Routine Microsecond Molecular Dynamics Simulations with AMBER on GPUs. 1. Generalized Born

By Andreas W. Götz, M. Williamson, Dong Xu, Duncan Poole, S. Le Grand, Ross C. Walker. Journal of Chemical Theory and Computation, 2012

1,977 citations recorded by the source corpus.

We present an implementation of generalized Born implicit solvent all-atom classical molecular dynamics (MD) within the AMBER program package that runs entirely on CUDA enabled NVIDIA graphics processing units (GPUs). We discuss the algorithms that are used to exploit the processing power of the GPUs and show the performance that can be achieved in comparison to simulations on conventional CPU clusters. The implementation supports three different precision models in which the contributions to the forces are calculated in single precision floating point arithmetic but accumulated in double precision (SPDP), or everything is computed in single precision (SPSP) or double precision (DPDP). In addition to performance, we have focused on understanding the implications of the different precision models on the outcome of implicit solvent MD simulations. We show results for a range of tests including the accuracy of single point force evaluations and energy conservation as well as structural properties pertainining to protein dynamics. The numerical noise due to rounding errors within the SPSP precision model is sufficiently large to lead to an accumulation of errors which can result in unphysical trajectories for long time scale simulations. We recommend the use of the mixed-precision SPDP model since the numerical results obtained are comparable with those of the full double precision DPDP model and the reference double precision CPU implementation but at significantly reduced computational cost. Our implementation provides performance for GB simulations on a single desktop that is on par with, and in some cases exceeds, that of traditional supercomputers.

Subfields: Materials Chemistry, Molecular Biology, Spectroscopy

View Paper: https://www.semanticscholar.org/paper/efdcb461e60d398accd1a5db0b4a24e45f06c80f

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Materials Chemistry

In Molecular Biology

In Spectroscopy

Citations in the corpus

Accelerated cryo-EM structure determination with parallelisation using GPUs in RELION-2 (2016, background): “for molecular dynamics simulations (Salomon-Ferrer et al., 2013; Abraham et al., 2015), quantum chemistry (Ufimtsev and Martı́nez, 2008), and machine learning (Jia et al.”
Heterogeneous Parallelization and Acceleration of Molecular Dynamics Simulations in GROMACS (2020, background): “9,13,21,33,37,38 A common approach is to assign different i -particles to each SIMT thread requiring a separate j particle loaded for each pair interaction.”
Molecular dynamics simulations and novel drug discovery (2018, background): “The simulation time can reach the microsecond (μs) and even millisecond (ms) level [167].”
Computational Physics on Graphics Processing Units (2012, background): “[38], and it was shown how just a few GPUs (and even just one) can outperform the same code running on 1024 CPU cores: the weight of the communication between nodes exceeds the benefit of having additional CPU cores, while the few GPUs do not”
SPFP: Speed without compromise - A mixed precision model for GPU accelerated molecular dynamics simulations (2013, methodology): “For GB simulations we have already shown [9] that these errors are sufficiently small such that the SPDP GPU implementation generates trajectories with numerical stability that is equivalent to the CPU implementation.”