High Performance Computer Cluster

The Dan L Duncan Comprehensive Cancer Center’s Biomedical Informatics Group in the Biostatistics and Informatics Shared Resource maintains an enterprise-quality high performance compute cluster for utilization by Duncan Cancer Center members and collaborators to develop new analytical algorithms and to analyze next generation sequence and other ‘omics “big data.”

If you would like to utilize our HPC environment or include cluster utilization in your grant, please contact cluster-support@breastcenter.tmc.edu.

The highly-available cluster has >600 physical CPUs in a single compute node architecture with a 10 Gb Ethernet connection to extensible Tier 1 (ie, direct-attached, rapid I/O) NetApp data storage. Queues are managed with the PBS scheduler. The system is maintained by an expert HPC system administrator in a Tier III data center under standard governance structures.

Diverse services including assistance with dbGaP and other restricted access data use requests are available to end users, and training is available from the HPC system administrator. An online wiki is available to Cancer Center customers behind the firewall for central management of user guides and other documentation.

Many investigators find it difficult to accurately predict how much compute capacity they will need over the course of any one project. BIG offers a tiered pricing model in which investigators can purchase batches of CPU/hours and monthly Tier 1 storage allowances via our iLabs online store. Our prices are extremely competitive compared to external academic benchmarks and are well below that of popular cloud service providers.