Jump Over Left Menu
Minutes of the meeting held on 16/ 3/1973 at State House, London
- Professor H Bondi (Chair)
- Professor D W Barron
- Professor P G Burke
- Mr D W Davies
- Dr G E Thomas
- Professor K W Allen
- Professor B H Bransden
- Professor G D S MacLellan
- Mr A E Seddon (NERC)
- Mr B R Taylor (DTI)
- Miss V E M Bowell (Engineering Division, SRC)
- Mrs J O Patton (Science Division, SRC)
- Dr J Howlett (Director of ACL)
- Mr J E Hailstone (Head, User Services Group, ACL)
- Dr V Reddish
- Mr E C Appleyard (Computer Board)
1. Dr Howlett voiced the congratulations of the Committee on the New Year's Honour awarded to the Chairman.
2. Apologies were received from Dr V C Reddish of Royal Observatory Edinburgh and HI' E C Appleyard of the Computer Board.
3. Dr Howlett said that with the approval of the Chairman he had invited Mr Hailstone, Head of the User Services Group at ACL, to attend the meeting because of his knowledge of the Allocation procedures.
2 Minutes of Last Meeting
4. The minutes of the first meeting were confirmed as amended.
3 Matters Arising
5. Mr Seddon referred to paragraph 6 and said that he had brought some notes about the arrangements between NERC and SRC. He would give these to the Secretary to circulate.
6. Mr Fossey agreed to discover if copies of the report by Dr Francis mentioned in paragraph 7 could now be distributed.
7. Dr Howlett commented on the request minuted in paragraph 9. The Council had decided upon the allocation of 20% of the IBM 370/195 to the Atlas Computer Laboratory. He could not seek a ruling on a hypothesis: the facilities would have to be demonstrably insufficient before Council could be asked to re-examine the current allocation of IBM 370/195 time between ACL and RHEL.
8. Mrs Paton said that the financial delegations suggested in paragraph 10 were not in line with those in use within SRC. The Committee had delegated authority up to Â£20,000. Thus items costing up to Â£20,000 could be approved by the Committee and purchased on the authority of the proper official within SRC. Above this level, items costing up to Â£50,000 had to be submitted to the Science Board, up to Â£100,000 to the Council, and above Â£100,000 to the Department of Education and Science (DES). The choice of these levels ensured that the Committees, Boards, and Council had, as they wished, a proper understanding of the number and volume of proposals brought forward.
9. Mrs Paton said that the Council would expect that most proposals should be discussed at Committee meetings. However, urgent proposals might be dealt with by correspondence, since the Committee did not meet very frequently.
4 Progress Report - ACC/73/1
10. Dr Howlett said that there would be a closing ceremony for Atlas on 30 March 1973. Invitations had been issued to a number of people, including all the past and present members of the Atlas Computer Committee.
11. Dr Howlett said that the Atlas computer had been working very satisfactorily except for some difficulties with the tape system. It was hoped that all outstanding work would be cleared, except that the users of the MVC system would be embarrassed. There would be some hiatus on the transfer of this work to the 1906A.
12. Dr Howlett continued by saying that the 1906A was being run on three shifts. An important change in the Operating System was expected soon with the introduction of Mark 7 of GEORGE 4. This would lead to the removal of many of the inefficiencies of the present Mark 6. The overall performance was comparable to that on other makes of machine. The filestore was an attractive feature of the system, but the traffic: entailed in its use was a significant factor in lowering the efficient utilisation of the machine.
13. He said that the 370/195 service was going exceptionally we1l. Recent usage figures had started to show signs of pressure from the Nuclear Physics work upon the work submitted from ACL. The 195 was serving a very mixed computing load, so that longer jobs such as those from ACL were starting to be relegated to night and weekend working. The trend could only produce users less well satisfied with the service.
14. Dr Howlett said that the new integrated graphics software system should come into operation at the start of April 1973. Generally statistics on the use of the SD4020 microfilm recorder showed that more than one million frames of film were produced in a year, and that the graphics systems were still in highly active use.
15. The Chemistry Committee had asked ACL to house a microdensitometer and to provide a service for its use. The machine had been installed, and was now in use in some cases with the approval of the Committee by users outside the crystallography field. It illustrated the flexibility of this kind of equipment.
16. Dr Howlett said that the Laboratory had now received two of the five GE2050 terminals, and that a later paper would discuss their distribution. He wished to make it clear that although the choice had fallen on the GE2050 on this occasion it did not follow that other terminals might not be more appropriate in other circumstances.
16. He was pleased to report that the cost of the new building had been contained within the initial estimate by dint of careful attention to costs. The chairman commented that this had been achieved only at the expense of a lot of hard work.
