The STK Parallel Computing Server allows you to do more in less time. As a parallel computing solution, the STK Parallel Computing Server distributes computations across multiple cores/machines on a network to take advantage of their resources and improve performance time for what could otherwise be lengthy calculations.
Using the STK Parallel Computing Server, STK already has built-in support for completing the following computations in parallel:
- Coverage and Figure of Merit
- Deck Access
- Chain Access
In addition to the built-in computations, you can use the STK Parallel Computing Server API in conjunction with STK Engine, Components, and ODTK to parallelize and distribute your own analysis applications (see The Power Behind Automating Custom STK-Based Analytics). The STK Parallel Computing Server API is available for .NET, Java, and—new for the 1.5 release—Python (beta)!
Whether you use the STK Parallel Computing Server API on its own or in conjunction with STK, it is useful to monitor machine resources and the progress of jobs submitted to the parallel cluster. Previously, this was only possible using the included Windows tray applications available to users with access to the central Coordinator machine on the cluster, which is often located remotely or can have restricted user permissions. Included in the release of STK Parallel Computing Server 1.5 is a new Scalability Portal—a web application for configuring and monitoring the activities of the central Coordinator and worker Agent machines on a cluster. Unlike the tray monitoring applications, the Portal is accessible from anywhere on the network.
The Portal includes a dashboard showing an overview of the status of the Coordinator and its registered Agents. It also includes a Job Queue displaying what parallel computations have been submitted to the system, how far along they are if the job is currently running, or whether the job has successfully completed. In the case of a failure, the Portal makes it easy to view the standard output/error and host log for any parallel process to aid in debugging.
The Scalability Portal comes with its own account system and is also compatible with Lightweight Directory Access Protocol (LDAP) and OpenID Connect (e.g. Keycloak) to work with existing user management infrastructures.
Parallel computing can require a lot of resources and produce a lot of output. Our goal with the new STK Parallel Computing Portal was to make that information easier to access, view, and manage from anywhere. Give it a try today!
Other new features included in the STK Parallel Computing Server 1.5 release include support for UTF-8 character encoding and a switch to workstation garbage collection for reduced memory consumption. To see what else is new, check out the STK Parallel Computing Server release notes.