Parallelization of a computational fluid dynamics solver
Abstract
This thesis is an effort of accelerating an in-house CFD code, using HPC (High Performance Computing) systems. The profit that is expected is the reduction of solution time and the ability to solve larger problems. In the introduction, previous work is reported and the most important technologies and programming techniques are analyzed. Understanding these technologies lead to the optimal exploitation of the available tools. Next the serial code is described and analyzed before the analysis of the parallelization steps, the memory management and the communication involved that are described in deep afterwards. The parallel code uses the domain decomposition method and the workload is balanced over the processors depending on their computational power. The domain decomposition is applied on simple and complex geometries using one- and multi-block topologies. Investigation of the Altivec vector unit also takes place to achieve extra acceleration. Finally, the results of the parallel code ...
show more
![]() | |
![]() | Download full text in PDF format (7.94 MB)
(Available only to registered users)
|
All items in National Archive of Phd theses are protected by copyright.
|
Usage statistics

VIEWS
Concern the unique Ph.D. Thesis' views for the period 07/2018 - 07/2023.
Source: Google Analytics.
Source: Google Analytics.

ONLINE READER
Concern the online reader's opening for the period 07/2018 - 07/2023.
Source: Google Analytics.
Source: Google Analytics.

DOWNLOADS
Concern all downloads of this Ph.D. Thesis' digital file.
Source: National Archive of Ph.D. Theses.
Source: National Archive of Ph.D. Theses.

USERS
Concern all registered users of National Archive of Ph.D. Theses who have interacted with this Ph.D. Thesis. Mostly, it concerns downloads.
Source: National Archive of Ph.D. Theses.
Source: National Archive of Ph.D. Theses.