Skip to main content
decorative
decorative
Home > Physics and Engineering > News > high-performance cluster for computational physics
Using view 'Single'
Version HistoryVersion History

Title

high-performance cluster for computational physics 

Story Author

Blake Laing 

Story Image

 

Story Date

8/6/2013 

Story Abstract

The "Narnia" cluster will provide Southern students experience with high-performance computational physics in classes and in research projects.

Story

The Department of Physics and Engineering has deployed a High-Performance Computing (HPC) cluster for use by students in their research and class work. This cluster consists of two Dell PowerEdge R620 servers, each with two 8-core Intel Xeon processors (Intel Xeon E5-2650 2.00GHz, 20M Cache, 8.0GT/s) and each with 32GB RAM (1600 MHz). The nodes of the "Narnia cluster" are named after characters from "The Chronicles of Narnia", and access to the cluster starts with the head node "Tumnus". The cluster runs the Ubuntu distribution of the GNU/Linux operating system and is administered by Professor John Beckett of the School of Computing.
 

Why is this a department priority? Because Moore's law has failed--processor speeds no longer double every 18 months. Instead, processors are coming with more cores and the only way to continue to increase computational capabilities is to learn to program in parallel. The Narnia cluster may be small, but it has a similar topology and the GNU/Linux operating system used by nearly all of the top supercomputers in the world. We expect that experience with this resource will open doors for our students.


The use of the physics "Narnia" cluster has gained momentum over the last academic year.

  • Students Sean Bryant and Greg Ngirmang worked with Professor Ken Caviness to use the cluster as a "Remote Kernel" to run some parallel Mathematica calculations for a research project.
  • Student Joshua Barrow worked with Professor Blake Laing to implement some FORTRAN codes which use MPI to calculate integrals on multiple cores in parallel.
  • Student Jonathan Sackett wrote an integration code to calculate a property of cold fermions in a box for a class project in Modern Physics (he actually developed this code on his own computer, but I include this as an example of an increasing emphasis on computation).

We expect to extensively utilize the cluster in the coming academic year and will soon report on some very interesting student projects in the works! For instance, Professor Blake Laing will use the cluster to investigate a model of above-threshold ionization and to calculate properties of three trapped particles with an electric dipole moment.

 

 
Approval Status Approved 
 
Attachments
Version: 4.0 
Created at 8/6/2013 4:31 PM  by Blake Laing 
Last modified at 8/6/2013 4:36 PM  by Blake Laing 
high-performance cluster for computational physics
by Blake Laing
August 06, 2013

The Department of Physics and Engineering has deployed a High-Performance Computing (HPC) cluster for use by students in their research and class work. This cluster consists of two Dell PowerEdge R620 servers, each with two 8-core Intel Xeon processors (Intel Xeon E5-2650 2.00GHz, 20M Cache, 8.0GT/s) and each with 32GB RAM (1600 MHz). The nodes of the "Narnia cluster" are named after characters from "The Chronicles of Narnia", and access to the cluster starts with the head node "Tumnus". The cluster runs the Ubuntu distribution of the GNU/Linux operating system and is administered by Professor John Beckett of the School of Computing.
 

Why is this a department priority? Because Moore's law has failed--processor speeds no longer double every 18 months. Instead, processors are coming with more cores and the only way to continue to increase computational capabilities is to learn to program in parallel. The Narnia cluster may be small, but it has a similar topology and the GNU/Linux operating system used by nearly all of the top supercomputers in the world. We expect that experience with this resource will open doors for our students.


The use of the physics "Narnia" cluster has gained momentum over the last academic year.

  • Students Sean Bryant and Greg Ngirmang worked with Professor Ken Caviness to use the cluster as a "Remote Kernel" to run some parallel Mathematica calculations for a research project.
  • Student Joshua Barrow worked with Professor Blake Laing to implement some FORTRAN codes which use MPI to calculate integrals on multiple cores in parallel.
  • Student Jonathan Sackett wrote an integration code to calculate a property of cold fermions in a box for a class project in Modern Physics (he actually developed this code on his own computer, but I include this as an example of an increasing emphasis on computation).

We expect to extensively utilize the cluster in the coming academic year and will soon report on some very interesting student projects in the works! For instance, Professor Blake Laing will use the cluster to investigate a model of above-threshold ionization and to calculate properties of three trapped particles with an electric dipole moment.

 

decorative
decorative