Computational Systems & Services
As part of its ongoing commitment to developing a regional cyberinfrastructure, the Institute has acquired and, beginning in April of 2009, deployed a High Performance Computer Cluster (HPCC), designated hpc01.mkei.org, is a key resource in hosting high-fidelity modeling, simulation and visualization applications and services for regional .edu, .com and .gov institutions.
The system, located at Tushaus Computer Services (TCS), 10400 Innovation Drive, Milwaukee, WI is fully operational, having grown from its original 16 Dell PowerEdge 1950 III servers to 29 servers today. The first 16 compute nodes each contain two, quad-core Intel Xeon 2.83Ghz E5440 processors, 1333 MHz FSB with 32 GB RAM. The latest addition of 13 nodes leverage Intel’s Nehalem QPI architecture with each node having two, quad-core Intel Xeon 2.26Ghz E5520 24GB RAM, providing a combined 232 core system.
The HPCC runs the Red Hat Linux v5.2 operating system with Dell's Platform OCS cluster management suite. The 29 compute nodes are interconnected by a QLogic 20GB/s 9040 Infiniband Switch. The cluster has a theoretical capacity of over 3 teraflops and a measured Linpack Benchmark of over 2.5 teraflops.
The cluster’s head node is a Dell 1950 with two, quad-core Intel Xeon 2.83Ghz processors, 32GB RAM, dual 300GB drives in a RAID 1 configuration powered by redundant power supplies.
The network and shared storage system consists of a GigE Ethernet and a MD1000 RAID 5 a shared storage subsystem with 8.5 terabytes of available disk space.
Software available on the cluster currently consists of GNU C, C++, Fortran and MPI compilers and libraries, Intel C, C++, Fortran and MPI compilers and libraries, MathWorks' MATLAB R2009a Distributed Computing client workstation and cluster tools, and Wolfram Research's Mathematica 7 and gridMathematica client workstation and cluster tools.
Usage of the Institute HPCC is currently available to selected academic, educational, governmental and industrial users for development of non-commercial modeling, simulation and visualization applications.
Requests for access can be made to Jonathon Gladieux ([email protected], 414.727.6418) or Jay Bayne ([email protected], 414.727.6425). Applications also require website login credentials. The application requests may also be made at https://www.mkei.org/Logon.aspx.