GENESIS: a hybrid-parallel and multi-scale molecular dynamics simulator with enhanced sampling algorithms for...

Area of Science:

• Computational chemistry and biophysics
• Software development for scientific simulations
• Molecular dynamics and enhanced sampling techniques

Background:

• Molecular dynamics (MD) simulations are crucial for understanding macromolecular behavior.
• Existing MD software may face limitations in handling very large systems and diverse simulation ensembles.
• Efficient parallelization strategies are essential for modern supercomputing architectures.

Purpose of the Study:

• Introduce GENESIS (Generalized-Ensemble Simulation System), a novel software package for MD simulations.
• Highlight the capabilities of its two MD simulators, ATDYN and SPDYN, for various models and ensembles.
• Demonstrate the performance and utility of GENESIS for large-scale biological simulations.

Main Methods:

• ATDYN utilizes atomic decomposition for all-atom and coarse-grained models.
• SP DYN employs domain decomposition for highly parallelized, large-scale simulations on supercomputers.
• Hybrid OpenMP and MPI schemes are implemented for multicore architectures.

Main Results:

• SP DYN achieves high parallel performance for systems exceeding one million atoms.
• GENESIS supports multiple replica exchange molecular dynamics (REMD) algorithms across various ensembles.
• Demonstrated REMD simulation of a mixed lipid bilayer using GENESIS.

Conclusions:

• GENESIS provides a powerful and flexible platform for advanced molecular dynamics simulations.
• Its high performance and diverse REMD capabilities benefit the study of large biological systems.
• GENESIS is freely available open-source software, facilitating further development and application.