A General Hardware Design Model for Multicontext FPGAs

We propose a general multicontext hardware design model to promote the use of multicontext FPGAs in wide range of applications. There are two major problems to be overcome in multicontext FPGAs programming. One is the complexity of programming an application in multicontext. The design model provides the context scheduling mechanisms to free the programmers from the low-level context management. The other is the difficulty in scheduling non-algorithmic applications. Our distributed demand-driven dynamicsc heduler facilitates their implementation. We have applied the design model to a network interface controller Martini. The result supports the effectiveness of our design model.