In this thesis the feasibility of a GPGPU (general-purpose computing on graphics processing units) approach to natural feature description on mobile phone GPUs is assessed. To this end, the SURF descriptor [4] has been implemented with OpenGL ES 2.0/GLSL ES 1.0 and evaluated across different mobile devices. The implementation is multiple times faster than a comparable CPU variant on the same device. The results proof the feasibility of modern mobile graphics accelerators for GPGPU tasks especially for the detection phase in natural feature tracking used in augmented reality applications. Extensive analysis and benchmarking of this approach in comparison to state of the art methods have been undertaken. Insights into the modifications necessary to adapt and modify the SURF algorithm to the limitations of a mobile GPU are presented. Further, an outlook for a GPGPU-based tracking pipeline on a mobile device is provided.
This thesis imparts a general view of the mechanics and implementation of latest voxelization strategies using the GPU. In addition to established voxelization procedures using the rasterization pipeline, new possibilities arising from GPGPU programming are examined. On the basis of the programming language C++ and the graphics library OpenGL the implementation of several methods is explained. The methods are compared in terms of performance and quality of the resulting voxelization and are evaluated critically with regards to possible use cases. Furthermore, two exemplary applications are detailed that use a voxelized scene in such a way that the augmentation of established techniques of real time graphics are facilitated. To this end, the concepts and the implementations of Transmittance Shadow Mapping and of Reflective Shadow Mapping utilizing a voxel based ambient occlusion effect is explained. Finally, the prolonging relevance of voxelization is put into prospect, by addressing latest research and further enhancements and applications of the presented methods.
In this bachelor thesis a code for astrophysical self-gravitating fluid
simulation is developed. The code runs mainly on the GPU. Minimal
simplifications of the physical model and some parameters for accuracy
and tuning allow simulations to be performed at interactive framerates
on most modern consumer grade computers that feature a dedicated
graphics card. It is used to simulate the birth of stars from a turbulent
molecular cloud. Multiple features of star formation, like accretion
discs and fragmentation, can be observed in the simulation, even when
low particle counts are used.
Ray Tracing enables a close to reality rendering implementation of a modelled scene. Because of its functioning, it is able to display optical phenomena and complex lighting. Though, numerous computations per pixel have to be done. In practice implementations can not achieve computer graphics" aim of real-time rendering close to 60 frames per second. Current Graphics Processing Units (GPU) allows high execution parallelism of general-purpose computations. By using the graphics-API OpenGL this parallelism can be achieved and it is possible to design and realize a Ray-Tracer, which operates entirely on the GPU. The developed approach will be extended by an Uniform Grid - a Ray-Tracing acceleration structure. Hence, a speed-up is expected.
This thesis` purpose is the implementation of Ray-Tracer, which operates completely on the GPU, and its expansion by integrating an Uniform Grid. Afterwards, the evaluation of maximum achievable performance takes place. Possible problems regarding GPU-programming will be identified and analysed.
