| |
|
|
|
|
NCSA Intel 64 Tesla Linux Cluster Lincoln Technical Summary |
Production date: February 2, 2009
[lincoln.ncsa.uiuc.edu]
[login-lincoln.ncsa.teragrid.org]
| COMPONENT |
DESCRIPTION |
| Hardware & Networking |
Architecture |
Heterogeneous
- Dell PowerEdge 1950 servers
- NVIDIA Tesla S1070 Accelerator Units
- Each server connected to 2 Tesla processors
via PCI-e Gen2 X8 slots
|
| Peak FLOPs |
47.5 TF |
Number of Servers
Number of CPUs (cores) |
192
1536 |
| Number of Accelerator Units |
96 |
| Processor |
- Intel 64 (Harpertown) 2.33 GHz dual socket quad core
- 1333 MHz front side bus (per socket)
- 2x6 MB L2 cache
|
| Memory |
- Per node (per core): 16 GB (2 GB)
- Type: DDR2
|
| Accelerator Units |
- 4 Tesla Processors
- 4.147 teraflops single-precision performance
- 345.6 gigaflops double-precision performance
- 16GB memory
|
| Network Interconnect |
InfiniBand SDR (linked with Abe) |
| Parallel Filesystem |
Lustre (400 TB total; shared with Abe) |
| Software |
Operating System |
Red Hat Enterprise Linux 4 (Linux 2.6.19)
|
| Compilers |
- Intel: Fortran77/90/95 C C++
- GNU: Fortran77 C C++
|
| Batch System |
Torque (based on PBS)/Moab |
| Environment Control |
SoftEnv |
| Grid Software |
Provided via
Coordinated TeraGrid Software and Services (CTSS) |
| Policies/User Limits |
Home directory disk quota |
50 GBytes |
| Interactive scratch quota |
None currently |
Charging Algorithm
(charging as of March 2 2009) |
# SUs = 8 * #Nodes * Time
where
SUs = Service Units
Time = Total Wall Clock Hours
|
The Lincoln cluster is intended as a resource in the use of heterogeneous processors for scientific computing.
|
|
|
|
|
