GSI Forum
GSI Helmholtzzentrum für Schwerionenforschung

Home » PANDA » PandaRoot » General » Simulation crashes after "singular matrix" in digitization macro
Simulation crashes after "singular matrix" in digitization macro [message #9321] Fri, 04 September 2009 14:16 Go to next message
Christian Leitold is currently offline  Christian Leitold
Messages: 35
Registered: August 2009
Location: Stefan Meyer Institute
continuous participant
From: *smi.oeaw.ac.at
Hello,

I have discovered a serious problem in our full simulation: So far, in the last few days I have only done short simulations with a limited number of events, never more than about 50. Now, after doing a little bit more, I was very surprised that the simulation suddenly crashed. I think the reason is in some way to related to an error in the digitization run:

 DIGI EXECUTION *********************
Hit array contains 56 hits
PndEmcMakeCluster, event: 81
***************** PndEmcMakeBump, event: 81 **************
 Digi at (250, 262) was a local max. Energy = 0.822928
 Digi at (248, 281) was a local max. Energy = 0.260981
 Digi at (234, 249) was a local max. Energy = 0.507599
EMC header: fired crystals= 56, digi= 31, Total energy= 2.13144 [GeV], Reconstru
cted clusters= 4, Total energy in clusters= 2.10211 [GeV]
-I- PndSttTrackFinderIdeal: all 114, acc. 1, rec. 1
-I- PndSttMatchTracks: rec. 1, quota 100 % 
FIT xy ********************
hitcounter: 8
Error in <TDecompLU::DecomposeLUCrout>: matrix is singular
-E- pre prefit FAILED 0 24.8538
-I- PndSttHelixHitProducer: 0 track 8 SttHits, 8 HelixHits created.
----------------------------------------

This is followed by fitter exceptions and errors in the Kalman task:

FitterException thrown with excString:
findpca failure
in line: 354 in file: /home/cleitold/pandaroot/fairsoft/trunk/trackrep/GeaneTrac
kRep.cxx
with fatal flag 0
FitterException Info Output

  *** S/R ERPROP   IERR =    2


     *** Error in subr. TRPROP   2  called bysubr. ERPROP

And then I end up with a tcands.root that does not contain a cbmsim tree. The problem occurs as soon as I do a simulation with about 80 to 100 events, even though it might not occur every time I run the simulation, so I suspect there might be a certain per-event probability for it.

Any ideas how to get rid of that kind of problems?

Thanks
Christian

[Updated on: Fri, 04 September 2009 14:16]

Report message to a moderator

Re: Simulation crashes after "singular matrix" in digitization macro [message #9322 is a reply to message #9321] Fri, 04 September 2009 14:45 Go to previous messageGo to next message
StefanoSpataro is currently offline  StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino
first-grade participant

From: *to.infn.it
Can you please write the whole error message?
Re: Simulation crashes after "singular matrix" in digitization macro [message #9323 is a reply to message #9321] Fri, 04 September 2009 14:57 Go to previous messageGo to next message
Christian Leitold is currently offline  Christian Leitold
Messages: 35
Registered: August 2009
Location: Stefan Meyer Institute
continuous participant
From: *smi.oeaw.ac.at
Alright, I have attached all the log files to this message. You might notice the out-of-bond errors in the sim log, but this error also occurs in successful simulation runs, not only in those which crash.
  • Attachment: 1-sim.log
    (Size: 99.41KB, Downloaded 217 times)
  • Attachment: 2-digi.log
    (Size: 103.69KB, Downloaded 211 times)
  • Attachment: 3-reco.log
    (Size: 79.16KB, Downloaded 207 times)
  • Attachment: 4-kalman.log
    (Size: 61.28KB, Downloaded 229 times)
  • Attachment: 5-microwriter.log
    (Size: 16.65KB, Downloaded 239 times)
Re: Simulation crashes after "singular matrix" in digitization macro [message #9324 is a reply to message #9323] Fri, 04 September 2009 15:30 Go to previous messageGo to next message
StefanoSpataro is currently offline  StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino
first-grade participant

From: *to.infn.it
Do you have tracks in the fit file (after kalman)?
Re: Simulation crashes after "singular matrix" in digitization macro [message #9325 is a reply to message #9321] Fri, 04 September 2009 15:37 Go to previous messageGo to next message
Christian Leitold is currently offline  Christian Leitold
Messages: 35
Registered: August 2009
Location: Stefan Meyer Institute
continuous participant
From: *smi.oeaw.ac.at
Well, there is a LheGenTrack branch in the fit file (attached).
Re: Simulation crashes after "singular matrix" in digitization macro [message #9326 is a reply to message #9325] Fri, 04 September 2009 15:41 Go to previous messageGo to next message
StefanoSpataro is currently offline  StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino
first-grade participant

From: *to.infn.it
Then the problem is that makeCand is trying to take information from LheGenTrack, even if it does not know how to do it.

