GPUs can perform batch processing of considerable chunks in parallel, providing extreme performance over traditional CPU based system. Their usage has shifted drastically from exclusive video processing to generic parallel processing. In the past decade, the need for better processors and computing power has increased due to technological advancements achieved by the computing industry. GPUs work as a coprocessor of CPUs and the integration of their architectures has raised many bottlenecks.
Multiple GPUs solutions have been developed by industry to tackle these bottlenecks. In this paper, we have explored the current state of GPU-accelerated database industry. We have surveyed eight such databases in the industry and compared them on eight most prominent parameters such as performance, portability, query optimization, and storage models. We have also discussed some of the significant challenges in the industry and the solutions that have been devised to manage these bottleneck.