Using SimGrid for Research in Large-Scale Distributed Systems
Arnaud Legrand, Martin Quinson
This tutorial will provide attendees with a clear understanding of current technology and best practice for exper- imenting parallel large-scale distributed computing research through simulation, and a first practical experience with the SimGrid framework.
The first part of the tutorial will present and contrast current experimental methodologies, giving attendees in-depth understanding of the scientific and technological issues at hand. The second part of the tutorial will focus on simulation, giving a state of the art of current simulation technology and discussing challenges for the development of sound simulation models. The tutorial will use the SimGrid simulation framework as an exemplar since it implements sophisticated and validated simulation models. The last part of the tutorial will provide a practical introduction to the SimGrid framework and a first hands-on experience to the users.
Over the years, SimGrid has emerged as one of the key tools for simulation of parallel and large-scale distributed systems. Its active user community counts over hundred members, mainly composed of researchers and students. In the last decade only, it grounded the experiments of 6 PhDs works, 25 journal papers, and over 100 conference papers. The community gathers every year during the SimGrid User’s Days workshop.
SimGrid is also used to test and tune two large applications by their authors: BigDFT (a massively parallel code computing the electronic structure of chemical elements developed by the CEA) and StarPU (a Unified Runtime System for Heterogeneous Multicore Architectures developed by Inria Bordeaux). Both of these programs enjoy large user communities themselves.
By the end of this tutorial, the attendees will understand what they could expect from SimGrid for their own work, and be able to join the flourishing SimGrid Users Community if they wish.