Software simulation of microprocessors can be implemented very efficiently using functional-first simulation, together with an attached cycle-approximate model. However, the modeling of cache memories can slow down simulation significantly. The Zynq 7000 series of FPGA devices from Xilinx incorporate a dual-core ARM Cortex A9 system, capable of running Linux, with an attached FPGA fabric. In this project the A9 cores run the software side of the simulator, and offload the cache simulation to the FPGA fabric. The aim is to discover the extent to which simulation speeds can be improved, in comparison with a software-only simulation running on a desktop.