Abstract <p>Scientific communities, stakeholders, and industry widely use high-performance computational resources for large-scale simulations and modeling. Due to high capital and operational costs make evaluating the return on investment critical. There is a growing interest in understanding whether proximity to these resources is linked to higher research productivity and quality, as measured by academic publications. The research question addressed in the article is whether there is a correlation between the volume of high-performance computational resources and the number of scholarly publications. The study utilized the Web of Science and TOP500 repositories to analyze correlations from 1993 to 2003, collecting more than 221 thousand publications based on twenty-four selected keywords related to high-performance computing and supercomputing. Pearson, Spearman’s rank, and cross-correlation analysis have been evaluated to determine the direction and strength. The high Pearson correlation coefficient shows a robust linear relationship. While noticeable Spearman correlation coefficients suggest some nonlinearities or outliers, other factors—such as the maturity of the research environment, regional differences, and institutional policies—could also shape the dynamics between the amount of computational resources and research productivity.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Supercomputing Centers and Their Impact on Research Output: Multinational Comparison

  • Sh. A. Sargsyan,
  • H. H. Baghdasaryan,
  • H. V. Astsatryan

摘要

Abstract

Scientific communities, stakeholders, and industry widely use high-performance computational resources for large-scale simulations and modeling. Due to high capital and operational costs make evaluating the return on investment critical. There is a growing interest in understanding whether proximity to these resources is linked to higher research productivity and quality, as measured by academic publications. The research question addressed in the article is whether there is a correlation between the volume of high-performance computational resources and the number of scholarly publications. The study utilized the Web of Science and TOP500 repositories to analyze correlations from 1993 to 2003, collecting more than 221 thousand publications based on twenty-four selected keywords related to high-performance computing and supercomputing. Pearson, Spearman’s rank, and cross-correlation analysis have been evaluated to determine the direction and strength. The high Pearson correlation coefficient shows a robust linear relationship. While noticeable Spearman correlation coefficients suggest some nonlinearities or outliers, other factors—such as the maturity of the research environment, regional differences, and institutional policies—could also shape the dynamics between the amount of computational resources and research productivity.