Adaptive Computing Unveils Moab 9.0 to Deliver Ease-of-Use Driven Productivity for the HPC Administrator and End-User
HPC Today  |  Wire  |  November 17, 2015
Note: the wired news below has been filtered but not edited by HPC Today.

Moab 9.0 Features an Updated Viewpoint with End-User Portal, Industry Leading NUMA Capabilities and Docker Container Support

Adaptive Computing, which powers many of the world’s largest private cloud and technical computing environments with its Moab workload and resource orchestration platform, today announced general availability of Moab HPC Suite 9.0 (Moab 9.0). The new features include:

  • Docker Container Support, which increases productivity, efficiency and security
  • A new end-user job submission portal, that delivers greater end-user productivity and decreased administrator overhead
  • Non-uniform memory access (NUMA) capabilities, which delivers unparalleled control and performance over any other competing solutions

Together, these new improvements deliver ease-of-use driven productivity for the HPC Administrator and End-User. Adaptive Computing will offer demos on the new Moab 9.0 features in booth No. 833 at Supercomputing ’15 in Austin, TX, November 16-19, 2015.

“Moab has always been a tremendous productivity enabler for the administrator but now Adaptive is offering tools to support the end-user,” says Marty Smuin, CEO of Adaptive Computing. “Helping make end-users become more self-sufficient improves their individual productivity and frees up valuable administrative time creating greater overall efficiency.

Key updates to Moab 9.0 include the following:
Docker Container Support
Moab 9.0 comes complete with Docker container support to increase productivity, efficiency and security for the administrator and end-user. For the administrator, Moab allows for easy creation of containers and automatically tears down the container once the job has completed. Containers provide a unique environment that is capable of running different flavors of Linux, custom library versions and exclusive applications. This provides an additional level of security, as this unique environment is self-contained with no threat of contaminating the rest of the data center. As for the end-users, Moab allows for easy container job submission. The end user simply submits their jobs and specifies the appropriate container from an easy drop down menu. Moab Docker containers make it very easy to checkpoint progress so the end user’s job can be requeued to pick up where they left off.

Viewpoint’s End-User Job Submission Portal
Moab 9.0 offers a significant enhancement to Viewpoint, Moab’s web-based graphical user interface (web GUI), by adding an end-user job submission portal. This new feature creates greater self-sufficiency for the end-user while reducing administrative overhead. In addition, improvements have been made to the database access performance, which allows real-time reporting analytics for even the largest systems.

Viewpoint retains its great administrative dashboard that monitors and reports workload and resource utilization but now offers end-users access to the:

  • Application Template Manager –Within the web GUI template creator the end user can create, publish, delete or modify templates and drag and drop script builders.
  • Workload Lists – end-users can obtain a more holistic view of system workloads through to determine if a job is active, completed, eligible, suspended, blocked or failed.
  • File Manager – end-users can create new directories, set permissions for file folders, compress and un-compress files for transport, preview, copy, move, rename, delete and download files
  • .

NUMA Capabilities
Adaptive’s new NUMA capability in Moab 9.0 maximizes efficiency of available hardware by specifying the location and configuration of the lower level resources their jobs will use. When Moab places workloads on a processor it requires it to access its own local memory or closest memory instead of randomly assigning memory to another processor or having memory shared between processors. This leads to consistent and predictable job run times, which makes Adaptive Computing’s scheduler more accurate and efficient in maximizing utilization and job completions. In addition, Moab uses NUMA groups, which allow users to better enforce limits and collect information about their jobs; thereby, providing greater insight into job resource usage per task. Moab has a unique ability within NUMA to allow for truly heterogeneous environments, handling hardware and accelerators from multiple vendors on the same cluster. These abilities help applications run faster as resource allocations are optimized to their specific needs.

Source : Adaptive Computing
with Business Wire

© HPC Today 2024 - All rights reserved.

Thank you for reading HPC Today.

Express poll

Do you use multi-screen
visualization technologies?

Industry news

Brands / Products index