HPC/Carbon Cluster - Overview: Difference between revisions
< HPC
Jump to navigation
Jump to search
mNo edit summary |
m (→Hardware) |
||
Line 13: | Line 13: | ||
=== Hardware === | === Hardware === | ||
[[Image:HPC Compute Rack-up.png|noframe|right|200px|]] | [[Image:HPC Compute Rack-up.png|noframe|right|200px|]] | ||
* | * 350 compute nodes, 8 cores/node (Xeon 2.5-2.7 GHz) | ||
* | * total: 2800 cores, 7 TB RAM, 40 TB disk | ||
* Infiniband interconnect | * Infiniband interconnect | ||
* | * 40 TB shared storage | ||
* Performance | * Performance 28 TFLOPS (aggregrate) | ||
* More [[HPC/Hardware Details|details on a separate page]] | * More [[HPC/Hardware Details|details on a separate page]] | ||
Revision as of 15:21, October 27, 2010
|
Primary Usage
- Modeling and Simulation
- CNM Theory and Modeling group
- CNM User Community
- Real-time on-demand data processing
- Nanoprobe Beamline
- E-beam lithography control
- Other high-data intensive instruments
Hardware
- 350 compute nodes, 8 cores/node (Xeon 2.5-2.7 GHz)
- total: 2800 cores, 7 TB RAM, 40 TB disk
- Infiniband interconnect
- 40 TB shared storage
- Performance 28 TFLOPS (aggregrate)
- More details on a separate page
Software
- Redhat Enterprise Linux 5/CentOS 5
- Moab Cluster Suite
- GNU and Intel compilers