June 5, 2007 Software
- Review minutes from May 22, 2007 Software meeting
- New FDC Geometry options
- Software tasks
- Action Items
Cebaf Center F324-325
To connect by telephone: 1.) dial:
800-377-8846 : US 888-276-7715 : Canada 302-709-8424 : International
2.) enter participant code: 39527048# (remember the "#")
We may have a VRVS connection depending if the equipment is in good working order. No guarantees.
Attendees: David L. (chair), Elton S., Elke A., Curtis M., Eugene C., Simon T. , Elliott W.
On the phone/VRVS: Mihajlo K., Matt S., Beni, Oojin K., Richard J.
- The first online/DAQ meeting will be next week. Elliott will send out an announcement.
New FDC Geometry options
Following the DC review, the FDC group came up with some alternative designs for the FDC geometry that significantly reduced the amount of material. Some changes were made to hdds by Richard to implement these. Later, the designs were refined to "option1" and "option2" and documented on the wiki. David implemented these in hdds and those versions were used for the studies presented at the meeting.
Mihajlo showed a couple of plots resulting from simulations done with a modified FDC geometry. The one showing the photon detection efficiency vs. angle for the original and modified (option 2) displayed some unexplainable structure. Namely places where the original geometry had better efficiency than the modified geometry which should never happen. The second plot showed very little difference between option1 and option2 of the new design which was also odd since option2 should have had significantly less material in the active region due to the lack of foam backing.
David Showed some plots of momentum resolution as a function of momentum and angle. These indicated roughly the correct trend that was expected, but nothing was shown to indicate the scale of the effect was correct.
It was deicided that Richard should go back and verify the last set of changes David made to hdds are correct. Also, we need to check radiation length plots to indicate that the changes are as we expect.
Improvements to Simulation
Quite a bit of discussion was had on how the output data could be modified to (optionally) include additional information.
Currently, special studies that require information not normally found outside of gustep.F are done mainly using Ntuples created by turning on preprocessor flags and recompiling hdgeant. The normal data output includes information about the particles handed to GEANT at the start of the event, hit information gathered during the event, and "truth tag" information that generally consists of 3D space points and track ids. In addition, there is a flag that can be set in the control.in file (TRAJ) that will write out detailed information for every step of every tracked particle in the simulation. This is useful for only certain types of studies since the output files can become quite large.
- Add option to insert the "birth" and "death" points of every particle in the output stream.
- Add a second flag to determine whether or not the same info for secondaries is also added
- Add some info as to the mechanism in which the final state particles were created in the generator, possibly including info. on intermediate particles not tracked by GEANT.
Since it was not clear to everyone what information is currently available in the output datastream, it was suggested that Richard give an intro to the simulation at a future software meeting. This would likely help new collaborators in particular who are unfamiliar with the current simulation package.
Elke brought up concerns that we may not be making the best use of our time at the software meetings due to a desire to present quick results with little supporting evidence to back it up. Some presentations have consisted mainly of 1 or 2 plots on the wiki without enough background information for a new collaborator to be able to follow the discussion easily. She suggested that more time be spent on the presentations with less of a sense of urgency to get to the final plot that may not be well understood and may also spark long discussions trying to explain what may turn out to be a bug in the code.
It was suggested that senior collaborators need to make it more of a priority for postdocs and students to attend the weekly meetings (like the software meeting) since it is a valuable part of their education and they are often on the front lines in doing simulation studies.
It was also suggested that the meeting minutes be sent out more regularly so people can stay aware of what is being discussed/presented even if they are unable to attend. (David said he'd try harder in this regard.)
- Put updated collaboration list in DocDB and commit list to repository (Zisis)
- Develop system for incorporating alternative b-fields in simulation (Richard and David)
- Look at recent FDC geometry changes (Richard)
- Create an easy way to produce radiation length plots and document it (David)
- Review the digitization scheme(s) we have now and figure out how to make them more coherent (Beni)