Resources Overview

These NCAR supercomputing, data storage, and archive systems support the development of climate models, weather forecasting, and other critical research. Access to them is available through allocation opportunities described here.

New HPE Cray XE cluster named "Derecho"

Hewlett Packard Enterprise (HPE) will build the next supercomputer to be installed at the NCAR-Wyoming Supercomputing Center in Cheyenne, Wyoming. The HPE Cray EX cluster – to be named "Derecho" – will be delivered in the fall of 2021 and will become operational in early 2022.

The 19.87-petaflops system is expected to deliver about 3.5 times the scientific throughput of the current Cheyenne supercomputer. The new system will get 20% of its sustained computing capability from graphics processing units (GPUs), with the remainder coming from traditional central processing units (CPUs).

Read about the selection of HPE to build the system.

Cheyenne: High-performance computing resource

Cheyenne is a 5.34-petaflops, high-performance computer built for NCAR by SGI. An SGI ICE XA Cluster, the Cheyenne supercomputer features 145,152 latest-generation Intel Xeon processor cores in 4,032 dual-socket nodes (36 cores/node) and 313 TB of total memory. In operation since 2017, Cheyenne is slated to be decommissioned in July 2022.

See Cheyenne system documentation.

Data analysis and visualization systems

The Casper cluster is a heterogeneous system of 22 specialized analysis and visualization nodes, 10 large-memory, multi-GPU nodes, and 64 nodes for high-throughput computing. The system’s dense GPU configurations support explorations in machine learning (ML) and deep learning (DL) in atmospheric and related sciences.

See Casper documentation.

Data storage and archives

Campaign Storage

The NCAR Campaign Storage file system is intended for use by the NCAR labs and universities as space for storing project data on publication timescales. The disk-based system has a five-year data-retention policy.

See Campaign Storage.

GLADE: Centralized file systems and data storage

This central file and data storage resource consists of file system servers and storage devices that are shared by the Cheyenne system and the Casper nodes. The GLADE scratch system features a total capacity of 15 PB and individual default quotas of 10 TB.

See GLADE system documentation

Stratus object storage

Stratus, a 5-PB object storage disk system, is for write-once, read-many (WORM), long-term data storage. The system supports an S3 interface for managing data. While Stratus storage is not yet generally available, interested users can contact CISL for information.

See Stratus object storage