Abstract. Checkpointing techniques are usually used to secure the execution of sequential and parallel programs. However, they can also be used in order to generate automatically a parallel code from a sequential program, these techniques permitted to any program being executed on any kind of ditributed parallel system. This article presents an analysis and a modelisation of an innovative technique -- CAPE -- which stands for Checkpointing Aided Parallel Execution. The presented model provides some hints to determine the optimal number of processors to run such a parallel code on a distributed system.