high performance computing on graphics processing units: hgpu.org

The goal of this thesis has been to develop a fast simulator for large-scale migration of CO2 in saline aquifers. We have also focused on being able to let the CO2 storage atlas from the Norwegian Petroleum Directorate specify the reservoir properties. In order to meet the demands of simulating on large data sets combined with high performance, we have investigated the possibilities of using graphic processing units (GPUs) to accelerate the computations.The Intergovernmental Panel on Climate Change, IPCC, has considered CO2 to be one of the main factors influencing the climate changes of today. Capture and storage of CO2 is one of the strategies which could reduce the amount of CO2 released into the atmosphere. However, there are still uncertainties related to flow of CO2 in saline aquifers. It is therefore necessary with fast simulators which can predict this behavior to minimize the risks involved in a storage project.GPUs are initially designed to accelerate graphic operations. As opposed to standard CPUs, where most of the transistor capacity is used on advanced logic, the GPU uses most of its transistors on floating point operations in parallel. This results in that the theoretical upper bound for floating point operations are 7-10 times higher on the GPU than the CPU. Thus, GPUs have shown to be a strong tool when solving hyperbolic conservation laws using stencil based schemes, as a large amount of the computations can be parallelized. In compliance with the storage atlases we have based our simulator on structured grids. Our numerical scheme consists of a finite volume method combined with an explicit Euler method.