Last month, I joined Addison Snell, CEO of Intersect360 Research, for a webinar to discuss the old conundrum of choosing any two of “fast, good, and cheap” into HPC concepts of high-throughput, reliable, and within budget. Adaptive Computing’s Moab HPC Suite can help you accomplish all three in your HPC solution.
Fast, Good, and Cheap HPC Concepts
Fast = High-Throughput. Adaptive Computing’s Moab helps you get more results faster and at lower cost. Moab achieves 90-99% resource utilization with massive scalability and workload-optimized resource allocation. Additionally, Moab’s topology-based scheduling and workload-optimized node provisioning ensure that throughput is as high as possible. Throughput isn’t just about the system, Moab also accelerates user productivity by simplifying job submission and management.
Good = Reliable. Moab ensures workload completes successfully without delays. Intelligent resource placement prevents job failures and Moab automatically responds to failures and events.
Cheap = Within Budget. Moab sets usage budget amounts for users, groups, and accounts. It utilizes resource allocations and scheduling checked against amounts. Moab tracks and reports system usage by each user, group, account, etc. for showback or pay-for-use chargeback.
HPC Technical Computing Trends
Addison and I also discussed today’s HPC and technical computing trends for enterprise. He compared technical computing vs. enterprise computing for HPC capabilities. Technical computing is driven by price performance, trying to get the maximum performance out of a given budget. Enterprise computing traditionally is driven by reliability, have things remain available and reliable.
Comment and Share: Moab gives you the best of all three areas: Throughput, Reliability, and Budget. That said, I am interested in which area is most important to you? Please comment with your thoughts.