17. Dr Howlett said that, with the imminent closure of the Atlas computer, the usual tables of Atlas usage had not been produced but would be issued after the closure. His report had included tables on the usage of the lCL 1906A and the IBH 370/195. He would welcome comments from the Committee on the kind of detail they wished to see.
18. Dr Thomas said that the Laboratory had spent a considerable amount of effort on quantum chemistry programs. He thougt this had been well received and would like the Laboratory to give some thought to more publicity for this. Dr Howlett mentioned that a joint conference was being arranged with the Chemical Society in the Summer of 1974 and that this would be an International meeting. Dr Thomas said that he was thinking in terms of a meeting rather sooner than this.
19. Mr Taylor enquired about the problems encountered by the Laboratory in mounting ASCOP on the 1906A. He was assured by Mr Fossey that the 1900 version of ASCOP had not previously been mounted under GEORGE 3. The system was now beginning to work and it looked as though a service could he introduced fairly shortly. The National Computing Centre had furnished all the help necessary.
20. Mr Seddon welcomed the assistance that had been given by staff at the Laboratory for work undertaken on behalf of NERC by the Institute of Geological Sciences. Much of the work was of a pioneering nature and was, as a result, highly valued.
5 Allocations Procedures Working Group
21. Nr Hailstone said that the APWG was set up in July 1972 by Mr Jolliffe. Its first aim was to present to the ACC a set of procedures for the allocation of time, together with provisional allocations between the various groups and divisions of SRC. The working group had chosen its own terms of reference and they now submitted a paper containing its recommendations together with an indication of the issues which needed attention in the future. Mr Hailstone then summarised the various documents attached to the report of the APWG.
22. Mr Hailstone drew attention to a problem associated with costing. To avoid the conflicts due to the disparity of the cost of time on the lCL 1906A versus that of the IBM 370/195, the group had decided to work in terms of computer power and had chosen to treat the 195 as 10 times more powerful than the 1906A. As a result, the provisional allocation between the Boards was shown in terms of power in Table 3 of the main report. Since the Laboratory had a 20% allocation, it was proposed to overcommit time to the various Boards in the expectation that on average, the time taken up would be less than the total commitment. The ACL allocation would be used as a buffer to prevent serious problems in the unlikely event of peak loading. He then commented on the current state of commitment.
23. With regard to the future, the paper reviewed various topics needing review in the future. Perhaps the most important of these was that a careful watch should be kept on the commitment of time between the Boards to ensure that the allocations in future took account of trends in scientific research.
24. The Committee drew attention to the fact that already the allocation to the ASR Division had been overcommitted. It felt that the group should review the distribution. Professor Burke thought it likely that there would be a large increase in demand from Engineering. Mr Hailstone said that interest in three-dimensional fluid flow indicated a likely growth in Engineering demand which could cause pressure on the allocations by the end of the year.
25. Dr Thomas said that the Committee would find it valuable to see figures for usage and forward demand against the allocations for the Boards. He thought that these comparative figures might be very revealing and might "show how much need there would be to alter the allocation between the Boards. Mr Hailstone agreed to produce these figures.
26. Professor Bondi thought that the ACC was not in a position to provide continuous review of the allocations of the Atlas Computer resources. He thought that an Allocations Working Group (AWG) was required. Mrs Paton thought that the working group should be set up by the Committee to furnish it with advice on the allocations of the various resources at the disposal of the Laboratory. The group would attempt to resolve conflict over the allocation of resources between the various Boards. The channel for dealing with unresolved conflict was clear; the matter would go ultimately to the Council to be settled. The Committee agreed to the formation of the AWG.
27. The Committee discussed the problem of the disparity of the costs in using the two machines. Some members felt that to obtain proper management control the computer resources should be accurately costed, whereas others felt that the criteria for choosing to use one machine rather than the other might not be based simply on the grounds of cost. Mrs Paton pointed put that SRC were considering some method of charging and that the Committee should bear this in mind before coming to any decision. The Committee, however, highlighted the difficulty that, without a proper charging mechanism, the Council's Boards and their specialist committees could approve work for the Atlas Laboratory in excess of their allocations and might not have a convenient yardstick for assigning priorities between the projects.
28. Mr Seddon said that the allocation of 5% of the ACL resources was an initial allocation. NERC's requirement~ might prove to be larger than were envisaged at the outset and he wished to know how this might be satisfied. Professor Bondi said that this was a matter for central discussion between NERC and SRC but he thought it appropriate for NERC to have a representative on the AWG to indicate changes in demand. Mrs Paton commented that the guaranteed allocation to NERC was separated from the time allocated to the Boards and reviewed by the Allocations Working Group. The Committee agreed that NERC representation on the AWG was unnecessary and suggested that the AWG formed its terms of reference to undertake the work itemised in the report.
