Bit stream compression using Self-Adaptive EFDR coding

摘要

With the rapid increase in the scale and complexity of circuits and systems, the storage of large amounts of output bit stream, which is the sequence of 0s and 1s, has become a severe problem. As the cost of data storage is high, typically for FPGA reconfiguration stream, a particular compression algorithm needs to be introduced to minimize the data size. However, current compression schemes either compromise to hardware overhead or lack adaption to different cases. This paper is to propose an improved scheme. Dictionary compression scheme based on bitmask is implemented on the source stream to find the analogous segments in the bit stream. Meanwhile, we make the selection and use run-length coding to achieve further compression. This paper is to present a novel Self-Adaptive Extended Frequency-Directed Run-Length (SA-EFDR) coding to improve the compression ratio by 15% in best case compared to Golomb coding. Also, our coding can make self-adaption to different cases. Improvement has been verified with typical benchmarks from ISCAS'89 benchmarks.