Accelerating CUDA graph algorithms at maximum warp
Proceedings of the 16th ACM symposium on Principles and practice of parallel programming
CULZSS: LZSS Lossless Data Compression on CUDA
CLUSTER '11 Proceedings of the 2011 IEEE International Conference on Cluster Computing
A universal algorithm for sequential data compression
IEEE Transactions on Information Theory
Parallel suffix array and least common prefix for the GPU
Proceedings of the 18th ACM SIGPLAN symposium on Principles and practice of parallel programming
Pipelined Parallel LZSS for Streaming Data Compression on GPGPUs
ICPADS '12 Proceedings of the 2012 IEEE 18th International Conference on Parallel and Distributed Systems
Practical Parallel Lempel-Ziv Factorization
DCC '13 Proceedings of the 2013 Data Compression Conference
Hi-index | 0.00 |
The need for data compression has grown for better utilization of network bandwidth and data storage space. LZ77 is the most widely used data compression method, which has many variants in practical applications. The biggest obstacle that prevents data compression from being used in high-speed applications is its high computation overhead. In this paper, we focus on parallelizing LZSS that is a derivative of LZ77 on GPU using the NVIDIA CUDA framework to improve the compression speed. Based on in-depth understanding of LZSS's dictionary-based compression mechanism and GPU's architectural features, we propose an effective method to parallelize LZSS compression algorithm on GPU. The biggest merit of this method is that it eliminates threads serialization by carefully redesign the algorithm process. Experiments on an NVIDIA GTX 590 machine with 13 benchmark files from real world demonstrate the effectiveness of our method, which achieves 2x speedup over existing work.