Problem 148 - negative currentInteractionLength
Summary: negative currentInteractionLength
Status: RESOLVED FIXED
Alias: None
Product: Examples/Basic and Novice
Classification: Unclassified
Component: N02 (show other problems)
Version: 1.0
Hardware: PC Linux
: P2 critical
Assignee: Michel.Maire
URL:
Depends on:
Blocks:
 
Reported: 2000-08-11 03:14 CEST by Gunter.Folger
Modified: 2007-04-08 18:46 CEST (History)
1 user (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this problem.
Description Gunter.Folger 2000-08-11 03:14:04 CEST
Problem report by :

 Daniel Elvira
 PAT/CD Fermilab
 daniel@fnal.gov
  (630)840-3604

> Dear Gunter,
>           a few days ago, I switched to GEANT4.2.0. (from 4.1.1). Until
> then, my simulation was running smoothly. Now the same user code is
> crashing with the following message:
>
> Process::SubtractNumberOfInteractionLengthLeft()  negative
> currentInteractionLength
>
> *** G4Exception: Aborting execution ***
> Abort (core dumped)
>
>           I looked at G4VProcess.cc, where
> SubtractNumberOfInteractionLengthLeft() is implemented. The way to
> get that negative value is if currentInteractionLength<0.0.
>           By the way, the crash goes away if I remove the EM physics
> (comment the line ConstructEM(); in the PhysicsList.cc file). I inspected
> my PhysicsList.cc file against the one in 4.2.0 N02 example and did not find
> any difference (except I am not implementing all the physics contained
> there). I then tried directly replacing my PhysicsList.cc file by the one
> in the 4.2.0 N02 example and still got the same crash.
>
>           Let me give you some details: I'm shooting a muon beam (a few
> hundred MeV), through liquid hydrogen, in a magnetic field. Ionization,
> Bremsstrahlung and pair production processes are turned on for muons.
> I'm running on a Linux PC.
>
>           Is it obvious to you what's going on? If I change the 4.2.0
> N02 example to shoot muons instead of protons I don't get any crash.
> Please let me know if you need more information, like my PhysicsList.cc
> for example. Right now I'm switching back to 4.1.1, but I would really
> like not to do that. Thanks for your help,
>
>
>                                                       Daniel
>
Comment 1 Michel.Maire 2001-11-26 05:14:59 CET
The problem would have desappeared in Geant4 V3.1, with corrections in
MuEnergyLoss and/or MuBremsstrahlung
Comment 2 scjohnson 2002-04-02 18:58:59 CEST
I've rediscovered this error while running the N02 example and including all of
the N04 physics processes.  In this case a negative currentInteractionLength
error is prolific (almost every event) when shooting 100 GeV protons into an
iron-scintillator calorimeter.  When shooting 1 GeV protons the error is much
less frequent.

I'm running the example on a linux box (7.2) with g++ 2.96.