GPU Accelerated Implementation of Density Functional Theory for Hybrid QM/MM Simulations.

Area of Science:

• Computational Chemistry
• Molecular Dynamics
• Quantum Mechanics

Background:

• Hybrid simulation tools, such as quantum mechanics/molecular mechanics (QM/MM), are essential for studying chemical reactivity.
• Complex environments pose significant challenges for traditional simulation methods.

Purpose of the Study:

• To present an optimized implementation of density functional theory (DFT) for hybrid simulations.
• To leverage graphic processing units (GPUs) for accelerating computationally intensive parts of QM/MM simulations.

Main Methods:

• Implementation of DFT for electronic structure calculations.
• Optimization using GPUs for exchange-correlation terms in QM/MM simulations.
• Extensive testing for numerical quality and performance.

Main Results:

• Achieved acceleration of 20-30 times faster than CPU versions for relevant computational portions.
• Demonstrated performance and numerical quality across systems of varying size and composition.

Conclusions:

• The GPU-optimized implementation significantly enhances the speed of QM/MM simulations.
• This development provides a more efficient tool for investigating chemical reactivity in complex systems.