Michigan: M. Abolins
UCL: R. Cranfield
CERN: D. Francis
Cracow: K. Korcyl, P. Golonka, A. Kaczmarska
NIKHEF : J. Vermeulen
1. Paper modelling: M. Abolins has found support for the assumption that the distribution of the RoIs is flat in eta and phi and had sent just before the meeting a spreadsheet with numerical information to J. Vermeulen on dsigma/dpT for jets for different pT and for four eta intervals, as calculated with the structure functions of the CTEQ collaboration. J. Vermeulen will produce plots. There are no new developments on the paper model, the new information of F. Wickens on the mapping of the LAr calorimeters has still to be taken into account, but it is expected that the effect on the paper model results will be small.
2. ROS modelling: The model of the RMU (ROS Modular Unit) seems to work, tests have been done with a test program from P. Golonka. The test program sends requests to the RMU model and receives responses. R. Cranfield will contact B. Gorini to obtain the correct values of the parameters to be used. The next step consists of using the RMU model in the testbed model.
3. Data Collection modelling, HLT modelling, at2sim: Multi-packet messages are now supported by the model, but there is a discrepancy between the measurement results and the predictions of the model. Further study is required. K. Korcyl is working on scripts for the complete ATLAS architecture. Variable fragment sizes are now almost supported, histograms of distributions provided by R. Scholte are used. Finally the new code from several developers has to be merged into the repository.
4. Simdaq: J. Vermeulen has implemented a model for a dual-CPU L2PU, with which the number of threads required as function of the utilization of the L2PUs can be studied. The model seems to work, but the first results are not yet completely understood. Also the maximum size of the queues in the switch ports connecting to the L2PUs can be studied as function of the utilization of the L2PUs. The anticipated results possibly can be described in chapter 14 of the TDR.
5. Workplan: In view of the deadline of 25 April for submission of the final drafts of the chapters of the TDR by the chapter editors and the Easter weekend there is not much time left. J. Vermeulen will send K. Korcyl a proposal for the section of chapter 14 on computer modelling next Thursday (17 April) with place holders for material not yet available. In the next meeting, after the deadline, the new draft which should be available at that time is to be discussed and an inventory of the missing parts is to be made.
6. Next meeting: Tuesday 29 April 15.30 h CERN time.
7. AOB: D. Francis mentioned the back-up documents, of which an outline
should be available in the middle of May and should be finished by June. In
the meeting of 29 April this issue (possible back-up documents on modelling) has to be addressed.
J. Vermeulen, 16 April 2003