|
Roger has a track record of pioneering work in the field of computer simulations, numerical methods and performance analysis of supercomputers. He established his reputation whilst at Stanford with the introduction of the Particle-In-Cell (PIC) method for modelling plasmas, galaxies, molecules and the universe. He is also responsible for some of the rapid methods for the solution of Poisson's equation where he broke new ground in developing the FACR algorithm. One consequence of his computer simulation work was that he became a leading authority on using and analysing the world's most powerful parallel computers. His N-half simple performance model has been widely accepted and used by the supercomputing community for more than 15 years. In addition, Roger was one of the initiators of several international benchmarking groups and initiatives being a co-founder of EUROBEN and the first chairman of the PARKBENCH Committee. After graduating in the Natural Science Tripos (Part II in Physics) from Cambridge University (UK) Roger spent three years as a Teaching Fellow at the University of Michigan. Then he joined English Electric as a nuclear physicist using early analogue and the first generation of UK electronic computers (EE DEUCE) for the design of nuclear power stations. This is the period when Roger fell in love with computers. From then on he would always have a computer in his hands for the rest of his life. At the beginning of the 60's Roger Hockney moved to California to work towards his PhD at Stanford University under Oscar Buneman, George Forsythe and Gene Golub. After completing his thesis entitled "The computer simulation of anomalous plasma diffusion and the numerical solution of Poisson's equation", Roger Hockney spent the late 60's working for NASA (Langley) and IBM Research (Yorktown Heights). In 1970 he became the first Professor in Computer Science at Reading University (UK), where he created a research group in computational physics. In 1985 Professor Hockney took early retirement and became a consultant in parallel computation while also holding Visiting Professorships at both Southampton and Westminster Universities. Roger has numerous published results in the area of numerical analysis, computational science and high-performance computing. He is author or co-author of three books: "Computer Simulation Using Particles" (with Jim Eastwood, McGraw Hill 1981; reprinted IOPP 1988), "Parallel Computers" and "Parallel Computers-2" (with Chris Jesshope, Adam Hilger 1981; 2nd edition IOPP 1988), and "The Science of Computer Benchmarking" (SIAM, 1996). His last public lecture was a joint tutorial on "Performance Analysis, Evaluation and Optimisation" with Tony Hey and myself last September at EuroPar'98 in Southampton. Roger Hockney was well liked by all those who worked with him, and will be missed both as a friend and colleague. His scientific work made him well known throughout the world from California, Michigan and England to Japan, Russia, Italy and Bulgaria as one of the world's leading computational scientists. Vladimir Getov
|
|