Mr Peter Davey is to be the new head of the Interactive Computing Facility. He will be a member of the Atlas Computing Division of the Rutherford Laboratory; this Division manages the Interactive Computing Facility.
A graduate in engineering from Cambridge University, Mr Davey has worked in the Nuclear Physics Laboratory, University of Oxford since 1966. He was responsible for building "PEPR", a DEC10-based instrumentation system for semi-automatic scanning and measuring of bubble-chamber film. More recently he has headed a group which has applied the basic ideas of PEPR to image analysis in a wide range of disciplines meteorology, hydrology, cartography and time-lapse photography. In 1969 he formed a company to exploit his original patents for industrial leak-testing equipment. Previously he had worked on instrumentation for nuclear power stations and had spent several years at universities in the USA and West Germany where he developed instrumentation and displays for particle physics.
The DN82 network node computer has been delivered to the Rutherford Laboratory and is now going through acceptance tests. It will provide remote access to the DEC10s at UMIST and Edinburgh for communication links from user sites which now terminate at the Laboratory. Initial tests showed some teething troubles which affected the performance of the DEC10s. We apologise to users for any inconvenience; networks of this type, connecting several major installations using advanced communication techniques, inevitably raise complicated technical problems. The staff of the Interactive Computing Facility are confident that DEC will overcome the present difficulties associated with the introduction of the DN82 into the network and that DECNET will then be a very satisfactory communications system.
Time on the UMIST and Edinburgh DEC10s is to be reserved for system development to try to solve the immediate problems and also to try out the various possible ways of linking the two machines to the 360/195s. Details are in Section 7, below.
The last issue of 'Rapid Response' mentioned that the GEC 4070 at Chilton could not be connected to the 360/195s. This may have given the impression that it was not possible to connect GEC 4000 series computers to the 360s at all. In fact, such connections have been in use for a long time. The particular version of the operating system used on the Interactive Computing Facility's 4070 has not supported the feature until quite recently. It should do so now; this will be tested shortly.
When the Interactive Computing Facility was set up, the Technical Group recommended a strong attempt to reduce unnecessary duplication of software packages and to provide adequately documented software for general use. There is a plethora of graphics packages and we are trying to introduce some standardisation for both batch and interactive graphics over all machines in the Facility. We have concentrated on packages which can be called from FORTRAN programs; this is in line with another of the Technical Group's recommendations.
The SMOG package, designed and written at the former Atlas Computer Laboratory (now merged with the Rutherford Laboratory), can access most hardware features of the FR80 Microfilm Recorder in a hygienic manner from the batch computers. It provides user coordinate mapping to each of the cameras and also line drawing and text output primitives. The package supports all the hardware character sizes of the FR80, variable spacing of characters, mono- and proportionally-spaced text oriented at any one of the 8 possible angles. SMOG is also available in a pilot form at Edinburgh on the DEC10.
GINO-F has been chosen as a standard high-level interactive graphics package. It is available at Chilton on the PRIME 400 and GEC 4070, and also on the DEC10 at Edinburgh; we are buying it for the UMIST DEC10. It can be used in batch mode on the 360/195s at Chilton. GINO-F is quite large and has poor facilities for defining the mapping from the user's problem space to the coordinate system of his output device; the user requiring unsophisticated graphics facilities can incur high overheads. None of the versions on the interactive machines can access the FR80 at present, though additional software to do this is at an early design stage.
Rutherford Laboratory staff have produced an interactive package called FINGS which gets round some of the limitations of GINO-F, and is available on the PRIME and GEC machines. It is likely that this package will be available on the DEC10s eventually. The package is aimed at Tektronix 4010 and 4014 displays. There is an FR80 back-end which supports the non-interactive routines. GINO-compatible routines are provided for 2-D line drawing, text (hardware) and symbol plotting. The Tektronix cursor and keyboard can be used as input devices. Window/viewport transformations have been introduced to ease the problems of coordinate system mapping. The package does make significant space and speed savings over GINO-F on the PRIME 400.
All these packages contain libraries of higher-level routines, eg for plotting histograms etc. In addition, many of the large and sophisticated applications packages on the interactive machines have their own graphics software facilities built in.
There is a good deal of international activity at present in defining standards for computer graphics. Professor Bob Hopgood of the Interactive Computing Facility is a member of various bodies set up by British and international standards organisations to do this. The limitations of GINO-F are well appreciated; the Computer-Aided Design Centre at Cambridge, which is responsible for GINO, has set up a Technical Committee to advise on the future development of the package. It will examine the compatibility of GINO with the standards currently proposed.
This is a series of low-cost microprocessor-based devices manufactured by Sigma Electronics Ltd. The Rutherford Laboratory has bought a model 5250. the Mini-GOC. This can be fitted to inexpensive Newbury Laboratories VDUs; the screen of the VDU can then be used as a graphical plotting area of 256 × 256 points. in black-on-white or white-on-black. It has a cursor for input of coordinate information and the selective erasure facility gives the user freedom to modify a displayed picture without re-drawing. The mini-GOC costs less than £2000; when the VDU is connected. the whole assembly is very cheap compared with eg a Tektronix 4010.
