Chapter 14. The Future of FPGA Computing
Throughout most of this book we've presented FPGAs as relatively small computing elements capable of accelerating key algorithms for embedded systems and, to some extent, for more esoteric scientific applications (the fractal image generator being one such example). FPGAs have become increasingly useful for such applications as the cost of the devices has fallen and as their densities have increased. You have also seen how the use of software programming methods, and C programming in particular, can simplify the creation of hardware-accelerated systemssystems in which the FPGA serves as a computational resource alongside other, more traditional processing resources.
Looking forward, it seems increasingly likely that applications requiring even higher levels of performancesupercomputing applicationswill come to rely on FPGAs or FPGA-like devices to provide greater levels of raw computing power. This trend toward FPGA-based supercomputing is increasing in practicality to the point where major supercomputing platform vendors and research teams are investing substantially in FPGA-related research. As of this writing, two major vendors of high-performance computing platforms (Cray and Silicon Graphics) have begun offering FPGA-based computers. This trend will only intensify as the benefits of FPGA computing become more apparent and more widely accepted.