KSSOLV

Official Resources

• Homepage: http://kssolv.org/
• Documentation: http://kssolv.org/ (and included manuals)
• Source Repository: Distributed via site / CPC Library
• License: GPL (General Public License) or similar open academic license

Overview

KSSOLV (Kohn-Sham SOLVer) is a MATLAB toolbox designed for solving the Kohn-Sham equations for electronic structure calculations. It utilizes a plane-wave basis set and pseudopotentials. While originally developed for prototyping and educational purposes, KSSOLV 2.0 has evolved into a capable tool for research, enabling algorithm development and direct comparison with standard codes like Quantum ESPRESSO.

Scientific domain: Algorithm development, educational DFT, electronic structure Target user community: Developers of DFT algorithms, students learning DFT, researchers needing a flexible prototyping environment

Theoretical Methods

• Density Functional Theory (DFT)
• Plane-wave basis sets
• Norm-conserving pseudopotentials
• Kohn-Sham formulation
• Self-Consistent Field (SCF) methods
• Linear response TDDFT
• GW approximation (in advanced modules)

Capabilities

• Ground-state energy and density
• Geometry optimization and atomic relaxation
• Ab initio Molecular Dynamics (AIMD)
• Time-Dependent DFT (TDDFT) - real-time and linear response
• Phonon calculations
• GPU acceleration (KSSOLV-GPU)
• Visualization of wavefunctions and densities (via MATLAB)

Key Strengths

Ease of Development:

• Written entirely in MATLAB
• Object-oriented design makes code readable and modifiable
• Ideal for testing new SCF mixers, preconditioners, or eigensolvers

Accessibility:

• leverages MATLAB's robust linear algebra libraries
• Immediate visualization and debugging capabilities within MATLAB environment

Modern Features:

• Version 2.0 includes significant performance and feature upgrades
• GPU support for accelerated calculations

Inputs & Outputs

• Input formats:

  • MATLAB scripts/functions defining the system
  • Pseudopotential files

• Output data types:

  • MATLAB data structures (energies, forces, density arrays)
  • Plots and visualizations directly in MATLAB figures

Interfaces & Ecosystem

• MATLAB: Fully integrated into the MATLAB ecosystem.
• Comparison: Can use compatible pseudopotentials to compare results with compiled codes.

Computational Cost

• Standard: Slower than compiled codes (C/Fortran) due to MATLAB overhead, though core linear algebra is optimized (BLAS/LAPACK).
• GPU: KSSOLV 2.0 GPU implementation offers significant speedups, making it viable for medium-sized research problems.
• Memory: Higher memory footprint than optimized Fortran codes.

Best Practices

Prototyping Workflow:

• Small Systems: Develop and test new algorithms on small molecular systems (e.g., silane, benzene) before scaling up.
• Profiling: Use MATLAB's built-in Profiler (profile on, profile viewer) to identify bottlenecks in new implementations.

Performance:

• Vectorization: When extending KSSOLV, ensure custom code is fully vectorized to maintain MATLAB performance.
• GPU Usage: If available, utilize the KSSOLV-GPU branch for production-grade SCF cycles.

Community and Support

• Hosting: Distributed via academic sites and CPC Library.
• Support: Limited to academic correspondence with authors (C. Yang group).
• Documentation: Includes PDFs and inline MATLAB help comments.

Performance Characteristics

• Speed: Generally slower than optimized Fortran/C++ codes (VASP, QE) for production runs, but KSSOLV-GPU offers significant speedups.
• Efficiency: Very high "developer efficiency" for implementing new ideas.
• Scaling: Limited compared to massive MPI codes, but sufficient for small-to-medium systems and method development.

Limitations & Known Constraints

• Language: Requires MATLAB license (proprietary).
• Performance: Not intended for massive high-throughput production runs on supercomputers.
• Scope: Feature set is smaller than mature packages like VASP/QE.

Comparison with Other Codes

• vs Quantum ESPRESSO: KSSOLV is for prototyping algorithms that might later be implemented in QE. QE is for production.
• vs DFT++: Both emphasize modularity/education, but KSSOLV uses MATLAB while DFT++ uses C++.

Verification & Sources

Primary sources:

1. Official Website: http://kssolv.org/
2. Yang et al., "KSSOLV 2.0..." (arXiv/Journal publications)
3. "KSSOLV - a MATLAB toolbox for solving the Kohn-Sham equations"

Confidence: CONFIRMED - Well-documented academic software.

Verification status: ✅ VERIFIED

• Existence: CONFIRMED
• Domain: DFT/MATLAB
• Key Feature: Prototyping/Education