DATE: Tuesday, 23 January, 1996
VENUE: VCPC, Vienna
The members of the Working Group present were as follows:
Apologies were received from Richard Kaufmann of DEC Ltd, Peter Dzwig of LPAC and Russel Winder of UCL.
The minutes of the plenary meeting in Brussels were distributed and agreed without amendment.
HL observed that the draft document on the Web site and that which had been circulated with the minutes of the previous meeting were different. AM confirmed that the version circulated with the minutes was the correct one and that on the Web needed to be updated.
Following discussion on the subject, it was considered useful to submit a paper for the EUROPAR conference in August, 1996.
HL referred to Michael Arentoft's comments in respect of links with manufacturers. HL mentioned that Peter Dzwig was trying to involve Tandem, and went on to say that the Commission was particularly interested in industry rather than purely research.
DC gave a report on the Supercomputing conference in the States and stated that DEC in the States were particularly interested in what EUROPA was doing.
DG spoke about the HPC++ standardisation and went on to say that he felt that EUROPA's approach to the problem addressed gaps which existed in the HPC++ standardisation.
HL introduced the subject of CORBA to the meeting. AM stated that this was something which he was going to look at in due course and was still unsure how EUROPA should approach the use of CORBA. DG said that he found CORBA interesting but that there were intrinsic flaws. DG went on to say, however, that commercially it might be useful, as an interface to CORBA ought to be able to be written fairly easily.
JP suggested that IONA would be worth approaching in respect of CORBA.
HL commented on Agnes Bradier's interest in EUROPA's work and stated that she would make contact to find information about the forthcoming call for Accompanying Measures.
It was suggestd that a priority was to prepare the final version of the paper for the HPCN conference. The preliminary version had been accepted for presentation.
DG raised the subject of the G7 project. He observed, however, that there were no promises of funding, but it was an opportunity to demonstrate what EUROPA could do. DG went on to say that a certain degree of funding might be available. DC stated that if there were a small project rather than a large one, then there might be a possibility of getting funding from this source.
It was felt generally that if there were a commercial or financial application for the standard this would be an opportunity to get international involvement. HL felt that the foreign exchange market might be the area in which to concentrate.
JP stated that a project proposal was going into EPSRC with regard to simulation of genetics. PM mentioned real-time database applications as a possibility. HL stated that the priority had to be to look for funding in the first instance.
HL stated that the Roadmap needed to be updated as most of the objectives had been superseded. She outlined certain areas which need to be revised and suggested that all members should look at their area and earmark the necessary revisions accordingly. HL suggested that the section on promotion should be updated. This was agreed by those present. HL stated that she would update the timetable. AM suggested that amendments should be mailed to the ArchSIG mailing list generally to facillitate discussion and that he would not make alterations without discussion.
JP referred those present to a list of variances which he had prepared previously. JP went on to say that there was still work to be done in ArchSIG and that he would not feel confident preparing an implementation based on what was already in the draft document.
HL reported that Peter Dzwig was being urged to work in this area and that Richard Kaufmann was seen as the best person to chair the Implementations SIG. HL went on to say that NAG were interested and that Clemens Helf of the University of Stuttgart might want to get involved in the Applications SIG.
DC suggested that a paper be put together in conjunction with the HPC++ group for presentation at the EUROPAR conference and this should be discussed with DG.
It was also suggested by IA that a Tools SIG might be developed as the functions inherent to such a group did not come within the remit of any of those mentioned.
Architecture Menu