Simulation of 1+1 dimensional surface growth and lattices gases using GPUs

Restricted solid on solid surface growth models can be mapped onto binary lattice gases. We show that efficient simulation algorithms can be realized on GPUs either by CUDA or by OpenCL programming. We consider a deposition/evaporation model following Kardar-Parisi-Zhang growth in 1+1 dimensions related to the Asymmetric Simple Exclusion Process and show that for sizes, that fit into the shared memory of GPUs one can achieve the maximum parallelization speedup ~ x100 for a Quadro FX 5800 graphics card with respect to a single CPU of 2.67 GHz). This permits us to study the effect of quenched columnar disorder, requiring extremely long simulation times. We compare the CUDA realization with an OpenCL implementation designed for processor clusters via MPI. A two-lane traffic model with randomized turning points is also realized and the dynamical behavior has been investigated.

💡 Research Summary

This paper explores the efficient simulation of 1+1 dimensional surface growth models and binary lattice gases using high-performance computing techniques on GPUs. The authors demonstrate that by mapping restricted solid-on-solid (RSOS) surface growth models onto binary lattice gases, they can achieve significant speedups through GPU parallelization. Specifically, for a deposition/evaporation model related to the Kardar-Parisi-Zhang (KPZ) growth dynamics and the Asymmetric Simple Exclusion Process (ASEP), simulations on a Quadro FX 5800 graphics card show up to a 100x speedup compared to a single CPU running at 2.67 GHz, when fitting into the GPU’s shared memory.

The paper compares two parallelization approaches: CUDA and OpenCL programming. The authors also present an MPI-based OpenCL implementation designed for processor clusters, allowing them to evaluate performance across different computing environments. This comparison helps in understanding how various parallelization strategies can be optimized on GPUs versus CPU clusters.

Furthermore, the study investigates the dynamical behavior of a two-lane traffic model with randomized turning points, which is another complex system that benefits from GPU acceleration due to its computational intensity and need for long simulation times. The paper highlights the importance of such simulations in studying quenched columnar disorder effects, where extremely long simulation durations are required.

Overall, this work showcases how advanced computing techniques can significantly enhance the efficiency of simulating complex systems, particularly those involving surface growth dynamics and lattice gas models. It provides valuable insights into optimizing GPU usage for scientific computations and offers a comparative analysis between different parallelization methods, contributing to the broader field of high-performance computing in materials science and statistical physics.