This secondment consisted in a collaboration with two faculties of the UTFSM in Valparaiso: physics and computer science. In particular, our contacts were professor William Brooks for the former and professor Luis Salinas for the latter. The group led by prof. Brooks joined the ATLAS collaboration in 2007 contributing among other topics to the heavy ions research program of the experiment and administrating the UTFSM Tier-2 site. The group of professor Salinas focuses on computational methods in science and engineering as well as computational finance.
The task undertaken was to give a series of frontal lectures and hold hands-on sessions about software engineering and parallel programming, including the characterisation of standard tools and techniques useful to enhance the productivity in those fields. The length of the course was 20-25 hours. Particular attention was dedicated to the hands-on exercises, inspired from real use-cases of HEP software, which accounted for half of the total duration of the course. The didactic program had many common points with the one proposed at USP and was organised in three main modules:
1. Software design in the many-cores era
2. Parallel programming
3. Tools and techniques
In addition, the availability to contribute to local software projects was pledged, for example providing expert advice or even concrete code prototyping. The CERN ER who made the two weeks visit to Valparaiso was Danilo Piparo, a member of the PH-SFT group, specialist of high performance scientific computing and parallel software development.
The audience of the course was composed of undergraduate and PhD students as well as experienced researchers coming from the field of HEP, mathematics, telecommunications and computer science. On average the attendance was of ten to fifteen people, all of them participating on a voluntary basis (for example, no credits valid for the academic curricula of the students could be recognised by the university). Quite some visibility was given to the lectures, advertised even on the university newspaper, and the overall level of interest was definitively high. The most important concepts related to parallel software engineering and development were successfully received by the audience. In order to establish a future reference, all the material of the course, both slides and exercises, was published on the web. The attendants interactively participated to the lectures, also influencing the proposed academic program, requiring to further develop some aspects treated during the lectures. An example was an additional lecture about advanced floating point arithmetic in a multithreaded environment, which was given even if not initially present in the program. It was indeed this interest generated in the students that allowed us to get in touch with the local research projects which were related to software technology. Guidance was indeed provided in the case of very concrete numerical and performance issues linked to two diploma (high energy physics and telecommunication) and one Ph.D. (environmental physics) theses.
In summary, all goals that were set before the visit were achieved. The visit was also a success in terms of contacts made and didactic material and experience accumulated within the SFT group thanks to the preparation of such an ambitious academic program. It should be noted that it is thanks to the effort of the host groups (and in particular of Y. Ivanov, F. Prokoshin and P. Arce) that the evaluation of this experience can so positive.
Issue of the PH news letter: