Application specific co-processor architectures can increase the performance for compute intensive algorithms and tasks. High-level design tools have a great potential for rapid development cycles, so they can save valuable human resources and cause faster time-to-market for specific designs and products. The OpenCL is an open standard for heterogeneous computation systems, thus gaining more and more attention and popularity in that field.
At the beginning of the thesis we gain an insight of high performance computing. This is followed by the detailed description of the OpenCL standard. After that, I investigate the capabilities of the Altera and Xilinx OpenCL development environment. Afterwards, I measured the performance of some application examples and I implemented a complex lane tracking driver assistance algorithm accelerated by FPGA. The solution is completely integrated on a Terasic DE1-SoC development board.
The purpose of the thesis is to get an overview of applying OpenCL technology on FPGA hardware in the means of performance and efficiency.