 |
» |
|
|
|
|
|
| High performance systems and clusters traditionally achieve very high utilization levels, fostered by the use of workload managers and schedulers that are designed to support multiple users and multiple HPC applications. Effective resource management utilizes an efficient queue management system that enables jobs, both batch and interactive, to be presented to the system based on site-defined policies and priorities. It also utilizes a scheduling system that can make the most efficient use of the hardware resources available. Schedulers allow administrators to set user priorities and policies, track usage, and provide a queuing system for job submissions. HP collaborates closely with the leading developers of schedulers and resource managers, to ensure simple deployment on our Cluster Platforms. In addition, the offerings have been enhanced for deployment with Insight CMU, such as integration into the management GUI and workload triggered automated provisioning. The solutions noted below are available directly from HP, in most countries.
|
|
|
|
|
|
|
Adaptive Computing offers two editions of Moab, optimized for HPC workloads.
Moab HPC Suite – Basic Edition is a policy-based HPC workload manager that integrates scheduling, managing, monitoring, and reporting of cluster workloads. It guarantees service levels are met while maximizing job throughput. It includes the patented Moab intelligence engine that integrates with existing middleware for unified administrative control, optimization, and reporting of cluster workloads. Its graphical management interfaces and flexible policy capabilities result in decreased costs and increased ROI.
Moab HPC Suite – Enterprise Edition provides enterprise-ready HPC workload management that self-optimizes the productivity, workload uptime, and meeting of SLAs and business priorities for HPC systems and HPC Cloud using the patented Moab intelligence engine. It is designed to meet the mission-critical HPC workload management needs for commercial businesses and business-oriented research and academic organizations. Enterprise customers will benefit from a single integrated solution that brings together the key enterprise HPC uses cases and capabilities as well as implementation, training, and 24x7 support services to speed the realization of the benefits for their HPC system and business.
|
|
|
|
|
|
|
|
PBS Professional® is a very robust and mature workload manager. Originally developed at NASA Ames in the early 1990s it now runs on many of the world’s largest HPC systems, including TSUBAME 2.0. PBS Professional captures the policies that define the service level agreements of an HPC service and helps ensure maximum system utilization and job throughput as well as shortest execution time. Balancing the business and political pressures from end-users, system administrators and service providers is a complex challenge. PBS Professional delivers the most capable workload management solution to enforce the business and political rules and provide a scalable, robust, deterministic service.
PBS Professional provides full support for the latest GPGPU systems. To optimally handle GPGPUs, PBS Professional is able to capture the physical relationships between GPGPUs and their closely coupled CPUs. This ensures that when physical resources are being assigned to jobs the closest GPGPUs and CPUs are chosen which helps reduce execution time by reducing latency between components. PBS Professional has unique capability to decompose complex systems into smaller entities, vnodes, which are able to run jobs. This provides benefits to both exotic shared memory systems and GPGPU systems.
PBS Professional and HP Datasheet
|
|
|
|
|
|
|
Platform LSF is a very popular and powerful workload manager, ideal for demanding, distributed high performance computing environments. It includes a rich set of policies for scheduling and prioritizing jobs based on combinations of static and dynamic system attributes. An extremely reliable, and richly configurable network-based queuing system provides maximum flexibility and central administration of all work in the cluster while ensuring that all work submitted to the cluster runs to completion. The software also includes comprehensive job management such as tracking and control. Platform LSF offers full compatibility with HP Insight CMU for simplified job management
Platform LSF is available standalone, but also as part of the specially-priced HP HPC Linux Value Pack which also includes HP-MPI and Platform’s Application Portal, and available with support from HP Services.
|
|
|
|
|
|
Printable version
