Comparison of an Explicit and Implicit Time Integration Method on GPUs for Shallow Water Flows on Structured Grids

The accuracy, stability and computational efficiency of numerical methods on central processing units (CPUs) for the depth-averaged shallow water equations were well covered in literature. A large number these already developed compared. However, graphics (GPUs), such comparisons are relatively scarce. In this paper, we present results comparing two time-integration structured grids. An explicit a semi-implicit time integration method considered. For method, performance several iterative solvers was implementation GPU study novel approach equations. This also holds repeated red black (RRB) solver that found to be very efficient GPU. Additionally, both compared with CPU-based software systems flows On GPU, simulations 25 75 times faster than CPU. Theory predicts an best suited due higher level inherent parallelism. It ran efficiently applications, preferred because condition step not restrictive. deep expect preferred.