Software Call Graph Analysis And Its Applications In Optimizing Simulation Engines

This paper presents the analysis of software call graphs for object-oriented systems,based on the software call graphs in function level(function call graph)and class level(class dependency graph).Compared with existing software call graphs,this paper presents the specific call graphs suitable for object-oriented systems,where object-oriented concepts(e.g.,objects,classes,and namespaces)are considered,which could describe the program executions more accurately.Static call graphs are used to represent the static features of the original systems,and dynamic call graphs are used to represent the dynamic executions of the systems.Then a series of graph analysis methods are presented to analyze the software systems in class level and system level.With the consideration of dynamic executions of systems,the analysis methods belong to Dynamic Analysis of Object-Oriented Systems(OODA).Based on the analysis of three different topology graphs,including static call graphs,dynamic call graphs and class dependency graphs,the presented methods can be used in object-oriented program analysis,system understanding,optimization and testing.This paper introduces the methods with the example of a resource analysis system of simulation engines.From the view of system analysis,the presented software call graphs enable researchers to get a more accurate understanding of system design.Simulation engine plays an important role in digital performance project,and the optimization is often accompanied with the project progress.System optimization can make the engine more efficient in specific simulation scenarios.However,it was a challenging task to analyze simulation engine systems considering the complexity of the systems and the lack of introduction documents.This paper applies the presented methods to analyze the simulation engine systems,and it proves that the proposed methods can be used to analyze the design of the engine systems and improve the efficiency of simulation engines.This work was successfully applied to design the digital stages in the 2016 and 2017 CCTV Spring Festival projects.