An OpenCL-based parallel acceleration of aSobel edge detection algorithm Using IntelFPGA technology

This paper examines the feasibility of using commercial out-of-the-box reconfigurable field programmable gate array (FPGA) technology and the open computing language (OpenCL) framework to create an efficient Sobel edge-detection implementation, which is considered a fundamental aspect of image and video processing. This implementation enhances speedup and energy consumption attributes when compared to general single-core processors. We created the proposed approach at a high level of abstraction and executed it on a high commodity Intel FPGA platform (an Intel De5-net device was used). This approach was designed in a manner that allows the high-level compiler/synthesis tool to manipulate a task-parallelism model. The most promising FPGA and conventional implementations were compared to their single-core CPU software equivalents. For these comparisons, local-memory, pipelining, loop unrolling, vectorization, internal channel mechanisms, and memory coalescing were manipulated to provide a much more effective hardware design. The run-time and power consumption attributes were estimated for each implementation, resulting in up to 37-fold improvement of the execution/transfer time and up to a 53-fold improvement in energy consumption when compared to a specific single-core CPU-based implementation.