| Title |
Design of a Scalable High-performance Pipeline Offload Engine for Lossless Decompression |
| Authors |
김영일(Youngil Kim) ; 천종민(Jongmin Chun) ; 최승도(Seungdo Choi) ; 이중걸(Jungkeol Lee) ; 송용호(Yong Ho Song) |
| DOI |
https://doi.org/10.5573/ieie.2019.56.8.28 |
| Keywords |
FPGA; Accelerator architectures; Data decompression; Zlib; Inflate algorithm |
| Abstract |
Lossless data compression improves the efficiency of data storage and transmission by reducing the size of the data. Unlike the compression process, decompression has a lot of difficulties in parallelization due to inter-symbol dependency. In this paper, we propose a high performance hardware offload engine for lossless decompression. The proposed offload engine accelerate decompression through parallelization of decompression operations. We use a reordering technique To solve the dependency problem in decompression algorithm that hinder parallelism. The effectiveness of the proposed scheme was evaluated via the implementation of an hardware architecture in the field-programmable gate array (FPGA) platform. Experimental results show that when the proposed offload engine is optimally configured for performance per resource usage, performance is increased by about 116% and hardware resources by about 12.9% compared to the offload engine that processes decompression tasks sequentially. |