The Distributed Computing Systems (DCS) Programme continues to be co-ordinated and supported by RAL. Within the programme, the degree of co-ordination appropriate to each research project varies according to its nature. Effort is concentrated on the following aspects:
The Programme continues to foster industrial interest in its research with the objectives of promoting collaborative projects and facilitating a high degree of technology transfer. This is achieved via:
During 1980/81 19 research grants and 3 Visiting Fellowships were awarded by the Council acting on the DCS Panel's recommendations, worth £1.4 million. A total of 74 grants has now been awarded, and there are 46 distinct, active projects in the Programme. Two Co-operative Awards were also made to DCS investigators for cooperative projects with UK companies.
Work on coupling the UNIX troff text formatting software to the SERC's FR80 microfilm recorder is now complete and has been used successfully by several research groups.
There is widespread demand for the portable compiler for the language PASCAL PLUS, funded by SERC through an EMR agreement between the Laboratory and the Queen's University of Belfast which was made available to the academic research community in pre-release form in June 1980.
The University of York, under an EMR agreement from the Laboratory, has been developing software to link UNIX (TM) (Western Electric Company) systems to British Telecom's PSS network and SERC's own X25 network. The first phase of this work is now complete and field testing of the software is commencing.
RAL staff have continued to support the usage of Cambridge Ring equipment by research groups in the DCS Programme both through the provision of hardware expertise and the development of software to support the Cambridge Ring in both the UNIX (TM) and UCSD PASCAL (TM) (Softech Microsystems Inc and the Regents of the University of California) operating systems. This work is being done in conjunction with research groups at the Universities of Cambridge and Kent. The software is already in use by a number of DCS research groups. The 10 Rings, of 6 nodes each, ordered from Logica VTS Ltd have now been delivered and installation is imminent.
Interest and experience of using the Cambridge Ring continues to grow in both the research and service communities. A number of highly successful Special Interest Group meetings covering performance modelling, protocols and operating system interfacing to the Ring have been held during the year. RAL support staff work closely with the Joint Network Team to foster the usage of local area networking and distributed computing techniques in the wider university community.
A bibliography of papers relating to the Cambridge Ring has been produced, and this together with reports on the Special Interest Group meetings has been distributed in the mailshot. The bibliography will be updated periodically.
Significant progress has been made in the research projects during the year. The multiprocessor systems at UMIST (Cyba-M) and Loughborough (NEPTUNE) have been the subjects of highly successful one day workshops. These machines are now being used by groups other than the originators for the investigation of the applicability of distributed computing techniques to specific problems of both industrial and academic interest. The University of Sussex POLYPROC system was also described in detail at a two-day workshop in June.
Detailed descriptions and progress reports on each of the projects in the programme are contained in the DCS Annual Report, which is available from SERC's Central Office.
A SERC Panel, chaired by Derek Roberts, recommended that SERC mount major new initiatives in several fields of Information Engineering. One of the new technologies recommended for encouragement was Software Technology.
The Roberts Panel reported that:
Already the cost of software is frequently greater than that of the associated hardware, and this trend will be accentuated by the continued reduction in (silicon) hardware costs. Software production must be one of the few industries where no adequate tools exist for specification, design, production updating and re-engineering. Despite the high cost, and long development cycle of most large systems, no serious attempt is being made to develop new software methods and standards which could reduce both cost and timescale. Add to that the need for improved hardware independence and more user-orientated approaches to high level language development and it is clear that this should be a major area for SRC support. The Panel is aware that the Computing Science Committee has devoted significant funding to this area but considers that more should be done, particularly in exploiting existing research and applying this in industry. As with Silicon Chip Design, there may be a need for a mechanism to bring Universities, Software Houses and Industry together and maximise the benefit of academic research, taking full account of the major contribution industry is able to make in this area.
SERC's Computing and Communications Sub-Committee (CCSC) has acted on the Roberts Panel's recommendations and launched a Software Technology Initiative on 1 November, which will run for approximately 5 years. The CCSC has identified three major objectives for the Software Technology Initiative:
RAL is co-ordinating and supporting the Software Technology Initiative. The Coordinator is responsible for implementing the CCSC's policy, stimulating new grant applications, co-ordinating existing work and encouraging technology transfer. The RAL support work involves the operation of an equipment pool of PERQ single user systems. These are part of a common base policy of orientating software development technology towards PASCAL/FORTRAN, UNIX operating system, PERQ single user computers, Cambridge Ring local area networks and X25 wide area networks.
The PERQs are being loaned to university investigators as part of the Software Technology Initiative's plan, organised by RAL, to:
It is envisaged that by setting up software technology centres, co-ordination, collaboration between funding bodies, an increased level of research activity and the creation of software engineering development and user communities, the British academic community will be drawn into close association with British industry to effect two-way transfer of software technology which is so vital in a modern industrial society.
