Computational Cluster Implementation at King Juan Carlos University

This case study details the integration of enterprise switches into a new computing cluster at King Juan Carlos University, aimed at enhancing biomedical research focused on cell ageing and reprogramming. The project, funded by the European Union and the Madrid Community, addresses the computational demands of deep learning and artificial intelligence techniques, which require extensive CPU nodes and GPUs for parallel computing. The cluster, named TALOS, features 512 execution threads in CPUs and 221,184 GPU cores, achieving over 20,000 Tflops of power with a storage capacity of 576 TB. The implementation utilized D-Link enterprise switches, specifically the DXS-3610-54T/SI and DGS-1210-52 models, to ensure high-speed connectivity and management capabilities. The study outlines the technical specifications and the operational requirements met by the infrastructure, highlighting the importance of reliable network management in high-performance computing environments. The successful deployment has significantly improved access and computational power for researchers at the university.