|
Dr. Alexei V. Vezolainen Dept. of Mechanical Engineering, June 14, 2017 |
 |
SEMINAR ABSTRACT:A parallel adaptive wavelet collocation method for solving a large class of Partial Differential Equations is presented. The parallelization is achieved by developing an asynchronous parallel wavelet transform, which allows to perform wavelet transform and derivative calculations on each processor without additional data synchronization on each level of resolution. The data are stored using tree-like structure with tree roots starting at sufficiently large level of resolution to shorten tree traversing path and to minimize the size of trees for data migration. Dynamic domain partitioning is developed for an efficient handling of non-simply connected partitioning of a computational domain. Dynamic load balancing is achieved via an automated domain repartitioning during grid adaptation step to ensure approximately the same number of wavelets on each processor. The parallel efficiency of the approach depends on the geometry of a specific problem. For a parallel Coherent Vortex Simulations of homogenous turbulence with linear forcing an effective non-adaptive resolutions up to 40963 has been obtained using as many as 4096 CPU cores. |
SPEAKER INTRODUCTION:Dr. Vezolainen conducts research focusing on High performance computing and visualization, Computational geophysics, Fluid dynamics, Shock wave physics, Planetary system formation. He also performs experimental and numerical analysis in the fields of geotechnical engineering (geological and environmental site investigations for construction). MS in Applied Mathematics and Physics, Moscow Institute of Physics and Technology (1997), PhD. in Physics, New Mexico State University (2003), Research Associate at the Department of Mechanical Engineering, University of Colorado at Boulder. Since 2011 Technical director, co-founder of “NPO “GeoVed” Ltd., Saint-Petersburg, Russia. |