Links on Android Authority may earn us a commission. Learn more.
UMC and TSMC to move to FinFET at 20nm
Apparently, the planar FinFET technology that these foundries intend to use is better suited for mobile SoC’s, like the ones from ARM, and it should bring a performance improvement of at least 15%-20% over Intel’s own FinFET technology. Intel wants to make a 22nm Atom processor using their own FinFET technology next year, but ARM CEO Warren East thinks Intel Atom is not competitive with ARM chips for mobile use. “An 800MHz ARM delivers the same performance as a 1.6GHz Atom,” says East.
“Bearing in mind that it is easier to make rectangular SOI FinFETs than rectangular bulk FinFETs, moving from triangular Intel bulk FinFETs to rectangular SOI FinFETs can deliver approximately 20% performance improvement,” says Professor Asen Asenov of Glasgow University who is the CEO of Gold Standard Simulations..
This is bad news for Intel as UMC has licensed the FinFET technology from IBM, just like Samsung and GlobalFoundries have, while TSMC is presumably building their own.
“We are happy to engage with a recognized technology leader such as IBM for this technology advancement effort. UMC’s position as a world-leading foundry involves timely introduction of leading-edge processes to enable next generation customer chip designs. Leveraging IBM’s technology expertise to shorten our 20nm and FinFET R&D cycle will create a win-win situation for UMC and our customers,” said IC Chen, vice president of advanced technology development at UMC.
TSMC expects the move to the 20nm FinFET technology to offer them a 1.9x increase in density over the 28 nm process, -25% power consumption, and 15-20% increase in performance.
The year 2014, when these FinFET chips will arrive in the market should be very exciting, because we’ll also get the new 64 bit ARMv8 architecture, and the next-gen CPU that comes after Cortex A15, which will be built with the 20 nm process. That’s the year when Intel should really start to fear ARM, because those chips will be used in a lot more types of devices and machines than just smartphones and tablets.