Esoteric Twist: An Efficient in-Place Streaming Algorithmus for the Lattice Boltzmann Method on Massively Parallel Hardware

Martin Geier,Martin Schönherr

doi:10.3390/computation5020019

Abstract

We present and analyze the Esoteric Twist algorithm for the Lattice Boltzmann Method. Esoteric Twist is a thread safe in-place streaming method that combines streaming and collision and requires only a single data set. Compared to other in-place streaming techniques, Esoteric Twist minimizes the memory footprint and the memory traffic when indirect addressing is used. Esoteric Twist is particularly suitable for the implementation of the Lattice Boltzmann Method on Graphic Processing Units.

Highlights

Esoteric Twist is suitable for the implementation of the Lattice Boltzmann Method on Graphic Processing Units
The lattice Boltzmann method (LBM) is a simple algorithm used in computational fluid dynamics for solving the Navier–Stokes equations
We presented the Esoteric Twist data structure for efficient thread safe execution of the LBM on massively parallel hardware

Summary

Introduction

The lattice Boltzmann method (LBM) is a simple algorithm used in computational fluid dynamics for solving the Navier–Stokes equations. The efficient implementation of the lattice Boltzmann method has received considerable attention from the computer scientist community [1,2,3]. A particular feature of the LBM in distinction to finite difference schemes is that the number of input variables of the local (i.e., node-wise) time integration scheme equals the number of output variables. This is important as it implies that each input datum is required only once. Most algorithmic and hardware optimizations of recent years focused on the reuse of data

Methods

Results

Conclusion