Improving soft real-time systems in games by deadline constraints

Abstract

In recent years, video games have seen a rise in popularity. This is partly due to the current situation of the Covid-19 pandemic which led to people finding a new source of entertainment while confined in their home. In light of this, a considerable amount of people, who do not have the knowledge of hardware limitations, have found gaming to fill the void. However, for the user to have an optimal experience, the game must be optimised in a way that the system can handle it. This all boils down to the system’s specifications and the user’s knowledge on how to use the information presented. Furthermore, gaming technology is always advancing, such as Raytracing and Deep Learning Super Sampling (DLSS), which are being involved in the latest releases and are expected to increase the user’s immersion at the cost of more processing power. By saying this, users who have been avid gamers for a while will also need to optimise their settings to compensate for this without the need of upgrading system components. Due to the technology advancements, namely by Nvidia and AMD, game optimisation is expected to be done by the user to accommodate for personal requirements. However, other third-party software that offers data tracking of currently launched application do not take over and offer assistance on improving the quality of games. Therefore, this leaves an open solution for a system to offer descriptions of statistics while also taking control over the quality of the game. This project describes how real-time systems can help gaming by illustrating how real-time data acquisition can benefit the user’s immersion. A system that automatically detects video games running on the system and act upon the data generated is proposed. The features of the system include real-time tracking of Frames per second (FPS), Latency (in milliseconds), Packet Loss and actions based on the data gathered. These actions include notifying the user of frame and audio lag which can lead to desynchronisation, unstable internet connection and guidance to solve these issues. All the previously mentioned functionalities trigger when the system detects reduction in quality in any of the tracked data. Real-time algorithms are included in the system to give accurate data and assign deadline constraints. Considering there are no all-in-one systems like what is being proposed, functionalities of the system are tested against multiple other trusted solutions to see the difference in the results. These comparisons include real-time tracking of each measure (FPS, Latency and Packet Loss), efficiency of data gathering and conciseness of each functionality triggering. An approximation of accuracy is generated, and conclusions indicate that the solution presented is indeed more effective with helping the user understand the information presented and more accurate at optimising the user’s settings. Furthermore, general evaluation indicates that systems which do not consider user understandability and useability get less traffic. Finally, the solution proposed performs as good as specialised third-party software with the addition of guidance and deadline constraints to improve the user’s experience.