COMPUTING LABORATORY

Annual Report for the year 1968-9

  1. The Computing Service

    In last year's report it was anticipated that Oxford might have a new computer, and a building to house it, in the comparatively near future and in fact much earlier than seemed likely at this time last year. Events have confirmed this prediction, and in the early summer of 1969 it became certain that the Computer Board would grant some 0.9m for the first ICL 1906A, that subject to satisfactory performance this machine would be delivered towards the end of 1970, and that a new building in the Banbury Road/Woodstock Road "Tongue" area, together with the adaptation of the houses 17 and 19 Banbury Road, would be available to house this machine and its associated computing service staff.

    Meanwhile the KDF9, and its staff, tried hard to meet rapidly-increasing demands. The total output was improved by the introduction of a third operator-staffed shift and some extra weekend operation, and some rather strict checks on the use of the machine by individual clients helped to produce a maximum turn-round time of about twenty-four hours for most of the year. This would hardly have been possible, however, if many clients had not made more use of extra-mural computing facilities, in particular the ATLAS machine at Chilton. Others co-operated with the requests of the staff to rewrite some lengthy programmes, often at considerable inconvenience, with a view to reducing machine time and thereby enabling them to finish off projects which, through lack of machine time, would otherwise have been delayed. Some important projects, moreover, have almost certainly not been attempted through well justified fears of inadequate computing power, and there is no doubt that the new machine, eagerly awaited, will have plenty of new clients and new problems.

    The console facilities, introduced at the start of the previous year, have apparently been welcomed and used substantially to provide faster turn-round for the testing and brief running of smaller problems. The graph plotters have also had a significant amount of use, and it is clear that the present facilities in this respect will soon be inadequate.

    Another important task accomplished this year is the production of a detailed set of conditions, which the proposed new machine must satisfy at the various stages of its manufacture. This is particularly important with the installation of the very first model of a new machine, to guarantee its delivery dates and the performance of its hardware and software. To this end a benchmark set of computer runs, chosen to represent the nature of the work likely to be performed on the 1906A and effectively representing a one-shift KDF9 load, was prepared in detail and submitted to ICL. The time taken for this run on the 1906A will give an estimate of the effective power of this new machine. This benchmark set, incidentally, has attracted the interest of several other universities now engaged in various stages of a similar machine-replacing exercise.

  2. Teaching and research

    There is nothing particularly new to report under this heading, except a request from the Mathematics Faculty that all first-year mathematical undergraduates should in future be given some optional training in computation. This is likely to initiate a more substantial programme, since it is rumoured that the U.G.C. and the Computer Board would like all undergraduates to have this training.

    Nine students supervised in the laboratory gained the Diploma in Advanced Mathematics and M. El-Tom, G. Ford, and J. Williams achieved the Degree of D.Phil. The number of graduate students in residence is now about half the total of forty recently estimated as a viable number in a laboratory of this size and scope.

    The Programming Research Group applied successfully to the S.R.C. for a computer to carry out experiments on operating systems and other software. The machine, a Modular 1 system, was installed in 45 Banbury Road, and the staff have already completed and used with this machine an experimental operating system written in a high-level language.

  3. Laboratory reorganisation

    Relevant to the future prospects mentioned in last year's report, the General Board decided that the Laboratory, particularly in its computing service aspect, was now becoming so large that in addition to the Director there should be a Computer Manager, of professorial stipend and status, who would effectively have full and detailed charge of the computing service. Mr. F. A. Scott, an Oxford graduate, was appointed to this post in June 1969, and was expected to come into residence and to a Fellowship at Corpus Christi on 1 January 1970. His first duty would be to prepare a detailed submission to the Computer Board, on the lines already agreed by the General Board, for financial support for new staff, equipment, and other matters required in connection with the new computer. If all is approved the computing service will have a total staff approaching fifty in number, which gives an indication of how computing in Oxford has developed during the last decade.

    The General Board also considered that the original Computing Laboratory Committee could hardly supervise all the work of the Laboratory. I n future the computing service will be supervised generally by a Computing Service Committee, chaired by a member of the General Board and including the Computer Manager, the Director, and representatives from user departments who themselves have detailed knowledge of computing and the requirements thereof. The academic work of the Laboratory will be supervised by the Faculty Board of Mathematics, and for this purpose a Sub-Faculty of Computation has been formed which will not only report to this faculty board but will also attempt to co-ordinate the teaching and, where appropriate, the research in computation in all departments and faculties at present engaged in such work. It will also try to stimulate computer activity in new areas, particularly involving non-scientific problems and teaching.

    This is therefore the last report of the undersigned Computing Laboratory Committee.

    Professor G. E. Blackman (Chairman), Dr. M. G. Bulmer, Professor C. A. Coulson, Professor H. M. Powell, Dr. D. ter Haar, E. F. Jackson, J. S. de Wet, Dr. R. Shackleton, Professor L. Fox (Secretary)