Attendees: Michael Baird, Eric Flumerfelt, Lynn Garren, Chris Green, Herb Greenlee, Chris Jones, Tom Junk, Kyle Knoepfel, Rob Kutschke, Marc Paterno, Gianluca Petrillo, Paul Russo
by: Paul Russo
art v2.10.02 bug fix release coming soon, fixes some problems for artdaq with source module product registration. This is not expected to be a problem relevant to current experiments.
As of art v2.09.03 run products created during BeginRun must have an aggregation function. Unfortunately the event range information stored in the file is ambiguous between a full-run range and a partial-run range with no events, which requires the aggregation function to take special care. We will have a discussion of what can be done to lift this ambiguity in the on-disk format. At the moment there is no data corruption.
We had a discussion of whether or not –fast-math is a compiler option which should be turned on. People will investigate, results validation needs to be done to see what the effect is.
NOvA is continuing testing their art v2 upgrade. Two outstanding issues:
- Loading GDML, problem with the magnetic field map trying to load before the geometry service
- Need to changes some internal services to inherit the reconfigure() api
- Possible problems with trigger simulation with art v2, may have root 5 dependencies