Choose timezone
Your profile timezone:
Geant4 Hadronic Working Group Meeting
→
Europe/Zurich
14/4-030 (CERN)
Description
Monthly Geant4 hadronic group meeting
To join by phone, dial-in numbers are: 71400 (CERN), +41227671400 (outside CERN) + Vidyo meeting Numeric Extension: 109341461
Minutes of the hadronic group meeting on 08-Dec-2016, dedicated to the changes of hadronic model parameters.
- First of all, a reminder of the agreement that we reached on this subject,
which was approved by the Geant4 Steering Board on 19-Dec-2014
"
In selected cases, to be evaluated by the Geant4 hadronic group,
Geant4 supports changes of hadronic model parameters with partners.
- For "support" we mean documenting which parameters can be varied
and within what range, debugging in the case of problems, and
performing some physics validation.
- For "partners" we mean Geant4 users willing to help the developers
in this effort, to provide feedback and complete information, and
to perform physics validation requested by Geant4.
- In those cases in which Geant4 does not support changes of hadronic
model parameters, users are anyhow allowed to try do it themselves
given that code is open, or even seek the help of individual Geant4
collaborators, but no support by the Geant4 collaboration should be
expected.
- For any use of modified models, each public presentation and
publication of the results must clearly state that a modified version
of Geant4 has been used, and document the parameters changed and their
values.
- The Geant4 Steering Board will be informed of all the decisions taken
by the Hadronic Working Group on this regard.
"
- The meeting was dedicated to the technical aspects of changing hadronic
model parameters. However, a few comments and remarks were made on what
type of model parameter changes are expected. We agreed that we will
provide the possibility for users to change some selected model parameters,
but not "1-sigma" variation intervals. These will be, eventually, determine
by the users, e.g. the neutrino community. Some of the Geant4 developers
have anyhow expressed their doubts of the possibility for users to
determine meaningful "1-sigma" variation intervals for detector simulation,
although this practice exists for event generators.
- It is clear that providing UI (User Interface) commands for changing model
parameters would be the easiest solution for users. Implementing this
correctly for both sequential and multi-threaded executions of Geant4
applications is not trivial, but a working solution is provided by Geant4
electromagnetic physics. However, a consensus was not reached on providing
UI commands in hadronic physics. The main worry is that it would make it
possible for inexperienced or naive users to easily change hadronic model
parameters in reference physics lists, which can lead to crashes or,
even worse, unphysical results.
Therefore, we strongly discourage the use of UI commands to change hadronic
model parameters.
- The solution we agree on in order to allow users to change parameters
of hadronic models is via C++ "Set" methods.
How to call properly these methods, at the level of Geant4 applications,
for both sequential and multi-threaded executions, will be provided as
a set of recipes to the interested users upon request from their side.
This allows to have a first "filter" of the users, and to put us in
contact with those who have good scientific arguments to change parameters
of hadronic models.
- For the hadronic developers, the task is to select and document those
model parameters that are meaningful to be changed. In particular, the
interval of physical validity for each of these parameters should be
carefully specified.
It is a good practice, but not a mandatory request, to collect all the
parameters that can be changed for a hadronic physics model into one
class. An example of this is provided by Precompound/de-excitation,
as illustrated by Julia's presentation.
- Regarding the C++ interface to access the hadronic model parameters
that are allowed to be changed by users, in particular which methods
(e.g. "Set...", "Get...", "Dump", "GetDefaults") and which signatures
should we recommend, we agreed to form a small group of interested
developers that should study it and come up with a proposal early
next year. Tatsumi Koi will coordinate this small group of developers:
please send an email to Tatsumi (tkoi@slac.stanford.edu) for those
who want to join.
There are minutes attached to this event.
Show them.