IGERT Brochure

Past Events:

March 4th,2009

IGERT Monthly Students Meeting
351 Johnston Hall
11:00 am

March 18th, 2009
 "Large Scale Computation Using LONI & TEZPUR" Speaker: Mr. Wes Even
(IGERT PhD Student)
2520 Patrick F. Taylor Hall
11am.

April 1st, 2009
 "GPU Programming and
Performance"
Speaker: Mr. Kevin Tubbs
(IGERT PhD Student)
2520 Patrick F. Taylor Hall 11:00 am

April 15th, 2009
 "Visualization Tools, VISIT and VISH and their Capabilities vs. Other Tools Available"
Speaker: Mr. Farid Harhad
(IGERT PhD Student)
3515 Patrick F. Taylor Hall
11:00 am

April 28th, 2009
"Compensation of Long
Input Delays for Unstable
Nonlinear and PDE Systems"
Guest Speaker:
Dr. Miroslav Krstic
338 Johnston Hall
3:00 pm

High Performance Computing

LSU will make its high performance computing resources hosted at the Center for Computation & Technology (CCT) available for all members of this project for software development and testing. Resources include :

• SuperHelix

  a 128 node dual processor (256 total processors) Red Hat Linux v7.3 IA32 Cluster (Xeon 2.0 GHz processors), 2 GB RAM, Myrinet interconnect

• MiniMike

  a 16 node dual processor (32 total processors) development Red Hat Enterprise Linux (RHEL) v3 cluster (Xeon 1.8 GHz processors), 2 GB RAM, Myrinet interconnect

• Nemeaux

  a 32 node dual processor (64 total processors) Mac OS X v10.4.1 (Darwin v7.9.0) 64-bit cluster (G5 2.0 GHz & 2.3 GHz 64-bit processors), 2 GB RAM, 80 GB HD, 3.5 TB RAID 5 Array attached to head node, GigE interconnect

• Santaka

  a single node multi-processor (32 total processors) SUSE Linux Enterprise Server (SLES) v9 IA64 cluster (Itanium 1.5 GHz 64-bit processors), 128 GB RAM, NUMA-link interconnect with 10 GigE

• GumboGrid

  an 8 processor Linux development environment for prototyping and testing Grid scenarios and infrastucture (dual 2.4GHz Xeon processors, Ethernet interconnect)

• Storage

  15 Terabytes of ad hoc storage used for /home/, /usr/local/compilers/, and /usr/local/packages/. There is also 32 TB of high-performance Panasas disk, accessible either through NFS or a proprietary parallel filesystem using a DirectFlow client on Linux systems.

The SuperMike and SuperHelix clusters are members of several Grids, including the heterogeneous EU GridLab testbed, based on Globus middleware. CCT research groups have large allocations on many supercomputers and different architecture machines around the world, which can be exploited for developing and testing portable software components.