In section 3, one algorithm for rendering irregular grids was presented. The same algorithm was parallelized to run on a network of IBM RS/6000 workstations [30]. The idea is to be able to use a set of heterogeneous workstations, each one with enough memory and power to run the whole algorithm alone.
The parallel version of the algorithm is well suited for a small number of workstations and its implementation has been done using the PVM (Parallel Virtual Machine) package developed at the University of Tennesee [51]. The algorithm works by using one workstation as the host computer that will distribute the dataset and work for the other processors. First the host gives each processor a copy of the whole database, then it proceed by giving each processor one part of the image to render, once the a processor is done it sends back to the host its output. If the task is not completed, the host will keep sending new chunks of work to each processor. This algorithm is a coarse parallel algorithm and its performance greatly depends on the size of the work being given to each processor by the host.