GSI Forum
GSI Helmholtzzentrum für Schwerionenforschung

Home » PANDA » PandaRoot » Bugs, Fixes, Releases » Emc digi crashes
Emc digi crashes [message #7695] Mon, 19 January 2009 17:29 Go to next message
Ralf Kliemt is currently offline  Ralf Kliemt
Messages: 507
Registered: May 2007
Location: GSI, Darmstadt
first-grade participant

From: 141.30.85*
Hi everyone,

I took the snapshot of the emc codes from rev.4383 to find it broken.

the gdb output gives:
Toggle Spoiler

As it turns out the "cosmetic" changes removed some skin...
Although I went to rev. 4314, which does contain the last change on PndEmcExpClusterSplitter, it seems to work (like for Dima and his post). So I think something is deleted outside of the PndEmcExpClusterSplitter class.

Kind greetings from the flu-infected Dresden,
Ralf.
Re: Emc digi crashes [message #7696 is a reply to message #7695] Mon, 19 January 2009 23:40 Go to previous messageGo to next message
mpeliz is currently offline  mpeliz
Messages: 51
Registered: February 2008
Location: Ruhr-Universitaet Bochum
continuous participant
From: *dip.t-dialin.net
Hi Ralf,

thanks for reporting this error. How can I exactly reproduce this
to have a look to it, e.g. which events should be used and which reco macro should be runned? Is the error observed on all platforms or just at Dresden?

One question for my understanding: Is rev.4383 used for the emc packages only or for all other packages?

Cheers,
Marc
Re: Emc digi crashes [message #7698 is a reply to message #7695] Tue, 20 January 2009 11:30 Go to previous messageGo to next message
M.Babai is currently offline  M.Babai
Messages: 46
Registered: January 2008
Location: Netherlands
continuous participant
From: *KVI.nl
Hi,

I'm running rev. 4383. And the following macro's work fine for me:

full_emc.C
digi_emc.C

They both finish with no error.

Cheers.
Re: Emc digi crashes [message #7700 is a reply to message #7698] Tue, 20 January 2009 20:29 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
Dear all,
indeed, by running 1000 events with the macro

full_emc.C

the application crashes with a same/similar error. I fixed the bug and tried to check in the new code. But it seems that I am not allowed to commit anything. What is the reason?
In addition I would like to get back to Marc's question. It's also not clear to me how one can exactly reproduce the same bug. Which tools are available for this? Is it possible to start the application directly with the event where the crash has been observed? This would help to debug the code in an very easy way.

Thanks in advance for your help and best regards,
Bertram.
Re: Emc digi crashes [message #7701 is a reply to message #7700] Wed, 21 January 2009 08:37 Go to previous messageGo to next message
Florian Uhlig is currently offline  Florian Uhlig
Messages: 424
Registered: May 2007
first-grade participant
From: *gsi.de
Hi

Maybe you have never applied for an account. Only with an account and the correct rights you're able to commit to the repository. Please have a look at
http://fairroot.gsi.de/General/SVN%20account.htm

Ciao

Florian
Re: Emc digi crashes [message #7703 is a reply to message #7695] Wed, 21 January 2009 11:05 Go to previous messageGo to next message
Ralf Kliemt is currently offline  Ralf Kliemt
Messages: 507
Registered: May 2007
Location: GSI, Darmstadt
first-grade participant

From: 141.30.85*
Hi all,

Just to clarify one thing.
You can start a run at a certain interesting (buggy) event. Just use CbmRunAna::Run(Int_t Ev_start, Int_t Ev_end)
(i.e. fRun->Run(26,28); ).

However there are no additional fancy tools for simulation steering or bug condition reproduction there, as some people are used to. Until now it worked well trough communications on the forum here.

Kind greetings from Dresden,

Ralf.

