[Nuclear] Undergraduate posters at Fall DNP meeting

[Jerry Gilfoyle]

Hi Folks,

I have several summer students that will be presenting posters at the fall
DNP meeting as part of the Conference Experience for Undergraduates (CEU)
program (http://www.lanl.gov/dnp/undergraduates.htm). Abstracts for the
posters are due Aug 1. Below are three abstracts from students. Please let
me know what you thiink of them.

Cheers,

Jerry

-------

C.J. Musalo, G.P. Gilfoyle, J. Carbonneau

Simulation of the CLAS12 Forward Electromagnetic Calorimeter

            The primary mission of Jefferson Lab (JLab) is to reveal the
quark and gluon structure of nucleons and nuclei and to deepen our
understanding of matter and quark confinement. At JLab there is a need for
high-performance computing for data analysis and simulations. The precision
of many experiments will be limited by systematic uncertainties and not
statistical ones; making accurate simulations vital. A robust CLAS12
simulation is currently being developed called gemc, where Geant4 is used to
simulate the components of CLAS12.  We have added the electromagnetic
calorimeter (EC) detector to the gemc simulation. The EC is a sampling
electromagnetic calorimeter made up of alternating layers of lead and
scintillator used to detect electrons, photons, and neutrons. We have
streamlined the mathematical model of the EC geometry.  The geometry is
stored in a mysql database on a server at JLab.  Using perl, we modified
this geometry database with our streamlined equations. We tested the new
geometry by sending straight tracks (no magnetic field) through the edges of
specific layers using the geantino, a Geant4 virtual particle that doesn’t
interact with materials.

-------

M. Moog, G.P. Gilfoyle, and J. K. Carbonneau

Simulating the Neutron Detection Efficiency of the CLAS12 Detector

We have studied the expected performance of the CLAS12 detector that will be
built at Jefferson Lab as part of the 12-GeV Upgrade. The Upgrade hopes to
further our understanding of the internal structure of nucleons by studying
nucleon properties such as form factors and generalized parton
distributions. The initial round of experiments for the 12-GeV Upgrade
include ones that require neutron detection and we are studying the neutron
detection efficiency in preparation for such experiments. A precise
knowledge of the neutron detection efficiency is required to keep systematic
uncertainty low in these experiments. Previously we studied the CLAS12
performance by generating the four-momenta of an electron and neutron after
a relativistic, elastic collision and passing these data into the
GEANT4-based simulation program gemc. The code uses the four-momenta of
these particles and simulates their interaction with the components of the
detector. We then reconstructed the events with the program Socrat. By
comparing the number of measured elastic, electron-neutron coincidences to
the number of elastic electrons detected in the simulation we extracted the
efficiency of one of the outer time-of-flight scintillator panels. We built
on this previous research by using the full array of time-of-flight
scintillators in the simulation and expanding the range of the neutron
momentum.

-------

J. K. Carbonneau, G. P. Gilfoyle, and E. F. Bunn

Development of a Computing Cluster At the University of Richmond

The University of Richmond has developed a computing cluster to support the
massive simulation and data analysis requirements for programs in
intermediate energy, nuclear physics, and cosmology.  It is a 20-node,
240-core system running Red Hat Enterprise Linux 5.  We have built and
installed the physics software packages (Geant4, gemc, MADmap...).  The
system has a theoretical processing peak of about 2500 GFLOPS. Testing with
the High Performance Linpack (HPL) benchmarking program (one of the standard
benchmarks used by the TOP500 list of fastest supercomputers) resulted in
speeds of over 900 GFLOPS.  The difference between the maximum and measured
speeds is due to limitations in the communication speed among the nodes;
creating a bottleneck for large memory problems. As HPL sends data between
nodes, the gigabit Ethernet connection cannot keep up with the processing
power.  We will show how both the theoretical and actual performance of the
cluster compares with other current and past clusters, as well as the cost
per GFLOP.  We will also examine the scaling of the performance when
distributed to increasing numbers of nodes.

-- 
Dr. Gerard P. Gilfoyle
Physics Department
University of Richmond, VA 23173  USA
e-mail: ggilfoyl at richmond.edu
phone:  804-289-8255
fax:    804-484-1542
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailman.jlab.org/pipermail/nuclear/attachments/20100716/90eea846/attachment.html