Class of numerical systems for pipeline bit sequential development of multiple optoelectronic data streams
暂无分享,去创建一个
One of the approaches to organization of pipeline processing of multiple flows of sequential codes is use of superfluous numerical systems, which allow executing all code operations digit-by-digit, beginning from higher orders. A numbers of such numerical systems are known. The basic advantage of these systems is limitation of carry length in arithmetic operations. It determines hardware expenses for organization of pipeline digit-by-digit processing. Theoretically there can be a variety of such systems. The class of superfluous numerical systems with additive and multiplicative relationships between orders is described in this article. It class generalizes known numerical systems for digit-by-digit pipeline processing and allows to create new numerical systems for digit-by-digit performance of arithmetic operations, beginning from higher orders.
[1] Algirdas Avizienis,et al. Binary-compatible signed-digit arithmetic , 1899, AFIPS '64 (Fall, part I).