High-Level Overview

KRONOS (Kohn-Residual Optimized Numerics Over Silicon) is a research-grade, ab initio plane-wave pseudopotential DFT engine targeting periodic crystalline systems. It solves the Kohn-Sham equations self-consistently using norm-conserving pseudopotentials, LDA and GGA exchange-correlation functionals, and a GPU abstraction layer for CUDA/HIP offloading. This page describes the outputs, scope, and internal conventions of the engine before diving into the SCF flowchart, component diagram, and algorithms in subsequent pages.

KRONOS (Kohn-Residual Optimized Numerics Over Silicon) is a research-grade, ab initio plane-wave pseudopotential Density Functional Theory (DFT) engine. It solves the Kohn-Sham equations self-consistently to compute:

• Ground-state total energy and its decomposition (kinetic, Hartree, XC, local PP, nonlocal PP, Ewald ion-ion)
• Kohn-Sham eigenvalues and band structure
• Electron density on real-space and reciprocal-space grids
• Hellmann-Feynman ionic forces
• Density of states

KRONOS targets norm-conserving pseudopotentials (NCPP) with LDA and GGA exchange-correlation functionals. It reads standard UPF v2 pseudopotential files and YAML input, and writes JSON summaries and HDF5 binary output.

Within the DFT ecosystem, KRONOS occupies the same niche as Quantum ESPRESSO's PWscf module -- a plane-wave code operating in reciprocal space with periodic boundary conditions. It is designed for periodic crystalline systems (bulk solids, surfaces, 2D materials with vacuum padding) using the pseudopotential approximation to replace core electrons.

All internal quantities use Rydberg atomic units (energies in Ry, lengths in bohr). The code is written in C++20 with a GPU abstraction layer for future CUDA/HIP offloading.