The Distributed Organization for Scientific and Academic Research (DOSAR)
sponsored a day-and-a-half High Performance Computing/High Throughput Computing
(HPC/HTC) School as part of the Fifth African School of Physics (ASP2018) at the
University of Namibia (UNAM) in Windhoek, Namibia, from June 24 to July 14. 
Seventy students participated in the event, from all over Africa and beyond. 
The ASP runs every two years, with previous schools in South Africa, Ghana, 
Senegal, and Rwanda.

ASP2018 agenda: https://indico.cern.ch/event/656460/
HPC/HTC School agenda: https://opensciencegrid.org/dosar/ASP2018/ASP2018_Schedule/

This year, the HPC/HTC instructors were Julia Gray <julia.ann.gray@gmail.com>
from Geneva, Switzerland, and Horst Severini <severini@ou.edu> and Patrick
Skubic <pskubic@ou.edu> from the University of Oklahoma -- that means, 2/3 of 
the teaching staff of the computing school were from OU. Much of the training
material for the school was originally created by Scot Kronenfeld and Alain Roy,
from the University of Wisconsin-Madison, with additions and upkeep by Rob Quick
<rquick@iu.edu> and Kyle Gross <kagross@iu.edu> from Indiana University.

Skubic and Severini, with the help of Chris Walker (also from OU) and 
Dick Greenwood (from Louisiana Tech University) spent considerable time
revising and updating the school material and all the exercises before
the school began. Severini also spent a week at UNAM earlier in the year,
as well as more time remotely later, to install all the software required 
to run all computing tutorials.

The computing school started on Tuesday morning, July 10, with an introduction
to HPC, HTC, and research computing in general. After that, and for the rest of
the day, the students were exposed to HTCondor jobs and exercises, followed by
more advanced concepts like workflows and application profiling, as well as more
advanced research computing applications. HTCondor is a batch submission 
software that is used widely around the world for High Throughput Computing.

On Wednesday morning, there was another tutorial based on the ROOT scientific
computing framework, followed by a lecture on how to use ROOT for a real-life
particle physics data analysis.

On both days, the students were given plenty of time to go through tutorials of
various complexities, and learn first hand how to submit jobs to HTCondor for
analysis and retrieve job output. The exercises were the most valuable part of
the computing school, since they allowed the students to get a more detailed
feel of the material. They were also very well received, so much so that the
students kept working on them even during lunches and coffee breaks.

We would like to acknowledge the OU Division of Administration and Finance
as well as the Homer L. Dodge Department of Physics and Astronomy for their
financial support which covered the travel costs for Skubic and Severini.