<html>
<head>
</head>
<body>
<font size="6" face="Times New Roman, serif">
<div><font color="#000080"><b>Old Dominion University</b></font></div>
<div><font color="#000080"><b>Department of Physics</b></font></div>
<div><font color="#000080"> </font></div>
<div><font color="#E36C0A"><b>Fall Colloquium Series</b></font></div>
<div><font size="3"> </font></div>
<div><font color="#17365D" size="5"><b>Tuesday September 13, 2011</b></font></div>
<div><font color="#000080" size="3"> </font></div>
<div><font color="#E36C0A" size="5"><b>"New Multiscale Code for Simulation of Coherent Synchrotron Radiations)"</b></font></div>
<div><font size="3"><br>
<font color="#1F497D" size="5"><b>Dr. Balsa Terzic</b></font></font></div>
<div><font color="#1F497D" size="5" face="Arial, sans-serif"><b>Jefferson Lab (CASA)<br>
</b></font></div>
<div><font size="3">Coherent synchrotron radiation (CSR) is an effect of curvature-induced self-interaction of a microbunch with a high charge as it traverses a curved trajectory. It can cause a significant emittance degradation, as well as fragmentation and
microbunching of the beam bunch. The development and optimization of the new designs for the existing and next-generation light sources crucially depends on accurate, high-resolution numerical simulations of this effect.</font></div>
<div><font size="3"><br>
Direct computation of CSR wakefields in 2D and 3D are prohibitively costly in terms of efficiency and memory requirements, as they require integration over the entire history of the bunch. Consequently, the present CSR codes employ a number of approximations
and simplifications that are often inadequate for resolving essential physics in many realistic situations. These situations where existing CSR codes fail are expected to become commonplace as the design of next-generation light sources commences. This provides
a strong impetus for the development of the new CSR codes that are both accurate and efficient.</font></div>
<div><font size="3"><br>
In this talk, I will present the progress report on the development of the fundamentally new, particle-in-cell, multiscale code for modeling CSR. The code will exploit advantages afforded by the mathematical formulation of the problem in wavelet basis: (i)
retaining information about the dynamics over the hierarchy of scales spanned by the wavelet expansion; (ii) natural removal of numerical noise (denoising) by thresholding of the wavelet coefficients; (iii) compact representation of relevant data sets and operators. The
resulting algorithm will be numerically optimized to run efficiently on graphical processing units (GPUs), and will be capable of modeling a number of different machines.</font></div>
<div><font size="3"><br>
As a proof-of-principle, we present the early benchmark result -- a comparison against the analytical results for a rigid-line bunch.</font></div>
<div><font size="3"> </font></div>
<div><font size="2"> </font></div>
<div><font size="3"> </font></div>
<div><font size="4">Presentation: Physical Sciences Building II 1100 @ 3:00 pm</font></div>
<div><font size="4">Refreshments: 1st Floor Atrium @ 2:30 pm</font></div>
<div><font size="3"> </font></div>
<div><font size="4">More details at <a href="http://www.physics.odu.edu">http://www.physics.odu.edu</a></font></div>
<div><font size="4">All are Welcome<font size="3">!</font></font></div>
<div><font size="3"> </font></div>
<div><font size="3"> </font></div>
</font>
</body>
</html>