Minutes of the Computing meeting held on 26 June 2000

Present: J.Closier, M.Cattaneo, J.Harvey, K.Harrison, M.Frank, M.Needham, T.Cass, G.Gracia, I.Belyaev,
              P.Bartalini, F.Ranjard, E.van Herwijnen, O.Callot, P.Mato, I.Korolko, R.Hierck, F.Fiedler, I.Videau,
              O.Schneider, A.Jacholkowska, G.Corti, J.L, B.Hay 

Minutes : Florence Ranjard

     1. SICBMC v233 (Flr)

The version does not run on NT: it is in a loop in the first event. May be a linking problem. Paolo will look at it with Eric.

It is very important that, as soon as a new version is on test, it is tested on the various platforms used for production.

    2. SICBDST v234r1 (Flr)

• The test version fa now the version number v234r1: some banks are dropped before writing the DST to reduce the
  event size.
• the DST2 problem is not yet solved: Bruce, Andrei and Agnieszka are working on it:
  • the VRHT bank seems to be corrupted  but we don't know yet the cause of the corruption
    • is it due to the pileup procedure, to the VELO code or to an unfortunate combination of both.
    • We hope to have an explanation soon.
  • In the meantime the RICH productiuon should start because RICH does not need DST2.
• the number of tracks stored in ATMC is limited to 10000.
  • we need a mechanism to reduce the number of tracks in ATMC,
  • this mechanism is implemented: all tracks with hits in trackers with E> Ecut   are stored in ATMC
  • to reduce the number of stored tracks it is enough to increase the Ecut from 0.Gev (today) to something reasonable.
    • The physics group should take a decision.

    3. Brunel v1r1 testing status report (M.Cattaneo)

• New version v1r1 to increase the Zebra common block length.
• The following subdetectors have checked the 100 events sample produced with Brunel v1
  • tracking, VELO and L1 are already checked out.
  • Karl has checked the MUON part.
  • no news from RICH and CALO
• Still problem to get new jobid from old database.

    6. A.O.B.

• staging disk space  has been increased by 400Gytes.

• Bookkeeping database
  • RUNSETS table has no room for the 2nd input file when making DST2.
    • that should be forseen in the next version
    • At the same time keep the log file address.

• COCOTIME questionnaire
  • has been sent this week
  • Joel will collect the replies.
  • It is important to mention the peaks of usage, not only a mean average over the year.
  • Tony will give the usage made by LHCb during the past year during the SW week.

===============================================================

next meeting :  3 July 2000 at 14:00 in bldg. 1/1-025           <<<====== EP conference room    

provisional agenda: - Brunel v1 testing status (M.Cattaneo)
                              - SICBDST v234 status (Flr)
                              - A.O.B