Why HPC workload management? Companies need to ensure all their costly HPC computing resources are being used wisely, from clusters to supercomputers to on- or off-premise cloud resources. In addition, these resources must be easy to use and manage. For this reason, HPC workload management is a critical component of any HPC system.
The necessary elements of an HPC workload management solution are:
- Job scheduling software to intelligently assign workload to resources based on availability, priorities, policies, job requirements and other data
- Workload manager to execute jobs (whether a large number of single-core jobs or a single, massively parallel job) and monitor progress on nodes in the system, managing completion of jobs and aggregating results
- User-friendly graphical interfaces (GUIs) for submitting jobs, monitoring progress and displaying and analyzing job output
“PBS Works is a key ingredient for HPC workload management.”See Press Release
-Intel Director, HPC and Many-Core Computing
delivers a superior range and quality of workload management capabilities for HPC users. With the PBS Works suite, users can improve ROI on cluster and supercomputing systems by optimizing how workload is scheduled, executed and managed across these resources; PBS Works intelligently automates processes and streamlines workflow, increasing system utilization and productivity. Using PBS Works, HPC users can be sure their policies, requirements and SLAs will be honored, while job speed and system performance is maximized.
PBS Works also simplifies the process of adopting and using HPC via powerful yet easy-to use portals for job submission and management, remote visualization
– easing cluster management and supporting the transition to HPC for small-medium businesses (SMBs).
Read more about PBS Works for Workload Management.