Linear algebra operators for GPU implementation of numerical algorithms
ACM SIGGRAPH 2003 Papers
Scalable Parallel Programming with CUDA
Queue - GPU Computing
Stencil computation optimization and auto-tuning on state-of-the-art multicore architectures
Proceedings of the 2008 ACM/IEEE conference on Supercomputing
Tuned and wildly asynchronous stencil kernels for hybrid CPU/GPU systems
Proceedings of the 23rd international conference on Supercomputing
A CUDA-Based Implementation of Stable Fluids in 3D with Internal and Moving Boundaries
ICCSA '10 Proceedings of the 2010 International Conference on Computational Science and Its Applications
Programming Massively Parallel Processors: A Hands-on Approach
Programming Massively Parallel Processors: A Hands-on Approach
Parallel SOR for solving the convection diffusion equation using GPUs with CUDA
Euro-Par'12 Proceedings of the 18th international conference on Parallel Processing
The performance model for a parallel SOR algorithm using the red-black scheme
International Journal of High Performance Systems Architecture
| Hi-index | 0.00 |
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.