Skip to content
Center for Integrated Research Computing (CIRC)

Resources

Home/ Resources
bluehive

BlueHive Cluster

480 nodes, 14,392 CPU cores, 72 TB RAM, 5.2 PB Storage, and 1,440 TeraFLOPS

The BlueHive cluster is CIRC’s primary Linux cluster for demanding computations, which provides approximately 1,440 TFLOPS of computing capacity. This system consists of 480 nodes with a high-speed, low-latency, InfiniBand interconnect. The most recent addition to the BlueHive Cluster houses dual-processor systems with up to 96 cores per node and ranges in memory from 20 GB to 3 TB. A number of compute nodes have dedicated coprocessing cards, including NVIDIA’s K80 (Kepler), V100 (Volta), and A100 (Ampere) GPUs. In addition, some nodes of the cluster are dedicated to running “big data” analytics applications, such as Hadoop and Spark.

Example FastX Desktop

Contributed hardware for priority access

The entire BlueHive cluster has an InfiniBand-attached storage system providing approximately 5.2 PB of configurable raw disk within a GPFS file system Over 200 nodes with varying capacity have been integrated into the BlueHive cluster for faculty investigators who have purchased additional priority-base compute capacity for the environment. CIRC runs the SLURM resource scheduler and queuing system to optimize usage and to support multiple users of the BlueHive Linux cluster environment. Additionally, users have access to a browser-based technology to connect to BlueHive using a resumable web browser interface.

Services

Affiliated Organizations

Return to the top of the page