PhD Dissertation Proposal by YE Chang | Accelerate community analysis by GPUs

Community analysis, which focuses on uncovering the properties of networks, has garnered significant attention. It provides valuable insights into the structure, dynamics, and functionality of networks, enabling us to make informed decisions, design interventions, and improve the efficiency and resilience of various systems. However, traditional CPU approaches become less competitive for community analysis.The exponential growth of big data and the increasing demand for efficiency necessitate the utilization of GPU for processing community analysis tasks because GPU offers higher levels of parallelism and bandwidth compared to CPU. But it is non-trivial to enable efficient community analysis tasks in GPUs because there are unique challenges which wastes GPU’s computation resources: (1) Limited memory; (2) poor memory access pattern; (3) imbalanced workload. In this proposal, we study two community analysis tasks, introduce the challenges of deploying them on GPUs, and propose our optimizations for those challenges.

Chang Ye is a PhD student with the School of Computing and information Systems, Singapore Management University(SMU), He receives his BSc and MSc degrees in the area of Electronic Engineering from Xidian University, in 2014 and 2018 respectively. His research interests are graph analytics and heterogeneous computing.