Key points are not available for this paper at this time.
Quantum mechanics permits certain kinds of non-local effects. This paper demonstrates how these can be used for distributed computation with minimal communication between various processors. The problem considered is that of estimating the mean of N items to a certain precision. First a serial quantum mechanical algorithm for this is presented that is faster than any classical algorithm. Next it is shown how this can be efficiently parallelized with quantum mechanical processors that are remotely located. These processors consist of coupled EPR particles. Each processor has just to communicate one bit of classical information to a central location at the end of its local computation.
Lov K. Grover (Mon,) studied this question.