Michigan: M. Abolins
CERN: S.Wheeler
UCL: R.Cranfield
Cracow: K.Korcyl, P.Golonka
NIKHEF : J.Vermeulen
1. Paper modelling: J.Vermeulen: new developments: paper model updated with support for various headers (ROD, ROBIn, ROS, raw Ethernet), calculation of number of frames in messages and CPU time associated with I/O dependent on number of frames, the number of ROBIns is no longe required to be an integer multiple of the number of ROS PCs (referred to as "ROBOuts" in the paper model).
2. ROS modelling: Nothing to report.
3. Data Collection modelling, HLT modelling, at2sim: P.Golonka referred to slides produced for the Data-Flow meeting of two days earlier (Monday 3 March), showing that the model for the Data Collection message passing, using raw Ethernet, produces correct results. The results for the DFM model show that this model is also in good shape. Models for the supervisor and the ROB emulator are also OK, the work on the SFI model is not yet finished. J. Vermeulen will send a spreadsheet with results from an analysis of measurements on the SFI done by C. Haeberli to P.Golonka and K.Korcyl. R. Cranfield is looking at sequential processing and asked for information on final aceptance factors. J. Vermeulen will try to provide this information. R. Cranfield also observed discrepancies between numbers provided by J. Vermeulen and numbers found in the first draft of the TDR. Finally R. Cranfield mentioned that all times in at2sim are specified in nanoseconds, microseconds is more convenient.
4. Simdaq: No news.
5. Chiba City: No news.
6. Plans: K.Korcyl will be at CERN the week after the T/DAQ rreview week, A. Kaczmarska probably can take care of a presentation during the week.
7. Next meeting: Wednesday 19 March 16.00 h CERN time.
J. Vermeulen, 18 March 2003