High-performance computing enhancements bolster artificial intelligence/machine learning research

As conversations about artificial intelligence (AI) and machine learning (ML) sweep the world, Clemson University remains committed to research exploration and excellence in this space. Palmetto, Clemson’s high-performance computing cluster, has recently benefited from institutional investments to ensure researchers have access to the top technologies needed for AI/ML work. 

Clemson’s Computing and Information Technology (CCIT) division has invested over $1 million in new Palmetto equipment to meet growing AI/ML research demands. This investment includes 40 H100 80G GPUs across five nodes for a total of 56 H100 80G GPUs and seven nodes available on the Palmetto cluster. For those outside of technical spaces, these numbers describe how much computing power is dedicated directly to AI/ML, and how these machines can manage the large datasets required in this research. Investments did not stop at new nodes; a dedicated high-speed Infiniband network between all H100 hosts further enables strong scale-out interconnections with reduced latency. 

“We designed these resources to meet the most challenging demands of today’s AI/ML workloads,” said Scott Groel, executive director of research computing and data infrastructure. “These new nodes feature the latest technologies from our partners, Dell and Nvidia, to provide unparalleled compute capability.”

The new AI/ML nodes and network upgrades align with Clemson Elevate’s pillar II (doubling research by 2035), providing cutting-edge resources for faculty, graduate students and partners to advance research.

“Our goal is to position Clemson researchers to pursue grant opportunities and their potential to contribute to important discoveries in artificial intelligence and machine learning,” said Brian Voss, vice president and chief information officer. “Clemson leadership continues to be a strong supporter in providing research computing and data resources across all fields as a part of the research ascension pillar of Clemson Elevate.”

Ultimately, AI and ML research requires additional power and machines uniquely designed for that purpose. Clemson is well-positioned with Palmetto to make such additions possible and available. These tools can foster breakthroughs across the academic spectrum. 

“The new compute nodes will enable faster simulations in materials science, accelerate molecular modeling and drug discovery in chemistry, and enhance genomic data analysis in bioinformatics, significantly reducing time to discovery and advancing research across these fields and across more disciplines,” said Jill Gemmill, associate vice president for Research Computing and Data.

Those interested in learning more about Palmetto and its now available AI/ML resources can visit this website.