Exhibitors & Sponsorship
Industrial representatives & agencies

Contact
Phone: +33(0)5 34 32 31 19
Email:
cimi-hpc@inp-toulouse.fr

WORKSHOP - PARALLEL IN TIME METHODS

Toulouse, January 11th-12th 2016
IMT (Institut Mathématiques de Toulouse), building 1R3, Toulouse, January 11th-12th 2016

Programme and organization:
Click here 

Slides

Here are the slides for talks given at the workshop (available soon) :

M. Emmett  Time-stepping on the exascale and time-parallel integration
M. Gander Direct time parallel solvers
D. Hupp  A parallel-in-time Navier--Stokes solver
T. Lunet  Von Neumann analysis of the Parareal algorithm applied to advection-dominated problems
I. Niyonzima
Time-parallel FE method for field computation of eddy currents problems excited by PWM sources
M. Olm  On scalable space-time balancing domain-decomposition solvers
R. Speck   Fault-tolerant parallel-in-time integration with PFASST

In this workshop, we want to address the challenge of large time dependent problems with time parallel methods. With the emergence of architectures with tens of millions of cores before the end of the decade, current parallelization strategies usually based on parallelization in space, inevitably reach their intrinsic limits. To use present and forthcoming massive levels of concurrency, additional directions of parallelism are thus required. For time-dependent problems, time-parallel methods have recently been demonstrated to be a promising approach. The idea of concurrency in the time domain is attracting growing interest in many different communities ranging from applied mathematics and computer science to climate and earth sciences as well as engineering sciences. The goal of the workshop is to gather international experts in numerical linear algebra and high-performance computing to present current research trends in this field, and to generate new ideas for future research activities and collaborations.

Confirmed invited speakers are currently Matthew Emmett (Computer Modelling Group Ltd, Calgary, Canada), Martin Gander (University of Geneva, Switzerland) and Robert Speck (Juelich Supercomputing Centre, Germany).

For further information please contact Xavier Vasseur (vasseur@cerfacs.fr) and Serge Gratton (serge.gratton@enseeiht.fr)

 

 
 

Choose a langage : French / English