Cluster Info
Math Cluster
The Math Department has an experimental computational cluster. The main purpose is for configuring and testing distributed computing applications so that they can be run on larger clusters. However, since our recent upgrades, the computational cluster is very fast for certain workloads.
The cluster is 64 machines with i7 processors. The first 54 are 6th-generation processors, and the last 10 are 7th generation processors.
Jobs can be sent to all nodes in the cluster, or a subset of the nodes.
At this time, the cluster is working with ssh MPI, Pari/gp using MPI, and Mathematica. We're in the process of getting it working with other applications including Matlab, Maple, and Magma. Many applications can be used for 'Pleasingly Parallel' workloads by running independent simultaneous programs using pdsh. Macaulay2 is available on all of the nodes for this purpose. Check back on this page because support is rapidly evolving.
Also, if you can get your own application running on the cluster, please write up a short description and put it on this wiki! You can edit these pages, and create a page with your instructions, linked from this page.
First Step: Setting up your Cluster SSH access, including testing your access with pdsh commands.
Next: Launching Mathematica Remote Kernels.
MPI: Trying out the MPI 'Hello World' program, which is a step to running many types of parallel jobs. MPI Hello World
Here is some info on running remote workers in Magma: Magma Cluster This is a work in progress since we don't have a nice, working example yet.