Finite Element Parallel Computations on a Beowulf Cluster CDF Applications
Abstract
It is well known that to solve the 21 st century challenge problems we need to increase the computational resources In the developed countries this difficulty is circumvented by using parallel computation. Due to our limited budget we have no access to supercomputers to
follow the fast development in the computional mechanics field. One of the few feasible current alternatives to follow that tendency is introduced by the concept of cluster,
specially that based on domestic microprocessor, called Beowulf cluster, due to its high performance/prize ratio. Due to this fact we have developed our cluster at CIMEC. In previous papers we have shown the main features of our cluster including a performance analysis to measure the efficiency of this parallel architecture. This work focus on simple CFD applications in order to validate our parallel codes.
follow the fast development in the computional mechanics field. One of the few feasible current alternatives to follow that tendency is introduced by the concept of cluster,
specially that based on domestic microprocessor, called Beowulf cluster, due to its high performance/prize ratio. Due to this fact we have developed our cluster at CIMEC. In previous papers we have shown the main features of our cluster including a performance analysis to measure the efficiency of this parallel architecture. This work focus on simple CFD applications in order to validate our parallel codes.
Full Text:
PDFAsociación Argentina de Mecánica Computacional
Güemes 3450
S3000GLN Santa Fe, Argentina
Phone: 54-342-4511594 / 4511595 Int. 1006
Fax: 54-342-4511169
E-mail: amca(at)santafe-conicet.gov.ar
ISSN 2591-3522