Analysis And Optimization Of MMORPG Memory

In recent years,with the development of the game market,the image quality of computer games has been continuously improved,and the gameplay has become increasingly rich.The consequent problem is that the memory requirement of game programs is also increasing.Even today,with the rapid advancement of hardware devices,game memory management is still a thorny issue.Massive Multiplayer Online Role-Playing Game(MMORPG)has a large user group.It has complex scenes,diverse gameplay,exquisite modeling.Compared with other types of games,it needs to deal with larger game data,more frequent allocation and release of memory,so its memory management is particularly important.This thesis takes "Revelation",a game developed by NetEase,as a use case to establish a complete MMORPG memory management and optimizations solution during game development.The following are the work finished in this thesis.1.Developed toolchain for memory diagnosis during game running for performance analysis.Found and detected some memory issues to be improved.2.Improved Memory allocator in game engine by replacing the Windows native memory allocation algorithm with the new one and performed statistical analysis and checking during memory allocation.3.Optimized and refactored data structure of native Python dictionary object to decrease memory usage while maintain the same access efficiency.4.Designed Memory DB to solve the problems of frequent I / O operations caused by the huge amount of game data.This solution uses the design of Protocol Buffers-like serialization algorithms and string pools to optimize game data in memory.After optimizing the above problems,a complete test is carried out,and good optimization results are obtained.The results of this research are expected to provide experience for memory management of similar game projects,some of which have inspiration and reference significance.