14nm shrink now going smoothly. http://www.bit-tech.net/news/hardware/2014/11/03/intel-core-m-broadwell/1
So are we going to see even greater problems come the next node shrink, or was it something specific to do with a design change. In other words was this just a bump in the road, or is moore's law starting to look a bit rough around the edges.
That's a good question, and I'm interested to hear about this too. Though it appears intel is a few paces ahead of other companies in terms of fabrication size, it's nice having a company that can use $100 bills as napkins for their employees to be able to venture into something that might be physically impossible. It seems to me intel has been doing things very incrementally, when they should have the money and knowledge to jump straight to something like 10nm. But I'm sure these increments are done because the physics of something this small isn't that simple and they need to test the boundaries of silicon. Like I said, it's nice they're the ones doing this, because other companies like TSMC can save their own money and jump straight to what they know is physically possible.
My memory could be failing me, but isn't Broadwell the first time the "flaw" caused a delay in the schedule.
Why should they 'jump' to 10nm? Each shrink offers performance and efficiency gains which means that end uses will want to upgrade their CPUs at the very least which in turn offers Intel a chance to make more money. The move to 10nm can be spread out over 10 years, all the while they are making money so why would they 'jump' to it in 2 years and then have to progress further to cover the other 8 years worth of money making?
