An MPI-CUDA implementation for the compression of DEM

A high performance terrain data compression method is proposed based on discrete wavelet transform (DWT) and parallel run-length code. But the implementation of the schemes to solve these models in realistic scenarios imposes huge demands of computing power. Compute Unified Device Architecture (CUDA) programmed, Graphic Processing Units (GPUs) are rapidly becoming a major choice in high performance computing. Hence, the number of applications ported to the CUDA platform is growing high. Message Passing Interface (MPI) has been the choice of high performance computing for more than a decade and it has proven its capability in delivering higher performance in parallel applications. CUDA and MPI use different programming approaches but both of them depend on the inherent parallelism of the application to be effective. In this approach, MPI functions as the data distributing mechanism between the GPU nodes and CUDA as the main computing engine. This allows the programmer to connect GPU nodes via high speed Ethernet without special technologies. We tackle the acceleration of the compression of digital elevation models (DEM) by exploiting the combined power of several CUDA-enabled GPUs in a GPU cluster. This implementation overlaps MPI communication with CPU-GPU memory transfers and GPU computation to increase efficiency. Several numerical experiments, performed on a cluster of modern CUDA-enabled GPUs, show the efficiency of the distributed solver. Our speed-up was over 20 compared to two thread CPU version.