Re: Tracks in the data structure

From: Matthew Worcester (mworcest@hep.uchicago.edu)
Date: Thu Dec 01 2005 - 15:10:02 CST


Hi all,

Sorry I missed the meeting; I was unexpectedly called away. However, it
sounds like some very good discussions were had. As to the new code
(which sounds great and will be very useful), it seems that our limited
goals for the first production release can be accomplished without it, and
not start us down a potentially rocky path if there are unexpected
problems. My vote, although it is not a strong preference, is to push the
new stuff to 0.2.

Matt

On Thu, 1 Dec 2005, Stan Seibert wrote:

> Another item from the meeting:
>
> I have code now which gathers track information during the GEANT4 simulation
> and puts it into the data structure. It's not as detailed as I like (some
> optical photon interactions are not correctly tagged), but it is much better
> than no information, like we have currently. I'm finishing up with some
> helper classes that will allow ROOT macros to load a track list from disk and
> build a data structure in memory that is actually useful for answering
> physics questions. (Who is my parent? Which daughters were produced? Go to
> the track start. Etc, etc.)
>
> Josh said it might be worthwhile to break the "freeze" and add the track
> feature to the code for the 0.1 release. Are there any strong opinions
> either way? It should not be a major issue since it does not affect anyone's
> current macros. Collection of track information is by default off. However,
> if someone adds
>
> /tracking/storeTrajectory 1
>
> to their macro, then GEANT will collect tracks, and Gsim will automatically
> put them into the data structure. So, unlike most of my recent changes, it
> is backward compatible and there should hopefully be few surprises.
> (Prepare for large files if you turn this on. Each IBD event is something
> like 8 MB on my test machine with all 40000 tracks. Clearly we will need to
> extend the pruner with a "cut optical photon" option.)
>
> So, what are people's views on this? Bump it to version 0.2 or delay 0.1?
>
> ---
> Stan Seibert
>



This archive was generated by hypermail 2.1.6 : Mon Dec 05 2005 - 00:01:01 CST