Chinese Academy of Science Institute of Computing Technology Key Laboratory of Computer System and Architecture Understanding Parallelism in Graph Traversal on Multi-core Clusters

There is an ever-increasing need for exploring large-scale graph data sets in computational sciences, social networks, and business analytics. However, due to irregular and memory-intensive nature, graph applications are notoriously known for their poor performance on parallel computer systems. In this paper we attempt to present a deep understanding of parallelism in graph traversal on distributed memory architectures with state-of-the-art multi-core processors. We study three parallel breadth-first search (BFS) algorithms with different programming models: MPI-only, BSPstyle hybrid MPI/OpenMP and asynchronous hybrid MPI/Pthreads. The third algorithm is a new one we propose, featuring with (i) overlapping computation and communication by separating them into multiple threads, (ii) maximizing multi-threading parallelism on multi-cores with massive threads to improve throughputs, and (iii) exploiting pipeline parallelism using lock-free queues for asynchronous communication. Experiments show our algorithm is 1.9× faster than the MPI-only program and 2.7× faster than the BSP-style hybrid one, and capable of processing 1.45 billion edges per second on a 32-node SMP cluster.