Maintainable, Sustainable, and Generalizable Data Layouts and Vectorization for Rigid-Body Molecular Dynamics

ls1-MarDyn (ls1) is an MD simulator designed for large-scale simulations of multi-site molecules and has been successfully used in a variety of scientific studies. It represents molecules as rigid bodies composed of multiple interaction sites that each exert forces on their neighbours, which are determined by those molecules with centre-of-masses within some cutoff distance. Past works have integrated the modern algorithm-selection particle simulation library, AutoPas, into ls1, replacing its older particle container and force calculation mechanisms. However, the limitations of AutoPas confine this integration to only single-site molecules. Extending these concepts to multi-site molecular representations introduces additional challenges, such as enabling efficient vectorization. At the same time, this highlights a broader issue, namely how to maintain high performance while ensuring long-term code sustainability across different hardware architectures. To address this, we focus on improving the maintainability and portability of vectorized force calculations. Instead of relying on manually specialized implementations for each instruction set, we introduce Google Highway as a portable SIMD abstraction layer. This approach significantly reduces code complexity while retaining competitive performance. We present initial results and discuss practical limitations.     «

ls1-MarDyn (ls1) is an MD simulator designed for large-scale simulations of multi-site molecules and has been successfully used in a variety of scientific studies. It represents molecules as rigid bodies composed of multiple interaction sites that each exert forces on their neighbours, which are determined by those molecules with centre-of-masses within some cutoff distance. Past works have integrated the modern algorithm-selection particle simulation library, AutoPas, into ls1, replacing it...     »