29. Dr, Howlett asked if the Committee would approve the distribution of terminals proposed in ACC 73/2. The Committee agreed to the distribution subject to the overall recommendation of the SRC Secretary's Computer Review Panel. The Committee agreed that the AWG should consider the allocation of terminals and recommend the placement for a period normally not exceeding a year. Where the AWG felt that provision of a terminal was essential to fulfil the terms of a grant of longer guarantee than one year, it could recommend the provision up to a maximum of 3 years.
6 Enhancements to the ICL 1906A System - ACC/73/2 and ACC/73/3
30, Dr Howlett said that the Science Board had turned down in October 1972 the enhancements requested in the paper ACC 72/4. At about that time it became c1ear that ICL wished to replace the Fixed Disc Store on the 1906A by an EDS 60 system. As a result he had returned to the Science Board with a request for the provision of extra magnetic tapes in December 1972, and the Science Board had agreed to this. The Laboratory had now an opportunity to examine the consequences of replacing the FDS with an EDS 60 system and, as a result, were now asking that the EDS 30 system at the Laboratory be enhanced by the provision of two further spindles costing Â£15,OOO.
31. The Committee approved the expenditure of Â£15,000 on the purchase of two extra EDS 30 disc drives.
7 New Microfilm Recorder - ACC/73/5
32. Dr Howlett said that the demand for use of the SD4020 had been considerable, but clearly fluctuated. The figures shown in the report were possibly affected by the erratic performance of the 4020. Overall the figures showed that there was no shortage of demand and that there was a very distinct increase in the use of 35 mm film. He explained that there were a number of large users who had been making extensive use of the 4020 for the manufacture of films as an integral part of their work. The Committee had asked him to investigate the refurbishing of the 4020. A factory refurbished 4020 would cost Â£20,000. It would have none of the features of the newer machine which had been highly thought of in the previous paper. It was made of out-of-date circuitry and would inevitably lead to awkward maintenance problems.
33. Dr Howlett said that the possibility of getting a very much more precise and efficient piece of recording equipment would make the graphical presentation of results very much more attractive to large users. The FR80 was not simply a recorder, it had a built-in processor whose use could have a dramatic effect on the amount of main computer time used. By way of illustration, Mr Hailstone reported that a run to produce a fully formatted tape for use on the 4020 took 82 seconds, Whereas a run to produce a tape with the FR80 format took only 48 seconds.
34. Professor Bondi said that, although he believed in the greater use of graphical forms of output, he was not convinced that the highest accuracy was required for this work. For this reason he felt that there was insufficient demand to justify the spending of Â£200,000 at this stage and he asked the Committee not to proceed with the proposal. Professor Allen said that no proper scientific case had been made for the new microfilm recorder, although he believed that there was an indisputable case. He thought that refurbishing the 4020 at a cost of Â£20,000 would be a complete waste of money. He wished to know how the Committee could assess demand.
35. Professor Bondi summed up the discussion by saying that the Committee looked with interest at equipment which enhanced the capability of the Laboratory to satisfy its users. In the case of the FR80, much clearer and more tangible evidence of need and interest on behalf of both users and the Boards would be required before the Committee would ask the Council to equip the Laboratory with the FR80.
8 Network Panel Report - ACC/72/6
36. Dr Davies summarised his report verbally and promised to provide a copy to Mr Fossey for circulation to the Committee.
9 Any Other Business
37. Dr Howlett said that there was a need to dig a trench between the Rutherford Laboratory and ACL to provide additional cabling between the Laboratories. He was therefore seeking the authority of the Committee to spend Committee Â£6,000 on this necessary work. The Committee approved this.
38. Professor Bondi reported that the Council had set up a Board to find a successor to Dr Howlett as Director of the Atlas Computer Laboratory. The Board had failed to find a suitable candidate and in the circumstances the Council had asked Dr Howlett to remain as Director of the Laboratory, and he was happy to say that Dr Howlett had agreed to do so for about a year. A consequence of this was that the Council was likely to set up a small panel to review the future work of the Laboratory.
39. Professor Bondi reported that ICL had approached the Council with an offer to install at ACL in the Summer of 1974, the first of their New Range P4 machines. A letter containing the proposal had been delivered to Dr Howlett shortly before the meeting. Placing the computer at ACL involved the SRC in providing considerable help to the British Computer Industry and this was a major policy matter for SRC to decide. He felt it best not to discuss the matter further at the Committee.
10 Date of Next Meeting
40. Mr Fossey said that the date of the next meeting was still fixed for 24 May] 973. Although the date should be kept free, the Chairman wished members to note that it might be advisable to defer the meeting.