Admin Login • Workshop Attendee Login

A Tutorial on OWL - A Suite for Scalable Monte Carlo Simulations

Host(s):
Ying Wai Li, ORNL
Valentino Cooper, ORNL
Markus Eisenbach, ORNL
Krishna Pitike, ORNL

When: Sunday, February 17, 2019 9:00 am - 5:00 pm
Location: CSP Conference Room (322)
There's no cost to attend this event.

Lunch will be provided.
Lunch Cost: $10.00

Lunch Options:
Turkey and Cheddar Cheese with Lettuce and Tomato on sourdough luna bread
Roast Beef and Sharp Cheddar with Lettuce and Tomato on onion roll
Hummus and Pickled Vegetables on multi-grain luna bread
Drink Options:
Coke
Diet Coke
Sprite
Water

Event Details

OWL is a scientific software for large-scale Monte Carlo simulations for materials.  Originally developed at Oak Ridge National Laboratory to implement Wang-Landau sampling with first-principles calculations for the study of finite temperature materials properties (hence its name OWL - "Oak-ridge Wang-Landau"), it has now extended to provide a collection of commonly used classical, modern and parallel Monte Carlo algorithms.  OWL is written in C++ with an object-oriented, modular software architecture, adopting the "MPI+X" programming model for parallelization.  It provides two modes of simulation: the stand-alone mode for simulating user-implemented model Hamiltonians; and the driver mode that utilizes an external package as a library for the calculations of physical observables.  Today, OWL interfaces with two open-source density functional theory codes, Quantum Espresso and Locally Self-Consistent Multiple Scattering, to perform first-principles based statistical mechanics simulations.  It is therefore highly suitable for running (flying) on high performance computers.

OWL is under active development with proper software engineering practices.  Capability extension are driven by ongoing research activities.  We intend to make OWL open source on Github.  Our goal is to build a community code and create a platform to facilitate algorithm advancements and knowledge exchanges.  We welcome comments, suggestions and contributions from the scientific community.

In this tutorial, we will demonstrate the use of OWL and highlight a few research applications.  Attendees will obtain hands-on experiences with package building and carrying out serial and parallel Monte Carlo simulations for simple spin systems, as well as first-principles based Monte Carlo simulations using OWL.  Bring a laptop with Linux or Mac operating system, or a laptop with access to a workstation or computing cluster.  Be ready to get your hands dirty!

Event Registration

Registration is required to attend this event. This registration is separate from the Annual Workshop Registration.

If you have already registered for the Annual Workshop, then You can log in to register for this event more quickly.