GSI Forum
GSI Helmholtzzentrum für Schwerionenforschung

Home » PANDA » PandaRoot » Bugs, Fixes, Releases » [Inconsistency?] ROOT 6.06 vs. 6.10
[Inconsistency?] ROOT 6.06 vs. 6.10 [message #21900] Tue, 05 December 2017 17:42 Go to next message
Ralf Kliemt is currently offline  Ralf Kliemt
Messages: 501
Registered: May 2007
Location: GSI, Darmstadt
first-grade participant

From: *gsi.de
Dear all,

We found a severe inconsistency between a minor step-up from ROOT 6.06 to 6.10, effecting our simulations. I tested a FairSoft oct.17 release with FairRoot 17.10a. I installed a second version with the older ROOT version, in order to pin the issue down to ROOT.

From the release notes of ROOT I found that the geometry classes were subject to a larger change in ROOT 6.08.
The effect is, that the track length in the STT (see attached pdf, page 2, upper row, middle) is capped at about 1000.

Did anyone see that effect, too?

Cheers!
Ralf
  • Attachment: comparing.pdf
    (Size: 49.63KB, Downloaded 106 times)
Re: [Inconsistency?] ROOT 6.06 vs. 6.10 [message #21902 is a reply to message #21900] Wed, 06 December 2017 08:48 Go to previous messageGo to next message
Radoslaw Karabowicz is currently offline  Radoslaw Karabowicz
Messages: 108
Registered: June 2004
Location: GSI
continuous participant
From: *gsi.de
Extremely interesting. Is the discrepancy also present in the MCTrack branch?
Number of secondaries? energies?
It is very strange that not all detectors are affected.
Re: [Inconsistency?] ROOT 6.06 vs. 6.10 [message #21905 is a reply to message #21900] Wed, 06 December 2017 10:58 Go to previous messageGo to next message
Ralf Kliemt is currently offline  Ralf Kliemt
Messages: 501
Registered: May 2007
Location: GSI, Darmstadt
first-grade participant

From: *gsi.de
Hi Radek,

Yes, we have that already in the MCTrack banch. Good call.
Plots in the pdf below.

This is really worrying.
Re: [Inconsistency?] ROOT 6.06 vs. 6.10 [message #21910 is a reply to message #21900] Wed, 06 December 2017 16:49 Go to previous messageGo to next message
Ralf Kliemt is currently offline  Ralf Kliemt
Messages: 501
Registered: May 2007
Location: GSI, Darmstadt
first-grade participant

From: *gsi.de
It seems ROOT does not like the DIRC support structures. The GEANT stepping aborts very often with
Quote:
************************************************************************ *********************************
* G4Track Information: Particle = mu-, Track ID = 4, Parent ID = 0
************************************************************************ *********************************

Step# X(mm) Y(mm) Z(mm) KinE(MeV) dE(MeV) StepLeng TrackLeng NextVolume ProcName
10002 343 -288 446 1.86e+03 0 0 632 DrcBarSupportS Transportation
10003 343 -288 446 1.86e+03 0 0 632 DrcBarSupportS Transportation
10004 343 -288 446 1.86e+03 0 0 632 DrcBarSupportS Transportation
10005 343 -288 446 1.86e+03 0 0 632 DrcBarSupportS Transportation

I contacted Roman to investigate further. Attached you'll find my comparison plots with and without the whole DIRC.

Ralf
Re: [Inconsistency?] ROOT 6.06 vs. 6.10 [message #21911 is a reply to message #21910] Thu, 07 December 2017 10:58 Go to previous messageGo to next message
Roman Dzhygadlo is currently offline  Roman Dzhygadlo
Messages: 12
Registered: July 2013
Location: GSI
occasional visitor
From: *gsi.de
Hi guys,

the issue is connected to the TGeoTubeSeg volume which has problems with new fairsoft packages (root6.10/geant10.2 ?)
I replaced it with TGeoTube with some volume subtraction. With new geometry it looks like everything is fine again.

cheers,
Roman
Re: [Inconsistency?] ROOT 6.06 vs. 6.10 [message #21960 is a reply to message #21900] Fri, 15 December 2017 10:13 Go to previous messageGo to next message
Ralf Kliemt is currently offline  Ralf Kliemt
Messages: 501
Registered: May 2007
Location: GSI, Darmstadt
first-grade participant

From: *gsi.de
There was another drc geometry update.
The problems become less. That's good.

You can find the comparison between the root versions below.

Ralf
Re: [Inconsistency?] ROOT 6.06 vs. 6.10 [message #21965 is a reply to message #21960] Fri, 15 December 2017 10:28 Go to previous messageGo to next message
Tobias Stockmanns is currently offline  Tobias Stockmanns
Messages: 477
Registered: May 2007
first-grade participant
From: *ikp.kfa-juelich.de
The largest descrepancies seem to be for the straw tube detectors. In Root 6.10 they have much more entries for long track length.
In addition the number of secondaries seems to be less in Root 6.10 (at least the MCTrack.fMotherID does not extend to numbers beyond 800).

What is the suggestion to proceed?

Cheers,

Tobias


Re: [Inconsistency?] ROOT 6.06 vs. 6.10 [message #21970 is a reply to message #21900] Fri, 15 December 2017 10:53 Go to previous message
Ralf Kliemt is currently offline  Ralf Kliemt
Messages: 501
Registered: May 2007
Location: GSI, Darmstadt
first-grade participant

From: *gsi.de
Hi.

Since we're looking at a "fresh" bug by root I favor to make the release with Root 6.06, i.e. the "old" FairSoft/FairRoot combination, which is running well, so far. The bug in root seems to be addressed already in their master branch (thanks Radek!) and we shall simply skip the latest FairSoft release until we get the fix.

In the future we'll be more flexible with the alibuild structure!

Cheers!

Ralf
Previous Topic: PID selection
Next Topic: New Release dec17
Goto Forum:
  


Current Time: Thu Aug 13 02:01:59 CEST 2020

Total time taken to generate the page: 0.02387 seconds