Implementation of a fast algorithm to find data clusters

In this paper we present a very fast general algorithm to find clusters in a stream of data. The algorithm is implemented in FPGAs (Field Programmable Gate Array) using VHDL language. In its most general form, the algorithm is capable of finding n consecutive elements characterized by some common parameters (a cluster) and reporting the parameters of the cluster such as its location and width within the data stream. This algorithm is used in the Dzero Trigger System to find clusters in the PreShower detector and is capable of finding the clusters within a set of 1280 elements and reporting their width and address in 65 clock ticks.