作者: Shuang Qiu , Zonghao Feng , Qiong Luo
DOI: 10.1007/978-3-030-18590-9_68
关键词:
摘要: De Bruijn graph traversal is a critical step in de novo assemblers. It uses the structure to analyze genome sequences and both memory space intensive time consuming. To improve efficiency, we develop ParaGraph, which parallelizes on cluster of GPU-equipped computer nodes. With effective vertex partitioning fine-grained parallel algorithms, ParaGraph utilizes all cores each CPU GPU, CPUs GPUs node, nodes cluster. Our results show that able traverse billion-node graphs within three minutes six an order magnitude faster than state-of-the-art shared based assemblers, more five times current distributed