Present in person: Burin Asavapibhop, Will Brooks, Volker Burkert,
Alan Coleman, Kyungseon Joo, John Price, Dennis Weygand.
Present via speaker phone: Larry Dennis, Maurik Holtrop, Allena Opper,
David Rowntree, Jianguo Zhao.
Agenda (as distributed):
ITEM 1 1:30 - 1:35 Revise agenda, comments/changes regarding
distributed
minutes, other administrative issues - All
ITEM 2 1:35 - 1:55 Proposed overview of GSIM environment -
including all
planned auxilliary programs. Possible modules
include event generators, pre-GSIM event
filters, GSIM 'knock-out' ('GSIMKO'), trigger
simulator, etc. - Maurik Holtrop (via
speakerphone and the web)
ITEM 3 1:55-2:20 Define the role and operating characteristics
of the trigger simulator - Allena Opper (via
speakerphone) - All
ITEM 4 2:20-3:00 Report on progress on existing projects,
associate
names/deadlines with new projects
-------------------------------------------------------------------------------
ITEM 1 - REVISE AGENDA, COMMENTS ON MINUTES, ADMINISTRATIVE - All
No changes, no comments.
Note that the clas_gsim mailing list is archived at:
http://www.cebaf.gov/ccc/hypermail_archives/CLAS/CLAS_GSIM/
If you are aware of CPU's we can use for GSIM, let me know! So far we
have partial use of farms at MIT and UVA, and full use of at least one
cpu at OU.
ITEM 2 - OVERVIEW OF GSIM ENVIRONMENT - Maurik
Maurik showed a diagram from the last collaboration meeting showing the
overall
flow of analysis. An item not on his diagram was discussed: a 'filter'
program which would pre-filter event generator data to remove events
(and/or particles) which are well outside of the acceptance. Reaction
to this idea was mixed; it will be discussed at a later meeting. (The
alternative is to carefully generate only the events that you want, in
any event generator you use. This may be complicated at small theta
near the torus coils.) A second discussion centered on whether
'GSIMKO' and the trigger simulator should be combined; for logistical
reasons (GSIMKO is far advanced, per Dennis Weygand and Alan Coleman)
it was decided that we should not combine them.
Maurik's global web page is at
http://www.physics.unh.edu/Maurik/gsim_info.shtml
and the analysis flow diagram is available from this page.
ITEM 3 - DEFINE ROLE AND CHARACTERISTICS OF TRIGGER SIMULATOR - Allena
Summary of the discussion: In the short-term, we need a very simple
trigger
simulator which will be relevant for the e1 runs. Since the Cerenkov
counters seem to be highly efficient (no known malfunctions and the
thresholds are extremely low), the main effect of the trigger on the
acceptance is that the energy threshold on the calorimeter limits and
modifies the W acceptance of the detector. Therefore the first
priority is begin with a near-trivial Cerenkov trigger coupled with a
first attempt at the calorimeter trigger.
Allena will aim to produce a working prototype of this level of
simulation by 8 April 1998.
The second aspect of this discussion was on the longer-term functions
of the trigger simulator. The prototype effort mentioned above will
give some guidance on how the code should be structured; other
elements which may shape the structure of the code:
1) Ability to read in the same trigger files as the hardware uses
2) Ability to include the low-level (~600) trigger bits and
to compare them to the measured data for these bits (this capability
will be implemented soon - needs a student to work on it! The hardware
to do this has been purchased but needs to be installed)
3) Ability to include time information if needed
We may or may not want to do 1-3 above, but we should structure the
code to make it straightforward to do so.
Some people who have worked on simple trigger simulators in the past:
Dave Doughty's students, Alex Vlassov, Joe Manak. Allena may want to
borrow parts of their codes.
ITEM 4 - PROGRESS ON EXISTING PROJECTS - All
STUDY TIMING AS A FUNCTION OF VOLUME TYPE AND PARAMETER SETTINGS -
David Rowntree
David studied the effect of 5 parameters on the speed of the
code. I quote from his email messages:
TMAXFD - The maximum angular deviation due to the magnetic field
permitted in one step (degrees).
STEMAX - The maximum step size permitted (cm).
DEEMAX - The maximal fractional energy loss permitted in one step
(0<DEEMAX<=1).
EPSIL - The boundary crossing precision (cm).
STMIN - The minimum allowed value for the maximum step imposed by
energy loss, multiple scattering, Cerenkov or magnetic field
effects (cm).
David went through the code and checked the values put in, and found
they
varied a lot by detector and material. The ranges:
TMAXFD: 0.3-5 degrees
STEMAX: 0.006-1 cm
DEEMAX: 0.05-0.5
EPSIL: 0.005-0.1 cm
STMIN: 0.003-0.1 cm
Some variation certainly makes sense, since the characteristics of the
components vary a lot. GEANT has some built in defaults for these
values:
TMAXFD: 20 degrees
STEMAX: 10,000,000 m
DEEMAX: 0.25 if the material is not `sensitive' and X0 (the radiation
length) is less than 2 cm.
0.25-0.2/sqrt(X0) otherwise.
EPSIL: No default.
STMIN: 2R/sqrt(X0) if the material is not `sensitive' (R is the range
of an electron of energy CUTELE+200kev).
5R/sqrt(X0) if the material is `sensitive'.
David's conclusions from the speed tests:
There is at most a factor of eight to gain in the speed via these
games. Changing these parameters will have effects on the accuracy of
the simulation, and I suspect I went out of acceptable ranges in my
tests. The next step is to look into the negative effects of these
changes, to try to find a reasonable middle ground.
