Exploring Hardware Software Partitioning on the Example of a Fingerprint Verification System

Modern FPGA technology opens tremendous new pos-sibilities in system design by facilitating a hardware soft-ware codesign approach. This partitioning of a system in hardware and software components is a delicate process which can either boost the system’s performance or lead to poor results when using an unfavorable partitioning. In this paper we investigate the impact of different hard-ware software partitioning strategies on a fingerprint ver-ification system which were evaluated in the course of a university lecture. The students were given a work-ing software-based solution and they had to perform two tasks. First they had to adapt the software which was orig-inally written for workstations to run on an FPGA softcore platform and second they had to find and implement suit-able hardware modules to further enhance the system’s performance. The different software and hardware opti-mization techniques are described and performance en-hancement as well as hardware overhead are given. Fi-nally, the system is evaluated using several case studies.