The NREL High Performance Computing Center Located at the Energy System Integration Facility (ESIF) has an integrated set of resources, including computational, scientific data management and storage, and networking capabilities and relevant expertise to support the mission of the Department of Energy, Office of Energy Efficiency and Renewable Energy (EERE). Research and development projects that are aligned with this mission can request access to the ESIF supercomputer (Peregrine) and data storage systems.
To use NREL HPC resources, you will need access to an allocation.
To run jobs on Peregrine compute nodes, you will need to specify a valid allocation. An allocation is a portion of the HPC center resources planned for a particular research project's use. Resources allocated are Peregrine compute node hours on Peregrine and project storage space (in units of Terabytes) on Peregrine and Gyrfalcon. Project storage space on Peregrine is on a high speed parallel file system called /projects. It is used to store files that are in active use and are shared by members of a project. Project storage space on the Gyrfalcon mass storage system (/mss/projects) is intended for longer term storage of files that are shared by members of a project.
Each job has exclusive access to the nodes it requests so if your job runs on 2 nodes for 3 hours, your allocation will be charged for 6 node hours. Please note that each node has either 16 or 24 processors (cores) and your allocation request should be made in node hours rather than processor core hours.
NREL High Performance Computing Center Systems Data Security Policy
The NREL High Performance Computing Center computer systems are operated as research systems and may only contain data related to scientific research. These systems are categorized as low per FIPS-199 and protected to the NIST 800-53 low security control baseline.
The NREL HPC center computer systems control data access via username and password authentication for network access, and UNIX directory and file permissions for data storage. Network and storage systems provide no explicit encryption. Users are responsible for protecting their data files and agree that NREL’s HPC security control implementation is sufficient for their work. In addition, these systems may not store or process export controlled data.
For more information about what this means and to find out whether your data may be put on these systems, please go to the NREL High Performance Computing Center Systems Data Security Policies page.
FY17 allocation cycle
Peregrine is fully allocated through October 31, 2017. We may be able to make additional allocations in early January once we understand how existing projects are using their allocations.
Allocations are principally done in response to an annual call for allocation requests, which are due in early August each year. For the FY17 allocation cycle, a call for allocation requests will be made on June 15, 2016 and requests are due August 1, 2016. Out of cycle requests are accepted but these are drawn from a much smaller pool of resources. Pilot project allocations of up to 20,000 node hours may be requested at any time.
Resources to be allocated for this allocation period include approximately 21 million node hours, 750 terabytes of shared data storage (/projects), and up to 1 petabyte of long-term data storage.
The FY17 allocation period runs November 1, 2016 through October 31, 2017.
NOTE: Please see the FY16 allocation request page to request an allocation that will start before November 1, 2016.
Request an allocation
To requested an allocation, fill out the HPC-Requests@nrel.gov.