System for interpreting Rent's Rule for ultralarge-scale integrated circuit designs
Original Publication Date: 2003-Sep-08
Included in the Prior Art Database: 2003-Sep-08
Computer hardware components have changed significantly since the 1960's, 1970's, 1980's, and even since the early 1990's. Work concerning Rent's Rule prior to the present paper has been based on a 1971 interpretation of two unpublished memoranda written in 1960 by E. F. Rent at IBM, even though today's computer components are significantly different from those in 1960 and 1971. However, because of the significant changes in design and implementation of computer hardware components since 1960-1971, a new interpretation of Rent's Rule is needed for today's components. The authors have obtained copies of Rent's two memos; in these 1960 memos, E. F. Rent describes the method that he used to deduce an empirical relationship between properties of IBM 1401 and 1410 computer hardware components. The authors have studied these memos carefully in order to understand Rent's original intent. Based on our careful reading of these two memos, the personal knowledge of one of us (R. Rand) with the 1401 and 1410 computers, and our experience with ULSI circuit design for high-performance microprocessors, the authors have derived an historically-equivalent interpretation of Rent's Rule suitable for today's computer components. The purpose of this disclosure is to present this new interpretation of Rent's Rule and its application to wirelength distributions of ultralarge-scale integrated (ULSI) circuits. This disclosure will: (1) describe the contents of the memos and Rent's method, (2) provide an historically-equivalent interpretation of Rent's Rule for today's computer components, and (3) apply this new interpretation to actual ULSI circuit designs. This new interpretation provides improved wirelength distribution models with better qualitative agreement and more accurate estimates of wirelength distributions and wirelength requirements in ULSI designs compared with prior methods.