The paper presents a study on performance improvement of an evolutionary computing system built on the basis of JavaScript technology. The system consists of web browsers performing computational tasks and a web server based on Go programming language (Golang technology) that controls the computing environment. The newest web technologies such as HTML5 Web Workres and JavaScript WebAssembly are used to maximize the utilization of the resources provided by web clients. Computational experiments are performed on large instances of popular NP-hard problems: travelling salesman problem (TSP), flowshop scheduling problem (FSP) and recourse constrained project scheduling problem (RCPSP).