TurboGenius: Python suite for high-throughput calculations of ab initio quantum Monte Carlo methods.

Area of Science:

• Computational Chemistry
• Quantum Mechanics
• Materials Science

Background:

• Ab initio quantum Monte Carlo (QMC) methods are crucial for accurate electronic structure calculations.
• High-throughput computational approaches accelerate scientific discovery.
• Controlling complex QMC jobs requires specialized software tools.

Purpose of the Study:

• To introduce and demonstrate the capabilities of the open-source TurboGenius Python package.
• To facilitate high-throughput ab initio QMC calculations.
• To validate and benchmark the TurboRVB package through integrated workflows.

Main Methods:

• Development of the TurboGenius Python package for QMC job control.
• Implementation of the TurboWorkflows package for constructing calculation workflows.
• Validation of Density Functional Theory (DFT) and QMC drivers against established quantum chemistry packages (PySCF, Quantum Package).
• Benchmarking of Diffusion Monte Carlo (DMC) calculations on standard datasets (G2, S22, A24, SCAI, cubic crystals).

Main Results:

• TurboGenius enables seamless high-throughput QMC calculations.
• DFT and Hartree-Fock energies computed with TurboRVB show excellent agreement with PySCF and Quantum Package.
• DMC calculations using the LDA nodal surface provide satisfactory atomization energies, binding energies, and lattice parameters, consistent with reference data.

Conclusions:

• The TurboGenius and TurboWorkflows packages offer a robust platform for automated, high-throughput QMC studies.
• Validation confirms the reliability of TurboRVB's DFT and QMC implementations.
• DMC calculations with the LDA nodal surface are accurate for predicting molecular and solid-state properties.