Jump Over Left Menu
Atlas Computer Laboratory
Some Research Applications of the Computer
Atlas Computer Laboratory
In the autumn of 1966 the Atlas Laboratory published a collection of short papers under the title Some Aspects of Current Operation and Research, with the aim, to quote from the Foreword, of showing the range of activities that go on in a big and busy computing centre. Two years later the Laboratory is bigger and busier and the range of activities still wider, and it seemed a good time to produce a second volume. The papers presented here are again, for the most part, short and self-contained and we hope they will be found interesting in themselves as well as emphasising the remarkable breadth of application of the digital computer.
The papers of the 1966 collection were all written by members of the Laboratory; this time we felt we should draw also on our users. We have about 1,200 users on our books, all of whom are research workers in universities, government establishments and research institutes supported directly or indirectly by government funds. They work in every conceivable field and their projects make demands on the machine running from under an hour to several hundred hours' computing time. Thus there is no such thing as an average or typical user (except in some crude and meaningless arithmetical sense) and any selection will be arbitrary. I made the present selection after looking through our records, with the idea of a wide range of kinds of application in mind. I am exceedingly grateful to the authors for agreeing so readily to write the papers - and, what is more, for actually writing them so promptly. Several of them presented me with a slight problem by ending their papers very charmingly with acknowledgements for the services they had had from us: should these be retained, in a volume published by the Laboratory itself? After some thought I decided it would be churlish to ask the authors to remove them, so they are here, exactly as they were written. Anyone who supplies any kind of service feels better for knowing that he has some satisfied customers; this applies very strongly to people who run computing services, for whom the receiving of more kicks than ha'pence is a known and accepted occupational risk.
The 1966 publication gave a fair amount of detail about the Laboratory's equipment and organisation. There is no need to repeat this here, but a short note may be found helpful and a background to the individual papers.
The Atlas Laboratory is one of the establishments of the Science Research Council and has been operating since October 1964 as a national facility for scientific computation. Its services are available, without charge, to research workers in all British universities and also to government and associated bodies at a charge which equals the actual cost of the work done. At the time of writing about 75% of the work-load comes from universities, 15% from government and similar users and the remaining 10% from the Laboratory itself.
The main equipment is a large Atlas I system made by ICT Limited - now renamed ICL, after the merger with English Electric Computers. The word length is 48 bits, the main store is l44K words distributed 48K words on cores of cycle time 2 µsecs and 96K on magnetic drums, with an automatic page-turning system which allows the user to address this as a continuous one-level store. There is a disc store (Data Products Model 5045) of capacity l6!M words with two independent read-write mechanisms, a main magnetic tape system of 16 Ampex TM-2 decks, using 1-inch tape and a subsidiary system of two IBM 729 Mark IV decks, using 7-track, ½-inch tape. The IBM system is temporarily reduced to one deck, the other being reserved for the SC 4020 microfilm recorder described below.
There are four independent input peripherals, two card readers and two paper tape readers, the latter allowing the use of 5, 7 or 8 track tapes; and for output two line printers (Analex 1,000 lines a minute), two card punches and one paper tape punch. In addition there is an SC 4020 magnetic tape-to-microfilm recorder, with which output can be produced directly on 16mm or 35mm film or on hardcopy photographic paper, in the form of printed text and tables, point plots, curves, diagrams and cine films. There is the standard equipment for punching cards and paper tape, and for reproducing (i.e. copying), interpreting and sorting punched cards.
A multi-access console system is being developed. An SDS Sigma-2 computer (marketed in England by G.E.C with the name S-2) with 32K core store of l6-bit words communicates on the one hand with Atlas through the disc and also through an independent channel, and on the other with a number of Teletype consoles. The principle is that the Sigma-2 will take care of all organisational work associated with the consoles, making only computational demands on the Atlas central processor and thus interfering very little with the normal batch-processing load. The hardware is installed and working and the system should be operating in a rudimentary way by the end of 1968.
The machine is run continuously on three shifts for six days a week.
The Laboratory sets out to provide a complete operating service including, if desired, punching of cards or tape from users' manuscripts. Users are required to write their own programs, but the Laboratory, through its Programming Group, provides basic programs such as compilers for the main languages (e.g. Fortran, Algol), library routines for all the important processes of mathematics and numerical analysis, and the large-scale general-purpose packages such as are required for statistical processing, analysis of survey data, determination of crystal structures and others. And through its Support Group it helps the user to make the best use of the equipment, services and programs that are available. The experience of the past four years has shown that making available these really powerful and general programs - writing any of which is a big undertaking - together with an organisation to help users to exploit them is one of the most valuable things the Laboratory can do.
The papers in this volume are placed around the site in the most appropriate place. Links only are provided here.
- Feasible Computability: A O L Atkin
- A New Conjecture Related to the Riemann Hypothesis: I J Good, R F Churchhouse
- Computing with Character Tables of Finite Groups: J K S McKay
- Computable Error Bounds for Some Numerical Problems: Joan E Walsh
- Time Series Analysis on Atlas - BOMM: P Kent
- Satellite Data Processing - Ariel III: Barbara Stokoe
- The Production of Articulated Subject Indexes by Computer:Janet E Armitage
- Manipulation of Chemical Graphs by Computer: Judith M Harrison
- COCOA - A Word-Count and Concordance Generator: D B Russell
- The N-Body Problem: S J Aarseth
- A Ten-Level Atmospheric Model for Forecasting Rain: G R R Benwell
- Calculation on the Nuclear Three-Body Problem -TAPEWORM: L M Delves
- New Methods in Transport Theory: I P Grant
- Element Pattern Generation for Finite Element Techniques: I J MacDonald
- Models of Organic Evolution in the Atlas Computer: P O'Donald
- Archaeological Classification at Chilton: F R Hodson
- Bode's Law and the Process of Statistical and Scientific Inference: I J Good