Solving Partial Differential Equations with Monte Carlo / Random Walk on an Analog-Digital Hybrid Computer

Inhalt:
Current digital computers are about to hit basic physical boundaries with respect to integration density, clock frequencies, and particularly energy consumption. This requires the application of new computing paradigms, such as quantum and analog computing in the near future. Although neither quantum nor analog computer are general purpose computers they will play an important role as co-processors to offload certain classes of compute intensive tasks from classic digital computers, thereby not only reducing run time but also and foremost power consumption. In this work, we describe a random walk approach to the solution of certain types of partial differential equations which is well suited for combinations of digital and analog computers (hybrid computers). The experiments were performed on an Analog Paradigm Model-1 analog computer attached to a digital computer by means of a hybrid interface. At the end we give some estimates of speedups and power consumption obtainable by using future analog computers on chip.