MAGI Workshop
Scientific computing, cloud, big data
VILLETANEUSE 2019

The fifth MAGI Workshop days will be held on Campus de Villetaneuse from université de Paris 13.

In 2014, the meetings took place in the green setting of the Centre International de Séjour, in the city of Vichy. The program is available by clicking on this link .

Importante date ; Location

October 2, 2019
Room T204 -- Salle du conseil -- IUT Villetaneuse

Special points to notice

These meetings bring together researchers from the University of Paris 13 as well as their colleagues working on the topics of high performance computing, large masses of data (big data, datamass), networking (IoT) and cloud computing. We want to bring together all those interested in these topics and who use architectures that range from clusters to clouds and grids. They are open to all users of the MAGI cluster and they serve as a place for scientific exchange and information. The meetings will also be used to listen and discuss issues in Physics, Statistics, Image, Mathematics, Computer Science, Economics...

Non exhaustive list of topics

Distributed and parallel scientific computing;
Data management and machine learning;
Massive data management in warehouses;
Grids and Clouds;
Images;
Networking (IoT);

Scientific Programme

**Location :** Room T204, Institut Universitaire de Villetaneuse ; **General Chair:** Christophe Cérin
Date : October 2, 2019	Session name	Information
10:30 -- 11:00	Coffee corner (room L208)
11:00 -- 13:00	Big Data / Networks / Edge (Chair: Khaled Boussetta)	Mohamed Oualid Attaoui (LIPN): Subspace Clustering over Evolving Data Stream; Ogechi Nwogu, Gladys Diaz, Marwen Abdennebi (L2TI, Paris 13): A Combined Static/Dynamic Partitioned Resource Usage Approach for Random Access in 5G Cellular Networks; Gael Beck (LIPN, Paris 13): A Distributed and Approximated Nearest Neighbors Algorithm for an Efficient Large Scale Mean Shift Clustering; Houssemeddine Mazouzi, Nadjib Achir and Khaled Boussetta (L2TI, Paris 13): Elastic Offloading of Multitasking Applications to Mobile Edge Computing;
13:00 - 14:00	Lunch at your own
14:00 - 15:30	High Performance Computing (Chair: Claude Tadonki, Mines Paristech)	Jean-Batiste Montavon (LAGA, Paris 13): Parallelization of an implicit scheme based on finite volume method; Tarek Menouer, Patrice Darmon (UMANIS): New container scheduling strategy based on Machine Learning techniques to reduce energy consumption in a cloud computing environment. Nicolas Grenèche (Paris 13): News about MAGI;
15:30 - 16:00	Coffee break (room S104)
16:00 - 17:30	New openings and Discussions (Chair: Christophe Cérin)	Guests: Anna Kobusinska (Poznan University of Technology, Poland) Cezary Mazurek (Poznan Supercomputing and Networking Center, Poland) Ariel Olekziak (Poznan Supercomputing and Networking Center, Poland) Presentations by our guests: PSNC - General overview and R&D projects; Open Smart City - Poznan Case Study Round table about the present and future of High Performance Computing and Big-Data. Scientific and engineering perspectives. The questions for our panelists could be as follows: What are the new trends in HPC? What are the new trends in Big-Data? How to make the convergence of HPC & Big Data architectures? IA (neural networks) and supercomputers. What is it possible to do at present time? What do you observe in the usage of supercomputers? New practices? New skills? What should be the structuring (administrative, technical) of an IT resource center today for serving all the disciplines in an university? What are the tips for moving forward? How to welcome newcomers and what can we do for them? Dany Lang (CEPN, Paris 13): The ASCOPE 2020 project;

Training (in french)

Les formations à l'utilisation du cluster MAGI sont assurées par Nicolas Grenèche ;
Des accords de partenariat au niveau Master, portant sur les mathématiques, le calcul scientifique et la science des données, viennent d'être signés avec L'Université Polyechnique Mohamed 6 (Maroc) et l'USMBA de Fès (Maroc). Pour de plus amples informations, voir avec Fayssal Benkhaldoun.