Cluster Design
The UTEP Hopper Cluster is a new GPU computing resource comprised of an NVIDIA H100 DGX node with eight H100 GPUs and an NVIDIA HGX H100 node with two 56-core Intel Xeon Sapphire Rapids CPUs and four H100 GPUs. The NVIDIA Hopper H100 Tensor Core GPU is specifically optimized for machine learning and is also a highly effective platform for computationally intensive scientific computing workloads.
Software supported on the cluster includes the NVIDIA HPC Software Development Kit (SDK) and optimized deep learning frameworks and models from the NVIDIA NGC Catalog. Batch and interactive access to the cluster is provided by the SLURM workload manager.
The Hopper cluster will be available for high-performance computing (HPC) and machine learning research, as well as for educational uses. The cluster equipment has arrived at UTEP and we hope to have it available for use during the fall 2024 semester.