Heterogeneous Parallelization and Acceleration of Molecular Dynamics Simulations in GROMACS

Abstract

The introduction of accelerator devices such as graphics processing units (GPUs) has had profound impact on molecular dynamics simulations and has enabled order-of-magnitude performance advances using commodity hardware. To fully reap these benefits, it has been necessary to reformulate some of the most fundamental algorithms, including the Verlet list, pair searching and cut-offs. Here, we present the heterogeneous parallelization and acceleration design of molecular dynamics implemented in the GROMACS codebase over the last decade. The setup involves a general cluster-based approach to pair lists and non-bonded pair interactions that utilizes both GPUs and CPU SIMD acceleration efficiently, including the ability to load-balance tasks between CPUs and GPUs. The algorithm work efficiency is tuned for each type of hardware, and to use accelerators more efficiently we introduce dual pair lists with rolling pruning updates. Combined with new direct GPU-GPU communication as well as GPU integration, this enables excellent performance from single GPU simulations through strong scaling across multiple GPUs and efficient multi-node parallelization.

The following article has been submitted to the Journal of Chemical Physics

References in corpus (10)

• A flexible algorithm for calculating pair interactions on SIMD architectures
• Strong scaling of general-purpose molecular dynamics simulations on GPUs
• More Bang for Your Buck: Improved use of GPU Nodes for GROMACS 2018
• Best bang for your buck: GPU nodes for GROMACS biomolecular simulations
• Harvesting graphics power for MD simulations
• Accelerated weight histogram method for exploring free energy landscapes
• RUMD: A general purpose molecular dynamics package optimized to utilize GPU hardware down to a few thousand particles
• Efficiency of linked cell algorithms
• Fully Integrated On-FPGA Molecular Dynamics Simulations
• On the Feasibility of FPGA Acceleration of Molecular Dynamics Simulations

Cited by in corpus (21)

• GPUMD: A package for constructing accurate machine-learned potentials and performing highly efficient atomistic simulations
• GROMACS in the cloud: A global supercomputer to speed up alchemical drug design
• Tinker-HP : Accelerating Molecular Dynamics Simulations of Large Complex Systems with Advanced Point Dipole Polarizable Force Fields using GPUs and Multi-GPUs systems
• JARVIS-Leaderboard: A Large Scale Benchmark of Materials Design Methods
• Synaptic-Like Plasticity in 2D Nanofluidic Memristor from Competitive Bicationic Transport
• Constant pH Simulation with FMM Electrostatics in GROMACS. (A) Design and Applications
• MiMiC: A High-Performance Framework for Multiscale Molecular Dynamics Simulations
• Multiscale Biomolecular Simulations in the Exascale Era
• Application-Driven Exascale: The JUPITER Benchmark Suite
• GROMACS on AMD GPU-Based HPC Platforms: Using SYCL for Performance and Portability
• Breaking the mold: overcoming the time constraints of molecular dynamics on general-purpose hardware
• Phonon density of states of magnetite (\ce{Fe3O4}) nanoparticles via molecular dynamics simulations
• STable AutoCorrelation Integral Estimator (STACIE): Robust and accurate transport properties from molecular dynamics simulations
• Redesigning GROMACS Halo Exchange: Improving Strong Scaling with GPU-initiated NVSHMEM
• Stencil Computations on AMD and Nvidia Graphics Processors: Performance and Tuning Strategies
• Interaction-Driven Giant Electrostatic Modulation of Ion Permeation in Atomically Small Capillaries
• OpenMM-MiMiC Interface for Efficient and Flexible Multiscale Simulations
• Entropic modulation of divalent cation transport
• A Reversible Unwrapping Algorithm for Constant Pressure Molecular Dynamics Simulations
• Improving the Efficiency of Variationally Enhanced Sampling with Wavelet-Based Bias Potentials
• Phenylalanine modification in plasma-driven biocatalysis revealed by solvent accessibility and reactive dynamics in combination with protein mass spectrometry