A second Mini-GOC is on order and will be installed in a university. Also on order is a Multi-plane GOC, a more expensive model which when fitted to a high-precision colour TV monitor turns it into a colour graphics display.
Staff at Surrey University are writing software to enable the GINO graphics package to drive a GOC. Any interested reader can get more information about the GOC from Dale Sutcliffe at the Rutherford Laboratory).
One of the facilities at Chilton is the FR80 Microfilm Recorder. This is a high precision plotting device whose output is a variety of photographic media - microfiche, 16mm and 35mm film (black/white and colour) and 12in wide photographic paper (hardcopy). One postcard-sized microfiche can hold 250 graphical pictures or pages of lineprinter output.
The FR80 can plot graphics or simulate lineprinter output. It is used off-line to plot jobs from magnetic tapes produced by a variety of graphics packages such as SMOG, FINGS, GINO and SPROGS.
The FR80 hardware consists of a small computer for control purposes, a precision cathode ray tube, a variety of cameras, a colour filter system, a vector generator and a character generator. Computer-generated plotting instructions are accepted on magnetic tapes via 7-track (556 and 800 bpi) tape drives and a 9-track 1600 bpi tape drive; 9-track 800 bpi will be available soon. The white phosphor cathode ray tube has a 16K by 16K raster defined over a 3.33" square tube face allowing 80 line pairs per mm to be resolved on 35mm film. A full screen vector can be drawn in 1, 2, 4 or 8 msec at one of 256 intensities, and 8 spot sizes through various combinations of filters. Characters can be plotted at up to 40,000 per second at anyone of 8 (45 degree) angles, and at anyone of 64 sizes. A software simulator in the FR80 allows characters of any size to be plotted more slowly. A variety of fonts is available.
The FR80 is an established facility with a wide range of users whose applications range from full colour animated movies, through graphs and tables of experimental results in a variety of scientific disciplines, high quality text and graphics output for publication, to compaction of line-printer output onto microfiche. The text of 'Rapid Response' is recorded on FR80 hardcopy by instructions generated using the SMOG package. The hardcopy is then "camera-ready copy" for printing.
The FR80 is managed by the Computing and Automation Division of the Rutherford Laboratory as part of the Laboratory's batch computing facility (see below). For further details about the FR80 contact Paul Nelson. Readers wishing to use the FR80 should discuss the details with him at an early stage.
The Rutherford Laboratory batch computing facility ranks as one of the largest available to scientific users in the UK. Its prime purpose is to provide bulk batch facilities for those SRC-supported research workers who cannot obtain such facilities from their University or Regional Centre.
Available since January 1972, the service has now expanded to cover most UK Universities via direct links. In addition a link has been available to the American ARPANET for two years, and to the Daresbury Laboratory's IBM 370/165 since last year.
The hardware available centrally is based on two IBM 360/195 processors, each approximately ten times the power of an ICL 1906A. Bulk storage is provided by 4000 Mbytes (total) of on-line disk-drive capacity, and high density tape drives (most at 6250 bpi). Some 30,000 magnetic tapes are held on site. Each CPU has 2 Mbytes of main memory, although after deducting space for the system there is 1,450 and 700 Kbytes respectively left for user programs.
To the user the two machines look like one system: usually programs run on either CPU as space becomes available. For those familiar with IBM systems the 195s run under IBM's OS/MVT, plus HASP. A locally written file editing and data preparation facility is available on-line from more than 200 terminals situated both at the Laboratory and at most UK Universities. Some 80% of the workload is now submitted via this facility, which also allows batch lineprinter and graphics output to be retrieved and inspected at any suitable terminal.
The batch service is used by customers from all the SRC's supported areas, ranging from astronomy and space satellite data analysis through engineering, nuclear physics, quantum chemistry to plasma physics. The facility is intended primarily for large projects, each of the 300 projects currently taking on average the equivalent of 200 ICL 1906A CPU hours each year.
About 30% of the users require some form of graphics service. The software systems GINO and SMOG are provided. Output may be displayed on a variety of terminals (eg Tektronix 4000 series), and/or sent to the FR80.
The computer centre is run by the Computing & Automation Division of the Laboratory as part of the SRC's central computing facility, and the Division Head is Mr W Walkinshaw. The Division also operates an ICL 1906A (due to close down in March 1978), the FR80, and several smaller machines. Within the Division, the Telecommunications Group (Group Leader: Mike Jane) is responsible for both the existing large networks around the 195s and 1906A and the commissioning of the network for the Interactive Computing Facility (see 'Rapid Response' No 4). A large development program is under way to provide further network facilities, and enhanced remote workstations (incorporating file editing facilities at the remote site).
