A hardware architecture is proposed which allows direct mapping of des
ign simulation topology onto an acceleration platform. In order to cla
rify architectural principles, the simulation is confined to functiona
l verification of unit delay, binary valued gate level logic designs.
Under this approach, a rank ordered design description is executed on
a massively parallel processor grid which implements an efficient and
direct model of the design, similar to prototyping. Architectural inno
vation reduces logic complexity and execution time of boolean evaluati
on and fanout switching circuits, while large scale parallelism is int
egrated at die level to reduce cost and communication delays. The resu
lts of this research form the basis for a multiple order of magnitude
improvement in reported state-of-the-art cost-performance merit for ha
rdware gate level simulation accelerators.