Intel has been cranking out desktop and mobile processors based on 22-nm tri-gate transistors for some time now. The company's Atom SoCs continue to be built using older 32-nm process technology, though. SoCs can require a mix of low-power, high-speed, and high-voltage transistors, making them somewhat more complex than traditional CPUs. At the International Electron Devices Meeting yesterday, Intel presented a paper detailing its plans for 22-nm SoCs.
The paper is highly technical and hasn't been published online, but EE Times has a good summary of the particulars. Intel's 22-nm SoC process includes several transistor types in addition to multi-layer interconnects, precision resistors, high-density capacitors, and spiral inductors. Chips built using the process should be suitable for mobile devices like smartphones and tablets, and Intel Director of Process Architecture and Integration Mark Bohr claims high-volume production will begin next year.
According to Bohr, the process gives Intel "a significant lead over [its] competitors." Sounds like he's particularly proud of the process' low leakage power, which should be especially desirable for mobile applications. Bohr also says the performance of Intel's 3D SoC transistors should be higher than that of competing solutions.
The 22-nm SoC process is likely to be used first for Silvermont, an architectural refresh of the Atom CPU core due out next year. The process reportedly scales down to 14 nm, as well, although there doesn't appear to be an official timeframe for that transistor shrink.
|Rumor: Intel Skylake-X and X299 will headline Computex 2017||55|
|Rumor: Nvidia to answer Radeon RX 550 with GeForce GT 1030||20|
|Samsung Galaxy Book tablets blend Windows 10 and Intel CPUs||17|
|Deals of the week: a mighty PSU, mid-range CPUs, and more||28|
|AMD board partners begin tricking out RX 560s and RX 550s||17|
|Dell shows off a pro-grade 4K HDR display and AIO machines||15|
|Rumor: Google to bake ad-blocking into Chrome browser||57|
|EpicGear's Defiant modular gaming keyboard reviewed||12|
|GeForce cards with faster RAM are inbound from multiple locations||19|
|Those power consumption numbers are very fermi-liar||+53|