Danilo Carastan-Santos Webpage

Associate professor (tenure) at University Grenoble Alpes. Researcher in lightweight, eco-responsible methods for parallel and distributed systems at the Grenoble Computer Science Laboratory (UMR 5217).

prof_pic.jpg

Welcome! In this webpage you will find useful professional information about me, news and updates about my professional activity. You can find information on how to reach me at the bottom of this page.

Research Topics

My research takes place in the context of large-scale parallel and distributed infrastructures such as supercomputers and computing clusters and their applications, with a recent emphasis on limiting their energy consumption and environmental impacts. Such infrastructures are complex in terms of high-level parallelism and heterogeneity, are shared by numerous users, and rapidly evolve over time. Such complexity raises many fundamental questions and technical challenges in order to exploit such infrastructures as efficiently as possible.

I organize my research activity in the following three research axes:

  • Data Science and Machine Learning applied to High-Performance Computing infrastructures’ resource management

  • Frugal methods to limit the energy consumption and environmental impacts of distributed computing

  • Analysis and development of parallel and distributed infrastructures applications and architectures

Selected Publications

  1. CCGrid 2019
    One Can Only Gain by Replacing EASY Backfilling: A Simple Scheduling Policies Case Study
    Danilo Carastan-Santos, Raphael Y. Camargo, Denis Trystram, and Salah Zrigui
    In 19th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing, CCGRID, Larnaca, Cyprus 2019
  2. SC 17
    Obtaining dynamic scheduling policies with simulation and machine learning
    Danilo Carastan-Santos, and Raphael Y. Camargo
    In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, SC, Denver, CO, USA 2017