Adaptive Work-Efficient Connected Components on the GPU

2016/12/04 by Michael Sutton, Tal Ben-Nun, Sutton, Michael +7
Computer Science · #Data Structures and Algorithms (cs.DS) #Distributed #FOS: Computer and information sciences #Parallel #and Cluster Computing (cs.DC) #cs.DC #cs.DS

paper · pdf · doi:10.48550/arxiv.1612.01178

4 pages

arxiv created 2016/12/04 · arxiv updated 2016/12/06

Abstract

This report presents an adaptive work-efficient approach for implementing the Connected Components algorithm on GPUs. The results show a considerable increase in performance (up to 6.8×) over current state-of-the-art solutions.

Related

• Overcoming Limitations of GPGPU-Computing in Scientific Applications
• Improving Scalability with GPU-Aware Asynchronous Tasks
• Dynamic Load Balancing in GPU-Based Systems - Early Experiments
• An Adaptive Load Balancer For Graph Analytical Applications on GPUs
• CTA-Pipelining: A Latency-Oriented Spatial Scaling Method for Multi-GPU Systems