PWDFT

Official Resources

• Homepage: https://github.com/ebylaska/PWDFT
• Source Repository: https://github.com/ebylaska/PWDFT
• Developer: Eric J. Bylaska (Pacific Northwest National Laboratory)
• License: Open Source (BSD-style or similar research license implied)

Overview

PWDFT is a plane-wave density functional theory (DFT) code developed by Eric J. Bylaska at Pacific Northwest National Laboratory (PNNL). It serves as a research platform and mini-application for exploring high-performance computing algorithms, particularly in the context of plane-wave basis sets and pseudopotentials. It is related to the development of NWChem (where Bylaska is a key developer) but exists as a standalone repository for testing and development purposes.

Scientific domain: Plane-wave DFT, electronic structure, high-performance computing (HPC)
Target user community: Method developers, HPC researchers, NWChem contributors

Theoretical Methods

• Kohn-Sham Density Functional Theory
• Plane-wave basis set
• Pseudopotentials (Norm-conserving)
• PSP (Pseudopotential) format support
• Parallel 3D FFTs
• Lagrange multiplier constraints
• Conjugate gradient minimization

Capabilities

• Ground-state total energy calculations
• Electronic structure minimization
• Wavefunction optimization
• Parallel execution using MPI
• Performance benchmarking for FFTs and parallel data structures
• Mini-app for testing HPC architectures

Key Strengths

HPC Research

• Used as a testbed for parallel algorithms
• Investigates scalability of plane-wave methods
• Optimized for memory-bound operations (FFT)

Connection to NWChem

• Developed by a lead author of NWChem's PSPW module
• Likely shares algorithmic heritage with NWChem's plane-wave implementation

Inputs & Outputs

• Input formats:
  • Input parameter files
  • Pseudopotential files (formatted)
• Output data types:
  • Standard output (Energy, convergence)
  • Wavefunction files

Computational Cost

• Research Code: Optimized for benchmarking FFTs and parallel communications.
• Scaling: Designed to test limits of HPC scaling; not optimized for production throughput like NWChem.

Best Practices

• Usage: Use for testing new algorithms or compiling on novel architectures (Cray/GPU).
• Production: Use NWChem for actual science production runs.

Comparison with Other Codes

• vs NWChem: PWDFT is a smaller, standalone research code/mini-app, while NWChem is a full production suite.
• vs PWDFT.jl: PWDFT.jl is a separate educational project in Julia by F. Fathurrahman; this PWDFT is the C++/Fortran research code by Bylaska.

Verification & Sources

• Primary Source: GitHub repository (https://github.com/ebylaska/PWDFT)
• Developer Info: Eric Bylaska (PNNL)
• Confidence: VERIFIED - Code exists and matches the Master List entry.