Abstract
This work presents our strategy, applied optimizations and results in our effort to exploit the computational capabilities of GPUs under the CUDA environment in solving the Laplacian PDE. The parallelizable red/black SOR method was used. Additionally, a program for the CPU, featuring OpenMP, was developed as a performance reference. Significant performance improvements were achieved by using optimization methods which proved to have substantial speedup in performance. Eventually, a direct comparison of performance of both versions was realised. A 51x speedup was measured for the CUDA version over the CPU version, exceeding 134GB/sec bandwidth. Memory access patterns prove to be a critical factor in efficient program execution on GPUs and it is, therefore, appropriate to follow data reorganization in order to achieve the highest feasible memory throughput.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
NVidia CUDA Reference Manual v. 3.1 NVidia (2010)
NVidia CUDA Toolkit 3.1 CUBLAS Library NVidia (2010)
NVidia CUDA C Best Practices Guide Version 3.1 NVidia (2010)
The OpenCL Specification Khronos group (2009)
OpenMP Application Program Interface version 3.0 OpenMP Architecture Review Board (2008)
Kirk, D.B., Hwu, W.W.: Programming Massively Parallel Processors. Morgan Kaufmann (2009)
Burden, R.L., Faires, D.: Numerical Analysis, 7th edn. Brooks Cole (2000)
Nickolls, J., Buck, I., Garland, M., Skadron, K.: Scalable Parallel Programming with CUDA. ACM Queue 6(2), 40–53 (2008)
Amorim, R., Haase, G., Liebmann, M., Weber, R.: Comparing CUDA and OpenGL implementations for a Jacobi iteration SpezialForschungsBereich F 32, 025 (December 2008)
Ha, L., Krger, J., Joshi, S., Silva, C.T.: Multiscale Unbiased Diffeomorphic Atlas Construction on Multi-GPUs. GPU Computing Gems. Morgan Kaufmann (2011)
Komatsu, K., Soga, T., Egawa, R., Takizawa, H., Kobayashi, H., Takahashi, S., Sasaki, D., Nakahashi, K.: Parallel processing of the Building-Cube Method on a GPU platform Computers & Fluids (2011) (in press, corrected proof)
Cecilia, J.M., Garcìa, J.M., Ujaldòn, M.: CUDA 2D Stencil Computations for the Jacobi Method Para 2010 - State of the Art in Scientific and Parallel Computing (2010)
Amador, G., Gomes, A.: A CUDA-based Implementation of Stable Fluids in 3D with Internal and Moving Boundaries. In: 2010 International Conference on Computational Science and Its Applications (2010)
Amador, G., Gomes, A.: CUDA-based Linear Solvers for Stable Fluids. In: International Conference on Information Science and Applications, ICISA (2010)
Venkatasubramanian, S., Vuduc, R.W.: Tuned and wildly asynchronous stencil kernels for hybrid CPU/GPU systems. In: 23rd International Conference on Supercomputing (2009)
Datta, K., Murphy, M., Volkovand, V., Williams, S., Carter, J., Oliker, L., Patterson, D., Shalf, J., Yelick, K.: Stencil computation optimization and auto-tuning on state-of-the-art multicore architectures. In: ACM/IEEE Conference on Supercomputing (2008)
Krüger, J., Westermann, R.: Linear algebra operators for GPU implementation of numerical algorithms. In: ACM SIGGRAPH 2003 (2003)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2012 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Konstantinidis, E., Cotronis, Y. (2012). Accelerating the Red/Black SOR Method Using GPUs with CUDA. In: Wyrzykowski, R., Dongarra, J., Karczewski, K., Waśniewski, J. (eds) Parallel Processing and Applied Mathematics. PPAM 2011. Lecture Notes in Computer Science, vol 7203. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-31464-3_60
Download citation
DOI: https://doi.org/10.1007/978-3-642-31464-3_60
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-31463-6
Online ISBN: 978-3-642-31464-3
eBook Packages: Computer ScienceComputer Science (R0)