<html>
<head>
</head>
<body text="#000000" bgcolor="#ffffff">
Curtis,<br>
<br>
The OSG is currently submitting a proposal for renewed 5-year
support, and has asked all of the collaborations that are using OSG
resources for a paragraph of support. I submitted something to her
on our behalf. You can see it along-side the paragraphs from a
number of other experiments and scientific collaborations in the
attached document.<br>
<br>
-Richard J.<br>
<br>
-------- Original Message --------
<table cellspacing="0" cellpadding="0" border="0" class="moz-email-headers-table">
<tbody>
<tr>
<th nowrap="nowrap" valign="BASELINE" align="RIGHT">Subject: </th>
<td>Re: paragraph on GlueX use of OSG</td>
</tr>
<tr>
<th nowrap="nowrap" valign="BASELINE" align="RIGHT">Date: </th>
<td>Thu, 3 Mar 2011 10:14:09 -0500</td>
</tr>
<tr>
<th nowrap="nowrap" valign="BASELINE" align="RIGHT">From: </th>
<td>Ruth Pordes <a href="mailto:ruth@fnal.gov"><ruth@fnal.gov></a></td>
</tr>
<tr>
<th nowrap="nowrap" valign="BASELINE" align="RIGHT">To: </th>
<td>Jones, Richard <a href="mailto:richard.t.jones@uconn.edu"><richard.t.jones@uconn.edu></a></td>
</tr>
</tbody>
</table>
<br>
<br>
<pre>Hi
Its somewhat a bit long compared to the examples we currnetly have. I attach the what I would propose to include for your comment.
Thanks!
Ruth
________________________________
From: Richard Jones <a href="mailto:richard.t.jones@uconn.edu"><richard.t.jones@uconn.edu></a>
Date: Thu, 3 Mar 2011 09:04:06 -0600
To: Ruth Pordes <a href="mailto:ruth@fnal.gov"><ruth@fnal.gov></a>
Subject: paragraph on GlueX use of OSG
The GlueX Collaboration is building a 12 GeV photon beam line and a dedicated spectrometer to study fundamental issues in strong QCD at Jefferson Laboratory. Our primary aim is to identify gluonic resonances in meson photoproduction by detecting their decays into exclusive final states in a hermetic detector with high acceptance and good resolution for both charged and neutral particles. Evidence for an exotic resonance is seen in its interference with known non-exotic resonances and non-resonant background amplitudes. Understanding these data requires high-statistics, both in the data and the Monte Carlo samples, and the capability to explore the interplay between resonant and non-resonant processes in a given final state. When the experiment ramps up to full intensity running, it is expected to collect about 1 PB of data per year. This data set consists of many different final states, each of which will require ability to generate Monte Carlo samples of comparable
size to the reconstructed event sample. The dominant source of error in analyzing these data is expected to be systematic in nature. The more dynamic the Monte Carlo simulation procedure can be, the more issues can be addressed during analysis by exploring systematic effects.
The Open Science Grid provides a high-throughput computing platform that is uniquely suited for providing dynamic simulation results for GlueX. The GlueX collaboration joined the OSG consortium in 2009, and was granted status as the Virtual Organization Gluex. A compute element and a storage element were set up at the University of Connecticut to provide resources for the needs of the Gluex VO during the initial period. Once sufficient experience has been gained in operating an OSG site, several additional sites at GlueX member institutions will be enabled for OSG access. Currently the UConn-OSG site provides a CE with 400 compute nodes and a SE with 10 TB of networked storage, both of which are made available to a broad range of users across the spectrum of virtual organizations for opportunistic use. Priority on this site is given to Gluex users, but with the GlueX experiment still in the construction phase, most of the available cycles are currently being shared wit
h other VO's. Experience with operations at the UConn site has shown that a small university group without dedicated cluster operations personnel can reliably operate an OSG site with over 95% uptime and very limited manpower commitment. Initial experience with simulations have shown that throughput can be readily obtained on the OSG in excess of 5 times over what would be available using only local cluster resources. Work is now underway on extending the adaptability and portability of the Gluex software stack, so that it will be capable of running on more sites, which will lead to a significant increase in the throughput gain factor to greater than 5. As a result of the success seen so far, the GlueX collaboration has adopted grid computing on the OSG as a foundational component of its computing plan, and is actively pursuing further integration of reources at member institutions into the OSG.
</pre>
</body>
</html>