The year has seen the amalgamation of various engineering applications computing activities into the work of a single group at RAL. The group is responsible for providing coordination, software and support for the Special Interest Groups which report to the Engineering Board Computing Committee. The Special Interest Groups are currently concerned in Tools for Interactive Programming, Artificial Intelligence, Control Engineering, Finite Elements, Electromagnetics and Electric Circuit Design. RAL also provides software associated with the Roberts Panel initiative for integrated microcircuit design, for CAD support of the Electron Beam Lithography Facility and for industrial collaboration, both in the UK5000 Gate Array project and in Electromagnetics. Progress in these various activities is summarised below.
This Special Interest Group met for the first time in November 1980. It has taken on the steering role for the Pre and Post-Processor Working Party programme. The aim is to provide a set of tools for the construction of interactive software, particularly aimed at the interactive CAD and pre and post-processing areas. These tools will be used by researchers who need access to software which is not central to their research activity. The Group's major interests will be in graphics input/output, command processing, data management, shape calculations, geometric modelling and mesh generation. Some of these areas are already covered by other SERC-supported groups.
The work for this Special Interest Group is carried out at Edinburgh University under a SERC contract, and the computer programs run on the DEC10 there. The major achievement during the year has been the completion of development on a new version WPOP of the POP-2 language, and this is now fully supported, together with PROLOG and three versions of LISP.
This Group has two main objectives:
A proposal for a machine-independent version of POP-2 suitable for installation on the PERQ has been accepted, and a proposal for a machine-independent version of PROLOG is being discussed.
The SIG-supported software is used extensively at Edinburgh and Sussex Universities, with other projects at Exeter, Warwick, London and Liverpool.
The main activity of this Special Interest Group is the development of a library of control engineering subroutines, which is carried out on the UMIST PRIME 750 by programming staff at Kingston Polytechnic supported by a SERC contract. The first version of the library, containing about one quarter of the algorithms originally proposed for inclusion, is now available. Keen interest in this work is being shown in America, Sweden and the Netherlands, with possibilities for international collaboration.
Other software now supported includes the Cambridge Linear Analysis and Design Package on the PRIME 750 at UMIST and the GEC 4090 at Cambridge, and the simulation languages ACSL on the PRIME computers at UMIST, Warwick and Sussex. An interactive data-analysis package IDA is being developed at Warwick with the support of the Group.
The Finite Element method has been a powerful computing tool for stress analysis for many years; only recently however has the technique been widely recognised as a suitable means for solving a broad range of partial differential equations, with applications in physics, electronics, medicine and fluid dynamics as well as engineering. With such a wide user group it is not surprising that a range of software is needed.
For the designer interested in stresses, the ASAS and BERSAFE programs have been provided; usage of these programs is steadily growing. Applications have been as disparate as the vibration of trombones and the quenching of metals. For advanced analysis in the stress area, and for analyses which regard finite elements as a tool for solving partial differential equations, a Finite Element Library has been provided. This is not a package, but a two-level library. The lower level (Level 0) consists of a set of subroutines which are necessary to build a finite element program. The Level 1 library consists of a set of small programs which show how the Level 0 library can be used to solve a wide range of problems, with examples from both stressing and non-stressing areas. The Finite Element Library has been under continuous development throughout the year, and the next version (release 2) will be distributed by the NAG Library organisation so that non-SERC users can take advantage of the software. One-week courses on how to use the software have also been given.
Generation of data in the finite-element mesh, and analysis of results, cause users major problems due to the large quantities of data involved. Version 7 of the interactive graphics mesh generator FEMGEN has become available during the year; this overcomes the size restrictions present in the previous version and also gives the user some powerful transformation facilities. Interfaces to the finite element software exist. Post-processing results following a finite element analysis requires a powerful 3-dimensional viewing and display tool, and this is provided by the FEMVIEW program which became available towards the end of the year. A result of its use is shown in Fig 3.3.
Specialist user meetings were held on Plate and Shell Analysis and Geotechnical Applications. Both meetings were well attended and more are planned for 1982.
The Engineering Board is supporting the development of computer programs for eddy current calculation in three dimensions. Several novel methods have been developed at the Laboratory and results have been compared with an experiment performed at the University of Bath. Fig 3.4 shows the comparison of eddy currents for an experimental configuration which provided a severe test of the computer method.
Release 6 of the PE2D software package (Poisson equation solver in 2 dimensions) now has a facility for plotting particle trajectories through magnetic or electrostatic fields. A version of the solver which can handle space charge beam problems has also been written and will be available shortly. At present the latter can deal only with electrostatic fields but further development and also extension of this work to three dimensions is planned.
The software under this heading has two main functions:
This has proved to be most effective in reducing the need for researchers to write specific programs to help in their projects - in most cases the team at RAL can advise on the use of an existing software package or technique to the demonstrable benefits of the research.
