At Computex yesterday, Intel President Renée James showed off a new 14nm tablet design that showcased Intel’s Broadwell CPU. Dubbed the Core M, this chip could be a fundamental game changer for Santa Clara, for one simple reason — it will allow the company to finally build big-core x86 devices that fit into form factors thin enough to compete with current Android, Apple, and Windows tablets based on its own Bay Trail.
According to James, the Broadwell prototype (dubbed Llama Mountain for an unknown reason) is a razor-thin tablet with a 32 watt-hour battery, a Broadwell Y CPU (branded as the Core M), a sisterboard with integrated WiFi radio, a dockable configuration that adjusts performance depending on system orientation, a 12.5-inch screen, and a total weight (minus keyboard) of just 670 grams. (1.47 lbs).
Total thickness? Just 7.2 millimeters. The Surface Pro 3, for comparison, is 9.1mm thick and weighs 800 grams — which is Intel’s indirect way of implying that it’s getting some serious improvements out of its 14nm technology, despite the fact that Broadwell is debuting much later than originally anticipated. Chipzilla is claiming 60% lower TDP for Broadwell with 20-40% better performance, 10-45% better SoC power, and a 50% reduction in package footprint. Typically these figures don’t apply simultaneously — the system might draw 10-45% less power depending on workload, with a lower TDP overall and better performance depending on the form factor.
It looks as though Broadwell will be the architecture where Atom and the Core family finally meet in the middle — which means we can expect significant price differentiation between the two. If Atom and Broadwell can fit into the same form factors, Intel has a golden opportunity to position its lower power x86 architecture as the mass-market chip while keeping the high-end Broadwell as a de facto luxury component.
The only remaining question is whether or not Intel can hit its own adjusted end-of-year ship date for Broadwell and whether it’ll lead with these ultra-low power processors. The company committed to a Q4 rollout last week, but chips at the bottom of the power consumption stack are typically the most difficult to fabricate. After the problems Intel has had with 14nm deployments, all eyes will be on the company to see if it can make this goal. If successful, Intel should have 14nm silicon in market before upcoming 20nm designs from Qualcomm and other players in the ARM market.