A computing structure for data acquisition in high energy physics
A review of the development of parallel computing is presented, followed by a summary of currently recognised types of parallel computer and a brief summary of some applications of parallel computing in the field of high energy physics. The computing requirement at the data acquisition stage of a particular set of high energy physics experiments is detailed, with reference to the computing system currently in use. The requirement for a parallel processor to process the data from these experiments is established and a possible computing structure put forward. The topology proposed consists of a set of rings of processors stacked to give a cylindrical arrangement, an analytical approach is used to verify the suitability and extensibility of the suggested scheme. Using simulation results the behaviour of rings and cylinders of processors using different algorithms for the movement of data within the system and different patterns of data input is presented and discussed. Practical hardware and software details for processing equipment capable of supporting such a structure as presented here is given, various algorithms for use with this equipment, e. g. program distribution, are developed and the software for the implementation of the cylindrical structure is presented. Appendices of constructional information and all program listings are included.