The UJ Research Cluster Training Workshop will take place from :
Workshop :
Thursday 9 April - Tuesday 14th
April.
Saturday 11 April and Sunday 12 April mornings are also included.
Development : Following the workshop, 15th April - 5th May, there will be an additional three weeks of tutorials and development based on the actual research applications of the participants. The goal is to complete the three week post workshop tutorial period and have all the high performance computing applications functional and running.
The training and
development will be provided by Ben Clifford (
Overview
UJ has recently installed and configured a Research Cluster for high performance computing. The current configuration has 96 compute cores, and this will be almost tripled in the near future. The UJ-Research Cluster is therefore a very significant resource.
** UJ scientists can run applications on the Research Cluster locally.
** UJ scientists can access the USA GRID (OSG), the SA National GRID (G-Lite) as well as the European GRID (G-Lite). The UJ Cluster is already configured and visible on both OSG and G-Lite grids
This intensive workshop will introduce the techniques of distributed computing for the science and engineering fields (others also welcome), with hands-on training in the use of the UJ Research Cluster for high performance computing. The post-workshop programme will be used for code development, customization and the paralellisation of applications, aiming at a situation where all participants are affective users of the UJ Research Cluster by the end of the period.
Registration
Register
as soon as possible to ensure a place.
Send
email to Simon Connell <shconnell AT uj.ac.za>
Preparation
Participants should have at least intermediate programming skills (one to two semesters experience in C/C++, Java, Perl, and/or Python) and hands-on experience with UNIX / Linux in a networked environment.
You are also invited to mail Ben Clifford <benc AT ci.uchicago.edu> to begin a discussion of the applications you would like to port to the cluster. This will allow him time to prepare for the development component of the program.
