Paradigm Shift in Reconfigurable Computing

As per the Moore's law, the number of transistors that could be placed on a single integrated circuit has been increasing exponentially, and today has reached the stage where the implementation of multiple processor cores on a single chip has become a reality. As a result, the capacity of modern FPGA devices will soon support the implementation of hundreds to thousands of soft microprocessor cores on a single chip. With improvements in technology, Moore's law will start applying to processors rather than gates, and will lead to various heterogeneous programmable processors getting embedded on an FPGA. Hence, the complexity levels of the next-generation FPGAs will be simply too high to be handled by low-level hardware design skills and even C-to-gates languages combined with high-level synthesis tools. To date, the direction of the FPGA development has been solely focused on performance, rather than productivity. The theme of this paper is to emphasize the shift in paradigm from the old traditional approach, differentiating between hardware programmers and software programmers, to a new paradigm wherein these boundaries are to be overcome and a unified community of programmers are provided tools to unleash the benefits of the growing potential of reconfigurable hardware using a software controlled environment.