Accelerating molecular modeling applications with graphics processors

Abstract

Molecular mechanics simulations offer a computational approach to study the behavior of biomolecules at atomic detail, but such simulations are limited in size and timescale by the available computing resources. Stateof-the-art graphics processing units (GPUs) can perform over 500 billion arithmetic operations per second, a tremendous computational resource that can now be utilized for general purpose computing as a result of recent advances in GPU hardware and software architecture. In this article, an overview of recent advances in programmable GPUs is presented, with an emphasis on their application to molecular mechanics simulations and the programming techniques required to obtain optimal performance in these cases. We demonstrate the use of GPUs for the calculation of long-range electrostatics and nonbonded forces for molecular dynamics simulations, where GPU-based calculations are typically 10-100 times faster than heavily optimized CPU-based implementations. The application of GPU acceleration to biomolecular simulation is also demonstrated through the use of GPU-accelerated Coulomb-based ion placement and calculation of time-averaged potentials from molecular dynamics trajectories. A novel approximation to Coulomb potential calculation, the multilevel summation method, is introduced and compared with direct Coulomb summation. In light of the performance obtained for this set of calculations, future applications of graphics processors to molecular dynamics simulations are discussed.

Physicochemical bases for protein folding, dynamics, and protein-ligand binding

“…Of particular note is that the enhanced sampling methods of MD can be used to calculate free energy differences and then reconstruct the FEL of a protein-solvent system. Computational simulations, coupled with improved computer power and advanced algorithms [23][24][25], have greatly facilitated protein dynamics studies, including studies on the protein folding mechanism, conformational space sampling, molecular motions, and conformational transitions, which in turn enhance significantly our understanding of the relationship between protein dynamics and functions.…”

Fast multipole methods on graphics processors

“…This force calculation is part of a time stepping procedure in which the accelerations are evaluated, while the potential is necessary for energy computations. Recently several researchers have reported the use of GPUs, either in isolation or in a cluster to speed up these computations using direct algorithms, in which the interaction of every pair of particles is considered (see e.g., [18,16,22,28]). While impressive speedups are reported, these algorithms suffer from their O N 2 nature for large N .…”

The impact of accelerator processors for high-throughput molecular modeling and simulation

“…Although suitable for coarse-grained simulations, the lack of support for an atomistic, biomolecular force field limits the applicability of the code. Stone et al [28] demonstrate the GPU-accelerated computation of the electrostatic and van der Waals forces obtaining 10 − 100 times speed-up compared to heavily optimised CPU-based implementations, but because the remainder of the MD simulation remains performed by the host CPU, the net speed-up is reduced to around 5 times. Preliminary results for a fully atomistic MD program called aceMD [29] show a 50 times speed-up measured at the peak performance of a parallel multi-GPU MD code.…”