DSC Lab Computing Cluster Enhanced with GPU-Accelerated Nodes

The Data Science & Computing Lab (DSC Lab) at IMCS has expanded its high-performance computing cluster by integrating new GPU-accelerated compute nodes.

The four new Intel Sapphire Rapids compute nodes feature 2x32 cores per node plus one NVIDIA H100 Tensor Core GPU per compute node, complementing the lab’s existing CPU-based infrastructure, which includes 480 Intel Xeon Skylake-SP and 832 Intel Xeon Cascade Lake cores. This enhancement markedly boosts the cluster’s capabilities in high-performance computing, GPU-accelerated simulations, and AI-optimized workflows.

"The integration of H100 GPUs paves the way for computational research across all focus areas at IMCS, enabling advances and innovation in scientific machine learning, multi-scale simulation, mixed-dimensional modeling, and scalable preconditioners", explains Dr.-Ing. Matthias Mayr, head of the Data Science & Computing Lab at IMCS. The upgrade also strengthens the lab’s infrastructure for ongoing research projects such as RISK.twin funded by dtec.bw.

For more details about the lab’s computational infrastructure and research activities, visit: www.unibw.de/imcs-en/dsc-lab