Dissertation > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device

Research and Design Ray Tracing Accelerated Algorithm Based on Heterogeneous Multi-core.

Author LiYong
Tutor CaoXueHong
School Nanjing University of Posts and Telecommunications
Course Signal and Information Processing
Keywords Ray Tracing Heterogeneous multicore KD-tree GPGPU
CLC TP391.41
Type Master's thesis
Year 2011
Downloads 57
Quotes 0
Download Dissertation

With the development of multi-core technology , more and more researchers began to pay attention and to proceed with the research and application of multi-core technology . Multi - core technology , multi-core processors and many-core processors (GPU can be seen as one of the manycore processor ) increasing interest and use by researchers . With GPU computing power and programmability improve , has become a cost-effective parallel computing coprocessor , particularly suitable for intensive , high parallel general purpose computing . Researchers began to introduction of intensive calculations based on the CPU to the GPU , GPU huge computing power to speed up the calculation process . In computer graphics , ray tracing algorithm widely used as an important method to generate a realistic image , in the movie and entertainment industry . Ray tracing with a high degree of parallelism conditions , therefore , the use of GPU-accelerated ray tracing algorithm has become a trend. Firstly, built on top of the CPU, GPU and heterogeneous multicore three platforms accelerated ray tracing algorithm structure KD- tree , and that heterogeneous multicore platform is more suitable for building KD-tree acceleration structures . Secondly , in a heterogeneous multi - core platforms to achieve the dynamic ray tracing algorithm completely rebuilt and the analysis of its adaptation . Finally, two dynamic ray tracing algorithm to analyze the scope of adaptation . The main content of this paper is as follows : 1) the basic principles and processes of ray tracing , choose the appropriate acceleration structure , build support CUDA GPU ray tracing model system . 2 ) were built on top of the CPU, GPU and heterogeneous multicore three platforms the acceleration structures KD -tree ; prove the effect of the KD-tree in a heterogeneous multicore platform built by testing the different static three-dimensional scene . 3 ) research and realized after the the completely rebuilt dynamic ray tracing algorithm , proposed the partial reconstruction of dynamic ray tracing and light coordinate transformation dynamic ray tracing algorithm . And the advantages and disadvantages , and heterogeneous multi-core platforms to achieve the performance of three dynamic ray tracing algorithms are analyzed and compared .

Related Dissertations
More Dissertations