CyGraph: A Reconfigurable Architecture for Parallel Breadth-First Search
Osama G. Attia, Tyler Johnson, Kevin R. Townsend, Philip Jones, Joseph Zambreno · 2014
Large-scale graph structures are considered as a keystone for many emerging high-performance computing applications in which Breadth-First Search (BFS) is an important building block. For such graph structures, BFS operations tends to be memory-bound rather than compute-bound. In this paper, we present an efficient reconfigurable architecture for parallel BFS that adopts new optimizations for utilizing memory bandwidth. Our architecture adopts a custom graph representation based on compressed-sparse raw format (CSR), as well as a restructuring of the conventional BFS algorithm. By taking maximum advantage of available memory bandwidth, our architecture continuously keeps our processing elements active. Using a commercial high-performance reconfigurable computing system (the Convey HC-2), our results demonstrate a 5× speedup over previously published FPGA-based implementations.