RynDMC

A package for doing DMC in a black-box, flexible way, implementing the tricks and utilities we use as a group

Definitely still in the very beta of beta stages, but will build out piece by piece

Roadmap

DMC

• Implement an AbstractDMC class to be subclassed by smarter implementations
• Implement a DiscreteWeightingDMC subclass to handle discrete weighting DMC simulations
• Implement a ContinuousWeightingDMC subclass to handle continuous weighting DMC simulations
• Implement an ImportanceSampledDMC subclass of ContinuousWeightingDMC to add in importance sampling

Potential Handlers

• Create C-level potential handler library that can take just a pointer to a potential
• Write better communication paradigm for doing more work at the C level before passing back to python
• Create multiprocessing based potential handler for automatic parallelization of potential calls
• Write little wrapper to Fortan and C-types based hook-ins for simple potentials

Customizations

• Add standard job config template for setting up a simulation
• Add

Dependencies

At the moment this depends on the McUtils and Coordinerds These might find their way in a git module level dependencies, though.