Parameter Extraction of PSP MOSFET Model Using Particle Swarm Optimization - SoC Approach

Abstract

System on Chip (SoC) architecture offers performance acceleration by offloading compute-intensive functions in FPGA logic together with application specific instruction set processor (ASIP). In this paper, we report a novel approach of SoC implementation for MOSFET parameter extraction. Xilinx’s Zynq 7000 SoC supplied with AVNET ZedboardTM is used in this work. Extraction of PSP MOSFET Model parameters for 65 nm technology devices has been carried out using Particle Swarm Optimization (PSO) algorithm. PSP MOSFET Model code is executed by ARM Cortex A9 processor available on Zynq 7000 SoC. Parameter extraction of PSP MOSFET model is carried out for two different cases of implementations of PSO algorithm using: (1) ARM Cortex A9 and (2) FPGA. In both cases, excellent agreement between measured data of 65 nm technology MOSFET devices and PSP MOSFET model is observed. We have measured time taken by PSO algorithm in both cases for its execution. Implementation of PSO algorithm using FPGA is found to be 3.71 times faster on average compared to that on ARM Cortex A9. In both cases, the RMS error between measured and PSP MOSFET model is found to be less than 10%.