ILVES: Accurate and efficient bond length and angle constraints in molecular dynamics

2025-03-17

AI Summary

Existing SHAKE, LINCS, and P-LINCS algorithms for bond-length and bond-angle constraints in molecular dynamics suffer from slow convergence, limited numerical precision, and—particularly for LINCS-based methods—inability to handle general angular constraints, thereby restricting the permissible time step and compromising simulation fidelity. This work introduces ILVES, a novel parallel constraint-satisfaction algorithm that achieves hardware-precision convergence for the first time. ILVES unifies support for arbitrary bond-length and bond-angle constraints—including anharmonic and many-body angles—overcoming fundamental limitations in both convergence behavior and constraint expressivity inherent to prior methods. By formulating constraints within a consistent dynamical framework and integrating deeply with GROMACS, ILVES delivers ≥2× speedup over state-of-the-art baselines across most systems. It substantially increases the stable integration time step, thereby enabling more accurate and efficient simulation of long-timescale physical processes, such as conformational transitions and the evolution of weak intermolecular interactions.

Technology Category

• Machine Learning: Hardware-aware ML
• Constraint Satisfaction and Optimization: Solvers and Tools
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems

Abstract

Force field-based molecular dynamics simulations are customarily carried out by constraining internal degrees of freedom. The de facto state-of-the-art algorithms for this purpose, SHAKE, LINCS and P-LINCS, converge slowly, impeding high-accuracy calculations and limiting the realism of simulations. Furthermore, LINCS and P-LINCS cannot handle general angular constraints, which restricts increasing the time step. In this paper, we introduce ILVES, a set of parallel algorithms that converge so rapidly that it is now practical to solve bond length and associated angular constraint equations as accurately as the hardware will allow. We have integrated our work into Gromacs and our analysis demonstrates that, in most cases, our software is superior to the state-of-the-art. We anticipate that ILVES will allow for an increase in the time step, thus accelerating contemporary calculations by a factor of at least 2. This will allow the scientific community to increase the range of phenomena that can therefore be simulated.

Problem

• Improves convergence of bond length and angle constraints
• Enables accurate molecular dynamics simulations with larger time steps
• Accelerates simulations by at least a factor of two

Innovation

• ILVES enables rapid convergence for bond constraints.
• ILVES supports general angular constraints in simulations.
• ILVES integrated into Gromacs enhances simulation efficiency.