Case Study of Finite Resource Optimization in FPGA Using Genetic Algorithm

Modem Field-Programmable Gate Arrays (FPGAs) are becoming very popular in embedded systems and high performance applications. FPGA has benefited from the shrinking of transistor feature size, which allows more on-chip reconfigurable (e.g., memories and look-up tables) and routing resources available. Unfortunately, the amount of reconfigurable resources in a FPGA is fixed and limited. This paper investigates the mapping scheme of the applications in a FPGA by utilizing sequential processing (e.g., Altera Nios II or Xilinx Microblaze, using C programming language) and task specific hardware (using hardware description language). Genetic Algorithm is used in this study. We found that placing sequential processor cores into FPGA can improve the resource utilization efficiency and achieve acceptable system performance. ln this paper, three cases were studied to determine the trade-off between resource optimization and system performance.