Coupling Complementary Simulations for Combined Performance and Energy Optimization
2026-06-08 • Distributed, Parallel, and Cluster Computing
Distributed, Parallel, and Cluster Computing
AI summaryⓘ
The authors improved how polymer simulations are done by combining two methods: one that looks at large-scale concentration changes (UDM) and another that handles small-scale particle movements (SOMA). They optimized both methods to run faster on GPUs and created a system to coordinate them working together efficiently across many GPUs. Their approach made the simulations up to 13 times faster and used 96% less energy without losing accuracy. This work shows how combining different tools carefully can save a lot of computing time and power in scientific research.
Polymer simulationsUneyama-Doi Model (UDM)SOft coarse-grained Monte Carlo Acceleration (SOMA)GPU optimizationKernel fusionMonte Carlo dynamicsHigh-performance computingEnergy efficiencyCoarse-grained modeling
Authors
Adel Dabah, Gregor Häfner, Sonja Happ, Simon Pickartz, Marcus Müller, Andreas Herten
Abstract
Polymer simulations are among the most computationally demanding workloads in soft-matter research, often requiring days of execution and high energy consumption to achieve physically meaningful results. In this work, we address these challenges through the coupling and optimization of two complementary simulation frameworks: the Uneyama-Doi Model (UDM) and the SOft coarse-grained Monte Carlo Acceleration (SOMA). UDM efficiently propagates concentration fields at the continuum level, while SOMA resolves chain-scale thermal fluctuations via particle-based Monte Carlo dynamics. Each model was individually optimized for GPU execution using kernel fusion, memory coalescing, asynchronous random-number generation yielding up to 70% (UDM) and 80% (SOMA) performance improvement. The coupling is performed through our proposed coordinator library that orchestrates data exchange and synchronizes time-stepping across multiple GPUs. Further management of coupling workload distribution enabled a 13x overall speedup and 24.5x reduction in total energy usage compared to the SOMA baseline, i. e., 96% energy saving. The proposed hybrid approach maintains the same scientific fidelity while drastically reducing the computational and energy footprint, showcasing the potential of energy-aware, cross-application co-design for sustainable high-performance simulations