Motivated by the general non-linear eigenvalue problem, we present a parallel algorithm to find the zeros of a complex analytic function in a given region. The algorithm is based on a two dimensional version of the bisection algorithm and is implemented in parallel using a master-slaves programming model. The master is responsible for organising the slaves while the slaves are responsible for determining if a given region contains any zeros. The results from the test calculations show that this algorithm achieves good efficiency provided that the number of processors does not exceed four time the number of zeros in the initial region.
