Return to Services

Cluster Computing

The Duke Compute Cluster (DCC) is a high throughput, high performance, parallel computing cluster for research support. Storage for data under analysis and access to run jobs on the common partition are provided at no cost to Duke research labs. If analysis needs exceed what can be handled under the common partition, researchers contribute node(s) through inclusion in grant proposals.

Dedicated compute clusters are also maintained for computationally intensive research labs that require specialized equipment or software. The Duke HPC Team supports the research computing infrastructure for:

  • Bartesaghi Lab – Purchased in 2018, the cluster consists of 56 high performance Cisco UCS C220 rack server hosts with 36 cores and 768 GB RAM each and utilizes a 150 TB all Flash storage array.
  • Randles Lab – Purchased in 2017, the cluster consists of 30 infiniband connected high performance UCS C220 rack server hosts with 44 cores and 512GB RAM each and 30┬áC240 rack server hosts with 44 core, 512GB RAM and 2 P100 GPUs each.