Home / Engineering Computing / Moab HPC Suite 7.1 Speeds HPC Performance

Moab HPC Suite 7.1 Speeds HPC Performance

By DE Editors

Adaptive Computing’s Moab HPC Suite 7.1 optimization solution accelerates high performance computing, visualization, and big data workloads. The solution’s topology-based scheduling, streamlined workload management for Cray systems, and visualization workload optimization help speed job processing by as much as 200%, according to the company.

The solution’s new topology-based scheduling schedules nodes in close proximity to each other. The new Moab topology node allocation policy plugin capability intelligently schedules parallel-processing jobs on inter-communicating nodes that are close to each other in a cluster based on the tailored topology of each specific high performance computing system. The ability to create custom Moab topology node allocation plugins using the Moab Workload Manager Plugin Developers Kit allows customers or partners to optimize and customize node allocation placement for their parallel or big data jobs based on their specific system topology, whether a customized generic topology or vendor specific such as Cray or SGI.

Moab HPC Suite now works in concert with the NICE EnginFrame and Desktop Cloud Visualization (DCV) products to centralize individual visualization workstations to visualization servers in the data center, transfer visualized pixels instead of large data sets for users to interact with, and intelligently place and manage visualization workloads as well as regular HPC cluster workloads. Multiple users can access and collaborate on interactive 3D application data from any location and almost any device.

An enhanced, streamlined architecture between Moab HPC Suite and Cray ALPS enables Moab and TORQUE to now schedule faster from outside the Cray cluster domain on dedicated, more powerful nodes. The new architecture enables faster deployment on Cray systems as well as simplified ongoing configuration and management for admins, the company says. This includes reduced node management, with Moab now able to auto-detect Cray nodes and GPGPUs for scheduling management, and with Cray ALPS node inventory and status reporting now managed in the Moab HPC Suite TORQUE resource manager. Management, utilization and available capacity is also improved with better automated clean-up of ALPS reservations after jobs have run to ensure resources are made available for new jobs.

For more information, visit Adaptive Computing.

Sources: Press materials received from the company and additional information gleaned from the company’s website.

About DE Editors

DE's editors contribute news and new product announcements to Digital Engineering. Press releases can be sent to them via DE-Editors@digitaleng.news.