Run-time reconfigurable FPGAs are powerful platforms for realising software-defined radio systems. This paper introduces a parametric approach to designing such systems based on application and device parameters. We analyse the potential for reconfiguration in several software-defined radio components and demonstrate how the degree of parallelism in a reconfigurable module influences reconfiguration time and performance. In a case study with a reconfigurable FIR filter, our method increases the performance by a factor of 2.4.
Tobias Becker, Wayne Luk, Peter Y. K. Cheung