Linear Collider Forum



Home » Analysis and Reconstruction » EUDET Telescope » A Marlin processor for pedestal calculation
Re: A Marlin processor for pedestal calculation [message #681 is a reply to message #678] Fri, 02 February 2007 09:35 Go to previous messageGo to previous message
gaede
Messages: 233
Registered: January 2004
Location: DESY, Hamburg
Hi Antonio,

the event loop in Marlin right now is really strictly linear, i.e. every event (up to maxRecord) is presented to every processor in the order they are defined in the steering file - similar to what you call 'case 1'. This is the natural choice for normal reconstruction/analysis where you add additional information to the event with every processor.

By using conditions in the execute section one can influence the porogram flow, e.g. only call certain part of reconstruction or analysis if certain pre-conditions are fullfilled. One can also of course have different sets of processors for different events, e.g. calibration and recontruction. But in principle every event will be processed exactly once by one 'path' of processors.

If your calibration will run on the full data sample of the job you can of course specify the set of input files twice and use a flag (isCalibration=true) for the first set and a flag (isCalibration=false) for the second set and so achieve what you want.

However if you want to run a calibration typically on the first N events and then process the full, larger set with these constants one needs to 'rewind' the data set when the calibration is finished and then start from the beginning for the rest of the processing.

I have written some experimental code that would allow to do so.
Please check out cvs tag fg20070202-rewind and look at $MARLIN/src/Testprocessor.cc.
When using the attached steering file the processor loops three times over the first three events for 'calibration' and then the rest (in this case only output processor) is called for all events. If this works for your case of pedestal calibration we could make a relaese of this 'feature' soon.
It migh be usefull also for other groups that want to do 'injob calibration'

Cheers, Frank.

Cheers, Frank.

  • Attachment: test.xml
    (Size: 2.64KB, Downloaded 1837 times)
 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic:Important steps toward the EUDET telescope analysis
Next Topic:ilcinstall: configuration file for Eutelescope
Goto Forum:
  


Current Time: Tue Dec 11 12:01:18 Pacific Standard Time 2018
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.1.
Copyright ©2001-2010 FUDforum Bulletin Board Software