School of Computer Science Intranet
W.J.Bainbridge, W.B.Toms, D.A.Edwards, S.B.Furber
m-of-n codes can be used for carrying data over self-timed on-chip interconnect links. Such codes can be chosen to have low redundancy, but the costs of encoding/decoding data is high. The key to enabling the cost-effective use of m-of-n codes is to find a suitable mapping of the binary data to the code.
This paper presents a new method for selecting suitable mappings through the decomposition of the complex m-of-n code into an incomplete m-of-n code constructed from groups of smaller, simpler m-of-n and 1-of-n codes. The circuits used both for completion detection and for encoding/decoding such incomplete codes show reduced logic size and delay compared to their full m-of-n counterparts. The improvements mean that the incomplete m-of-n codes become attractive for use in on-chip interconnects and network-on-chip designs.