United Devices announced the first High-Performance Computing (HPC) management solution that provides a single, normalized view into capacity and utilization across all legacy scheduling utilities, including Sun SGE, Platform LSF, Altair PBS and Condor.
HPC Insight consolidates data from all HPC devices running legacy workload schedulers, normalizes the data, and presents it through feature-rich reports, a real-time dashboard and an integration path to meta-scheduling utilities like HPC Synergy. The ability to assess capacity and utilization across all enterprise HPC resources enables a number of important business benefits: Companies can now intelligently plan the HPC strategy and budget, justify the purchase of additional resources and avoid unnecessary hardware expenditures. Data delivered by HPC Insight can also be leveraged by HPC Synergy to automatically provision and align resources according to user and application priorities.
“HPC Insight is the first software product to effectively address the business side of HPC,” said Ben Rouse, chief executive officer of United Devices. “To properly define and evaluate the decision criteria for HPC strategy and budget planning scenarios, decision-makers need visibility across the entire HPC environment. It can't be done without a platform-agnostic capacity management tool that is cost effective to acquire and deploy. For companies who embark on grid or other global HPC initiatives, HPC Insight is a critical first step in implementing an automated optimization engine for enterprise workload scheduling.”
Before HPC Insight, IT professionals were forced to use traditional systems management tools that were too expensive and time consuming to deploy — or do nothing at all. HPC Insight, on the other hand, is affordable, easy to install and deploy, and requires no dedicated personnel to operate or maintain.
Deployment options scale from a non-intrusive implementation (with no device agents installed) to a fully managed environment. In the former case, data is gleaned from legacy scheduler artifacts such as log files and configuration files. In a fully managed scenario, UD agents interrogate system registries and other device-specific objects to collect a comprehensive set of data needed to fully express capacity and utilization trends and statistics.