PAGE 12 LIBRARY AND TECHNOLOGY SERVICES - - - - - - - - RESEARCH COMPUTING Research Computing at Lehigh continued to deepen and enrich its core services to the research community while looking to the future by building new initiatives on this solid foundation. The most mature, well-established research computing projects are hosted by our on-premises high-performance computing (HPC) cluster, which continues to provide millions of hours of compute cycles (measured in core-hours) to dozens of research groups, while offering hundreds of software packages and a high-performance file system. Our NSF-funded Hawk cluster provided 2.7 million corehours to external partners who use the Open Science Grid. We have improved our onboarding procedure while supporting screening and compliance for export-controlled projects and upgraded our operating systems to improve security and usability. Our group continues to build storage solutions for projects that use big data. This includes a new proof-of-concept plan to use upcom ing lifecycle upgrades to our storage system, which uses our two petabyte Ceph filesystem, to establish a new backup service along with faster storage tier for researchers who require higher-through put storage space. Alongside steady growth in the total number of cores funded in part by faculty “condominium” investments (i.e., researcher buy-in to our shared HPC cluster), we have also introduced new graphics process ing unit (GPU) architectures and begun planning to standardize and bundle faculty hardware investments in order to improve usability and maximize the number of core-hours that our community can access. As we work to incrementally improve the usefulness of our HPC cluster, we are also preparing infrastructure upgrades to the datacenter that support future growth. These include a pending project to upgrade our cooling systems and a feasibility study for adding liquid cooling. This latter project would provide a step change to our maximum capacity and an opportunity to scale our existing comput ing services in coming years. Cloud computing continues to provide critical support for expand ing our research computing capacity, particularly for secure data, which includes personally-identifiable information (PII) and personal health information (PHI). LTS has recently invited its first tenants to use the Lehigh Secure Research Cloud (SRC), a cloud computing environment built using Amazon Web Services which is capable of performing analysis of secure data. We expect this initiative to grow alongside the University research footprint, particularly as Lehigh attracts new faculty whose research depends on secure data, for example, new faculty in the College of Health. As always, teaching, direct consultation with faculty, and outreach to the community provides a glue that binds these services to novel and creative research projects. The Research Computing group provided ten advanced seminars in 2023 which followed an addi tional ten introductory seminars provided by peers in other groups within LTS, in order to provide an integrated seminar series that teaches students, faculty, and staff the best practices for building their unique research computing projects. Our HPC infrastructure supported three credit courses during the academic year along with a Quantum Chemistry workshop during the summer. In the spring, students from Southern Lehigh High School toured the data center to learn more about how Lehigh University infrastructure facili tates research across academic domains. In the coming year, the Research Computing group will look forward to working with many of our colleagues in LTS to apply our infrastructure and expertise to new and exciting research projects led by the faculty. Number of seminars/workshops taught last year, and the topics • A series of four Python seminars friendly to novice users for the staff, faculty, and students. • A series of four seminars that teach the basics of “R,” a statistics-friendly scripting language. • Six interrelated High-Performance Computing (HPC) semi nars in Spring 2023 delivered both in-person and remotely, covering the basics of HPC; optimizing our use of Lehigh resources; writing parallel scientific software; best practices for research data management; and building reproducible, collaborative workflows. • HPC training and support for the computational Quantum Chemistry workshop held in the summer, with participants from multiple departments and collaborators from other institutions. • Four summer consulting and office hours sessions for HPC projects. WHO’S USING HPC ► 199 active users 54 active project investigators
RkJQdWJsaXNoZXIy MTA0OTQ5OA==