ACADEMIA

Adaptive Computing announced that the University of Birmingham in Birmingham, England has selected Moab Adaptive HPC Suite as its intelligent automation software solution. Adaptive Computing’s innovative technology reduces air conditioning costs and input power for the Birmingham Environment for Academic Research (BlueBear). With Moab Adaptive HPC Suite, the annual cost savings on input power and air conditioning costs for BlueBear are estimated at 10% of the total power bill for running the cluster – totaling approximately £10,000 or $16,000 USD.

Used for leading research in bioinformatics, cognitive neuroscience, cancer clinical trials and robotics, to name a few, demand on the BlueBear cluster is often unpredictable. The University required a dynamic and flexible system that could operate under high and low demand, for jobs both big and small. Moab Adaptive HPC Suite, a dynamic workload and resource manager that schedules, monitors, manages, optimizes and adapts HPC workloads and resources to meet demand, SLAs and priorities, was selected to provide additional bandwidth for the data requirements of the University’s top researchers. Moab Adaptive HPC Suite consolidates workload onto fewer nodes and dynamically powers nodes on and off – matching capacity with workload needs and reducing power usage and energy costs.

“We are excited to take our engagement with Adaptive Computing to the next level after observing the reliable up-time of BlueBear and its reduced energy costs,” said Paul Hatton, HPC and visualization specialist at the University of Birmingham. “One of the key considerations that persuaded the University to commit to sustained HPC development was Moab’s proven ability to reduce overall power saving in our facilities.”

Moab Adaptive HPC Suite was able to deliver a system with the following product features:

• 384 mainly dual-processor dual-core (4 cores/node) 64-bit 2.6 GHz AMD Opteron 2218 worker nodes (a total of 1536 cores).
• Most nodes have 8 GB of memory with 16 of them having 16 GB. There are also some 8 core 32 GB memory nodes.

“We’re thrilled to help the University of Birmingham meet the computing demand of its students and academics,” said Michael Jackson, president and COO of Adaptive Computing. “Moab HPC provides the exact balance of flexibility and optimization for maximum resource utilization and increased ROI for new and existing clusters.”