QMC2

Official Resources

• Homepage: https://github.com/jorgehog/QMC2
• Source Repository: https://github.com/jorgehog/QMC2
• License: Open Source (Implicit/Check Repo)

Overview

QMC2 is a C++ implementation of quantum Monte Carlo methods, specifically focusing on an efficient Diffusion Monte Carlo (DMC) algorithm. Developed by jorgehog, it combines a high-performance C++ core with flexible Python scripting for setup and analysis. It is designed for calculating ground state properties of atoms, molecules, and extended systems with high precision.

Scientific domain: Quantum Chemistry, Electronic Structure, Many-Body Physics Target user community: Developers and researchers in QMC methods

Theoretical Methods

• Diffusion Monte Carlo (DMC): Real-space fixed-node DMC for projecting ground states.
• Variational Monte Carlo (VMC): Optimization of trial wavefunctions.
• Trial Wavefunctions: Supports Slater-Jastrow type wavefunctions.
• Basis Sets: Implements analytical potentials and basis functions (Gaussian/Slater).

Capabilities (CRITICAL)

• Efficient Solver: Optimized C++ implementation of the DMC random walk.
• Parallelization: MPI-based parallelism for scaling across multiple cores/nodes.
• Custom Potentials: Extensible class structure (Potentials subclassing) to define new physical systems.
• Orbital Management: OrbitalGenerator (via SymPy) and Orbitals classes for handling complex basis sets and mappings.
• Visualization: Integration points for DCViz and PySide for observing walkers/densities (optional GUI components).

Key Features

Hybrid Architecture:

• C++ Core: Heavy numerical lifting (random walks, wavefunction evaluation) in compiled C++.
• Python Layer: Simulation setup, parameter definition, and result analysis in Python.

External Libraries:

• Armadillo: For efficient linear algebra operations.
• Boost: For general utilities and mathematical functions.
• MPI: For parallel implementation (Optional).

Extensibility:

• designed for users to plug in new hamiltonians or basis sets without rewriting the core engine.

Inputs & Outputs

• Input formats:
  • Python scripts defining the system geometry and trial function parameters.
  • XML/JSON configurations (if applicable).
• Output data types:
  • Text/Binary logs of energies, variance, and acceptance ratios.
  • Checkpoint files for restarting walkers.

Interfaces & Ecosystem

• Dependencies: Armadillo, Boost, MPI, Python headers.
• Visualization: Connects with DCViz for data visualization.

Workflow and Usage

Users typically compile the C++ backend and then write a Python script that imports the QMC2 module. The script defines the atoms, basis set, and QMC parameters (time step, walkers), then invokes the VMC optimization followed by the DMC propagation.

Performance Characteristics

• Speed: Efficient C++ memory management and math kernels.
• Scaling: Linearly scalable with number of MPI processes (walker parallelism).

Comparison with Other Codes

Verification & Sources

Primary sources:

1. GitHub Repository: https://github.com/jorgehog/QMC2

Verification status: ✅ VERIFIED

• Source code: OPEN
• Status: Research code, active in past years.
• Focus: Efficient C++ DMC implementation.