GSI Forum: General » new release on the Grid

Home » PANDA » PandaRoot » General » new release on the Grid

Show: Today's Messages :: Polls :: Message Navigator

new release on the Grid [message #12055]

Fri, 17 June 2011 15:39

Gianluigi Boca
Messages: 177
Registered: March 2004

first-grade participant

From: *gsi.de

hallo especially to all the analyzers for the next week meeting;
Stefano is waiting for me for the 'final' version of the cleanup code of the Stt Pattern Recognition.

I promised to be ready by last night, but I am finishing the debug this afternoon.

Please be patient, it should not take much longer, I will notify the OK with another message

Gianluigi

Report message to a moderator

release of the Cleanup code [message #12063 is a reply to message #12055]

Fri, 17 June 2011 20:47

Gianluigi Boca
Messages: 177
Registered: March 2004

first-grade participant

From: *gsi.de

Hallo everybody,
I put in svn the code for the cleanup of the events.
I plan in the next future to improve it but for the time being this is enough.
This code is effective also on 'non-background-mixed' data since it cleans some of the spurious tracks that sometimes the PR of the STT (+Mvd) produces when there are high multiplicity.

I invite the analysers, IF THEY CAN AND THERE IS TIME, to try this new version : just update the directory

/sttmvdtracking

and

/stt

In case they found problems, please report to me ASAP, I'll be in the office this weekend.

Sorry for the fact that I finished only now, and thank you very much in advance to all the analysers and Stefano and for their collaboration.

Cheers Gianluigi

Report message to a moderator

Re: new release on the Grid [message #12068 is a reply to message #12055]

Sat, 18 June 2011 13:34

Dima Melnychuk
Messages: 213
Registered: April 2004
Location: National Centre for Nucle...

first-grade participant

From: *play-internet.pl

Hi Gianluigi,

Running reconstruction for eta_c channel with latest version of STT code in 2 from 11 runs by 1000 events I have the following crash.

===========================================================


The lines below might hint at the cause of the crash.
If they do not help you then please submit a bug report at
http://root.cern.ch/bugs. Please post the ENTIRE stack trace
from above as an attachment in addition to anything else
that might help us fixing this issue.
===========================================================
#5  0x00002ba910db1264 in PndSttMvdTracking::AssociateFoundTrackstoMCquater (
    this=0x2ba91572d010, keepit=0x7fffb9f2c390, info=0x7fffb9ea9290,
    Ox=0x7fffb9f2b430, Oy=0x7fffb9f2b2f0, R=0x7fffb9f2b1b0,
    X1=0x7fffb9f2a7b0, Y1=0x7fffb9f2a670, X2=0x7fffb9f2a530,
    Y2=0x7fffb9f2a3f0, X3=0x7fffb9f297b0, Y3=0x7fffb9f29bf0,
    nTracksFoundSoFar=44, nHitsinTrack=0x2ba91572d5a0,
    ListHitsinTrack=0x2ba915759920, nSkewHitsinTrack=0x2ba91572d5f0,
    ListSkewHitsinTrack=0x2ba915771020, nPixelHitsinTrack=0x2ba91572d640,
    ListPixelHitsinTrack=0x2ba91572d6e0, FromPixeltoMCTrack=0x7fffb9be9030,
    nStripHitsinTrack=0x2ba91572d690, ListStripHitsinTrack=0x2ba915737320,
    FromStriptoMCTrack=0x7fffb9be8e10, daTrackFoundaTrackMC=0x7fffb9f2c150)
    at /d/panda02/dmelnych/pandaroot/pandaroot/sttmvdtracking/PndSttMvdTracking.cxx:6372
#6  0x00002ba910dd9862 in PndSttMvdTracking::Exec (this=0x2ba91572d010,
    opt=0x2ba90a8caf00 "")
    at /d/panda02/dmelnych/pandaroot/pandaroot/sttmvdtracking/PndSttMvdTracking.cxx:2573
#7  0x00002ba900bd6bb0 in TTask::ExecuteTasks ()
   from /misc/cbmsoft/Lenny64/may11/fairsoft/tools/root/lib/libCore.so.5.29
#8  0x00002ba900bd5ecb in TTask::ExecuteTask ()
   from /misc/cbmsoft/Lenny64/may11/fairsoft/tools/root/lib/libCore.so.5.29
#9  0x00002ba90a81654b in FairRunAna::Run (this=0x8fb3d0, Ev_start=0,
    Ev_end=1000)
    at /d/panda02/dmelnych/pandaroot/pandaroot/base/FairRunAna.cxx:272
===========================================================

