Minutes of the 18^th Cosmic Rack Software meeting

Present: Michael, Alexander, Christoph, Ivan, Teppo, Wolfgang, Giacomo, Tapio, Veikko, Martin

Transparencies

Transparencies can be found at the agenda page.

Group status

There are four people actively contributing to the Cosmic Rack Software project, and there are three presentations.

Change of chairperson

Alexander Dierlamm will be the chairperson for the next meetings.

Test beam analysis

Alexander reported about inter-channel coupling. He found a value of 10% for room temperature and 3.5% at -15°Celsius. It is not yet understood from where this large difference comes. Additionally the cluster size is also different between the runs. During the discussion, there was the idea to check the APV modes and Alexander will investigate further in order to have results ready for the Physics TDR.

Alexander also showed residuals and efficiencies for the unaligned CRack geometry, which can be improved by switching to the aligned Geometry, which is not default but can be set via a .orcarc card.

Regarding the CRack software, Alexander stated that there are several drawbacks: ZEBRA output file size limitation, unavailability of chaining input files, and starting at a given event number

Giacomo said that the CARF RootAnalysisFramework could help to solve these issues. He offered help for this.

It was discussed how to upgrade the system to a recent ORCA release, and the conclusion was to wait until ORCA_8_10_0 is released and then to start the migration. Update:Alexander reports that the difference between the two inter-channel coupling values is due to the APV mode: In deconv mode it is 10% and in peak mode it is 3.5%.

Alignment

Tapio presented his work on alignment for the CRack. He coded the CRack alignement hierarchy, ported the alignment interface to the CRack software, and showed first results of an alignment. He reported about the "MonoHit" problem, where the track fit is biased towards the center of the rods in case two stereo hits are recorded, and the track is going in the wrong direction if one of the two stereo hits is missing. He reported that 38% of the tracks have two stereo hits and thus can be used for alignment.

Memory Leaks

Ivan Reid reported about his hunt for memory leaks. The largest memory leak was found in the function Readout::setNoiseInAdcCounts, where a vector was allocated for each event but never deleted. Memory leaks in TBSeedGeneratorFromNFarLayers::seeds() and ConcreteRecTrack were eliminated by using a DaqPersistentReconstructor, with the drawback of a speed loss in data processing. Ivan is now thinking about to correct the DaqReconstructor to regain the processing speed again. Update:Ivan reports that the processing time in fact is the same, with or without the "Persistent" reconstructor.

Action list

Study the difference in interstrip coupling in detail (Alexander). This is necessary input for the Physics TDR.
It was decided to move to ORCA_8_10_0 as soon as it is released (Ivan with help by Giacomo on changing the scripts). This will solve the problem that a lot of code is still copied from user directories.
Try out if it makes sense to use RootAnalysisFramework (Christoph with guidance and support from Giacomo). This would solve the ZEBRA related problems.
Implement the HIP method on the CRACK (Tapio). This is necessary for the Physics TDR.

Next meeting

The next meeting will be announced by Alexander Dierlamm on the Cosmic Rack Mailing list.

Author: Martin.Weber@cern.ch Last modified: Wed Aug 18 14:51:12 2010

Minutes of the 18th Cosmic Rack Software meeting