Ashok
reported that he has given Warren Shaw and Ian Gable a crash course
on how to work with Athena. They now know how to setup their
environment and run a job. They have done this at CERN.
The next step is to try and do a compile and run on our local machines.
At the moment there seems to be two problems: Some files seem to
be missing and there is some problem with read/write accesses.
It is unclear what the exact source of the problems is, but
Ashok is confident that the problem can be solved in a few days.
David Quarrie and Rob recommended that Athena release 8.0.1
might be a better release than 8.0.0.
Ian Gable pointed out that the local compile might be hampered by
an environment variable that isn't set right. If that is true,
the fix could be very simple.
Ashok also pointed out a very important piece of information:
When accessing CERN via the ssh protocol, one has to use ssh version 1
and not version 2, like everywhere else. Note however that if you
actually log on to CERN via ssh, then version 2 is the one to use....
Warren an Ian
spent most of the week with Ashok, learning about Athena as well
as getting ready for CERN (Warren).
Margret
untangled the G4 HEC geometry of the LArG4TB standalone version
and gave a
presentation (pdf)
on that. She cautions that the philosophy and
naming of the standalone geometry will not be the same in the Athena code, i.e.
it is a little outdated. But it sheds some light on how the detector is
implemented and what is involved in getting at the root of how to
obtain the energy deposited in LAr vs. energy deposited in the Cu plates.
Margret wants to move to the Athena version as soon as possible, when things
are up and running at UVic, so she can keep up with the latest developments
on this package.
Tamara
continued her work on the noise subtraction by looking at the signal/noise ratio
for muon runs. If she rejects cells with large signal/noise (replacing
them with a signal randomly sampled from a gaussian of width equal to
the noise estime of this cell obtained from a different muon run), she obtains a very
nice Gaussian for the cluster noise distribution. (The large S/N cells, if kept, would contribute
to a tail in this distrubution). The width of the Gaussian is an estimate of the noise
of the cluster of cells that is used for Tamara's analysis. The sigma
of this Gaussian, in the case of a fixed cluster, is
about 2.1 GeV. Tamara then showed plots of the resolution as a function of beam energy with
the noise subtracted in quadrature. Her task is now to determine the constant term and
then to do the whole analysis again for event-by-event clusters rather than fixed size clusters
she was using so far.
Tayfun
continued his investigation on why he sees a slope when he plots x(mwpc)-x(cal) vs. x(cal).
He looked at an electron run that hits the calorimeter at (x,y)=(0,0). When he
looks at the difference between the calorimeter and the beam chamber measurement, he
sees a small offset in x (about 0.3cm) and a large offset in y (about 3.7cm).
He applies this offset to all of his measurements as a correction and finds
that it eliminates the slope.
This reminded Michel of a talk given by Sven June last year on exactly the same thing.
Sven also determined an average offset in x and y. Michel would like Tayfun's
final numbers on the offset, so he can include them into TBRootAna.
Tayfun has now started to write his thesis.
Michel
had some final points for discussion: In Athena versions 7.8.0 and higher,
the code no longer deals with unconnected channels. See this
email for details.
The conclusion among the group was to at most do a very quick fix or even do nothing to
deal with it, but get Athena version 7.6. to UVic, make sure it runs
and freeze it, so that Ntuples can be reproduced locally.
-----------------------------------------------
Reminder:
The next testbeam meeting will not be on Friday 11:00, but on Monday, May 17th at 10:00