MPI-based MCMC sampler for TDC II

This repository hosts my personal work-in-progress MCMC sampler for the Taiji Data Challenge II (TDC-II). The core design centres on MPI parallelisation: each independent Markov chain runs as a MPI process. See the documentation for demos.

Installation

Before installation, ensure that the following dependencies are satisfied:

• mpi4py: requires a working MPI implementation that supports at least the MPI‑3 standard. I have used Open MPI to develop the project.

• h5py: parallel I/O features are required. See the official guide for setup instructions.

Once the dependencies are in place, navigate to the project root directory and install with:

pip install .

Contribution

Before contributing to this repository, please share your ideas and discuss with me via issue, email, or other means. To be honest, I am not yet sure whether the current design is truly viable or effective for real‑world use. Therefore, I strongly encourage suggestions, constructive criticism, and alternative perspectives.