Research And Design Of A Distributed FPGA Coprocessor Platform

With the development of the Internet and computer technology, whether scientific research task or daily living needs of the masses, the requirements of calculation processing tasks need to be improved continuously and the stress of computer application software to deal with become heavier and heavier. At the same time, there higher requirements for the processing power of computer, with hardware coprocessor joined to be the trends of the times and the demand of social development.As the distributed technology is the dismantling of a large computing task into smaller tasks, with the capable of assign these tasks to multiple processors for processing, so that these tasks can be processed in parallel and the processing efficiency to be greatly improved. As FPGA with features such as high speed, low power, reconfigurable and rich parallel resources, the specialized FPGA coprocessors can take full advantage of FPGA resources and play its advantage of hardware acceleration to improve the overall system efficiency and responsiveness.Meanwhile, with using this distributed in research of FPGA, the hardware advantages of FPGA will make the FPGA coprocessor platform for maximum processing capacity.In this thesis, the FPGA coprocessor platform is launched on the computer network. Under the concept of distributed technologies, we integrate the specialized FPGA coprocessors with different functions to build up. The FPGA coprocessor platform allows the user to send control commands via the browser of PC, enabled AES encryption and decryption coprocessor and storage coprocessor to perform data encryption and decryption and storage capabilities. Construction of the platform is able to transfer some tasks from computer to hardware. On one hand, it can share the task of the computer and improve its processing performance. On the other, tasks of distributed parallel processing use FPGA coprocessor to improve overall efficiency. At last, the FPGA coprocessor platform for maximum processing capacity will be played.