Skip to main content Skip to navigation

SIMU/CCP5 Workshop

Accelerating Dynamical Simulations

Location
CECAM, Lyon, France
Dates
begins a.m. Monday 22 March 2004
ends p.m. Wednesday 24 March 2004
Organisers
Professor Mike Allen, Centre for Scientific Computing, University of Warwick, Coventry CV4 7AL, United Kingdom.
Professor Gerald Kneller, Centre de Biophysique Moleculaire, CNRS, Rue Charles Sadron, F-45071 Orleans Cedex 2, France.
Financial Support
SIMU
CCP5
Draft Programme and Participant List
program.pdf (65kB) (revised March 17).

Motivation

The intention is for this to be a workshop devoted to discussions of methodology, rather than a conference with presentation of scientific results. The problem of bridging the timescale gap lies within the SIMU area, and also within the current focus of CCP5 activities, and there are several aspects, which may be broadly segregated into two classes: (i) where the aim is to extract the long-time essential dynamics as faithfully as possible, and (ii) where the acceleration is simply applied to an equilibration stage, or phase-space exploration stage, which can be unphysical provided correct sampling of starting configurations for exact dynamics is guaranteed.

A list of the aspects typically covered in current research is:

  • rapid equilibration using possibly unphysical dynamics or models
  • parallel tempering with soft potentials, different densities, different temperatures
  • switching polymer entanglements on and off
  • transition path sampling methods for rare events
  • design of lower-order descriptions based on stochastic remodelling
  • links with signal processing
  • time-stepping methods
  • hyperdynamics
  • kinetic Monte Carlo schemes for multiple-minimum problems
  • time-temperature superposition rules
  • conformational flooding

The workshop will concentrate on algorithmic approaches to speeding up simulations, especially trying to assess how rigorous the different approaches are, and to judge the reasons for success (or failure) of the more empirical approaches. The formal mathematical and statistical-mechanical background to the various proposed methods would be the main focus of discussion. Ways of checking the success or failure of methods would also be a focus: how to extract the essential dynamics from a microscopic simulation and feed this information into the design of a successful approximate method. Hence, signal processing and slow-mode analysis are relevant.

The meeting will avoid, as much as possible, discussion of the much wider field of mesoscale modelling, spatial coarse-graining, effective many-body potentials and so on, but would concentrate on simulations which intend to model dynamical behaviour in a way consistent with a prescribed microscopic model. The workshop would assess the progress that has been made within and outside SIMU in the last few years in this area, clarify the current situation, and summarize outstanding problems that need to be tackled in the future.

Confirmed Participants

The total number of participants is limited to 30. Those who wish to attend should contact Mike Allen.

Practical Arrangements

Participants should aim to arrive at the weekend; the meeting will start early Monday morning and finish after lunch on Wednesday. Directions to CECAM may be found on the practical information page. Travel expenses supported by receipts will be refunded up to a maximum of

  • 150 euro (participants from Europe)
  • 300 euro (participants from USA)

Participants may make their own accommodation arrangements using the information on the hotels page, or may ask for help by emailing the CECAM secretary Emmanuelle Crespeau. An accommodation and subsistence allowance will be provided at a rate of

  • 80 euro / day, for 3 days = 240 euro (participants staying in hotels)
  • 30 euro / day, for 3 days = 90 euro (participants staying in ENS rooms)