Skip to main content

Powerful new GPU computing resources available to researchers across state

Faculty and student researchers at WVU and any other university or college in West Virginia can now request access to the new high performance computing cluster called Dolly Sods

This graphic processing unit cluster is ideal for supporting specific computational tasks requiring extra speed and power, such as machine learning, artificial intelligence, molecular dynamics for materials and drug discovery, image and video processing, and training neural networks.

This high-performance computing cluster was established with the support of an MRI grant awarded by the National Science Foundation under the leadership of former professor Blake Mertz. Following his departure from the University, Aldo Romero, director of research computing, has assumed responsibility for overseeing the installation and operationalization of this new GPU-based system.

“High-performance computing systems equipped with graphics processing units have become indispensable cornerstones in fields such as artificial intelligence, data analytics and classical molecular dynamics,” Romero said. “Before the availability of GPU-based systems, methods for training machine learning models predominantly relied on CPU systems, which are often slower and less suitable for the computational demands of modern applications. 

“The architecture of neural networks is inherently well-suited for GPU computation, making this technology invaluable. The availability of this state-of-the-art GPU system provides an unparalleled opportunity for researchers across multiple disciplines to access and leverage cutting-edge computational capabilities,” he added. 

Learn more about Dolly Sods specifications. 

To request access, visit and select “New User Account Request.” 

Direct questions about HPC resources to This will automatically create a ticket and assign it to the HPC team in the WVU Research Office.

This article is republished from MOUNTAINEER E-News — read the original article.