3.2 Introduction of Local batch clusters

There are three local batch clusters at IHEP:

• The HTCondor cluster.
• The Slurm cluster.
• The Hadoop cluster.

The HTCondor cluster is built to support High Throughput Computing, and HTCondor is adopted as the Workload Management System. Most jobs running in the HTCondor cluster are single-core and serial jobs.

The Slurm cluster is built to support High Performance Computing, and Slurm is adopted as the Workload Management System. Most jobs submitted to the Slurm cluster are multi-core and parallel jobs.

The Hadoop cluster is a newly founded computing platform againest HEP requirements and LHAASO experiment is supported by the Hadoop cluster at present.

Details about job submission and job query will be given in the following sections.