Evaluation of a spectral element CFD code on parallel architectures

We present an evaluation of a Spectral Element Navier-Stokes solver on three generations of parallel architecture. The parallel platforms used are a network of Transputers, an Intel iPSC/860 and a Meiko CS2. While performance on the older Transputer and iPSC/860 machines is fairly predictable, the b...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Parallel computing Ročník 21; číslo 7; s. 1137 - 1150
Hlavní autoři: Floros, N., Reeve, J.S.
Médium: Journal Article
Jazyk:angličtina
Vydáno: Amsterdam Elsevier B.V 1995
Elsevier
Témata:
Applied sciences
Computational fluid dynamics
Computer science; control theory; systems
Computer systems and distributed systems. User interface
Data storage equipment
Distributed computer systems
Distributed memory multiprocessor
Exact sciences and technology
Hierarchical systems
Hypercube
Incompressible Navier-Stokes
Multiprocessing systems
Navier Stokes equations
Problem solving
Simulation
Software
Transputer array
Transputers
Computational fluid dynamics
Distributed memory multiprocessor
Transputer array
Incompressible Navier-Stokes
Hypercube
Evaluation
Distributed memory multiprocessor system
Fluid dynamics
Transputer
Navier Stokes equation
Incompressible fluid
ISSN:0167-8191, 1872-7336
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:We present an evaluation of a Spectral Element Navier-Stokes solver on three generations of parallel architecture. The parallel platforms used are a network of Transputers, an Intel iPSC/860 and a Meiko CS2. While performance on the older Transputer and iPSC/860 machines is fairly predictable, the behaviour of the code on the SPARC-10 based Meiko CS2 demonstrated the influence of the internal memory hierachy on performance. These effects give the code problem dependent performance parameters that makes scaling properties difficult to quantify. Our code however, seems future proof, provided sufficiently large problems can be found to overcome the relatively large communications costs with respect to the processor speed.
Bibliografie:ObjectType-Article-2
SourceType-Scholarly Journals-1
ObjectType-Feature-1
content type line 23
ISSN:0167-8191
1872-7336
DOI:10.1016/0167-8191(95)00010-L