As Graphics Processing Unit (GPU) is emerging as the general purpose computing devices, it is becoming more and more popular and powerful computing paradigm for conducting the numerical computation tasks that are traditionally undertaken by using Central Processing Unit (CPU). In this paper, the accelerated water quality (WQ) analysis solver is developed for efficient water quality simulation on GPU. The GPU-accelerated solver is implemented by using Open Compute Language (OpenCL), an industry standard for heterogeneous computing, to ensure the portability of the parallelized solvers on various hardware vendors' devices. With the parallelized WQ solvers, water quality simulations can be accelerated with the massive computing threads on a GPU. The parallelized solver has been tested with the large water distribution system (WDS) models on different GPUs, including those available from the public cloud computing platforms such as Amazon EC2, which offers very affordable and appealing computing capability that may not be available with some low-end GPU-enabled PCs. The performance analysis and comparison are presented in the paper for both conventional CPU-based and the GPU-accelerated water quality solvers.
展开▼