Аннотация:In this paper we consider two parallelization schemes for parallel solving of binary decomposable problems by tree-like recursion procedures which can be adapted for GPU computing systems. We compare the efficiency of these schemes, using different parameters of parallelization.