Minutes of the UVic Atlas Meeting held Friday, March 18, 2005
(present at UVic: ML, RS, PP, AAst, MF, TI, WS)
(present at CERN: KV, RMcP)
Meeting Agenda

Rolf told us about his attempt to find random events in the 2002 data. According to the run logbook, most of the 2002 runs have only about 1-2 random events in 20000. However, one muon run was listed with 95 events. Rolf looked at it and printed out the number of events with particular trigger bits set. None of the events had a random trigger bit recorded, and there seems to be no obvious explanation as to `what happened' to the random events. So Rolf decided to go ahead now and resume the data conversion.
Rolf also made contact with Peter Loch, who provided him with a `SimpleNoiseTool' code, and Sven provided Rolf with an ASCII file of noise that he had obtained from his ntuple studies. Rob pointed out that this ASCII file contains the old serial FEB channel numbers, which will have to be converted into software offline identifiers. Michel mentioned that Tamara could also easily produce noise files in TBRootAna based on files with random noise. Rob thought it would be a good idea to pursue that and then compare it to Sven's file. (NB: the output should be in MeV).

Kai has put in a good amount of work for the SUSY group. He was asked this week to contribute to the `low mass' production. 100k events need to be produced, and Kai is already busily generating them. He produced a first run of 5000, ran the reconstruction to produce AODs (really slow: 5hrs on mercury2 for 5k events) and had them validated by the SUSY group. He also started to produce background datasets and sets with `broken channels'.

Rob gave an overview over some topics from the Mini Calorimeter Workshop.
He mentioned the H8 LAr+Tiles meeting, where Calibration issues were discussed. The H8 testbeam had some runs with very low energy e and pi. The beam of these runs however also contained high energy muons, which are difficult to deal with.
The general performance meeting dealt largely with preparations for the Rome meeting - there were lots of nice talks on the agenda.
A new Calorimeter Monitoring group has formed and met during this workshop. It has two conveners Imma Riu (online) and Bob Kehoe (Athena-based monitoring) for both aspects of the monitoring.
The meeting on Data Format and Unpacking discussed the feasibility of the code to deal with the incoming data. For example: The code of the high level ETmiss trigger accesses each and every calorimeter cell to make a decision. Clearly this will take too much time. A faster way will be by creating a vectorial sum to obtain an ET/FEB. The DSP on the ROD will do this calculation.

AOB: Michel brought up the question of `which meetings should we go to?'. With the detector construction nearing completion and the installation under way, the typical weeks of detector meetings (e.g. LAr week) are of less concentrated interest than they were in the past. Some meetings that are of interest don't happen consistently during particular `weeks'. However, it is important to maintain some kind of presence in the different areas we are working on. This makes it a little difficult for the non-CERN resident people to make a decision as to which meetings to go to and which ones to omit. Rob suggested that we should maintain some presence at the LAr weeks and the Hadronic Calibration meetings (which normally happen during software weeks). He also suggested the alternating Etmiss and Jet reconstruction meeting which happens on Wednesday evenings (CERN) / mornings (west coast). Srini would like to combine this into a Calorimeter Performance meeting. When this is close to an Atlas week - it is often put into the Atlas week.
Rob pointed out that for any upcoming travel, please do not book anything before April 3rd (because we don't know what our funds are before the budget split meeting)

Rolf mentioned that he installed release 10.0.0 ok on the fate (under slc3), but had trouble installing it on mercury, where the KitValidation totally failed.

Action Items:

1. Margret and Rolf to cross-check which identifiers are used in their packages.
2. Michel to create noise files with TBRootAna (in MeV).