High Performance Architecture for Reciprocal Function Evaluation on Virtex II FPGA

The fast and accurate evaluation of elementary functions is vitally important in many fields of scientific computing .We describe, in this paper, a method to calculate powering (X p ) which employs table lookups and polynomial approximation, a second-order Taylor series expansion. We also present the architecture for evaluating the reciprocal on virtex-II FPGA. The virtex-II FPGA family incorporates large Blocks memories SelectRAM and provides fast arithmetic carry logic capability; we exploit these resources to implement our architecture on the XC2V80(-5) FPGA circuit with operating frequency over 31 MHz.