HPC2002 Result: IBM Corporation eServer 325 Cluster

Benchmark	Reference Time	Runtime	Ratio
361.wrf_m	86400	200	432
SPECenvM2002	432

Benchmark

Reference
Time

Runtime

Ratio

361.wrf_m

86400

200

432

SPECenvM2002

432

Hardware

Hardware Vendor:

IBM Corporation

Model Name:

eServer 325 Cluster

CPU:

AMD Opteron (246)

CPU MHz:

2000

FPU:

Integrated

CPU(s) enabled:

120

CPU(s) orderable:

1 or 2 per node

Primary Cache:

64KBI + 64KBD on chip

Secondary Cache:

1024KB(I+D) on chip

L3 Cache:

None

Other Cache:

None

Memory:

6GB DDR333 CL2.5 Registered ECC per node

Disk Subsystem:

1 x 36GB SCSI per node

Other Hardware:

See Notes.

Software

Parallel:

MPI

Processes-Threads:

120

MPI Processes:

120

OpenMP Threads:

N/A

Operating System:

SUSE LINUX 8.0 SLES 64 bit Kernel
k_smp-2.4.21-60 (from Service Pack 3)

Compiler:

Fortran: Portland Group 5.1-3 Fortran 90
C: Portland Group 5.1-3 C

File System:

Linux/reiserfs

System State:

Multi-user

Other Software:

MPICH-GM 1.2.5..10
GM 2.0.6_Linux
NetCDF 3.5.0

Notes / Tuning Information
Flags: Fortran: -O3 -fast C: -O3 -fast Preprocessor: -E -I. -I$(CDFINC) -I$(MPIINC) Linker: -Bstatic Extra Libraries: -lnetcdf -lmpich -lgm -lpthread Submit command to run applications: mpirun.ch_gm -np 120 -machinefile n00x2.mf --gm-recv blocking --gm-no-shmem -wd `pwd` $command Cluster Configuration: Two CPUs per node, 60 nodes All benchmark files are on a shared file server Nodes and file server use NFS shared file system $(CDFINC) is the NETCDF include directory $(MPIINC) is the MPICH include directory Computation Network: Myrinet M3F-PCIXD-2 Adapters (one per node) Myrinet M3-E128 Switch Enclosure 8 x Myrinet M3-SW16-8F 8-Port Line Card x345 File Server (2): 2 x 2.8 GHz Xeon Processors 4GB DDR266 CL2.5 Registered ECC Memory 2 x 36GB 10K RPM SCSI Drives Software RAID 0 Red Hat Linux 7.3 Ext3 Local File System NFS Shared File System File Server Network: Cisco 6509 Gigabit Ethernet Switch 3 x 48-port Copper Built-in Gigabit Ethernet Adapters Set the following BIOS parameters: DRAM Interleave = AUTO Node Interleave = Disabled ACPI SRAT = Enabled

Notes / Tuning Information

  Flags: 
    Fortran:
      -O3 -fast
    C:
      -O3 -fast
    Preprocessor:
      -E -I. -I$(CDFINC) -I$(MPIINC)
    Linker:
      -Bstatic
    Extra Libraries:
      -lnetcdf -lmpich -lgm -lpthread
 
  Submit command to run applications:
    mpirun.ch_gm -np 120 -machinefile n00x2.mf
      --gm-recv blocking --gm-no-shmem -wd `pwd` $command
 
  Cluster Configuration:
    Two CPUs per node, 60 nodes
    All benchmark files are on a shared file server
    Nodes and file server use NFS shared file system
    $(CDFINC) is the NETCDF include directory
    $(MPIINC) is the MPICH include directory
 
  Computation Network:
    Myrinet M3F-PCIXD-2 Adapters (one per node)
    Myrinet M3-E128 Switch Enclosure
    8 x Myrinet M3-SW16-8F 8-Port Line Card
 
  x345 File Server (2):
    2 x 2.8 GHz Xeon Processors
    4GB DDR266 CL2.5 Registered ECC Memory
    2 x 36GB 10K RPM SCSI Drives 
    Software RAID 0
    Red Hat Linux 7.3
    Ext3 Local File System
    NFS Shared File System
 
  File Server Network:
    Cisco 6509 Gigabit Ethernet Switch
       3 x 48-port Copper 
    Built-in Gigabit Ethernet Adapters
 
  Set the following BIOS parameters:
    DRAM Interleave = AUTO
    Node Interleave = Disabled
    ACPI SRAT       = Enabled

First published at SPEC.org on 03-Mar-2004

Generated on Wed Mar 3 17:51:48 2004 by SPEC HPC2002 HTML formatter v1.01