On the convergence of the euler-jacobi method

The Euler-Jacobi method for the solution of the symmetric eigenvalue problem uses as basic transformations Euler rotations which diagonalize exactly 3 × 3 submatrices. We prove that the cyclic Euler-Jacobi method is quadratically convergent for matrices with distinct eigenvalues. We show that the cyclic Euler-Jacobi method is a relaxation method for minimization of a functional measuring the departure of a rotation matrix from being a diagonalizing matrix.