Low-power ‘flat-field’ transistor suits IoT, on-chip AI
The company was formed in 2017 by Professor Asen Asenov, former CEO of Gold Standard Simulations Ltd. (GSS), a leader in variability CMOS simulations and variability-aware technical CAD.
Semiwise’s first offering is something called the flat field transistor (FFT). It is applicable to bulk CMOS manufacturing processes down to 20nm and suited to Internet of Things (IoT) and on-chip artificial intelligence operations, Asenov claims.
Applied to 20 nm bulk CMOS the FFT has 30 percent higher speed performance and two orders of magnitude lower leakage current at equivalent drive current compared to its bulk counterpart the company said. The far field transistor configuration was designed using GSS – now Synopsys – variability-aware TCAD tools. It has low variability with factors in the range of 0.6mV/micro, which is much lower than the reported variability for 14nm FinFET and 28nm FDSOI manufacturing processes.
The low variability of the FFT means that guard-banding on numerous design parameters can be reduced and makes the transistor suited to near and sub-threshold logic and low-voltage tolerant SRAM design, the company claims.
See also: GaN HEMT addresses UHF radar market
See also: Transistors cannot enable very low-power IoT, could memristors do the job?
“The concept of variability resistant transistors has been around for a few years however despite significant venture capital investments many companies have failed to commercialise the technology due to lack of firm understanding of statistical variability and adequate simulations capabilities,” said Asenov, in a statement. “Armed with the most accurate and reliable simulation tools from GSS – now Synopsys – and more than 35 year of experience in the semiconductor industry Semiwise will succeed when others have failed.”
Semiwise has been operational since August 2017 and has engaged with a number of semiconductor manufacturers interested in evaluating the FFT technology. The company is also working on semiconductor IP for FDSOI and FinFET CMOS manufacturing processes.
www.semi-wise.com
www.nif-t.net
See also: Graphene transistor could enable low-power computers 1000x faster
See also: An X-band GaN PA MMIC for phased array radar applications
See also: Carbon nanotubes enable tiny transistors to build themselves
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
