In the world today, by the development of computer science, processors have to evaluate more, and harder computing tasks. Today, tasks like high-resolution, HD movie streaming, 4K picture processing and editing, furthermore – by development of virtual reality and the gaming economy – 3D picture processing appear in everyday, consumer electronics. Simple processors, using standard processing methods - despite the growing number of transistors in them – can’t fulfill these requirements. In recent years, one of the system engineer’s main tasks was to develop new methods and devices to correct this defect of processors.
Throughout the writing of this thesis, I get acquainted with parallel and heterogeneous programming’s definition, usage, and advantages. A modern, standardized form of heterogeneous programing is the OpenCL methodology, which describes the necessary requirements for heterogeneous programming’s hardware and software support. The final goal of the thesis is implementing a picture filtering task on the given devices and tools with FPGA support, in an OpenCL system. In practice I used Altera Corporation’s SoC circuit that supports OpenCL, and called Cyclon V, and the Altera’s development kit, Altera SDK for OpenCL. In the end of the thesis I sum up my experiences, and evaluate the practicality of using OpenCL in the modern picture processing tasks.