Purdue University in Indiana, US, has unveiled its Gautschi supercomputer.

Housed at the Rosen Center for Advanced Computing (RCAC), the supercomputing cluster was funded in part by a $50 million grant given by Lilly Endowment Inc. for the Purdue Computes initiative. Of that grant, $20m was explicitly allocated to the university’s Institute for Physical Artificial Intelligence.

Built in partnership with Dell, AMD, and Nvidia, the Gautschi cluster consists of 338 Dell compute nodes and two 96-core AMD Epyc Genoa processors, the latter of which has 384GB of memory. It also contains six large memory nodes with 1.5TB of memory, six inference nodes with two Nvidia L40S GPUs, and 20 AI nodes, eight of which are the new Nvidia H100 SXM GPUs.

All compute nodes have 200Gbps of NDR Infiniband interconnect and will be in service until 2030. The Gautschi cluster will be in early-user production by the end of this month.

Named after Walter Gautschi, professor emeritus of computer science and professor emeritus of mathematics at Purdue University, Gautschi is expected to place within the top 150 on the top 500 list of the world’s most powerful supercomputers. However, despite the claim, no information about the system’s compute power was detailed.

Professor Gautschi was in attendance at the dedication ceremony on October 21.

Purdue University’s RCAC is home to ten clusters, including Anvil, Negishim Geddes, Bell, Weber, Gilbreth, Scholar, Brown, Workbench, and Hammer. The Anvil cluster was installed in November 2021, offering a peak performance of 5.3 petaflops (double-precision) across 1,000 nodes, and was RCAC’s most powerful cluster.

In May 2023, the university added 104 Nvidia A100 GPUs to its Gilbreth cluster, increasing its GPU capacity by 46 percent and doubling the system’s AI performance.