Present: Giuseppe Cerati, Vito Di Benedetto, Lynn Garren, Krzysztof Genser, Thomas Junk, Gianluca Petrillo, Saba Sehrish, Erica Snider, Hans Wenzel
Remote: Erin Conley, Daniele Gibin, Gleb Sinev, Dorota Stefan, Robert Sulej, Filippo Varanini
Project status report [Erica Snider]
-
new release of
nutools
-
a bug on the newer version (with 3D vectors) of
geo:GeometryCore::WireIDsIntersect() fixed in v06_43_00
-
to our knowledge, it was not yet used; no assessment performed about how wrong the bugged result might be
-
need to hear from some of the experiments before proceeding with new Geant4 and Genie
-
within today, to have the new versions enter the coming release
-
[Thomas Junk] About floating point exception: how do we deal with them once found? [A] "Serial" debugging seems to be the only approach
Conclusions
-
ICARUS is also giving the all clear for updating to new Genie and Geant4
-
will solicit answers from the three remaining experiments about that
LArG4: What’s the plan? Repeal and Replace? [Hans Wenzel]
-
plan to replace
LArG4 local code (often ripped from ancient G4 versions) in:
-
photon management
-
physics lists
-
step limiter
-
there is still room for investigating the different available physics processes, following up on Isaac's work
-
[Erica Snider] how can we choose the new physics lists? it's possible that different beam lines would be simulated with different lists
-
patch in Geant4, reducing string lookup and conversions, will be available in a next Geant4 version
-
[Thomas Junk] we don't want to use step limiter anywhere else than in the active volume
-
[Thomas Junk] DUNE Dual Phase had trouble with too low electron cluster statistics: limiting the step even further than 0.3 mm will exacerbate the issue, in which case they should be aware of it
-
[Erica Snider] step limit setting in GDML might be less flexible than desirable
Optimizing the access to one-to-many associations [Saba Sehrish]
-
use case: the order of associations contains information
-
requirement: in the list of associations one-to-many
O -> M, all associations with each single object must be contiguous
-
added functionality to access not only the range of
M's for one O, but also the O itself
-
working on access to associations with metadata
There are minutes attached to this event.
Show them.