Project Team Moais

Members
Overall Objectives
Scientific Foundations
Application Domains
Software
New Results
Contracts and Grants with Industry
Partnerships and Cooperations
Dissemination
Bibliography
PDF e-pub XML


Section: New Results

Chimeric algorithms design

To reach provable muticriteria performance, we used the coupling of various algorithms that adapt in several contexts: recursive cascading of both sequential and parallel algorithms with work-stealing; coupling specific algorithms on heterogeneous platforms (eg CPU/GPU); interactive distributed computations; fault-tolerant computations by coupling both a trustfully platform with low computation bandwidth and and an unreliable computing platform with high bandwitdh. A unification work is currently developed for the design of a chimeric algorthms that is composed of the parts of multiple algorithms, interactively cascaded to achieve provable multicriteria performance.