Dima

Report message to a moderator

Re: new release on the Grid [message #12070 is a reply to message #12068]

Sat, 18 June 2011 17:12

Gianluigi Boca
Messages: 177
Registered: March 2004

first-grade participant

From: *gsi.de

Hi Dima, thanks for your message. I am looking into it.
Preliminary question : when you call the stt-mvd task you should use the
following constructor :

PndSttMvdTracking * SttMvdTracking = new PndSttMvdTracking();

or you can use also :

PndSttMvdTracking * SttMvdTracking = new PndSttMvdTracking(0);

or you can use also

PndSttMvdTracking *SttMvdTracking=new PndSttMvdTracking(0,false,false);

Is that what you do?
Gianluigi

Report message to a moderator

Re: new release on the Grid [message #12071 is a reply to message #12070]

Sat, 18 June 2011 20:12

StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino

first-grade participant

From: *8-87-r.retail.telecomitalia.it

Just check the macros in macro/run/tdrct/eta_c. Those are the ones Dima is using.

Report message to a moderator

Re: new release on the Grid [message #12073 is a reply to message #12055]

Sat, 18 June 2011 21:48

Dima Melnychuk
Messages: 213
Registered: April 2004
Location: National Centre for Nucle...

first-grade participant

From: *play-internet.pl

Hi Gianluigi,

Indeed the constructor

PndSttMvdTracking *  SttMvdTracking = new PndSttMvdTracking(0, false, true);

was used.
And it was used all the time before and before the recent update it didn't cause any problem.

As I understood the data produced with the last flag set true are valid. Only in addition track assosiation with MC tracks is performed. But I suppose that macro which produce data on GRID have the same constructor and similar problem is possible.

Dima

Report message to a moderator

Re: new release on the Grid [message #12074 is a reply to message #12073]

Sat, 18 June 2011 22:03

Johan Messchendorp
Messages: 693
Registered: April 2007
Location: University of Groningen

first-grade participant

From: *xs4all.nl

Hi,

Indeed, the data that are now produced on the Grid faces the same problems. There are a large number of jobs (30-40%) that do not pass the validation, e.g. crashed during reconstruction. In any case, we have now about 100k of signal events per physics channel produced on the Grid that didn't crash (only STT).

Greets,

Johan.

Report message to a moderator

Re: new release on the Grid [message #12075 is a reply to message #12073]

Sat, 18 June 2011 22:07

Gianluigi Boca
Messages: 177
Registered: March 2004

first-grade participant

From: *gsi.de

Dima Melnychuk wrote on Sat, 18 June 2011 21:48

Hi Gianluigi,

Indeed the constructor

PndSttMvdTracking *  SttMvdTracking = new PndSttMvdTracking(0, false, true);

Indeed, the last value in the constructor is set true when one
wants the comparison with MC.

Of course there is a bug somewhere in the code (that I cannot reproduce for the moment) but also from the backtrace I noticed it is in the non-essential part of the class (the comparison with MC, indeed).

So given the lack of time for now please set 0, false, false
in the constructor, later on I will fix the problem.

Thanks a lot
Gianluigi

Report message to a moderator

Re: new release on the Grid [message #12076 is a reply to message #12074]

Sat, 18 June 2011 22:12

Gianluigi Boca
Messages: 177
Registered: March 2004

first-grade participant

From: *gsi.de

Johan Messchendorp wrote on Sat, 18 June 2011 22:03

Yes, as I said to Dima, the error is in a part of the code which is not relevant. I don't think any of the results would be affected by this.
By setting 0,false,false
in the constructor the crash should go away.

So please let's try to do this for the moment. Later in a less frantic period I will fix it
thanks Gianluigi

Report message to a moderator