We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
图数据压缩格式对广度优先搜索算法 影响的特性化分析.
- Authors
任 含; 邓军勇
- Abstract
With the rapid growth of graph computing requirements, the same graph algorithm had different effects on memory access and performance when processing data in different compression formats. Five different data compression formats COO, CSC, CSR, DCSC and CSCI of the two graph data were used as the input of the BFS algorithm. Performance analysis and comparison were carried out, and different selections were based on the performance characteristics. Data in compressed format to achieve the best results. The performance analysis parameters for BFS processing in five different data compression formats mainly included execution time, data movement amount, calculation amount and power consumption etc., and correlation analysis was carried out for performance and power consumption. Experimental results showed that the BFS algorithm had different performance indicators for processing different data compression formats. Among them, CSR and CSC were relatively better in terms of data movement, calculation operation, execution time and power consumption. In terms of effectively improving the cache hit rate index, the DCSC compression format was the best. In terms of memory usage, the DCSC compression format had a relatively small memory usage. A basis for how to choose different compression formats according to different performance characteristics in the BFS algorithm was provided by this conclusion.
- Publication
Journal of Zhengzhou University (Natural Science Edition), 2021, Vol 53, Issue 2, p26
- ISSN
1671-6841
- Publication type
Article
- DOI
10.13705/j.issn.1671-6841.2020260