A text summary of each talk is given below. The transparencies can be
found at:
Han's
Sasha's
Salavat's
Pal's
Andrei/Jim's
Pam's
mehmet's
Minutes of the July 18, 2001 Jet/Met meeting
-------------------------------------------
Attendence: Sasha, Andrei, Irinia, Olga, Paris, Marcello, Salavat,
Stephan Wynhoff, Sarah Eno, Hans Wenzel, Jim Brooke, Pal Hidas,
Dan Green, Pam Chumney, Shuichi Kunori, James Rolfe, Mehmet
Agenda
------
production for 2e33 - Hans Wenzel
short talks (1-2 transparencies)
plans/what do we need to study tau's at 2e33 - Sasha Nikitenko
plans/what do we need to study MET's at 2e33 - Pal Hidas
plans/what do we need to study jet's at 2e33 - Andrei/Jim
plans/what do we need to study SUSY's at 2e33 - Salavat Abdullin
L1 Calorimeter trigger at 2e33 - Pam
Update on qqH->qqWW - Mehmet
------
production for 2e33 - Hans Wenzel
The good news is: it takes 2-3 weeks to proces the entire
HLT sample at low luminosity. Processing seems to go smoothly.
FNAL expects to have the complete fz files for the HLT sample
by the end of the week.
The bad news is production is halted due to a bug in CaloIter that
affects events with pileup. A fixed version of ORCA is expected
next week. Hans will install it himself at FNAL.
The priorities for production are:
0) ntuples from 10k events from each bin of the "buggy" low lum sample
for Andrei
1) low lum HLT sample
2) other low lum samples needed by Sasha
2) whatever portions of the HLT sample Pal, Shuichi, and Pam
say they need to check the MET calculation and to refine
the L1 Higgs trigger.
3) Mehmet's samples
4) we will wait until Salavat's new HCAL code is released in
mid-October to reprocess the HLT sample at high luminsoity.
(would take about 20 weeks, will need help from more regional
centers to complete this)
Hans mentioned that there is not much computing available at FNAL
for users now, but expects this to improve soon (September?).
However, users may soon need to get a kryptocard to log into FNAL.
Users who want an FNAL account should send an email to Hans.
Pam said she had good luck accessing the sample at FNAL from Wisconsin.
Sasha also seems to be having okay access from CERN. However, Jim
Brooke and Ritva both have reported problems with reliability of
the link, which is why they want FNAL accounts.
plans/what do we need to study tau's at 2e33 - Sasha Nikitenko
Sasha reminded us which are the most important HLT bins
to study backgrounds for the tau trigger. He also pointed
out 4 signal samples he would like given high priority
for the 2e33 processing. He would like to test Sridhara's new
L1 tau algorithm as part of this exercise. And, since the rate
to tape is now only about 4 hz for his higgs to tau channels, he
will investigate loosening the L2 cuts. He will need help from the
b/tau group to do the tracking digitization for his backgrounds
and signals.
plans/what do we need to study MET's at 2e33 - Pal Hidas
Pal has found evidence that the CaloItr bug does indeed affect
MET for the high luminosity sample. He will make up some plots
and distribute. He says he doesn't see any evidence that the
bug affects the low lum sample, and so thinks FNAL should proceed
to make ntuples from the "buggy" 10^33 data and Andrei can use
this to make jet energy corrections. He will then put these into
the ntuple maker, and use this on the new low lum data, once it
is produced. He will then produce rate and resolution curves
for uncorrected met, and for sasha type-2 corrections.
plans/what do we need to study jet's at 2e33 - Andrei/Jim
Andrei says to make the 2e33 calibration, he only needs a small
portion of the hlt data (10k events from each hlt bin)
Then, once the full HLT sample is available, he will do the jet rate
curves and turn-on curves. He will use the "branson" weights
to do the multi jet rates.
plans/what do we need to study SUSY's at 2e33 - Salavat Abdullin
Salavat says that we currently have 2 SUSY samples processed.
However, he thinks this will be enough to understand what we
need to trigger on SUSY at low luminosity. He says that the
trigger-ability depends mainly on the mass scale, because
you either get more jets, each with less Et, or few jets, but with
higher ET. Sarah said D0 had found it more difficult to trigger
off the type of SUSY event with more jets, each with less Et.
To do a full background analysis, he needed about 200,000 events
of "physics" background (which should be do-able to process at
low luminosity). However, he thought we needed 10^7 QCD events
above 200 GeV. We need to understand if this is really true, if
we really need that many QCD events to undersatnd the trigger rate
with acceptable errors. If so, we may need to start to understand
how to get trigger rates using a parameterization. For pure jet
triggers, this may be do-able. But, MET may be hard? We discussed
whether his old CMSJET studies could be used to estimate the rates,
but decided that they could not, because it has been shown that
it does not do a good job of predicting MET rates in samples
with low true MET?
L1 Calorimeter trigger at 2e33 - Pam
Pam said she is currently working on a low lum trigger table using
the old 10^33 data. She is finding it hard to fit everything
into 4 Hz. She expects a note in August.
Update on qqH->qqWW - Mehmet
Mehmet is gearing up to do a study of qqH->qqlnlnu. He is currently
studying the lepton distributions in these events at the generator
level, and comparing his results to results from Zeppenfeld. He
expects to be able to start ORCA production of his signals and
backgrounds before the release of Salavat's code in mid-October
(when alot of the computing resources should be turned over to
a reprocessing of the high luminosity HLT sample)