Title :
Computational Geometry on a Systolic Chip
Author :
Chazelle, Bernard
Author_Institution :
Department of Computer Science, Brown University
Abstract :
This paper describes systolic algorithms for a number of geometric problems. For the sake of realism we restrict our investigation to one-dimensional arrays whose communication links with the outside are located at the end cells. Implementations yielding maximal throughput are given for solving dynamic versions of convex hull, inclusion, range and intersection search, planar point location, intersection, triangulation, and closest-point problems.
Keywords :
Analysis of algorithms; VLSI; computational geometry; convolution; parallel computation; pipelining; real-time algorithms; systolic arrays; Algorithm design and analysis; Circuits; Computational geometry; Computer science; Concurrent computing; Costs; Pipeline processing; Systolic arrays; Throughput; Very large scale integration; Analysis of algorithms; VLSI; computational geometry; convolution; parallel computation; pipelining; real-time algorithms; systolic arrays;
Journal_Title :
Computers, IEEE Transactions on
DOI :
10.1109/TC.1984.1676494