A scalable interface-resolved simulation of particle-laden flow using the lattice Boltzmann method

• Published in: Parallel computing 2017-09, Vol.67, p.20-37
• Main Authors: Geneva, Nicholas, Peng, Cheng, Li, Xiaoming, Wang, Lian-Ping
• Format: Article
• Language: English
• Subjects: Code optimization; Data structure; Interface-resolved direct numerical simulation; Lattice-Boltzmann method; Particle-laden turbulent flow; Scalable performance
• ISSN: 0167-8191; 1872-7336
• DOI: 10.1016/j.parco.2017.07.005

•We examine the scalable implementation of the lattice Boltzmann method (LBM) in the context of interface-resolved direct numerical simulation of wall-bounded turbulent particle-laden flows.•Three distinct aspects relevant to performance optimization of our lattice Boltzmann simulation are studied: (a) fused implementation of the LBM core substeps; (b) impact of data structure; (c) communication strategy associated with the treatment of fluid-solid interactions.•Study of five different algorithms to carry out the LBM core substeps and optimization of data structure led to a 5–6 times speed-up of the single-phase flow code.•This combined with optimized communication strategy at the solid-fluid interface led to 4.0–6.75 times speed-up of the interface-resolved particle-laden flow code. We examine the scalable implementation of the lattice Boltzmann method (LBM) in the context of interface-resolved simulation of wall-bounded particle-laden flows. Three distinct aspects relevant to performance optimization of our lattice Boltzmann simulation are studied. First, we optimize the core sub-steps of LBM, the collision and the propagation (or streaming) sub-steps, by reviewing and implementing five different published algorithms to reduce memory loading and storing requirements to boost performance. For each, two different array storage formats are benchmarked to test effective cache utilization. Second, the vectorization of the multiple-relaxation-time collision model is discussed and our vectorized collision and propagation algorithm is presented. We find that careful use of Intel’s Advance Vector Extensions and appropriate array storage formats can significantly enhance performance. Third, in the presence of many finite-size, moving solid particles within the flow field, three different communication schemes are proposed and compared in order to optimize the treatment of fluid-solid interactions. These efforts together lead to a very efficient LBM simulation code for interface-resolved simulation of particle-laden flows. Overall, the optimized scalable code of particle-laden flow is a factor of 4.0-to-8.5 times faster than our previous implementation.