I have sayd that LheGenTrack works with the new code, while makecand works with the good. At the moment you cannot use makeCand together with LheGentrack... you should go back to the svn release of the tutorial, you cannot use makeCand with the newest version before Klaus does fix it.
Re: Simulation crashes after "singular matrix" in digitization macro [message #9327 is a reply to message #9321] Fri, 04 September 2009 16:10 Go to previous messageGo to next message
Christian Leitold is currently offline  Christian Leitold
Messages: 35
Registered: August 2009
Location: Stefan Meyer Institute
continuous participant
From: *smi.oeaw.ac.at
Ok, I went back to the revision 5813 from the tutorial. The original reco macro uses PndTpcLheTrackFinder/Fitter, even though we have STT instead of TPC here. If I turn that off, the Kalman task fails in finding the track array, and does not initialize. So, should I just keep using the PndTpcLheTrackFinder/Fitter?
Re: Simulation crashes after "singular matrix" in digitization macro [message #9328 is a reply to message #9327] Fri, 04 September 2009 16:46 Go to previous messageGo to next message
StefanoSpataro is currently offline  StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino
first-grade participant

From: *to.infn.it
In theory the kalman should work also with the old code. What is going wrong there?
Re: Simulation crashes after "singular matrix" in digitization macro [message #9329 is a reply to message #9328] Fri, 04 September 2009 17:05 Go to previous messageGo to next message
Bertram Kopf is currently offline  Bertram Kopf
Messages: 110
Registered: March 2006
continuous participant
From: *ep1.ruhr-uni-bochum.de
Hi,

Stefano Spataro wrote on Fri, 04 September 2009 16:46

In theory the kalman should work also with the old code. What is going wrong there?


There was already a discussion about this or a similar crash with the tutorial revision. It is probably helpful to look at the threads:
http://forum.gsi.de/index.php?t=tree&th=2535&start=0&rid=0&a mp;S=5c35e1ea3b0a3b4d2958aa245cc31420#page_top

and

http://forum.gsi.de/index.php?t=tree&th=2540&start=0&rid=0&a mp;S=5c35e1ea3b0a3b4d2958aa245cc31420

Cheers,
Bertram.
Re: Simulation crashes after "singular matrix" in digitization macro [message #9330 is a reply to message #9329] Fri, 04 September 2009 17:10 Go to previous messageGo to next message
StefanoSpataro is currently offline  StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino
first-grade participant

From: *to.infn.it
The first problem was coming from using a too new version of svn, and the solution was to sue the tutorial svn version.
I suppose the same solution applies to the 2nd thread, considering that was referrinf to a more recent svn version.
Re: Simulation crashes after "singular matrix" in digitization macro [message #9331 is a reply to message #9330] Fri, 04 September 2009 17:21 Go to previous messageGo to next message
Bertram Kopf is currently offline  Bertram Kopf
Messages: 110
Registered: March 2006
continuous participant
From: *ep1.ruhr-uni-bochum.de
Hi Stefano,

Stefano Spataro wrote on Fri, 04 September 2009 17:10

The first problem was coming from using a too new version of svn, and the solution was to sue the tutorial svn version.
I suppose the same solution applies to the 2nd thread, considering that was referrinf to a more recent svn version.


no, by going back to the tutorial revision the application crashed too. See:
http://forum.gsi.de/index.php?t=tree&goto=9199&rid=431&S=5c3 5e1ea3b0a3b4d2958aa245cc31420#msg_9199

Cheers,
Bertram.
Re: Simulation crashes after "singular matrix" in digitization macro [message #9332 is a reply to message #9331] Fri, 04 September 2009 17:23 Go to previous messageGo to next message
StefanoSpataro is currently offline  StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino
first-grade participant

From: *to.infn.it
Then I don't know,
maybe Klaus could help, I have no diea on what is inside the tutorials macro, I knwo only the tracking code.
Sorry.
Re: Simulation crashes after "singular matrix" in digitization macro [message #9333 is a reply to message #9321] Fri, 04 September 2009 17:37 Go to previous messageGo to next message
Christian Leitold is currently offline  Christian Leitold
Messages: 35
Registered: August 2009
Location: Stefan Meyer Institute
continuous participant
From: *smi.oeaw.ac.at
The problem with the old code was, that there the classes used in the macro were PndTpcLheTrackFinder/Fitter. So my concern was, that this were the "wrong" ones since we are using STT here, ...
Re: Simulation crashes after "singular matrix" in digitization macro [message #9334 is a reply to message #9333] Fri, 04 September 2009 18:45 Go to previous messageGo to next message
StefanoSpataro is currently offline  StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino
first-grade participant

From: *5-87-r.retail.telecomitalia.it
PndTpcLheTrack fitter is just the name of the class, but it works also for STT. What is important is to use the correct setup in the PndLheHitsMaker (SetSttMore(3), SetTpcMode(0)).
Re: Simulation crashes after "singular matrix" in digitization macro [message #9340 is a reply to message #9334] Mon, 07 September 2009 09:11 Go to previous messageGo to next message
Christian Leitold is currently offline  Christian Leitold
Messages: 35
Registered: August 2009
Location: Stefan Meyer Institute
continuous participant
From: *smi.oeaw.ac.at
Stefano Spataro wrote on Fri, 04 September 2009 18:45

PndTpcLheTrack fitter is just the name of the class, but it works also for STT. What is important is to use the correct setup in the PndLheHitsMaker (SetSttMore(3), SetTpcMode(0)).

OK, then everything is set correctly here.
Re: Simulation crashes after "singular matrix" in digitization macro [message #9343 is a reply to message #9332] Mon, 07 September 2009 11:04 Go to previous message
Klaus Götzen is currently offline  Klaus Götzen
Messages: 293
Registered: June 2006
Location: GSI
first-grade participant
From: *gsi.de
Hi,

I'm working on the necessary changes to the tutorials and rho to get it running in the newer revisions and hope to check in something to trunk tomorrow latest (cause I'll be away for a week then).

Don't know whether the crashes are related to wrong tutorial macros or rho, but I plan to get at least Fast Sim running to check the analysis part.

For the configuration of the detector modules I'm unfortunately no expert...

Cheers,

Klaus
Previous Topic: Dipole and transcient B field effect
Next Topic: Full sim: all detected particles have PDG ID zero
Goto Forum:
  


Current Time: Fri Mar 29 05:43:30 CET 2024

Total time taken to generate the page: 0.01500 seconds