New programs mounted during the year include DINAI (digital filter analysis), ADFP (amplitude and delay fitting) LCP2 (filter design) and SPICE 2G (circuit analysis particularly of MOS transistor circuits). Work has continued to improve both the existing software and the users interface to it. Three courses on SPICE and a NAP-2 seminar were held during the year.
Use of the software is steadily increasing, the program SPICE and ASTAP, for instance, both having been used over 3000 times in the period. In its role as the distributor of the PRIME version of NAP-2, the team has sent copies to 14 sites in the UK and Europe and two in the USA.
The principal event was the installation of the GEC 4090 computer at RAL in May. This new GEC product is aimed at competing with the PRIME 750 or VAX 11/780s. The machine delivered was the first preproduction model and because of this had a somewhat degraded performance. After installation a range of benchmarks were run confirming the GEC 4090 performance characteristics and also showing the machine to be highly reliable, despite being a preproduction model.
As a result of the success of the GEC 4090, four more 4090s have been ordered for installation at Cambridge, Cardiff, Bristol and a second one at RAL. A further 4090 has been ordered for Manchester on behalf of the physics community. These machines replace smaller GEC machines which are being used to replace GEC 2050 workstations. The Cardiff machine has already been installed and has attracted favourable comment from the users.
The Queen Mary College and Heriot-Watt machines were installed in August and November, respectively. These machines had been much delayed due to the financial problems at the end of 1980. All the installations have been very smooth, with the machines providing a service within a day or two of delivery.
Development of the GEC software has centred on performance. Our analysis showed that the principal cause of poor performance was the pressure of disk traffic. A series of changes reduced the traffic significantly and virtually doubled the performance. In addition, many improvements have been made in other areas, in particular the network code. UNIX is being implemented on GEC 4000 machines in such a way as to co-exist with OS4000 and share the same file store. It turns out that the GEC hardware design leads to a fairly efficient UNIX implementation.
The machines in the field have, as usual, been highly reliable and provided a very satisfactory service. The loading on the machines has risen steadily but the demand for file space has been insatiable and shown that the user requirements have outstripped the capacity of the two 70 Mbyte drives supplied with each machine.
The support and software maintenance has been severely strained with the large number of machines supported now over 20. However the quality of the work has been high, judging from the low level of expressed dissatisfaction. User support has now become more reliant on the site managers as they have become more experienced. This has allowed the support level to be fairly low. In fact, much of the effort at RAL is concentrated on providing good quality manuals and solving the more difficult problems.
Now that the computers have been available for almost three years it can be seen that the initial aim of supporting the machines from RAL has been successful and resulted in cost savings and a good user service.
No new PRIME computers have been installed during the year. However, hardware enhancements have been made on several machines. The size of physical memory has been increased on all PRIME 750s from 1.5 Mbytes to 2 Mbytes and on PRIME 550s from 0.5 Mbyte to 0.75 Mbyte. Two dual density magnetic tape decks have been purchased, one for Sussex University and the other for RAL. The old magnetic tape deck from RAL is to be installed at University College London. This will enable a tape archiving system to be introduced on all systems since all PRIMEs now have a magnetic tape unit. A new upper and lower case lineprinter has been installed at RAL to replace the worn upper case only printer.
Support for PRIME computers has been transferred to UMIST as the machines were added to SERCnet. UMIST is now responsible for all user support activities and for the support of system software, apart from the operating system and network utilities.
The Georgia Tech Software Tools subsystem has been purchased for all ICF PRIME computers. This subsystem provides a UNIX-like command environment with the Software Tools, Shell, pipes and standard input and output ports. A comprehensive range of utilities is provided, including a screen editor, and interfaces to the standard PRIME compilers. At present this subsystem runs at RAL and UMIST but it is hoped that other ICF machines may support it, subject to adequate memory and disk-space.
The appearance of cheap high powered single user computer systems with good interactive capabilities via high precision displays, linked together by high speed local area networks, heralds a completely new way for most SERC investigators to achieve the major part of their computing requirements. Within the next few years, many such systems will be available from different manufacturers. Consequently there is a likelihood of many different systems being purchased in the SERC environment leading to a great deal of duplication of basic software development.
SERC sees a need for a co-ordinated development plan to ensure that the UK makes the best use of its finances and of its limited manpower. The SERC has therefore decided on a strategy of creating a common hardware and software base for software development which will encompass all scientific subject areas. Briefly the common software base will be PASCAL and FORTRAN running under the UNIX operating system implemented on the common hardware base of PERQ single user computers linked locally by Cambridge Rings and nationally by the X25 wide area network systems (SERCnet and PSS).
