Inq, a Modern GPU-Accelerated Computational Framework for (Time-Dependent) Density Functional Theory.

Area of Science:

• Computational chemistry and materials science
• High-performance computing
• Quantum mechanics

Background:

• Density Functional Theory (DFT) and Time-Dependent DFT (TDDFT) are crucial for electronic structure calculations.
• Existing implementations often face limitations in performance and modularity on modern hardware.
• The demand for efficient computational tools in scientific research is ever-increasing.

Purpose of the Study:

• To present inq, a novel, from-scratch implementation of DFT and TDDFT.
• To leverage Graphics Processing Units (GPUs) and modern hardware for enhanced computational performance.
• To create a concise, modular, and open-source platform for DFT/TDDFT research.

Main Methods:

• Development of a new DFT/TDDFT code, inq, utilizing C++.
• Implementation of algorithms optimized for Graphics Processing Units (GPUs).
• Adoption of modern code design principles for modularity and conciseness.

Main Results:

• inq provides a complete DFT/TDDFT implementation in approximately 12,000 lines of C++ code.
• The code is designed for efficient execution on GPUs, utilizing modern hardware capabilities.
• The modular design facilitates community-driven development and adaptation.

Conclusions:

• inq represents a significant advancement in GPU-accelerated DFT/TDDFT software.
• Its open-source and modular nature fosters collaborative development for high-performance computing.
• The platform is well-suited for exploring emerging computational architectures in chemistry and materials science.