Circuits of current nanometer technologies suffer from great variabiliy both static (process
variations) and dynamic (temperature, voltage and aging). Designers can increase reliability
of these circutis by implementing specific mechanisms to deal with this variability. In this
project we first propose to design and implement robust sensors to be part of variability
monitoring networks. A key goal of this project is that the sensors must be aging and
radiation tolerant, because both variables significantly affect the performance of nanometric
circuits, even at ground level.
As pre-fabrication validation support we will develop a complex simulation framework that
can systematically and efficiently deal with the three views faced in the project: variability,
aging and radiation.
There can be also seen a positive side of variability. There are systems that require centain
randomness which can be found in the variability suffered by the circuit. This usually
happens in authentication or security systems. Another goal of this project is to design two
basic primitives in security systems: PUF (Physiscally Unclonable Functions) and RNGs
(Random Number Generators) based on variability sensors.
Finally, we will explore the potential use of emerging devices taking advantage of the
expertise of the research group on modeling and simulation of memristors. We make a step
forward and we propose the use of this kind of emerging devices to implement a variability