High-performance computing has always been fueled by the insatiable desire to solve the most complex problems confronting scientists and engineers – from academia to leading edge industries – HPC has powered nearly 50 years of innovation. Along the way, as HPC evolved beyond the data center and computing grids, a new method for exploiting HPC has begun to emerge: technical HPC ‘clouds’.
Technical computing clouds build on the accepted trends for driving the costs out of the delivery of computing while increasing the rate at which new services can be created. Cloud computing has the potential to redefine how we create, deploy, scale and support technical computing infrastructures. To the extent that technical cloud computing helps maximize the productive use of all HPC assets, helping to increase your ability to cost-effectively exploit the bottom-line business benefits from HPC investments, cloud computing will continue to evolve to support future applications that we cannot imagine today.
Powerful software tools are needed to manage and match the dynamic, quickly changing workloads with available resources – including energy.
PBS Professional is a core component of the PBS Works suite that efficiently distributes HPC workloads across all types of computing infrastructure – from workgroup clusters to clouds. Monitoring how efficiently resources are being used – from energy to software licenses is key to utilizing clouds wisely. Green Provisioning, which is part of PBS Professional, helps users extract the most from their investments in HPC clouds by shutting down and restarting underutilized systems, often lowering energy use by 20%.
PBS Analytics provides HPC administrators with a suite of simple, customizable tools for analyzing and viewing complicated job data as well as software license assets – providing the visibility into how well their expensive HPC resources are being utilized. PBS Analytics turns oceans of data into insight, knowledge and intelligence.
PBS Catalyst provides a simple and productive 'drag and drop' user interface for submitting, monitoring and managing jobs using HPC clouds. End user productivity improves since there are no complicated commands to learn, allowing users to spend more time creating better models and simulations.