AI-for-Science Systems (AIS²) Lab

MATCSSI: Opening the Quantum Toolbox for Everyone

For those working in computational chemistry/physics, it is clear that nowadays we have an immense zoo of quantum softwares and approaches, and these are constantly growing everyday. Moreover, several of these softwares can interact via often complex APIs, or requiring several pre- and post-processing steps. What if we could perform complex many-body quantum simulations workflows, from quasiparticle excitations to superconductivity, via a simple, unified, and reproducible architecture? That’s what MATCSSI (https://matcssi.tacc.utexas.edu/) is.

MATCSSI (Many-body Applications and Tools for Computational Software and Simulation Interoperability) was developed to give a fresh look to this challenge: an interoperable software ecosystem that wrap state-of-the-art many-body electronic structure codes within an intuitive, and open Jupyter-based environment.

MATCSSI Architecture

Why it matters?

  • The bottleneck of many-body simulations: Performing accurate many-body electronic structure calculations remains a challenge in computational chemistry. These methods are crucial for understanding excitons, polarons, superconductivity, and carrier dynamics.
  • Limitations of current workflows: Beyond DFT methods, such as GW, BSE, and electron–phonon coupling, typically involves multiple softwares, complex input structures, and HPC job dependencies, which makes it difficult for newcomers and even experienced users.
  • Our step forward: MATCSSI integrates leading open-source tools into interactive Jupyter Notebooks, turning complex workflows into reproducible pipelines that run on HPC resources.
  • Our philosophy: “Make easy things easy, and difficult things possible.”

How does MATCSSI work

  1. Notebook Framework

    • The foundation of MATCSSI is a collection of open-source Jupyter Notebooks designed to automate, teach, and document every step of the many-body simulation workflow, from setting up inputs to visualizing results.
    • Each notebook focuses on a physical property, ensuring simplicity and modularity.
    • The very core of MATCSSI, is the EPWpy python library (https://epwpy.org/), which will be discussed in depth in another post.

  2. Use Cases

    • Electronic band structures
    • Optical spectra and excitons
    • Charge carrier mobility and phonon-limited transport
    • Finite-temperature quasiparticles, polarons, and superconductivity

  3. Dual-Purpose Design

    • Learning tool — For self-learning or classroom teaching, on laptops or small servers.
    • Cloud/HPC engine — Designed for scalable deployment on systems like TACC supercomputers, lifting the barriers to large-scale calculations. We offer FREE computing time so you can test our platform!

The codes

  • EPW An open-source community code for ab initio calculations of electron–phonon interactions. Built on Density-Functional Perturbation Theory (DFPT) and Maximally Localized Wannier Functions (MLWFs). It enables accurate modeling of superconductivity, transport, optics and carrier dynamics.

  • BerkeleyGW A state-of-the-art implementation of many-body perturbation theory. It is capable of computing quasiparticle excitation energies and optical properties. Essential for studying excitonic effects and electronic excitations beyond standard DFT.

  • SternheimerGW An approach that uses time-dependent density-functional perturbation theory to compute quasiparticle energies without explicit unoccupied states. Both the Green’s function (G) and screened interaction (W) are obtained by solving linear Sternheimer equations — offering improved computational efficiency and scalability.

What we learned

  • Interoperability in action: By linking existing codes into a coherent workflow, MATCSSI allowed users to explore many-body properties without dealing with code-specific barriers. We have extensivelly employed the platform in international workshops and schools!
  • Scalable and modular: The same workflow can run on a laptop or scale to HPC environments, adapting seamlessly to the user’s resources.
  • Educational impact: Designed to serve as both a research tool and a teaching platform, allowing students and early-career researchers to learn by doing it.

Looking ahead

  • Extend notebook coverage to new codes and emerging many-body methods.
  • Potentially integrate other cloud-based HPC executions.
  • Integrate machine learning approaches.
  • Develop a community library of interoperable notebooks for reproducible materials simulations.

Final word

We believe that with MATCSSI, we’re taking a meaningful step towards accessible, interoperable, and reproducible many-body quantum simulations, to both experienced researchers and students.

Supported by the National Science Foundation (Award #2103991), MATCSSI embodies the vision of a modern computational ecosystem.