Rigid-CLL: avoiding constant-distance computations in cell linked-lists algorithms.

Area of Science:

• Computational Chemistry
• Molecular Dynamics
• Biophysics

Background:

• Efficient identification of nonbonded interacting atoms is crucial for molecular simulations.
• Standard cell linked-list methods, while asymptotically optimal, compute many unnecessary distances.
• Existing improvements focus on refining atomic influence volumes.

Purpose of the Study:

• To propose a novel improvement strategy for cell linked-list algorithms in molecular simulations.
• To reduce the number of distance computations by intelligently managing cell occupancy.
• To enhance computational efficiency, particularly for systems with rigid body dynamics.

Main Methods:

• A new cell linked-list technique is presented that avoids populating cells with atoms at constant distances.
• This method optimizes the order of operations within the standard algorithm.
• The strategy is particularly effective for simulations involving rigid molecular groups.

Main Results:

• The proposed technique can decrease distance computations by over an order of magnitude compared to standard methods.
• Significant computational savings are achieved without additional computational costs.
• The method complements existing approaches that bound atomic influence volumes.

Conclusions:

• The optimized cell linked-list approach offers substantial performance gains for molecular simulations.
• This method provides a computationally inexpensive way to accelerate simulations involving rigid bodies.
• The strategy is compatible with other cell linked-list enhancement techniques.