(Click here for printable version)

Article Detail:

Starting from AXIS 12.5.03.001, the ability of Standard Distributed Processing to work over multiple computers will be disabled. You will still be able to use distributed processing with up to 15 local automatic helpers in one computer. If you are already using Distributed Processing over multiple machines, we will continue to support your setup for just over a year to allow for a transition to a single box solution. For versions starting with 12.5.03.001 you will need to request a custom code from GGY to enable this part of the functionality. We will be happy to discuss in more detail how your farm is set up and suggest the most appropriate long term solutions over the next year. The "Standard Distributed Processing Spanning Multiple Machines" feature will be completely removed from the system in the February release in 2011.

For more information please refer to the following article:
http://www.ggy.com/support/kbase/kbdetails.asp?searchterm=&articleid=1066

Distributed Processing allows AXIS calculations to be run in parallel over multiple processor cores in one computer significantly reducing the runtimes.

The standard AXIS license allows you to run an AXIS batch job using the Distributed Processing option with up to 15 Helpers.

This Standard Distributed Processing feature is a single user solution for reducing AXIS runtimes when complex processes on large volumes of data are involved.  Helpers used in Standard Distributed Processing mode are automatically started by the master copy of AXIS on the same machine as the master utilizing multiple processor (CPU) cores of the computer. Standard Distributed Processing works very well for a one-batch-at-a-time scenario if a user is in full control of a multi-processor multi-core computer.

AXIS GridLink vs. Standard Distributed Processing

AXIS GridLink is designed to provide a secure, robust production environment that can scale over multiple dedicated computers up to 100's of processor cores running multiple distributed jobs in parallel or in succession. Consider AXIS GridLink as a solution when:

• many jobs in different datasets need to be run
• end-user workstation need to be free to develop models
• required processing resources exceed 15 CPU cores (e.g. for stochastic processing)
• multiple users need to share the processor farm
• multiple versions are in use
• access control to data partitions is required
• teams of users need secure production/testing/development environment

For more information on AXIS GridLink please refer to www.ggy.com\gridlink

• Changes To Standard Distributed Processing Functionality In AXIS
• AXIS GridLink Module
• AXIS Performance and Scaling Considerations