SERC Subject Committees will participate in the implementation of this policy by enabling the central purchasing of PERQ computers for grant holders to be done via the Central Computing Committee and by ensuring that investigators use the PERQ in all appropriate circumstances as well as encouraging them to follow the common base software development policy. The Common Base Policy is not the same as standardisation, however, and it will evolve as the state of the art improves.
The PERQ is a high powered, single user computer system with a high precision display system which provides a significant improvement in the quality and speed of interaction. Its main features are:
A high quality, superbly interactive computing system is created if each investigator has his own single user PERQ linked to his colleagues' PERQs and other departmental computing resources by a Cambridge Ring, with inter-university co-operation being fostered by the national X25 network connections.
The whole academic community, not just Computer Science, is a major user and developer of software and so the degree of ease with which software can be developed affects the scientific productivity of many researchers. SERC has approved a plan to increase the productivity of scientific research requiring computing by:
Currently the academic software technology base is very non-uniform in that the knowledge, experience, tools, techniques and equipment vary considerably between projects. The motivation to create a common hardware and software base is to bring together all of the best existing tools, packages and techniques into a uniform framework so that the whole is more effective than the sum of diverse parts. This will be achieved via EMR contracts to move existing software into the common base, specific purchases, the direct results of SERC research projects using the common base equipment and the snowball effort that will be generated as a natural consequence of providing a state of the art hardware base. A good example of the common base snowball effect is the widespread use of the UNIX operating system which has enabled a large number of software tools to be made available throughout the UK academic community.
The common base policy briefly is:
SERC wishes the common software base to be the UNIX operating system and the common hardware base to be the PERQ. The PERQs should be networked together via Cambridge Rings, SERCnet and PSS to allow widespread co-operation between users and developers. This combination of software and hardware is widely accepted as being the best combination for developing software in the coming years. A common base does not imply rigid standardisation however.
Computer technology develops at a rapid pace and it is expected that the next few years will see the cost of single user systems decline and their quality and capability increase. Therefore today's PERQ is seen as only the first machine forming the common hardware base. The common base will develop over the coming years.
RAL has been looking for a suitable single user computing system for several years. In August RAL was instrumental in effecting a marketing and manufacturing deal between ICL and the Three Rivers Computer Corporation of Pittsburgh who developed the PERQ. In September SERC signed a memorandum of agreement with ICL to collaborate on the further development of single user systems to effect the common base policy. RAL staff are transporting the UNIX operating system to the PERQ, building a PERQ to Cambridge Ring interface, mounting GKS - the new international standard graphics package - on the PERQ, mounting applications packages and co-ordinating and supporting university-based PERQ developments in association with the ICF, Distributed Computing Systems Programme and the Software Technology Initiative.
Late in 1979, a proposal was made to provide assistance to one local school (Wantage Sixth Form) in order to assess its potential as a mechanism for improving recruitment of junior programmers by RAL. This assistance took the form of an allocation of time on one of the PRIME computers at RAL, the loan of two terminals from a pool of old terminals and dial-up facilities from Wantage to enable the teachers and students access to the machine. The PRIME computer provided a more sophisticated programming environment than had been available on the school microcomputers. In particular it was possible to provide the PRIMOS operating system, a better version of BASIC, as well as FORTRAN and PASCAL and a graphics package. Five or six students were also offered computing work experience at RAL for five weeks during the summer holidays.
In the year 1980/81 there were 14 students in the upper sixth and 16 in the lower sixth studying the A-level computing course at Wantage. Two of these students who had also been on the work experience scheme at RAL the previous summer were offered programming jobs. Owing to the success of the scheme, it was decided to try to expand the project to provide similar facilities at other schools in the neighbourhood. Thanks to generous help from PRIME, RAL can now provide a PRIME 400 with 512 Kbytes of store and an 80 Mbyte disk. This should prove sufficient to enable the number of terminals at Wantage to be increased to four, and provide support for at least four more schools in other towns. Leased lines have already been installed between RAL and three local schools: Wantage Sixth Form, St Birinus School (Didcot) and Abingdon School. Leased lines have been ordered to link RAL with Wallingford Upper School and Abingdon College of Further Education.
For the year 1981/82 the number of students studying A-level computing at Wantage has increased to 58. They have therefore been provided with two extra terminals. The four new schools will be provided with two terminals each initially (from the pool of old terminals) but it will be possible to increase this number later if demand warrants it. It may be possible to provide links between the RAL schools computer and other educational machines to enable the easy transfer of educational software.
It is hoped that the schools themselves will take some part in managing the schools' computer and allocating resources amongst themselves so that they become aware of all aspects of a professional computing organisation. In addition to providing computing facilities, close links have been forged between the teaching staff at Wantage and the professional programming staff at RAL which has proved useful in providing the teachers with knowledge of the practical problems involved in bringing a large programming task to a satisfactory conclusion. It is hoped that similar links will be built up between RAL staff and the other schools involved in the project.
