The Batch System (SGE)

Several of the CIR compute systems require that all codes and applications be run via a batch system on those compute systems. In particular the CSF, Hydra, Redqueen, and Zrek all run their own batch system.

We run the Sun Grid Engine (SGE) batch system on the above CIR compute resources.

Why use a Batch System

Batch systems control exactly when and where jobs will run on a particular system. They also allow you to request specific resources (e.g., Intel vs AMD hardware, high-memory nodes). Only when those resources are available will your application, or job run. Batch systems also ensure fair usage – there are many users and jobs on the CIR systems, all making different demands on the resources (memory, CPU cores, networking) and so allowing the batch systems to choose exactly when to run your job is the only sensible way of running the system.

The following pages document commands and options found in SGE. These can be used on the CSF, Hydra, Redqueen and Zrek. Some options will be specific to certain systems and this will be indicated where appropriate.