week, Intel held its own ‘tech day’ during which it went over various technologies it is currently working on. One of the major points of the presentation was the upcoming 10nm process node, which Intel claims is a ‘full generation ahead’ of the 10nm technology competing chip makers are using, including Samsung, TSMC, Qualcomm and others.
So what is so different about Intel’s 10nm tech compared to its rivals? Well for starters, it uses third-generation FinFET technology, but there is a bit more to it than that. One of the key focusses for Intel when it comes to 10nm is Hyper Scaling, which allows Intel to “continue the benefits of Moore’s Law economics by delivering transistors that are smaller and have lower cost-per-transistor”. We can see Intel’s example of this in the graph below, which was published as part of the press materials for this event.
With Intel’s 10nm node, the company claims it can deliver a 2.7x improvement in transistor density compared with its 14nm products. This is achieved due to the ‘minimum gate pitch’ shrinking from 70nm down to 54nm and the minimum metal pitch coming down from 52nm to 36nm. This enables a logic transistor density of 100.8 mega transistors per mm squared, which is estimated to be double that of competing 10nm chips.
Aside from that, Intel’s jump from 14nm to 10nm will bring further improvements. The latest process node is said to deliver up to 25 percent better performance at a 45 percent lower power draw. An upcoming enhanced version of the 10nm process known as ’10++’ will boost this performance benefit by an additional 15 percent, while reducing power draw by a further 30 percent.
KitGuru Says: While other companies might already be out of the gate with 10nm products, it sounds like Intel’s own node is much more refined. If Intel’s 10nm CPUs can really deliver those performance advantages, coupled with reduced power draw, then this could end up being a big generational leap.