[Updated on: Wed, 21 January 2009 11:07]

Report message to a moderator

Re: Emc digi crashes [message #7704 is a reply to message #7703] Wed, 21 January 2009 11:15 Go to previous messageGo to next message
Florian Uhlig is currently offline  Florian Uhlig
Messages: 424
Registered: May 2007
first-grade participant
From: *gsi.de
Hi Ralf

You can start the event at the interesting event, but if the bug is due to a problem connected to random numbers it must not occor in this event. If the eroor occurs independend of the seed of the random number generator then starting with the interesting event will probably help.

Ciao

Florian
Re: Emc digi crashes [message #7705 is a reply to message #7704] Wed, 21 January 2009 11:20 Go to previous messageGo to next message
StefanoSpataro is currently offline  StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino
first-grade participant

From: *physik.uni-giessen.de
But of course one could set the same seed number...
Re: Emc digi crashes [message #7706 is a reply to message #7705] Wed, 21 January 2009 11:34 Go to previous messageGo to next message
Florian Uhlig is currently offline  Florian Uhlig
Messages: 424
Registered: May 2007
first-grade participant
From: *gsi.de
Hi Stefano

But you will have a different number. Think of a table of 100 numbers. If you start with event 1 you crash becaus you take number 99 from the list. If you start with event 26 you will maybe not crash because you take number 2 from the list.
To reproduce errors if random numbers come into play is sometimes a big problem. Normaly you reproduce the error with the same executable taking the same seed. If your crash happens in event 123456 then youhave bad luck.

Ciao

Florian
Re: Emc digi crashes [message #7707 is a reply to message #7706] Wed, 21 January 2009 11:58 Go to previous messageGo to next message
StefanoSpataro is currently offline  StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino
first-grade participant

From: *physik.uni-giessen.de
Ok,
understood, you are right Smile
Re: Emc digi crashes [message #7708 is a reply to message #7706] Wed, 21 January 2009 12:02 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
Dear Florian and Ralf,

thank a lot for your prompt help.
Concerning the random number issue I would like to ask you again to read the documentation which I posted already few days ago:

http://hepunx.rl.ac.uk/BFROOT/dist/releases/newest/RandControl/doc/Rando ms.ps

In particular section 5.0 "Reproducing single stages" is focused on this problem.

Best regards,
Bertram.
Re: Emc digi crashes [message #7718 is a reply to message #7700] Thu, 22 January 2009 10:45 Go to previous messageGo to next message
StefanoSpataro is currently offline  StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino
first-grade participant

From: *physik.uni-giessen.de
Which was exactly the bug?
Re: Emc digi crashes [message #7719 is a reply to message #7718] Thu, 22 January 2009 11:46 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,
in the destructor of EmcData/PndEmcSharedDigi.ccx the pointer "(PndEmcDigi*) fDigi" will be deleted. This pointer is not owned by this class. Once the pointer has been deleted there and will be still used at another place, the application will crash. Therefore one has to remove this line there. The object "EmcData/PndEmcCluster.*" instead takes care of the deletion of these pointers.
Unfortunately, I did't get write permissions for the PandaRoot repository yet. Therefore it is right now not possible for me to commit this small bug fix.

Cheers,
Bertram.


Re: Emc digi crashes [message #7720 is a reply to message #7719] Thu, 22 January 2009 11:52 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,
I got just now the write permissions and checked in the changes already. Please test revision 4398 (trunk) whether everything is working fine now.

Cheers,
Bertram.
Re: Emc digi crashes [message #7724 is a reply to message #7720] Thu, 22 January 2009 12:49 Go to previous message
StefanoSpataro is currently offline  StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino
first-grade participant

From: *physik.uni-giessen.de
Hi,
I have just tried the new code with an event sample that had the crash, and now it runs till the end.
Therefore, it seems this bug is solved. Probably it was introduced in one of the latest changes, considering that last year I was able to run thousand of events without any problem.
Thanks for the fix.
Previous Topic: PndTpcPadResponseTask Warning ?
Next Topic: PndHypHit build error
Goto Forum:
  


Current Time: Wed Nov 27 23:10:38 CET 2024

Total time taken to generate the page: 0.01082 seconds