Home › Scientific Tools DB › 1. GROUND-STATE DFT › 1.8 Educational / Lightweight DFT

PyFock

**PyFock** is a modern, pure-Python implementation of Density Functional Theory (DFT) and Hartree-Fock (HF) methods. Designed with transparency and hackability in mind, it leverages **JIT compilation (Numba)** and optionally **GPU accele…

1. GROUND-STATE DFT 1.8 Educational / Lightweight DFT VERIFIED
Back to Mind Map Official Website

Overview

**PyFock** is a modern, pure-Python implementation of Density Functional Theory (DFT) and Hartree-Fock (HF) methods. Designed with transparency and hackability in mind, it leverages **JIT compilation (Numba)** and optionally **GPU acceleration (CuPy)** to achieve performance comparable to compiled C++ codes while maintaining the simplicity of a Python codebase.

Reference Papers

Reference papers are not yet linked for this code.

Full Documentation

Official Resources

  • Homepage: https://github.com/manassharma07/PyFock
  • Source Repository: https://github.com/manassharma07/PyFock
  • Developer: Manas Sharma
  • License: MIT License

Overview

PyFock is a modern, pure-Python implementation of Density Functional Theory (DFT) and Hartree-Fock (HF) methods. Designed with transparency and hackability in mind, it leverages JIT compilation (Numba) and optionally GPU acceleration (CuPy) to achieve performance comparable to compiled C++ codes while maintaining the simplicity of a Python codebase.

Scientific domain: Algorithm Development, Education, Specialized Python-based QM. Target user community: Developers, students, and researchers prototyping new methods.

Theoretical Methods

  • Methods: Restricted and Unrestricted Hartree-Fock (RHF/UHF), DFT (RKS/UKS).
  • Functionals: LDA, GGA (PBE, BLYP) via LibXC.
  • Integrals: Obara-Saika scheme implemented in Python/Numba.
  • DIIS: Direct Inversion in the Iterative Subspace for convergence.

Capabilities

  • Transparency: Every step of the SCF cycle is visible in readable Python.
  • Acceleration: Can switch between NumPy (CPU) and CuPy (GPU) backends.
  • Visualization: Built-in simple GUI for results (PyFock-GUI).

Key Strengths

  • Educational Value: Excellent for understanding how a DFT code is built from scratch.
  • Modern Stack: Demonstrates high-performance Python computing (JIT/GPU).
  • Open Source: MIT licensed and actively hackable.

Comparison with Other Codes

  • vs PySCF: PySCF is the industry standard for Python QC, but it relies on C modules for heavy lifting. PyFock is pure Python (facilitated by JIT).
  • vs Psi4: Psi4 is C++ with Python bindings.

Performance Characteristics

  • Precision: Validated decision better than $10^{-7}$ Hartree compared to PySCF.
  • Scaling: Near-quadratic $O(N^{2.05})$ for ERIs due to Cauchy-Schwarz screening.
  • GPU Speedup: Can achieve 14x-20x speedup over CPU (4-core) execution using CuPy backend for large systems.

Limitations & Known Constraints

  • Features: Currently lacks analytical gradients (geometry optimization is limited/numerical), periodic boundary conditions, and hybrid functionals (roadmap items).
  • Scope: Primarily for single-point energy and algorithm demonstrations, not for production dynamics or solid-state physics.

Best Practices

  • Backend: Use numpy backend for small debugging cases (traceable), switch to cupy (GPU) for exploring larger system performance.
  • Memory: Density fitting is enabled by default to save memory; disable only for debugging raw integrals.

Community and Support

  • Support: Community-driven via GitHub Issues.
  • Tutorials: Author provides detailed blog posts (BragitOff.com) explaining the code structure.

Verification & Sources

Primary sources:

  1. Repository: PyFock GitHub
  2. Author Blog: BragitOff.com

Verification status: ✅ VERIFIED

Related Tools in 1.8 Educational / Lightweight DFT

PyFock tinydft DFT++