This Python package aims at providing tools to study stochastic processes:
- numerical integration of SDEs
- numerical solver for the Fokker-Planck equations
- first-passage time computation
- instanton computation
- rare event algorithms
The code is available on Github or Framagit, with documentation on ReadTheDocs.
I am the main developer of stochrare, with some help from Thibault Lestang (Oxford). Contributors are welcome!
GHOST (Geophysical High-Order Suite for Turbulence) is a pseudospectral code for numerical simulations of turbulent flows developed by Pablo Mininni (University of Buenos Aires, Argentina) and Duane Rosenberg (Colorado State University, USA). It solves the Navier-Stokes equations in a tri-periodic cubic domain, and allows for adding transverse fields such as rotation, density stratification or magnetic field. GHOST can be used for direct numerical simulations, and also includes a variety of subgrid-scale models. The code is parallelized with a hybrid MPI/OpenMP method (the code also supports some GPU acceleration with CUDA), and runs on a large range of platforms, from laptops to supercomputers. More information can be found on Pablo Mininni's webpage or on GitHub.
In addition to running the code for numerical simulations, I have contributed a few additions and improvements to the code. These include for instance developing, with Duane Rosenberg, wave/vortex analysis tools for rotating-stratified turbulence, or writing a vortex tracking algorithm for 2D turbulence.
Idealized atmospheric GCM
I have been using two models based on the GFDL dynamical core: fms-idealized, developed by the group of Tapio Schneider at Caltech, and more recently, Isca, developed by the group of Geoff Vallis in Exeter.
I have also used the CliMT framework, developed in particular by Joy Monteiro and Rodrigo Caballero in Stockholm.
Here is some information on software which is not directly related to my research, but which I use to make my life easier.
I am in charge of organizing the colloquium for the physics department at ENS de Lyon. This implies a lot of tasks which can benefit from automation or assistance from a computer program: organizing the schedule, announcing the seminars, etc. This python project aims to do that.