Graphics floating point coprocessor having matrix capabilities

摘要

A graphics coprocessor designed to work in conjunction with a host graphic processor in a graphics system. The coprocessor is adapted to perform arithmetic calculations including matrix calculations. The matrix size is such that the intermediate results require more registers than are practical to include in the coprocessor. This has been solved by arranging for certain selected ones of the intermediate results to continue within the program execution from stage to stage and avoiding intermediate storage.