Cluster Descriptions

~200 processor Windows based cluster~
(in development)

The latest and greatest in the BCRL! Microsoft donated this "super computer" in the summer of 2009. Current power restrictions will soon be overcome, allowing BCRL staff to power the five large racks of computers. The goal for this system is to implement Microsoft's brand new High Performance Computing (HPC) software. Once completed, this system will serve distrbuted database and processing technologies. Capable of running MPI, and the more recent, OpenMP message passing implementations. Graduate students recieve hands-on experience programming and analyzing data flows. The system divides large jobs and sends the pieces to the indivdual computers to be processed. Upon completion of each machines tasks, the system then recompiles the data for study.

Physical aspects of this system include 100 rackmount nodes, each with 2 AMD Opteron 2.0 Ghz 64bit processors. Each node also houses 8 GB of RAM and 4 250 GB Hard Disks, bringing the total storage space to 100 Terabytes.
Gigabit ethernet connections interconnect the nodes, ensuring high speed connectivity.

~96 processor Ubuntu based cluster~
(in development)

At one point, this cluster was built to run Sun Enterprise's Sun Grid Engine (SGE) cluster resource management software. Now, the hardware is being reconfigured with Ubuntu, a linux distibution. Once configured, researchers plan to create a large Hard Disk pool. This "Mass Storage" capabilty will allow the research staff to further their investigation in both distributed encryption and database technologies.

Physical aspects of this cluster include 48 Intel nodes, each with two 550 Mhz PIII processors. They also have 512 MB of RAM and a 40 GB Hard Disk. 100 Mbs ethernet interfaces link the nodes.

~36 processor Fedora based cluster~

This cluster provides entry-level programmers a basic understanding of the message passing implementation (MPI). Although MPI is becoming somewhat outdated. This system is a great starting point for both undergraduate and graduate-level researchers. Students can write their own MPI programs, designed to 'split' themselves and run on multiple computers. Once written, they can then execute the program on the cluster. The result: programs that may take years to run on one machine can have their runtime drastically reduced.
This cluster is configured with Fedora, a powerful linux distribution, and MPICH2 message passing implementation

Physical aspects of this cluster include 18 Intel nodes, each with two 550 Mhz PIII processors. They also have 512 MB of RAM and a 5 GB hard disk. 100 Mbs ethernet interfaces interconnect the nodes.

~30 processor Solaris based cluster~

This system, configured in 2005-06, is configured to run LAM MPI (message passing implementation). For three years it has been used by Saint Cloud State University's chemistry staff to ananlyze molecular modeling and blood-flow. These technologies can be used to simulate stint designs. Studying these simulations, researchers can desing stints that increasing blood flow in bypass patients.

Molecular modeling simulations ( at the 'nanotech' level), provides scientists with data that they can use to better understand chemical-tissue reactions.

Physical aspects of this cluster include 30 Sun nodes, each with a 64 bit Sparc processor. 128 MB of RAM and 20 GB hard drives. 100Mbs ethernet connections interlink the nodes to complete the system.

~8 processor OpenSolaris based cluster~
(in development)
This cluster, currently being configured, is an experiment with both; the OpenSolaris operating system, and Carr-Parrinello Molecular Modeling (CPMD). Dr. Renat Sultanov is hopeful that this system will provide him and his students a powerful testbed for his molecular modeling research at the 'nanotech' level.

Physical aspects of this system include 4 rackmount nodes, each with 2 AMD Opteron 2.0 Ghz 64bit processors. Each node also houses 8 GB of RAM and 4 250 GB Hard Disks.
Gigabit ethernet connections interconnect the nodes, ensuring high speed connectivity.