ABSTRACT
In this chapter, we describe the choices that were made for a base level implementation of the different parallel graphics decomposi tions. The approach used for developing the parallel programs is to devise a single basis graphics rendering algorithm, and then build different parallel task partitioning and memory referencing schemes on top of it. This allows an equal comparison of a number of different approaches for parallelism since the underlying algorithm is the same. This basis algorithm is not compromised by the parallel algorithms since it can be modified to the specifics of each particular parallel approach. The first section of this chapter presents this basis algorithm by describing the underlying serial approach, as well as the choices that were made which were common to all of the parallel formulations. The second section of this chapter describes the measurement techniques used to obtain timings for the different programs. The last section gives the performance analysis measures used to analyze the different parallel implementations that are presented in the next chapter.
