A very fast CORDIC (coordinate rotation digital computer)-based Jacobi-like algorithm for the parallel solution of symmetric eigenvalue problems is proposed. It becomes possible by not focusing on the realization of an exact Jacobi rotation with a CORDIC processor, but by applying approximate rotations and adjusting them to single steps of the CORDIC algorithm, i.e., only one angle of the CORDIC angle sequence is applied in each step. Although only linear convergence is obtained for the most simple version of the proposed algorithm, the overall operation count (shifts and adds) decreases dramatically. A slow increase of the number of CORDIC angles involved during the runtime retains quadratic convergence.
展开▼
