The Ohio Supercomputer Center’s (OSC) newest, most advanced high performance computing (HPC) cluster is ready to support research and education in Ohio and beyond. Cardinal, a Dell Technologies-based cluster, has been designed to meet the rising computational needs of artificial intelligence and machine learning work.
“OSC collaborated with Dell Technologies, Intel and NVIDIA to incorporate advanced HPC technologies into this new cluster, which offers a 140% increase in computing power over all of OSC’s current systems combined,” said David Hudak, OSC executive director. “Cardinal demonstrates our commitment to continually evolving our HPC resources to ensure that Ohio’s researchers, students and commercial clients are working within a nationally competitive environment.”
The Cardinal cluster was ranked No. 211 on the Top500 List, which tracks the most powerful commercially available computer systems around the globe, in November 2024. In addition, the Green500 List, which evaluates the most energy-efficient systems, ranked Cardinal No. 15.
Cardinal is a 39,312-core, 378 node Dell PowerEdge with Intel CPU Max 9470 HBM machine with a theoretical system peak performance of 10.5 petaflops. The new cluster features:
-
326 nodes: 104 cores, 128GB HBM and 512 GB DDR5
-
32 nodes: 104 cores, 1 TB of memory and 4 NVIDIA Hopper H100 GPUs w/ 94GB memory and NVLink
-
16 nodes: 104 cores, 128GB HBM and 2TB DDR5
-
4 login nodes: 104 cores, 128GB HBM and 1TB DDR5
Detailed system specifications are available on Cardinal’s webpage.
Early users of the Cardinal cluster reported on the increase in processing power and speed of computations for their academic research and commercial projects.
Xiaotian Han, assistant professor in the Department of Computer and Data Sciences at Case Western Reserve University, has used the Cardinal cluster to accelerate research on large language model optimization. Han found that the new cluster’s H100 (94GB) GPUs significantly boosted the efficiency of his experiments, reducing the standard runtime of about 24 hours on the Ascend cluster to 9.5 hours on Cardinal.
“The capability of Cardinal will significantly facilitate my future research, allowing for faster iterations and the ability to test larger models in a shorter time,” Han said. “This not only saves time but also enables more ambitious research objectives, such as developing advanced techniques for reducing inference latency and improving memory usage in large-scale models.”
TotalSim, a Dublin, Ohio, company that performs computational fluid dynamics (CFD) simulations and modeling for its clients, also observed a major reduction in runtime for its jobs, a major asset for projects with tight deadlines.
“Our smaller core count jobs show a significant speed up compared to the Pitzer cluster, with a 30% to 65% reduction in run time,” said Greg Padgett, senior CFD engineer. “Our large parallel jobs show better scalability. What used to top out between 400-700 cores now scales up to 700-1500 cores.”
Cardinal joins the Pitzer (2018, expanded in 2020) and Ascend (2022) clusters on the OSC data center floor at the State of Ohio Computer Center in Columbus. Cardinal will replace the Owens cluster (2016) during a phased transition over the next few months. OSC, in collaboration with The Ohio State University, also will add significant computing resources to the Ascend cluster that will be available to the HPC community in early 2025.
Watch the build process of Cardinal unfold:
Partner statements:
“Combining the computing power of Dell PowerEdge XE9640 servers with the Ohio Supercomputer Center’s sophisticated approach to discovery will help the next generation of researchers and scientists solve tough questions across disciplines using AI, ML and HPC,” said Dave Lincoln, vice president of Compute Systems, Dell Technologies. “The Cardinal cluster will be a valuable addition to OSC’s resources, helping to ensure these critical technologies are accessible for Ohioans.”
“The Intel Xeon CPU Max Series is an optimal choice for developing and implementing HPC and AI workloads, leveraging the most widely adopted AI frameworks and libraries,” said Ogi Brkic, vice president and general manager of Data Center AI Solutions product line at Intel. “The inherent heterogeneity of this system will empower OSC’s engineers, researchers and scientists, enabling them to fully exploit the doubled memory bandwidth performance it offers. We take pride in supporting OSC and our ecosystem with solutions that significantly expedite the analysis of existing and future data for their targeted focus areas.”
“NVIDIA’s accelerated computing platform provides the groundbreaking performance academic supercomputer centers such as OSC require to run AI applications across a variety of fields,” said John Josephakis, global vice president of business development for HPC and supercomputing at NVIDIA. “The new Cardinal cluster with NVIDIA H100 Tensor Core GPUs and NVIDIA Quantum-2 InfiniBand networking will help the state of Ohio stay at the forefront of innovation and meet the growing demand for HPC and AI resources.”
The Ohio Supercomputer Center (OSC) addresses the rising computational demands of academic and industrial research communities by providing a robust shared infrastructure and proven expertise in advanced modeling, simulation and analysis. OSC empowers scientists with the services essential to making extraordinary discoveries and innovations, partners with businesses and industry to leverage computational science as a competitive force in the global knowledge economy and leads efforts to equip the workforce with the key technology skills required for 21st century jobs.
