The Department of Computer Science would like to invite you to the public talk on occassion of the habilitation of

 

Juan J. Durillo 

on

Multi-Objective Optimization of Software: An overview of current results and future challenges 


Date: Friday, 8th of June, 2018, 10-11am
Place: Room 3W04, 1st floor, ICT building

Abstract

As hardware continues to grow in complexity and new computing paradigms arise, designing software that optimally exploits the features exposed by the underlying computing platform becomes a tedious, error-prone, time-consuming task that usually leads to non-optimal running times or power consumptions.

The last few years have witnessed an ever-growing interest in developing auto-tuning methods to offload the traditionally human-guided approach of tuning programs to automatic search, intelligent techniques or machine learning approaches. Most of these systems are specific to optimize a single criterion or focuses on optimizing applications composed of only a few lines of code.

In this talk, I will introduce my view of a holistic approach to automatically optimize the execution of software regarding different criteria, present some obtained results and describe further research directions in the area.

CV

Juan J. Durillo received his PhD from the University of Malaga (Spain), where he worked on the design and implementation of parallel and distributed multi-criteria optimization methods, in particular from the field of Metaheuristics.

From May 2011 to October 2017, he was a PostDoc (Assistant Professor) at the University of Innsbruck (Austria), where he worked with Prof. Thomas Fahringer in the DPS group. During his time in Innsbruck, he contributed to some challenging problems in the field of auto-tuning parallel applications within the context of the Insieme compiler and multi-objective scheduling of workflows for IaaS clouds.

Since November 2017 he is a scientific employee at the Leibniz Supercomputing Centre of the Bavarian Academy of Sciences and Humanities in Munich (Germany).

Nach oben scrollen