More than 200 computer science research journals publish thousands of research articles yearly. The publishers maintain explicit repositories of research articles and the bibliographic databases are locked behind paywalls. The exponential growth in the number of articles makes it troublesome for computer science research scholars to search and retrieve relevant research articles quickly and precisely. The web search engines (WSEs) are used by scholars to ease their searching and retrieving operations. However, the enlarged list of web search engines and lack of availability of their performances information makes WSE selection difficult for the scholars. This paper presents an empirical performance comparative analysis of the top five WSEs (i.e., Google, Bing, Yahoo, Baidu and Yandex) to retrieve computer science research articles in response to domain-specific queries varying in complexity. The comparison and evaluation methodology, performance analysis using information retrieval metrics (i.e., response time, recall, and precision), and systematic comparison of results are presented to highlight strengths and weakness of the WSEs. The results have shown that Google is performance-wise better WSE for searching and retrieving computer science research articles. The findings could be helpful for computer science research scholars in using an appropriate WSE in their searching and retrieving processes.