Minutes of the UVic Atlas Meeting held Friday, March 11,
2005
(present in person: ML, RS, Alan, WS, PP, TI, MF; (RK, Asok: part time))
(present by phone: KV, RMcP)
Meeting Agenda
Rob gave an overview over the situation of the current athena releases.
For those of us who do testbeam analysis, release 9.1.2 is the one
to use. It is stable and it includes Rolf's Pool converter.
However it does not include LArROD-02-08-07, which is necessary
for the FCAL energies to be correct. This patch will be included
in release 10.0.1 (to be built on March 23).
Release 10.0.0 is currently being built. It might be an interesting
release to install at UVic, as it has some of Sven's latest
contributions in it. (Rolf will try to install it.)
However, none of the series 10 releases are going to be very
stable - the idea is to copy a working nightly roughly once every
3 weeks into a 10.x.0 version. Release 11, which is to come in September,
is supposed to be made stable again.
Kai presented a little study of the number of TopoClusters he found
as a function of cluster-cuts. He looked at 180 GeV pions in the H8 testbeam
and imposed 90 different cuts. His slide shows the average number
of found clusters vs. the cut applied. Four cut-groups can be seen.
They correspond seed cut values of 2,3,4,5. (The seed-cut is given in
terms of signal/noise.)
A discussion arose over how to tackle cluster building. Rob's feeling
was that we will likely end up with an approach that has a similar philosophy
as the one used in OPAL: First loose clustering requirements are applied and
then cluster-quality selections pick out the final choice.
Margret gave an update on her 2002 TB MC. She quickly gave an overview
over the contents of the generic Atlas CBNT ntuples that can be created
when producing and ESD file from Digits and when producing an AOD from an
ESD file. She concluded that for testbeam purposes it doesn't make
much sense to look at the AOD level, so she focused on the ntuple
produced when creating an ESD.
She has created ntuples at 3 energies (50, 100, 200 GeV) for electrons
and pions each. She showed plots of total summed EMEC+HEC energy
for all energies and found that while the energies for the EMEC are
systematically reconstructed a little too high, the energies for the
HEC are systematically too low. Margret will now look at her generated
files in more detail.
Rolf started the ESD production of the TB data on Monday here at UVic
on the muse cluster. When looking at the ESD data, he found that he could
not see any events with random triggers. It was unclear at that point, whether
the random events are not included in the ESD or whether the analysis
program Rolf used jumps over those events. Until the source of this
is found, the ESD production is on hold. Once Rolf has made sure that
random events are indeed included, the production can resume again.
Rolf estimates that the conversion from bytestream to ESD for the complete
dataset will take about 1 to 1.5 weeks. When that is done, it will have to
be transferred to castor.
AOB:
Michel pointed out that our next meeting will have to be 1h long
because the reading room required for faculty candidate meetings.