Research Initiation: Mechanical Synthesis of Systolic Arrays

用于实现仿射递推方程的收缩阵列的设计正在进行。本研究的目标是：(a)设计一个自动化系统，用于从仿射递归方程生成收缩架构，并与现有的VLSI布局工具接口；(b)收缩阵列的合成，允许设计由多个级联子阵列组成的复杂收缩阵列。新的心脏结构可以利用这一理论进行设计。设计的系统也可以作为新理论的测试平台。收缩阵列是一类重要的并行架构，在利用VLSI技术的进步方面显示出巨大的希望。由于它们是特定于问题的，因此开发用于从高级规范中综合它们的自动化工具非常重要。这项研究是针对仿射递推方程，它涵盖了一般类型的实际问题。首席研究员是一位新博士，他的研究已经取得了重大进展。研究集中在重要和及时的主题。强烈建议提供支持。

The design of systolic arrays for implementing Affine Recurrence Equations is being carried out. The goals of this research are (a) the design of an automated system for generating systolic architectures from Affine Recurrence Equations and for interfacing with existing VLSI layout tools; (b) the synthesis of systolic arrays that allows the design of complicated systolic arrays consisting of a number of sub- arrays in cascade. New systolic architectures can be designed using the theory developed. The system designed also serves as a testbed for the new theory. Systolic arrays are an important class of parallel architectures that have shown great promise for exploiting the advances in VLSI technology. Since they are problem specific, the development of automated tools for synthesizing them from high-level specifications is very important. This research is directed towards Affine Recurrence Equations, which covers a general class of practical problems. The principal investigator is a new Ph.D. who already has significant progress in his research. The research focuses on important and timely topics. Support is highly recommended.