| |
|
|
|
|
NCSA Dell NVIDIA Linux Cluster Forge Technical Summary |
[forge.ncsa.illinois.edu]
| COMPONENT |
DESCRIPTION |
| Hardware & Networking |
Architecture |
Heterogeneous
- Dell PowerEdge C6145 servers
- NVIDIA Fermi M2070 Accelerator Units
- 32 nodes connected to 6 Fermi processors
via three PCI-e Gen2 X16 slots
- 12 nodes connected to 8 Fermi processors
via four PCI-e Gen2 X16 slots
(as of March 2012)
|
| Peak FLOPs |
150 TF |
Number of Servers
Number of CPUs (cores) |
44
704 |
| Number of Accelerator Units |
288 |
| Processor |
- AMD Opteron Magny-Cours 6136 2.4 GHz dual-socket eight-core
- 1333 MHz front side bus (per socket)
- L3-Cache: 2x6 MB, shared
(only 10 MB is visible due to the HT Assist feature using 2 MB as a directory cache)
|
| Memory |
- Per node (per core): 48 GB (3 GB)
- Type: DDR3
|
| Accelerator Units |
- 448 CUDA Cores
- 1.03 teraflops single-precision performance
- 515 gigaflops double-precision performance
- 6GB memory
|
| Network Interconnect |
InfiniBand QDR |
| Parallel Filesystem |
GPFS (600 TB total) |
| Software |
Operating System |
Red Hat Enterprise Linux 6 (Linux 2.6.32)
|
| Compilers |
- Intel: Fortran77/90/95 C C++
- GNU: Fortran77 C C++
|
| Batch System |
Torque (based on PBS)/Moab |
| Environment Control |
Modules |
| Grid Software |
Provided via XSEDE |
| Policies/User Limits |
Home directory disk quota |
50 GBytes |
| Interactive scratch quota |
None currently |
| Charging Algorithm |
# SUs = 16 * #Nodes * Time
where
SUs = Service Units
Time = Total Wall Clock Hours
|
The Dell NVIDIA cluster is intended as a resource in the use of heterogeneous processors for scientific computing.
|
|
|
|
|
