Paral·lelització del codi del simulador de terratrèmols MudPy amb CUDA

MudPy is an earthquake simulator developed by a team of geologists and geophysicists. It works well an serves its purpose, but it has some weaknesses, the main one being its relative poor performance. This problem limits its usefulness, since a normal simulation can take hours or even days. The main...

Descripción completa

Detalles Bibliográficos
Autor: Coll Carrillo, Marc
Tipo de recurso: tesis de maestría
Fecha de publicación:2021
País:España
Institución:Universitat Oberta de Catalunya (UOC)
Repositorio:O2, repositorio institucional de la UOC
OAI Identifier:oai:openaccess.uoc.edu:10609/132646
Acceso en línea:http://hdl.handle.net/10609/132646
Access Level:acceso abierto
Palabra clave:algorismes paral·lels
computació d'altres prestacions
simulació de terratrèmols
algoritmos paralelos
computación de otras prestaciones
simulación de terremotos
parallel algorithms
high performance computing
earthquake simulation
Algorithms -- TFM
Algorismes -- TFM
Algoritmos -- TFM
Descripción
Sumario:MudPy is an earthquake simulator developed by a team of geologists and geophysicists. It works well an serves its purpose, but it has some weaknesses, the main one being its relative poor performance. This problem limits its usefulness, since a normal simulation can take hours or even days. The main goal of this project is to parallelize part of MudPy's code with CUDA, so that simulations can work faster in those machines where a GPU is available. Since MudPy is mainly written in Python, Numba will be used, which is a platform that allows using all the power of CUDA directly in Python, without having to rewrite the code to C or C++. Other possible strategies to increase the performance are also explored.