One idea David had was to possibly run with higher cuts for the
calorimeter showers, and instead of 'dumping' the invisible energy,
distribute it in a reasonable, parameterized way. I.e., if you run
with a 10 MeV cut, when a secondary gets down to that level, just add
10 MeV (times the sampling fraction) to the relevant volume.
Etienne Burtin did a study of the effects of cuts on calorimeter
resolution in 1995. This study indicated that the resolution begins to
worsen when the cuts are above 0.5 - 1 MeV.
The group decided we needed to establish criteria for the accuracy of
GSIM, to determine where we can afford to set some of these
parameters. Volker Burkert agreed to work on this for next week.
For the following week, David and Jianguo agreed to tackle the project
of 'turning off' various volumes to see where the time is spent in the
code. Maurik emphasized that the best way to do this is not to turn
off the geometry, but to turn off secondaries in the volume, to get
realistic results.
DETERMINE THE TIME PENALTY FOR ELECTRONS IN THE COILS - David
Rowntree, Will Brooks
Will generated 1 GeV electrons at 30 degrees in phi, and 0 degrees
in phi (these files are in /home/esc/gsim/mc_inputfiles). David timed
these two event types, and the electrons heading toward the coils took
5 times as long to process as electrons heading toward the fiducial
region. This indicates we should study what's going on in the region
of the coils to see if we can decrease the cpu time without losing
much accuracy.
Jianguo suggested we could also generate events only within the
fiducial region for the high-statistics running, and separately
generate events which go into the coils; make a library of these
'junk' events, and just add them into the clean events, just as we are
thinking of doing with the target-related background. I think this
might be a good approach....
GENERATE ELASTIC SCATTERING EVENTS TO EVALUATE TBT QUALITY -
Burin Asavapibhop
Burin successfully generated elastic scattered events and
reconstructed them with time-based tracking. His first result is to
find a 25 MeV width for the proton's W distribution. According to
Volker's estimate, this should be below 10 MeV, depending on whether
air or helium is used in and between the chambers. Burin will
investigate further. He also seems to lose ~1% of the events in going
from HBT to TBT.
Burin will be happy to give his data file away to others to study,
e.g. the types of studies being done by Stepan Stepanyan, John McNabb,
Volker Burkert, and others (major systematic errors in tracking-related
quantities) could benefit from using Burin's data file.
MISCELLANEOUS
Maurik pointed out that the Unix pipe function doesn't work for BOS
output. Dennis Weygand said he was very close to finishing a project
which will remedy this among other problems.
Maurik and John Price noted that CELEG output may need to be modified;
currently may write out the PART bank (I think?) and does write out
MCTK and MCVX banks, but there are some incompatibilities between
these two banks, needs to be sorted out (the contents are somewhat
different).
The issue of whether we should generate events uniformly, or
according to phase space, or according to best-guess dynamics,
surfaced yet again. We should discuss this in a future meeting in
detail.
Larry Dennis suggested another approach to the trigger simulator and
some other issues: we could remove the digitization stage from GSIM
altogether. The digitization would be performed by an external
program. The output files would probably be much larger, and some
code re-structuring might be required, however, this would decouple
all aspects of detector response simulation (slow, GEANT) from issues
of calibration and digitization technique (fast). This might be the
best way to do the job. It appears to be a longer-term issue, and
therefore not completely appropriate for the GSIM FOCUS Group; but we
may want to discuss it further. Anybody want to develop a proposal for
how to do this?
The meeting adjourned about 3:10 pm.
-------------------------------------------------------------------------------
GSIM Focus Group Members and other GSIM contacts:
- Burin Asavapibhop (U. Massachusetts, on-site at JLAB) 757-269-7322
- Will Brooks (JLAB) 757-269-6391 brooksw@cebaf.gov (core group, GSIM
Focus Group primary contact)
- Volker Burkert (JLAB) 757-269-7540 burkert@cebaf.gov
- Bryan Carnahan (Catholic U) 202-319-5317 or -5315 or 202-483-5205
08carnahan@cua.edu
- Larry Dennis (Florida State U) 850-644-1804
larry@fsulcd.physics.fsu.edu
- Steve Dytman (U. Pittsburgh) 412-624-9244 dytman@vms.cis.pitt.edu
- Laurent Farhi (Saclay, on-site at JLAB) 757-269-7680 farhi@cebaf.gov
(core group)
- Rob Feuerbach (Carnegie-Mellon U.) 412-268-2749(CMU#),
757-269-?(JLAB#) feuerbac@ernest.phys.cmu.edu
- Maurik Holtrop (UNH) 603-862-2019 holtrop@cebaf.gov (GSIM
coordination)
- Kyungseon Joo (UVA, on-site at JLAB) 757-269-5307 kjoo@cebaf.gov (core
group)
- Joe Manak (JLAB) 757-269-5829 manak@cebaf.gov
- Si McAleer (Florida State U) phone? mcaleer@cebaf.gov
- Allena Opper (Ohio U) 740-593-1982 opper@akopper.phy.ohiou.edu
- John Price (RPI) ? pricej@cebaf.gov
- David Rowntree (MIT) 617-258-5442 tree@mitlns.mit.edu (core group)
- Dave Tedeschi (U. South Carolina) 803-777-1132 tedeschi@sc.edu
- Mike Vineyard (U. Richmond) 804-289-8257 mvineyar@richmond.edu (core
group)
- Dennis Weygand (JLAB) 757-269-5926 weygand@cebaf.gov (GSIM
coordination)
- Jianguo Zhao (MIT) 617-258-5438 jzhao@mitlns.mit.edu (core group)