 |
|||||||||||||||||
 |
|
 |
|||||||||||||||
 |
|||||||||||||||||
|
Because the present technologies are simply insufficient for real-world applications like real time solutions of Navier-Stokes partial differential equations. Today, using the best supercomputer that we have, CFD and MHD simulations are still thousand to million times slower than real time. Simulation time in the order of hours means that researchers cannot easly try their ideas. For instance, you cannot optimize "trial-and-error" the form of a rearview mirror of car in order to reduce the air friction. High fidely models of controlled fusion reactors needs weeks of simulation time for few seconds of real time. With these limitations, we cannot use model based optimal control techniques. Just to give you an idea of the potential of the quantum computing technology, consider the present (2016) state-of-the-art of silicon high performance microprocessors build by Intel. Consider two figures: transistor lenght (L=10 nm) and maximum frequency (FMAX=5 GHz). Taking the Feynman paper as reference (1984/85), using single silicon atom as quantum transistor, the new numbers are L=0.1 nm and FMAX=150THz). This means an improvement factor of 100 on linear dimension and 30,000 on operating frequency. Combining these two improvement factors, the global figure of merit of the quantum computing technology vs the actual state-of-the-art is F = 3.000.000 Therefore, a quantum computer can be three million better of a classic silicon IC dinosaur. Evolution, my friends. In ten years the top performance Intel workstation that I'm using right now, it will be an IT dynosaur, good only for a museum. |
 |
 |
|
| Site Map |