We investigate the performance of an algorithm to detect selfcollisions in a tubular object approximated by spheres. The approach utilizes the Bounding Volume Hierarchy (BVH) to arrange the spheres and it was implemented using sequential and parallel algorithms. The tubular object has a snake-like motion, and the algorithm calculates the closer pairs of spheres considering the hierarchy and the parallelism. Experiments were carried out to analyze the performance of the implementations with different object’s motions.