Parallel and Efficient Optical MSD Arithmetic Processing

修正符号数系统由于其固有的弱位数间依赖性，被提出作为一种有效的快速并行数字算法。近年来，利用各种光学逻辑和存储处理技术，提出了光学MSD算术运算。然而，在大多数现有的光学方案中，采用了多阶段算法。对于多级方法，由于光级联，额外的信号能量损失和速度降低是不可避免的。为了保持快速的处理速度，提出了一种基于单级全息光学内容可寻址存储器（CAM）的MSD算法。然而，由于其复杂的三步相开关记录，其实验实现困难。在本研究项目中，提出了一种基于非全息光电CAM的MSD快速加法处理体系结构。我们的一阶原理证明实验验证了所提出的概念。作为优点的数字，还对这种方法和其他现有方法进行了比较。基于这个关键的光电CAM元件，实现更复杂的MSD算术运算成为可能。最后，为了完成MSD算术处理，传统二进制数和MSD数之间的转换，这是其他光学方法没有解决的问题，也在研究中。

The modified signed-digit (MSD) number system, because of its inherent weak interdigit dependance, has been suggested as a useful means for a fast and parallel digital arithmetic. Recently, using various optical logic and memory processing techniques, optical MSD arithmetic operations have been proposed. However, in most existing optical schemes, a multiple-stage algorithm is utilized. For a multi-stage approach, because of optical cascading, additional signal energy loss and speed reduction is unavoidable. To maintain a fast processing speed, a single-stage holographic optical content-addressable memory (CAM) based MSD algorithm was suggested. However, due to its sophisticated three-step phase-switching recording, its experimental implementation is difficult to achieve. In this research program, a novel non-holographic opto-electronic CAM based fast MSD addition processing architecture is developed. The proposed concept has been verified with our first-order proof-of-principle experiments. As figure of merit comparison of this and other existing approaches has also been made. Based on this key opto-electronic CAM element, implementation of more sophisticated MSD arithmetic operations are possible. Finally, for the completion of a MSD arithmetic processing, the conversion between the conventional binary and MSD numbers, an issue that other optical approaches did not address, is also under investigation.