Search

Recently, a new trend has emerged in the field of parallel and high performance computing, the hybrid implementation using CPU-GPU modules. In such implementations, the computational load is shared between the CPU and GPU, in order to improve the computational efficiency. However, the task of sharing the computational load between the two modules is a rather difficult one, with a number of limitations being imposed. This paper extends our recent work on community detection, which is based on transforming a network of nodes into a set of threaded binary trees. In this work, we share the computational load between the two units: the CPU takes specific samples of the network communities and organizes them in the form of threaded binary trees. The GPU takes over the heavy load of reading this data and transforming it into a path-matrix. Finally, this matrix is sent back to the CPU for analysis, community detection and overlaps, as well as network information upgrades. Our simulation results show significant improvement over our previous strategy and other known community detection strategies found in the literature.

Several synchronous applications are based on the graph-structured data; among them, a very important application of this kind is community detection. Since the number and size of the networks modeled by graphs grow larger and larger, some level of parallelism needs to be used, to reduce the computational costs of such massive applications. Social networking sites allow users to manually categorize their friends into social circles (referred to as lists on Facebook and Twitter), while users, based on their interests, place themselves into groups of interest. However, the community detection and is a very effortful procedure, and in addition, these communities need to be updated very often, resulting in more effort. In this paper, we combine parallel processing techniques with a typical data structure like threaded binary trees to detect communities in an efficient manner. Our strategy is implemented over weighted networks with irregular topologies and it is based on a stepwise path detection strategy, where each step finds a link that increases the overall strength of the path being detected. To verify the functionality and parallelism benefits of our scheme, we perform experiments on five real-world data sets: Facebook®, Twitter®, Google+®, Pokec, and LiveJournal.

Community structures are formed in many real-world networks, e.g., biological or medical groups, student groups, and so on. Communities are perhaps the most important feature of today's networks, since the majority of people who join a network also tend to join one or more communities. Therefore, several researchers find that the detection of hidden communities is a very interesting and challenging research field. Communities are represented as the groups of nodes on a graph, corresponding to users with similar interests. This paper introduces a novel, interdisciplinary, approach for community detection, combining social networks and distributed systems, where remote access to shared files is offered in a networked environment. A new metric, based on data requests, is introduced and used as a measure of the belonging degree of a node in a certain formed community. Two sets of simulations are used to verify our scheme: simulation results on synthetic networks and results derived from real data.