Maybe in the next 5 years, when Moore's law may run out http://en.wikipedia.org/wiki/5_nanometer (unless Intel et al. have some extraourdinary technology up their sleeves) we will finally see the advent of FPGA's?
Without the annual growth of transistor density, the second best avenue to gain performance will probably be specialization, and reconfigurable specialized hardware looks more attractive than fixed one.
Without the annual growth of transistor density, the second best avenue to gain performance will probably be specialization, and reconfigurable specialized hardware looks more attractive than fixed one.