| 產品介紹 > 中介軟體
> 計算工作管理套件> Moab Adaptive Computing |
| Moab HPC Products |
New Moab® 6.1 Delivers Massive Scalability for High Throughput Computing and Maximized GPGPU Utilization to Accelerate Your HPC Results and Value
The new Adaptive Computing Moab 6.1 HPC product family delivers massive scalability for high throughput computing, productivity improvements for job arrays, and maximized GPGPU utilization for large-scale, heterogeneous HPC environments. The new 6.1 generation of the Moab HPC product family includes Moab Cluster Suite™ 6.1, Moab Grid Suite™ 6.1 and Moab Adaptive HPC Suite™ 6.1 for dynamic workload and resource management that self-optimizes HPC environments.
|
New enhancements and features in the Moab 6.1 product family include:
•Powerful 10x scalability improvements drive productivity by handling and running more jobs faster, supporting high throughput computing and providing faster command responsiveness. Moab 6.1 HPC products now scale to support the hundreds of thousands of job submissions demanded by high throughput computing and provides sub-second user command responsiveness. This will enable current and new customers to accelerate the job throughout and results they see from their HPC investments. And we've based these scalability improvements on our experience and expertise in managing the most scale-intensive and complex HPC environments in the world, including over one third of the top 50 and over one third of the compute cores in the top 100 systems in the Top500 supercomputers.
•Extended and simplified productivity controls over job arrays as Moab 6.1 HPC products now offer extended control over job array submissions to ensure SLAs and resource sharing policies are enforced with usage limit policies that administer number of job maximums by credentials or class. New, simplified reporting and management of job arrays make it easier for users with filtering that allows jobs to be summarized, tracked and managed at the master job level. Additionally, in Moab Grid Suite 6.1, enhanced grid workload management options for job arrays allow users to select optimal job placement across clusters or require single cluster placement. These enhancements build on Adaptive Computing's leadership in multi-dimensional analytic policies that model IT resources, workloads, SLAs, and priorities across domains and silos to automate real-time decisions and actions.
•Advanced GPGPU scheduling and management allowing Moab 6.1 to maximize GPGPU utilization with new automatic detection, scheduling, and policy-based management of individual GPGPUs for each job. Moab 6.1 is able to provide and intelligently use statistics and key metrics-such as error counts, temperature, fan speed or memory-on GPGPU utilization for optimal job placement as well as to enable issue diagnostics for administrators. These new GPGPU features leverage Adaptive Computing's unique management abstraction layer that integrates and organizes the chaos of complex heterogeneous IT environments to enable maximum control and optimization.
•Simplified installation for faster time to value with the new Moab Solution Installer that guides customers through the basic installation process and the ability to install only client commands on additional machines.
•Optional Moab Viewpoint™ portal is easier to customize to your organization's needs with new user interface framework enhancements that enable you to simplify job submission, reporting and training for users and administrators for improved productivity and visibility. This is also part of our goal for Moab Adaptive HPC Suite to enable organizations to leverage the best aspects of the cloud model for their traditional HPC environments today with optional integrated HPC cloud capabilities including this user self-service portal along with dynamic resource provisioning and usage billing & chargeback. |
| |
|
| |
|
|
|
| Adaptive Computing |
 |
|
