Extending Von Neumann uniprocessor architecture to exploit result sharing

The author describes how result sharing schemes can be applied to Von Neumann uniprocessor architectures in order to avoid redundant computations. He presents the modifications needed in a typical Von Neumann architecture to support result sharing. The extensions involved are simple and minimal. In order to fully exploit result sharing using the proposed scheme, the author discusses the issue of the machine code structure as it relates to result sharing potential. The simplicity of the proposed changes and the many real-life applications which will benefit from result sharing make the proposed scheme both useful and practical.<
>