Due to the wide variety of current and next-generation supercomputing architectures, the development of highperformance parallel visualization and analysis operators frequently requires re-writing the underlying algorithms ...