OU Supercomputing Center for Education & Research
University of Oklahoma   OSCER   OU IT

 

 

OSCER Resources


Cluster: Dell Linux Pentium4 Xeon
topdawg.oscer.ou.edu

Cluster Hardware

Item Description Quantity
Performance GFLOP/s 6,553.6 theoretical peak
Racks Dell PowerEdge 4210 42U 29
Nodes
Compute 512 PowerEdge 1850
Head (interactive logins) 2 PowerEdge 2850
Storage 8 PowerEdge 2850
Grid 2 PowerEdge 2850
Management 2 PowerEdge 2850
ITA 1 PowerEdge 1850
Lustre Metadata Server 1 PowerEdge 1850
CPUs Pentium4 Xeon "Irwindale"
3.2 GHz, EM64T, 2 MB L2 cache
2 per node
Motherboard/
Chipset
Dell PowerEdge/Intel E7520
Compute, spare, ITA, Lustre Metadata Server: PowerEdge 1850
Head, storage, grid, management: PowerEdge 2850
1 per node
Main Memory
DDR 400 MHz (800 MHz FSB)
Compute (512) 480 x 4 GB
32 x 8 GB
= 2,176 GB total
Head (2) 8 GB
Storage (8) 8 GB
Grid (2) 8 GB
Management (2) 8 GB
ITA (1) 4 GB
Lustre Metadata Server (1) 4 GB
Hard Disk
(SCSI except as noted)
Storage Area Network Data Direct
/home, /scratch: SATA; 18,626 GB (13,853 GB useable)
/home NFS 4,408 GB useable
/data NFS 4,408 GB useable
/scratch     Lustre     5,037 GB useable
EON RAID SATA; 11,175 GB (9,083 GB useable)
Compute (512) OS & local scratch:
1 x 73 GB each (62 GB useable),
37,376 GB total (31,744 useable)
Head (2) OS: 2 x 73 GB (RAID1) each (62 GB useable),
292 GB total (124 useable)
Storage (8) OS: 2 x 73 GB (RAID1) each (62 GB useable),
1,168 GB total (496 useable)
Management (2) OS: 2 x 73 GB (RAID1) each (62 GB useable),
292 GB total (124 useable)
Grid (2) OS: 5 x 300 GB (RAID5) each, 3,000 GB total
Note: 1 Grid node has a hot spare (906 GB useable),
1 Grid node has no hot spare (1,037 GB useable)
ITA (1) OS: 2 x 73 GB (RAID1) each (62 GB useable),
292 GB total (124 useable)
Lustre Metadata Server (1) OS: 2 x 73 GB (RAID1) each (62 GB useable),
292 GB total (124 useable)
GRAND TOTAL
(SAN, OS, local scratch)
72,513 GB (57,615 GB useable)
Interconnects
High Performance InfiniBand Topspin Communications/Cisco
120 switches and 270 switches
Benchmark: bandwidth 5 Gbps, latency 3 microseconds
I/O GigaBit Ethernet Force10 Networks E600/E1200 (576 ports total)
Benchmark: bandwidth 976 Mbps, latency 9 microseconds
Management 100 Mbps Ethernet Dell PowerConnect 3348 (768 ports total)
Programming Model Message Passing  

Cluster Software

Operating System Red Hat Linux Enterprise 4.0 (kernel 2.6.9)
Message Passing MPICH
Scheduler Platform LSF HPC
Compilers GNU: gcc, g++, g77, gfortran (Fortran 90/95)
Intel: ifort (Fortran 77/90), icc (C/C++)
NAG: f95 (Fortran 90/95; front end to gcc)
Portland Group: pgf90, pgf77, pghpf, pgcc, pgCC
Numerical Libraries BLAS, LAPACK, scaLAPACK
Parallel Debugger TotalView
Code Analyzer LintPlus for C, Fortran Lint
Other software purchases to be announced


This resource and its availability are subject to change without notice (although notice will be given if at all possible).

 

 


Copyright (C) 2005 OU Supercomputing Center for Education & Research