Skip to Main Content
This paper proposes GAPLA: a globally asynchronous locally synchronous programmable logic array architecture. The whole FPGA area is divided into locally synchronous blocks wrapped with asynchronous I/O interfaces. Data communications between synchronous blocks are controlled by 2-phase handshaking signals under bundled-data delay assumption. The size and shape of each locally synchronous block are programmable so that different modules in a design can be effectively implemented. By dividing the FPGA area into smaller blocks, the delays of long interconnect wires, which could easily dominate other delays in conventional FPGAs, only come into picture when there are communications between blocks. Therefore, each block could run at higher speed. The area overhead of adopting the GALS style in GAPLA architecture is estimated to be very small (about 7%). Experimental results show an up to 55% performance improvement compared to the conventional FPGAs.