scispace - formally typeset
Search or ask a question

Showing papers by "Richard Cole published in 1985"


Journal ArticleDOI
TL;DR: A deterministic algorithm for finding the k th smallest item in a set of n items, running in O log log n 2 parallel time on O(n) processors in Valiant's comparison model is given.

49 citations


Proceedings ArticleDOI
21 Oct 1985
TL;DR: Two new techniques for establishing lower bounds on the information flow in VLSI circuits are presented: an averaging technique, which is easy to apply to a variety of problems, and a technique for constructing fooling sets in instances where the averaging method is unlikely to provide an adequate bound.
Abstract: This work comprises two parts: lower bounds and upper bounds in VLSI circuits. The upper bounds are for the sorting problem: we describe a large number of constructions for sorting N numbers in the range [0,M] for the standard VLSI bit model. Among other results, we attain: • VLSI sorter constructions that are within a constant factor of optimal size for almost all number ranges M (including M = N), and running times T. • A fundamentally new merging network for sorting numbers in a bit model. • New organizational approaches for optimal tuning of merging networks and the proper management of data flow. The lower bounds apply to a variety of problems. We present two new techniques for establishing lower bounds on the information flow in VLSI circuits. They are: • An averaging technique, which is easy to apply to a variety of problems, including a long standing question regarding the AT2 complexity for sorting. • A technique for constructing fooling sets in instances where our averaging method is unlikely to provide an adequate bound.

14 citations


Proceedings Article
01 Jan 1985
TL;DR: VLSI sorter constructions that are within a constant factor of optimal size for almost all number ranges M (including M = N), and running times T are attained.
Abstract: This work comprises two parts: lower bounds and upper bounds in VLSI circuits. The u~per bounds are for the sorting problem: we describe a large number of constructions for sorting N numbers in the range (0, M) for the standard VLSI bit model. Among other results, we attain: • VLSI sorter constructions that are within a constant factor of optimal size for almost all number ranges M (including M = N), and running times T. • A fundamentally new merging network for sorting numbers in a bit model.

2 citations