PARALLEL JACOBI CALCULATION ACCELERATION IMPLEMENTATION METHOD FOR FPGA

摘要

A parallel Jacobi calculation acceleration implementation method for an FPGA. Said method comprises: initializing processing units: inputting data of an n × n dimensional matrix into an FPGA, and using parallel Jacobi calculation to perform rotation transformation processing, in the parallel Jacobi computation, plane rotation being performed using a CORDIC algorithm; a diagonal processing unit calculating a symbol set corresponding to a rotation angle and outputting same to an off-diagonal processing unit; updating elements of the diagonal processing unit; updating elements of the off-diagonal processing unit; and exchanging the elements of the processing units: after updating the elements of each processing unit, exchanging the updated elements of the processing units.