A GPU-Enabled Extension for Apache Ignite to Facilitate Running Genetic Algorithms

With the increasing rate of data generation in recent years, there is a need for modern tools to process these massive amounts of data. To that end, in-memory platforms are becoming increasingly popular, which can process high volumes of data at high speed and performance by utilizing Main Memory. Apache Ignite is one of the in-memory platforms that can process in parallel on multiple nodes. Although this platform provides many useful features, one of its limitations is the lack of utilizing the GPU's high processing power. Undoubtedly, using GPUs for operations that deal with heavy processing or high data volumes can be very beneficial, and significantly accelerate processing. One of the algorithms supported by Ignite is the Genetic Algorithm, which usually deals with large amounts of data, and might be very time-consuming. In this paper, we have provided an extension for Ignite in which users can utilize GPUs to run their Genetic Algorithm applications. Also, we have used various GPU-related optimization techniques to improve performance and finally evaluated our extension with three benchmarks. Our results proved the ease of use, and the high performance of the proposed work compared to Ignite.