An Integrate Taxonomy of Standard Indicators for Ranking and Selecting Supercomputers

With the expansion of industries and of new technologies that require more computing resources, supercomputer has attracted much attention as a platform for high performance computing services. Although various features and indicators for Testing and evaluation of supercomputers have been proposed, however, a comprehensive classification of these features has not been provided so that the developer, designer, or user can easily compare and select supercomputers by scoring. Therefore, selecting a supercomputer solution to meet the requirements is a complicated problem. This paper presents a new feature-based and integrated taxonomy (Hierarchical Supercomputer Performance Indicators Taxonomy), and their scoring and ranking routine which facilitates the process of supercomputer evaluation and selection. Also, this paper provides a case study and using the proposed framework presents a comparison between some commercial and research supercomputers including Fukaku's ideal supercomputer, Sharif supercomputer, Aramco supercomputer and ITU supercomputer. The ranking results show that Aramco supercomputer, ITU supercomputer and Sharif supercomputer have 65.9%, 57.6% and 48.2% of ideal supercomputer points, respectively. Also, the percentage of coverage of a customer's needs with supercomputers and the percentage of index coverage categorized separately for supercomputers are determined. In this article, the proposed method has been compared with the Top500 method, which shows that this method facilitates the ranking, comparison and selection of the appropriate supercomputer in various fields such as military and law enforcement systems by considering various aspects of design and implementation.