Leo Stocco, Electrical & Computer Engineering, UBC

EECE 597	Prof. A. Ivanov - ECE

To Apply:	email	For information & availability of specific projects.
ID	Status	Name
AI-1	Available	Optimization SW for Molecular Dynamics Simulations
In this age, almost every high-tech product has an electronic component. However, electronic chips tend to degrade over time. Degradation of CMOS device characteristics over time is an ever-growing and multifaceted problem. This problem is intensified considering the new applications as well as device scaling. The extent of irreversible damages caused by device degradation can better be understood, considering sensitive space and military applications of CMOS devices. Various numerical methods are in use to study the molecular and electronic phenomena responsible for degradation of semiconductor devices. The optimization software that you will help develop is an essential part to support reliability research in state of the art silicon chips.
AI-2	Available	Modeling Performance Degradation in FPGA SoC Devices
We have an ongoing project on CMOS degradation due to aging. As part of our multifaceted effort to understand the degradation phenomena and the restrictions it imposes on designers, we have been performing various experiments on FPGA devices as an inexpensive approach for collecting actual test data. In this project, our aim is to understand the delay degradation at the level of primitive circuit elements accessible to us on FPGA devices (LUT6 in Zynq 7000). Our group is very experienced in such experiments and we have the necessary hardware/ software platforms to perform them. We seek a passionate MEng student to join our team and help with the design, execution and analysis of the experiments.
AI-3	Available	FPGA-Based Molecular Dynamics Simulation HW
Dedicated computational hardware (aka accelerators) are becoming more common in such fields as quantum chemistry, computational material science and quantum monte carlo simulations. This project involves designing an RTL-level solution for molecular dynamics simulations using ReaxFF forcefelds. Ideally the solution would be benchmarked against CPU and GPU (but this is not necessary for this project). Knowledge of molecular dynamics is not required as the algorithms are simple and the C/C++ code is readily available. A good understanding of RTL design and experience with Xilinx toolchain is preferred.