Characterizing Molecular Dynamics Simulation on Commodity Platforms

Abstract

Molecular Dynamics (MD) simulation is an essential tool driving innovation in key scientific domains such as physics, materials science, biochemistry, and drug discovery. Enabling larger, longer, and more accurate MD simulations can directly impact scientific discovery and innovation. While domain-specific architectures for MD exist, they are not widely accessible, and MD performance on commodity platforms (i.e., CPUs and GPUs) remains critical for supporting broad and agile scientific progress.This paper aims at characterizing MD simulation on commodity platforms with a benchmark campaign on modern systems available in public cloud offerings. We focus on LAMMPS, one of the prevalent MD frameworks, and characterize several representative and diverse MD experiments. We find that the benchmarked CPU instance provides good scalability to many cores, while the reference LAMMPS GPU implementation struggles with scaling to multiple devices. Additionally, we evaluate the performance impact of application-specific parameters such as error threshold and arithmetic precision.Our study indicates that key drivers for further improvement of LAMMPS performance on commodity systems are: 1) improving scalability and offload efficiency in multiaccelerator systems and 2) reducing work imbalance in the CPU parallelization. Set initial positions and velocities Initial integration step: Get x(t+Δt), v(t+(½)Δt) Apply boundary conditions Update neighbor list data structures Compute non-bonded forces with an appropriate force field, compute bonded forces

A Bird's Eye View on Quantum Computing: Current and Future Trends

“…Regardless of the specific application's field, more advanced hardware in memory and resources is necessary [55]. Indeed, this problem affects all disciplines more broadly related to molecular dynamics [64], which could substantially benefit from quantum acceleration for the same reasons.…”

AutoParams: An Automated Web-Based Tool To Generate Force Field Parameters for Molecular Dynamics Simulations

“…Many fields of biochemical research have been augmented by the application of theoretical methods in recent years, including drug design and high-throughput compound screening, − protein engineering, − and development of new fluorescent molecules and photoswitches. − As computational resources become more powerful, larger scales of simulation and analysis become possible, whether through sheer quantity of compounds, which may be investigated by a given method, or by improved accuracy and scale of analysis available for any single compound. For example, improvements in the GPU architecture combined with expanded storage capabilities allow for solvated molecular simulations to be performed at the microsecond scale rather than the nanosecond scale, and simulations, which once took months to run, can be completed in less than a week. − Machine learning has further expanded the scope of potential data sets and simulation times available. These improvements to computational capacity lead to increased demand to exploit that capacity in pursuit of a better scientific understanding of complex biochemical problems.…”