Contribution to parallel and vector computation in multibody dynamics

In spite of the continuous and sustained increase in CPU-time performances of modern computer processors, one notes a real eagerness in the scientific community to investigate both the vector and parallel computational approaches. The goal is obviously to attain high-speed simulation whatever the envisaged numerical treatment may be. The aim of the present contribution is, firstly, to point out the advantage of splitting a multibody system into distinct parts in a `coarse grain level' parallelization context, and secondly, to stress the advisability of vectorizing multibody equations of motion in order to compute them in an adequate computer environment.