DocumentCode :
2173685
Title :
A Parallel Preconditioned Conjugate Gradient Solver for the Poisson Problem on a Multi-GPU Platform
Author :
Ament, M. ; Knittel, G. ; Weiskopf, D. ; Strasser, W.
Author_Institution :
VISUS Visualization Res. Center, Univ. Stuttgart, Stuttgart, Germany
fYear :
2010
fDate :
17-19 Feb. 2010
Firstpage :
583
Lastpage :
592
Abstract :
We present a parallel conjugate gradient solver for the Poisson problem optimized for multi-GPU platforms. Our approach includes a novel heuristic Poisson preconditioner well suited for massively-parallel SIMD processing. Furthermore, we address the problem of limited transfer rates over typical data channels such as the PCI-express bus relative to the bandwidth requirements of powerful GPUs. Specifically, naive communication schemes can severely reduce the achievable speedup in such communication-intense algorithms. For this reason, we employ overlapping memory transfers to establish a high level of concurrency and to improve scalability. We have implemented our model on a high-performance workstation with multiple hardware accelerators. We discuss the mathematical principles, give implementation details, and present the performance and the scalability of the system.
Keywords :
Poisson equation; computer graphic equipment; conjugate gradient methods; coprocessors; parallel processing; workstations; Poisson problem; SIMD processing; communication intense algorithms; hardware accelerators; heuristic Poisson preconditioner; high performance workstation; multi GPU platform; overlapping memory transfers; parallel preconditioned conjugate gradient solver; system scalability; Bandwidth; Character generation; Concurrent computing; Graphics; Hardware; Iterative methods; Jacobian matrices; Linear systems; Scalability; Sparse matrices; Conjugate Gradient; Multi-GPU; Parallel Preconditioning; Poisson Problem;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Parallel, Distributed and Network-Based Processing (PDP), 2010 18th Euromicro International Conference on
Conference_Location :
Pisa
ISSN :
1066-6192
Print_ISBN :
978-1-4244-5672-7
Electronic_ISBN :
1066-6192
Type :
conf
DOI :
10.1109/PDP.2010.51
Filename :
5452414
Link To Document :
بازگشت