Vermilion Computing System

NREL's Vermilion high-performance computing (HPC) system is allocated for NREL workloads intended for Laboratory Directed Research and Development, strategic partnership, or U.S. Department of Energy Office of Science projects.

Allocations

For Vermilion, NREL's Institutional Advanced Computing Allocation Committee makes allocation decisions through the annual allocation request process. 

System Architecture

Vermilion is an OpenHPC-based cluster configured to run compute-intensive jobs and has a mixture of node types which include AMD EPYC 7352 Rome CPU's and nVidia A100 GPU nodes. It is a cluster comprising 120 compute nodes and 30 GPU nodes.

The nodes run as virtual machines in local virtual private cloud (OpenStack) and are connected to each other and to storage by 25Gb Ethernet network utilizing RDMA. All nodes and storage are connected in a fat tree topology via ethernet. Partitions are flexible on Vermillion. A list of partitions can be found by running the sinfo command.

Learn more about Vermilion (GitHub).