Nearest Neighbors Search Using Multi-GPU

Meshless methods are increasingly gaining space in the study of electromagnetic phenomena as an alternative to traditional mesh-based methods. One of their biggest advantages is the absence of a mesh to describe the simulation domain. Instead, the domain discretization is done by spreading nodes along the domain and its boundaries. Thus, meshless methods are based on the interactions of each node with all its neighbors, and determining the neighborhood of the nodes becomes a fundamental task. The k-nearest neighbors (kNN) is a well-known algorithm used for this purpose, but it becomes a bottleneck for these methods due to its high computational cost. One of the alternatives to reduce the kNN high computational cost is to use spatial partitioning data structures (e.g., planar grid) that allow pruning when performing the k-nearest neighbors search. Furthermore, many of these strategies employed for kNN search have been adapted for graphics processing units (GPUs) and can take advantage of its high potential for parallelism. Thus, this paper proposes a multi-GPU version of the grid method for solving the kNN problem. It was possible to achieve a speedup of up to 1.99x and up to 3.94x using two and four GPUs, respectively, when compared against the single-GPU version of the grid method.

On applicability of MQ-RPIM and MLPG meshless methods with 3D extended-enriched base functions for estimation of mode I stress intensity factor and fatigue crack growth in cyclic tensile and bending load of an un-notched and notched shaft

In this study, a numerical meshless method is used to solve the weak form of the linear elastic equations in solid mechanics. Evaluation and comparison of the numerical meshless methods have been carried out via the radial point interpolation meshless method with multi-quadrics base functions (MQ-RPIM) and meshless local Petrov-Galerkin method (MLPG). Using these two methods, stress intensity factors in an elastic medium containing geometric discontinuities and cracks are estimated based on tensile and bending cyclic loading. The analysis domain has been identified via three-dimensional modeling of the notched and un-notched shafts with an initial surface semi-elliptical crack subjected to tensile or bending cyclic loadings. To enhance the accuracy of calculations, the RPIM meshless method is applied using polynomial and extended-enriched 3D base functions. Shape functions have been developed using standard and optimal parameters and values with Mono-Objective Function in PSO algorithm. In the MLPG meshless method with the extended-enriched functions, discretization is performed via direct and penalty factor methods, to reach more efficient results and meet the boundary conditions. Efficiency comparison of the selected numerical methods with the experimental findings and the numerical analysis of finite elements method indicates that in comparison with the MLPG method, MQ-RPIM enriched meshless method can be utilized with fewer nodes in the analysis domain while reaching the accuracy and convergence with lower stress intensity factors and gentler slope. However, the processing time of the MLPG meshless method is lower than that of the other methods.

Application of meshless methods to modeling of pistol bullets

The paper presents an analysis of applicability of two meshless methods to the modeling of pistol bullets on the example of a 9 mm Parabellum. The studies included the following methods: SPH and SPG. The results of computer simulations were confronted with ballistic test results in terms of shape-dimensional compliance of the deformed projectile. The relative error of the projectile diameter was 15 and 17% for the SPG and SPH methods, respectively. The deformation form for the SPH method deviated from the ballistic test results, while the SPG method faithfully reproduced the shape of the deformed projectile. Keywords: mechanical engineering, impact simulation, pistol bullet, SPH, SPG