Reconfigurable computing paradox
From Wikipedia, the free encyclopedia
The reconfigurable computing paradox, defined by Reiner Hartenstein, notes that replacing a supercomputer with a FPGA results in a speed-up factor of up to four orders of magnitude even though the clock speed of the FPGA is substantially lower and the effective integration density falls behind Moore's law by up to 4 orders of magnitude.[1] In addition, although FPGAs are power-hungry, the reduction of the electricity bill to less than 10% has been reported when migrating from a supercomputer to a FPGA application in the area of oil and gas. An even a higher potential in saving energy can be expected for the near future.
The von Neumann syndrome explains the reconfigurable computing paradox.
