Abstract: One or more domains are independently clocked with separate clocks. Each clock is an asynchronous stop/start clock implementing a self-tuning clocking methodology. Domain circuit speed is monitored and the clock adjusted to tune the domain to run at near maximum speed. Inter-domain data transfers are performed by a four-way handshaking method. In effect the clock period of the respective clocks during the data transfer becomes the slower period of the two domains' clock periods. An inter-domain arbiter is implemented at each domain for deciding which domain's request is to be granted during an immediate clock period. Data input to a domain is tracked to determine when data is present. When no data is present, the domain's clock is stopped.

Abstract: A field programmable gate array (FPGA) including both routing and logic blocks (RLBs) and routing and arbiter blocks (RABs) is disclosed. The RABs are periodically placed throughout the FPGA and operate either to arbitrate the arrival of simultaneous signals or to synchronize simultaneous signals. In addition, each of the RLBs are capable of operating in accordance with two clock signals and an asynchronous initialization. The combination of the RLBs and RABs allow the FPGA to operate synchronously and asynchronously.