2012
DOI: 10.1002/cpe.2917
|View full text |Cite
|
Sign up to set email alerts
|

A multi‐GPU shallow‐water simulation with transport of contaminants

Abstract: This work presents cost-effective multi-graphics processing unit (GPU) parallel implementations of a finitevolume numerical scheme for solving pollutant transport problems in bidimensional domains. The fluid is modeled by 2D shallow-water equations, whereas the transport of pollutant is modeled by a transport equation. The 2D domain is discretized using a first-order Roe finite-volume scheme. Specifically, this paper presents multi-GPU implementations of both a solution that exploits recomputation on the GPU a… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
10
0

Year Published

2013
2013
2018
2018

Publication Types

Select...
6

Relationship

1
5

Authors

Journals

citations
Cited by 17 publications
(10 citation statements)
references
References 12 publications
0
10
0
Order By: Relevance
“…Each cell can be dead or alive and following several behavior rules each cell changes its state during the game. Finally, SHWA is a shallow water simulator with pollutant transport presented in Viñas et al This iterative benchmark needs three stages/kernels to compute the evolution of a mesh of finite volumes: in the first one, the flux variation among the elements of the mesh is computed. Then, the global time step is computed for each iteration.…”
Section: Discussionmentioning
confidence: 99%
See 3 more Smart Citations
“…Each cell can be dead or alive and following several behavior rules each cell changes its state during the game. Finally, SHWA is a shallow water simulator with pollutant transport presented in Viñas et al This iterative benchmark needs three stages/kernels to compute the evolution of a mesh of finite volumes: in the first one, the flux variation among the elements of the mesh is computed. Then, the global time step is computed for each iteration.…”
Section: Discussionmentioning
confidence: 99%
“…For this reason, the kernel does not involve any template. Regarding SHWA, although it can benefit from local memory, the implementation using the generic stencil templates we developed would be less efficient than a manual implementation, and thus, the kernel used was written manually. There are two main reasons why the stencil templates are not a good match for the stencil kernel of this problem.…”
Section: Discussionmentioning
confidence: 99%
See 2 more Smart Citations
“…These structures need to be refreshed with the most recent version of the data they replicate after each update and before the next round of computations begins. When accelerators are used, this gives place to a data exchange between them [25]. Another example are pipelines, in which data proceed through a series of tasks that transform them and handle the results to the next task in the sequence.…”
Section: Multi-device Support In Hplmentioning
confidence: 99%