The Dogwood cluster is a Linux-based computing system available to researchers across the campus free of charge. With over 11,000 computing cores, a low latency, high bandwidth Infiniband EDR switching fabric and a large scratch disk space, it provides an environment that is optimized for large, multi-node, tightly coupled computational models typically using a message passing (e.g. MPI) or hybrid (OpenMP + MPI) programming model. Most of the cluster is comprised of nodes with Intel Xeon processors with the Broadwell-EP micro-architecture and 44 cores per node. There is a partition with some newer Intel Skylake nodes as well as a smaller knl partition, with Intel Xeon Phi processors (Knight’s Landing) for development purposes. Resources are managed through a fair-share algorithm using SLURM as the resource manager/job scheduler.

To learn more:
Getting started on Dogwood
Getting an account on Dogwood
Dogwood SLURM examples
Dogwood system details
Dogwood MPI modules
Dogwood partitions and limits

Sign up for the next “Using Dogwood” Research Computing Training Course:
Research Computing Course Calendar.

Review Research Computing Course Materials:
Using Dogwood