Using Moab's Policies to Create a More Intelligent HPC Compute Environment

Tue, Sep 21, 1:00 PM - 2:00 PM EDT

Is your HPC compute environment limiting what jobs can be run, and when? Or are your incoming and calendared jobs determining how your environment is provisioned and allocated to handle the workload? . . .

• Are some Linux jobs delayed because they need a different "flavor" of Linux and must wait until current Linux nodes are idle and can be reprovisioned, often while siloed Windows nodes sit idle?
• Are some Windows jobs delayed when all your Windows nodes are busy, even though a sufficient number of idle nodes are available but provisioned with other OSes?
• Do a lot of idle nodes stay fully powered for long periods of time because it's too much hassle to power them down and back up again just to save a little electricity?
• Have you resorted to siloing your environment in an effort to "protect" resources to satisfy job requests?

The solution to all these costly and inefficient scenarios is to put Moab's intelligent policy engine to work.

• With Moab, you no longer need to create silos to protect resources for upcoming jobs. Instead, allow Moab to dynamically modify the environment based on job requests.
• Need a Windows node, but they're all busy? Moab will find an idle Linux node and swap its operating system to Windows.
• Moab ensures that your compute environment is not restricting the types of jobs that can be run. Instead, upcoming jobs will determine how your environment is configured from moment to moment.

Join us for a free online demo of how Moab helps you intelligently put hardware resources to full use by employing on-demand multi-OS swapping, energy-aware computing, and other SLA-centric policies.