E-Book Content
Lecture Notes in Chemistry Edited by G. Berthier M.J.S. Dewar H. Fischer K. Fukui G. G. Hall J. Hinze H. H. Jaffe J. Jortner W. Kutzelnigg K. Ruedenberg J. Tomasi 44 Supercomputer Simulations in Chemistry Proceedings of the Symposium on Supercomputer Simulations in Chemistry, held in Montreal August 25-27, 1985, sponsored by IBM-Kingston and IBM-Canada Edited by M. Dupuis Springer-Verlag Berlin Heidelberg New York London Paris Tokyo Editor M. Dupuis IBM Corporation, Dept. 48B MS 428, Data Systems Division Neighborhood Road, Kingston, NewYork 12401, USA ISBN 978-3-540-17178-2 001 10.1007/978-3-642-51060-1 ISBN 978-3-642-51060-1 (eBook) This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically those of translation, reprinting, re-use of illustrations, broadcasting, reproduction by photocopying machine or similar means, and storage in data banks. Under § 54 of the German Copyright Law where copies are made for other than private use, a fee is payable to "Verwertungsgesellschaft Wort", Munich. © Springer·Verlag Berlin Heidelberg 1986 Softcover reprint ofthe hardcover 1st edition 1986 Printing and binding: Druckhaus Beltz, Hemsbach/Bergstr. 2152/3140-543210 Preface Awareness of the need and potential of supercomputers for scientific and engineering research has grown tremendously in the past few years. It has culminated in the Supercomputer Initiative undertaken two years aga by the National Science Foundation and presently under full development in the United States. Similar initiatives are under way in several European countries and in Japan too. Thus the organization of a symposium on 'Supercomputer Simulations in Chemistry' appeared timely, and such a meeting was held in Montreal (Canada) in August 1985, sponsored by IBM-Kingston and IBM-Canada, and organized by Dr. Enrico Clementi and Dr. Michel Dupuis. In connection with this, IBM's support of the Cornell University Supercomputer Center, several projects in the IBM Research Division, the experimental parallel engine (ICAP) assembled at IBM-Kingston, and the announcement (Fall 1985) of an add-on vector feature to the 3090 IBM mainframe underscore IBM's commitment to high-end scientific/engineering computing. The papers presented in this volume discuss topics in quantum mechanical and statistical mechanical simulations, both of which test the limits of computer hardware and software. Already a great deal of effort has been put into using vector supercomputers in these two areae. Much more is needed and, without doubt, ie bound to happen. To start, an historical perspective of computational quantum chemistry is provided by Professor Löwdin. The contribution by Ohno and co-workers gives an indication of the present status of Japanese supercomputers. Kutzelnigg et al., Bauschlicher et al., and Guest et al. deecribe three significant efforts dealing with various aspects of programming quantum mechanical methods on vector supercomputers, and advances already observed in applying these methods to important problems of chemical physics, and surface science. Finally Truhlar et al. discuss the state-of-the-art in the field of chemical dynamics where today's studies contribute significantly to the detailed understanding of gas phase collision dynamics. An alternative to vector supercomputers is discussed by Clementi et al., in their presentation of the loosely coupled array of processors (lCAP). This system is ideally suited to quantum and statistical simulations. Lie and et al., and Scheraga et al. present two studies carried out on the ICAP system. Indeed the approach to the determination of the three-dimensional structure of proteins is CPU intensive and inherently 'parallelizable.' Statistical simulations were born with the ad