QMC standards

From QMC

This should become a discussion on what it would take, e.g., to make it possible for group A to use in its own latest DMC algorithm the trial wave function optimization method of group B in combination with the trial functions of group C. This probably would be made much simpler if we can agree on some standardized interfaces and IO procedures.

I have some ideas on how this might be accomplished, and if there is any interest I'll supply them, but who knows it's not a problem at all. If this is not, then what is?

Here is a summary of what some of us consider important in a standard input/output format and a discussion of a specific input/output format that we use: StandardizationCriterion

[edit] Standards workshop

On Monday 11th September 2006 there will be a one day MC data format meeting at ETH Zurich.

Possible topics:

• discuss and define standard data fornats for Monte Carlo simulation

data

• XML formats for the main results and meta-information
• binary formats (e.g. HDF-5) for time series, binned data, or other

large data sets

• these formats can then be the basis for
  • visualization tools (e.g. for time series analysis)
  • evaluation tools (binning analysis, error evaluation, covariance, autocorrelations, equilibration, ...)
  • archiving of results
  • publishing of Monte Carlo data in an archive or as electronic auxiliary material of publications

Confirmed attendees: Matthias Troyer, D. Ceperley, D. Landau, J. Kim, F. Assaad, T. Schultess

Contact Matthias if you are interested in attending.