Biclique counting is a crucial challenge in algorithmic research with broad applications. The author proposes GBC, leveraging GPU parallelism and advanced techniques to achieve significant speedups in (p, q)-biclique counting. The HTB data structure and Border vertex reordering optimize memory usage and intersection computations. Load balancing strategies ensure equitable distribution of workloads among threads.
Counting bicliques presents challenges due to exponential growth concerning p and q. The proposed GBC algorithm outperforms existing solutions by achieving remarkable speedups. By introducing HTB and Border techniques, the author optimizes intersection computations and memory usage. Pre-runtime task allocation and runtime task stealing enhance load balancing for improved efficiency.
To Another Language
from source content
arxiv.org
Głębsze pytania