GSI Forum
GSI Helmholtzzentrum für Schwerionenforschung

Home » PANDA » PandaRoot » Bugs, Fixes, Releases » Bug in parameter handling?
Bug in parameter handling? [message #12082] Sun, 19 June 2011 11:50 Go to next message
Felix Boehmer is currently offline  Felix Boehmer
Messages: 149
Registered: May 2007
Location: Munich
first-grade participant

From: *gsi.de
Dear framework experts,

on our bug hunt we came across a strange behavior in the parameter management.

Apparently, when reading the parameters from the ASCII file, everything works, meaning up to the digi macro.
However, when reading in the root parameter files in the reco macro, with EXACTLY the same calls for first and second input, the parameter files are no longer found (e.g. the GAS file seems to be looked up in the wrong path)!

After 2 days of checking and re-checking every line of code in the fitting algorithms to find the problem I now have to conclude that something like a missing geometry must be the problem of the fit as well.


What is going on? Can somebody imagine how this could have been introduced with the latest upgrade? I have been working on 2 weeks old version of the rest-framework before ...

Can somebody confirm this behavior? I know that Dima saw the same thing for the TPC reco macro. Is somebody having likewise problems with the STT reco? If not, then something with the macro has to be off, but I can't find it. And I wonder why it has worked before ...


Cheers

Felix

[Updated on: Sun, 19 June 2011 11:52]

Report message to a moderator

Re: Bug in parameter handling? [message #12084 is a reply to message #12082] Sun, 19 June 2011 12:16 Go to previous messageGo to next message
Dima Melnychuk is currently offline  Dima Melnychuk
Messages: 213
Registered: April 2004
Location: National Centre for Nucle...
first-grade participant
From: *play-internet.pl
Hi Felix,

Talking about problem of GAS file, it's not a problem of parameter handling.
It's a problem that in PndTpcRiemannTrackingTask position of GAS file is taken not from parameters file but hardcoded value without path is used.


Calling PndTpcRiemannTrackingTask::Init at line 219

fnsectors= PndTpcDigiMapper::getInstance()->getPadPlane()->GetNSectors();

PndTpcDigiMapper is called.

If PndTpcDigiMapper was not properly initialised before, its constructor is called
PndTpcDigiMapper::PndTpcDigiMapper(bool autoinit) 

with autoinit=true

and hardcoded file names are used without any path

if(autoinit){
    // objects instantiated here may be replace with the init method!
    fgas= new PndTpcGas("NEON-90_CO2-10_B2_PRES1013.asc",400);
 	   
    //TODO: Get these things from Database!!!
   fgem=new PndTpcGem(5000,           // Gain
                       0.02);          // Spread
 	    
   fzGem=0.;
 	   
   fpadShapes = new PndTpcPadShapePool("2mmPads.dat",
 	                                        *fgem,
                                        0.5, // lookup range
                                        0.02, // Lookup Step
                                        0.01); // LookupIntegrationStep
    
    fpadPlane= new PndTpcPadPlane("padplane.dat", fpadShapes);
   
 }


When tpc reco macro contained PndTpcClusterFinderTask initialised before PndTpcRiemannTrackingTask
there at line 128 PndTpcDigiMapper was properly initialised with parameters from RTDB
PndTpcDigiMapper::getInstance(false)->init(fpadplane,fgem,fgas,fpar->getPadShapes(),fzGem,t0,sf);

and later in PndTpcRiemannTrackingTask this properly initialised instance was called.

Dima
Re: Bug in parameter handling? [message #12085 is a reply to message #12084] Sun, 19 June 2011 12:29 Go to previous messageGo to next message
Felix Boehmer is currently offline  Felix Boehmer
Messages: 149
Registered: May 2007
Location: Munich
first-grade participant

From: *gsi.de
Thank you, Dima!

The RiemannTrackingTask uses the digimapper to get information from the padplane, which is absolute nonsense! The parameter file should be used for this, getting the padplane directly! We are fixing it right now.

I didn't see this problem before since the clusterizing was done in the reco macro for a second time, and then the digimapper instance was still properly valid and initialized...

Thanks for the help!

I'll be keeping you updated!

Cheers

Felix
Re: Bug in parameter handling? [message #12086 is a reply to message #12085] Sun, 19 June 2011 12:48 Go to previous message
Felix Boehmer is currently offline  Felix Boehmer
Messages: 149
Registered: May 2007
Location: Munich
first-grade participant

From: *gsi.de
Hi Dima,

the problem is fixed in the new version of RiemannTrackingTask. However, the problem of broken Kalman fits still remains. I'll keep you updated.


Cheers

Felix
Previous Topic: TPC MVD and GEM correlators
Next Topic: TPC RECO fixed
Goto Forum:
  


Current Time: Sun Dec 01 00:04:28 CET 2024

Total time taken to generate the page: 0.00601 seconds