Research

Computer Science

The Computer Science activity supports long-term, basic research on the software, tools, and techniques that allow scientists to harness the potential of advanced computing and smart networking technologies and extreme-scale data, including machine learning. Advanced Scientific Computing Research (ASCR) Computer Science research leads the way in the R&D of the software infrastructure that is essential for the effective use of the most powerful high performance computing systems in the country, tools to manage and analyze data at scale, and cybersecurity innovation that can enable the scientific integrity of extreme scale computation, networks, and scientific Science. ASCR Computer Science plays the role of reducing risk when industry will not invest in the specialized software required for future Leadership Computers. Supercomputer vendors often take software developed with ASCR Computer Science investments and integrate it with their own software.

It is widely acknowledged that the computing industry is currently entering an era of unprecedented hardware architecture diversity — a veritable explosion of technology innovation that will greatly increase the complexity of Leadership Computing. Such diversity and complexity will be accompanied by major challenges at all levels of systems software, data management and analytics tools, heterogeneity, the cybersecurity environment, and applications software. To have Leadership Computing systems that are useful for science applications at the time of delivery, ASCR Computer Science research focuses its investments on software needed for emerging technologies long before they are in operational use. These efforts are essential to ensure DOE mission applications including data-intensive applications, are able to use commercially available HPC through the exascale era and beyond Moore’s law.

Last modified: 2/26/2019 11:01:16 AM