Problem 1813 - [Regression] Tracks get stuck due to unknown circumstances
Summary: [Regression] Tracks get stuck due to unknown circumstances
Status: RESOLVED INVALID
Alias: None
Product: Geant4
Classification: Unclassified
Component: tracking (show other problems)
Version: 10.2
Hardware: PC Linux
: P5 major
Assignee: Takashi.Sasaki
URL:
Depends on:
Blocks:
 
Reported: 2016-01-14 19:02 CET by o.freyermuth
Modified: 2016-01-17 01:44 CET (History)
1 user (show)

See Also:

Attachments
Verbose dump of stuck track (94.87 KB, text/plain)
2016-01-14 19:02 CET, o.freyermuth 		Details
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this problem.
Description o.freyermuth 2016-01-14 19:02:21 CET 
Created attachment 374 [details]
Verbose dump of stuck track

Dear Geanters, 

since upgrading from 4.10.1p02 to 4.10.2, many of our tracks stop moving and a maximum step number captured in our mc-application (based on geant-vmc) is exceeded. 
This essentially makes 4.10.2 unusable for production for us. 

The geometry in question is a Trap made from BGO-material, surrounded by a Trap from Carbon. Also removing the carbon basket does not fix that problem. 
It happens for various particle types. 

Since I fail to identify the underlying issue clearly, 
I attach a verbose dump of one of these tracks which got stuck. I can provide more information as-needed. 

Thanks for your assistance, 
Oliver
Comment 1 Takashi.Sasaki 2016-01-17 00:41:13 CET 
We have no relation with "geant-vmc" even it might be based on Geant4. Please have a support from them. 

Personally, I cannot use any of my time for an anonymized guy.
Comment 2 o.freyermuth 2016-01-17 01:44:18 CET 
(In reply to comment #1)
> Personally, I cannot use any of my time for an anonymized guy.

Not sure what you mean with anonymized... I'm from Bonn university and have also in the past made bug reports e.g. due to missing decay channels which were fixed just fine. 

I have luckily just about an hour ago identified the root of this problem: It's a regression in the geometry navigation of ROOT which we instructed Geant to use, so it is indeed unrelated to Geant itself, and was hidden by a parallelized upgrade of ROOT and Geant. 

Thanks for your time nevertheless.