Readers interested in using the batch computing facility should contact Bart Fossey at the Rutherford Laboratory.
For the reasons outlined in Section 2 (above), the period 8.30-10.00am on Mondays, Wednesdays and Fridays will be used for system development on the DEC10s. At these times no service will be guaranteed and the machines may not be available to users. If either is available. users are welcome to logon: this will help the system managers to assess how well any changes in the software stand up to "real live users". As the service is not guaranteed, any use of it in these periods will only be charged at the lowest "discount" rate. To keep the accounting systems simple, this rate will apply between 8.00 and 10.00am on all days. ie the charging rate will be low at these times on Tuesdays and Thursdays although normal service is being provided.
There are to be several enhancements to the two DEC10 installations. Both will acquire an additional 100Mbyte disk drive to bring the available filestore up to the level originally planned, an upgrade of the TM10A magnetic tape controller to a TM10B. This will reduce the heavy system overheads.
The Edinburgh machine will also get a replacement for the present seven-year-old lineprinter. It will have a 96-character set to aid document preparation. A CALCOMP 1051 drum-plotter and a 906 controller are to be installed, with a software interface to the GINO graphics package. The DC75NP communications processor will get extra memory so that it can handle more nodes.
UMIST will get a second TU10 tape drive: this will improve the security of archiving and make tape facilities more readily available to users, particularly for preparing FR80 tapes.
The main memory on the PRIME 400 is to be enlarged to 0.5 Mbytes. This will help to maintain the standard of the service while the number of users is increasing rapidly. Also, a 300 Mbyte capacity disk-drive is to be provided as well as the two existing 80 Mbyte drives. The increased on-line storage capacity will assist investigations into mounting large-scale scientific data-bases on such machines.
The disk-drives on the GEC 4070 have been upgraded from 35 to 70 Mbytes.
The PRIME at Nottingham University has been upgraded to a model 400; acceptance tests were completed on 15 November and the full computing service will be available early in 1978.
The Interdata 8/32 at Sheffield has been running multi-user operations since October. Acceptance tests for the upgraded software should be complete by the end of December 1977.
Most of the hardware for the upgrade of the PDP11/45 at Leeds University has arrived and commissioning has started. The local management at Leeds has produced a newsletter; anyone who wants to be put on the circulation list should contact Dr Alan de Pennington at the Department of Mechanical Engineering. Leeds University.
The Interactive Computing Facilities Committee held their 2 November meeting at the Rutherford Laboratory. They inspected the interactive computing hardware, saw demonstrations of it in action, briefly toured the batch computing installation and met the staff of the Interactive Computing Facility. They also met the chairmen of the Special Interest Groups. The discussion which took place should help in estimating software requirements over the next few years.
A meeting was held at the Rutherford Laboratory on 23 September 1977, and over 40 research workers attended. It was chaired by Mr R John Lansdown of Turner, Lansdown, Holt and Paterson.
The meeting opened with a description of the Interactive Computing Facility by Mr J E Hailstone. Mr E J Davies of LAMSAC (Local Authorities Management Services Advisory Committee) surveyed the available software for CAAD and came to some remarkably pessimistic conclusions about the existence of truly machine-independent program packages. Mr R W Witty then described the FR80 microfilm recorder and Dr D A Duce described the graphics software which is now available on the SRC's computers. Mr A Bijl and Dr D Rosenthal, both of Edinburgh University, discussed the available terminal equipment and described some specialist hardware developed by their group. Professor T W Maver (Strathclyde University) spoke on current trends in the organisation of the architectural profession and the impact of computer-aided design. He commented on the wide divergence between different schools of architecture in the extent to which they introduce computer-aided techniques to students.
A general discussion followed. It stressed the importance of graphics in CAAD and the need for an index of the available packages. There was considerable interest in computer animation and in the use of data-bases.
The meeting recommended that a Special Interest Group should be set up and this has now been approved.
A second open meeting, for research workers interested in electromagnetic applications, will be held at the Rutherford Laboratory on 25 January 1978. It will be chaired by Professor P J Lawrenson. Its main purpose is to describe what the Special Interest Group has done and to get the views of the research community about future requirements. There will also be talks on individual projects, and demonstrations. Any interested reader should contact John Collie at the Rutherford Laboratory.
The SIG is holding regional meetings to introduce R&D workers in digital circuit design to software developed in the various centres. One meeting has already been held at Edinburgh University and others are planned at Leeds, Manchester, Bristol and Brunel. The next (at Brunel) is on 6 January 1978. Interested readers should contact Peter Dewar at the Rutherford Laboratory.
A workshop on computer-aided digital circuit design is to be held at SRC's headquarters (State House. High Holborn. London) on 5 January. It will discuss priorities in development of CAD software for digital systems. There will be talks by specialists in various fields. Interested readers should contact Mrs Cynthia Davis at the Rutherford Laboratory.
A Special Interest Group for control engineering is to be set up.
