L1Calo Software Meeting at UCL on 3 August 2001

Before the meeting, Bruce had made a useful compilation of various mails, talks and other notes on this subject over the last few months.

Our discussions ranged quite widely and I didnt take many notes. The following paragraphs summarise the main points. Although the meeting was very useful, these are mainly the outstanding questions.

Initially only four of the new UK ROBins are being produced by Royal Holloway. Future production and how widespread their use is expected to be should be clarified with John Strong and/or Fred Wickens. Eg will they be used in detector integration tests?

New software for the ROBins is required. This may be written by Royal Holloway with help from Gordon (who wrote the version for the existing UK ROBin).

Bob explained the dynamic behaviour of the ROS. His UML diagrams should be useful in clarifying this (in particular the Draft Phase2A document).

If we use the "dumb Slink" (ODIN mezzanine on a PCI carrier board) it is not clear how many Slinks are handled by one ROB IOM and/or how many ROBs can be included in the collapsed ROS on a PC. This should be clarified with Jorgen Petersen.

If we implement our own ROD crate DAQ, we had thought it might be convenient if we could treat each crate of RODs as a ROB and run the EBIF on a separate single node. However although the message passing library allows IOMs to communicate via IP, it is thought that the data collection, handled by a different library, may not support IP. So this suggestion may not be feasible.

Questions about how to feed events into the monitoring system from ROD crate DAQ, or the present status of monitoring events at various places in the ROS software (eg ROB or EBIF) should be addressed to Giovanna Lehmann and/or Louis Tremblet.