Parallelization and improvements of the generalized born model with a simple sWitching function for modern graphics...

Area of Science:

• Computational chemistry
• Biophysics
• Molecular dynamics

Background:

• Simulating biologically relevant systems faces challenges in rapid solvation energy calculation and simulation trajectory length.
• The Generalized Born model with a Simple Switching function (GBSW) approximates Poisson-Boltzmann (PB) theory for efficient solvation free energy and force calculations without explicit solvent.

Purpose of the Study:

• To present a parallel refactoring of the GBSW algorithm for implementation on graphics chips.
• To significantly enhance the speed of solvation calculations and molecular dynamics simulations.

Main Methods:

• Parallel refactoring of the GBSW algorithm.
• Implementation on low-cost graphics processing units (GPUs) with thousands of cores.
• Validation of accuracy against previous implementations.

Main Results:

• Achieved speed increases of one to two orders of magnitude over previous GBSW implementations.
• Demonstrated linear scaling of the algorithm with system size, improving cost-effectiveness for large systems.
• Successfully utilized the GPU-accelerated GBSW model for folding studies of the chignolin system.

Conclusions:

• The GPU-accelerated GBSW model offers significant speed enhancements for molecular simulations.
• This advancement makes folding studies of peptides and small proteins more accessible.
• The model provides a cost-effective solution for solvating large biological systems.