SPEC® MPIM2007 Result

Copyright 2006-2010 Standard Performance Evaluation Corporation

Celestica

A2210 ("Serenade") -- Reference Platform

MPI2007 license: 0017 Test date: May-2007
Test sponsor: Advanced Micro Devices Hardware Availability: Jan-2004
Tested by: Chris Parrott Software Availability: May-2007
Benchmark results graph

Results Table

Benchmark Base Peak
Ranks Seconds Ratio Seconds Ratio Seconds Ratio Ranks Seconds Ratio Seconds Ratio Seconds Ratio
Results appear in the order in which they were run. Bold underlined text indicates a median measurement.
104.milc 16 1582 0.989 1550 1.01  1546 1.01  16 1582 0.989 1550 1.01  1546 1.01 
107.leslie3d 16 5113 1.02  5235 0.997 5231 0.998 16 5113 1.02  5235 0.997 5231 0.998
113.GemsFDTD 16 6617 0.953 6387 0.988 6569 0.960 16 6617 0.953 6387 0.988 6569 0.960
115.fds4 16 1970 0.990 1994 0.978 1994 0.978 16 1970 0.990 1994 0.978 1994 0.978
121.pop2 16 4172 0.989 4153 0.994 4100 1.01  16 4172 0.989 4153 0.994 4100 1.01 
122.tachyon 16 2781 1.01  2774 1.01  2785 1.00  16 2781 1.01  2774 1.01  2785 1.00 
126.lammps 16 2968 0.982 2910 1.00  2891 1.01  16 2968 0.982 2910 1.00  2891 1.01 
127.wrf2 16 7785 1.00  7814 0.998 7772 1.00  16 7785 1.00  7814 0.998 7772 1.00 
128.GAPgeofem 16 2099 0.984 1993 1.04  2084 0.991 16 2099 0.984 1993 1.04  2084 0.991
129.tera_tf 16 2754 1.01  2769 1.00  2756 1.00  16 2754 1.01  2769 1.00  2756 1.00 
130.socorro 16 3830 0.997 3825 0.998 3830 0.997 16 3830 0.997 3825 0.998 3830 0.997
132.zeusmp2 16 3101 1.00  3104 1.00  3099 1.00  16 3101 1.00  3104 1.00  3099 1.00 
137.lu 16 3526 1.04  3526 1.04  3562 1.03  16 3526 1.04  3526 1.04  3562 1.03 
Hardware Summary
Type of System: Homogenous
Compute Node: Baseline Cluster Node
Interconnect: Baseline Cluster Interconnect
File Server Node: Baseline Cluster Node
Head Node: Baseline Cluster Node
Total Compute Nodes: 8
Total Chips: 16
Total Cores: 16
Total Threads: 16
Total Memory: 64 GB
Base Ranks Run: 16
Minimum Peak Ranks: 16
Maximum Peak Ranks: 16
Software Summary
C Compiler: Pathscale 2.5
C++ Compiler: Pathscale 2.5
Fortran Compiler: Pathscale 2.5
Base Pointers: 64-bit
Peak Pointers: Not Applicable
MPI Library: MPICH2 1.0.3
Other MPI Info: None
Pre-processors: No
Other Software: ACML 3.6.0 (BLAS/LAPACK routines)

Node Description: Baseline Cluster Node

Hardware
Number of nodes: 8
Uses of the node: compute, head, fileserver
Vendor: Celestica
Model: A2210 ("Serenade")
CPU Name: AMD Opteron 848
CPU(s) orderable: 1-2 chips
Chips enabled: 2
Cores enabled: 2
Cores per chip: 1
Threads per core: 1
CPU Characteristics: --
CPU MHz: 2200
Primary Cache: 64 KB I + 64 KB D on chip per chip
Secondary Cache: 1 MB I+D on chip per chip
L3 Cache: None
Other Cache: None
Memory: 8 GB
Disk Subsystem: 36 GB SCSI disk [Seagate ST336607LC - 10000 RPM]
Other Hardware: --
Adapter: Broadcom BCM95704
Number of Adapters: 2
Slot Type: PCI (built into chipset)
Data Rate: 1 Gbps
Ports Used: 1
Interconnect Type: Gigabit Ethernet
Software
Adapter: Broadcom BCM95704
Adapter Driver: --
Adapter Firmware: --
Operating System: SLES 9 SP3
Local File System: ReiserFS
Shared File System: NFS
System State: multi-user
Other Software: --

General Notes

Only the first node serves as a head and fileserver node, in addition to
being a compute node.  All other nodes are compute nodes only.

Interconnect Description: Baseline Cluster Interconnect

Hardware
Vendor: Broadcom
Model: BCM95704
Switch Model: SMC TigerSwitch 10/100/1000 8624T
Number of Switches: 1
Number of Ports: 16
Data Rate: 1 Gbps
Firmware: --
Topology: star (single switch)
Primary Use: MPI and filesystem traffic

Base Compiler Invocation

C benchmarks:

 /opt/mpich2-1.0.3-pathscale-2.4/bin/mpicc 

C++ benchmarks:

126.lammps:  /opt/mpich2-1.0.3-pathscale-2.4/bin/mpicxx 

Fortran benchmarks:

 /opt/mpich2-1.0.3-pathscale-2.4/bin/mpif90 

Benchmarks using both Fortran and C:

 /opt/mpich2-1.0.3-pathscale-2.4/bin/mpicc   /opt/mpich2-1.0.3-pathscale-2.4/bin/mpif90 

Base Portability Flags

115.fds4:  -DSPEC_MPI_LC_TRAILING_DOUBLE_UNDERSCORE 
121.pop2:  -DSPEC_MPI_DOUBLE_UNDERSCORE 
126.lammps:  -DMPICH_IGNORE_CXX_SEEK(*) 
127.wrf2:  -DF2CSTYLE   -DSPEC_MPI_DOUBLE_UNDERSCORE   -DSPEC_MPI_LINUX 
130.socorro:  -fno-second-underscore 

(*) Indicates a portability flag that was found in a non-portability variable.

Base Optimization Flags

C benchmarks:

 -O2   -LANG:copyinout=OFF 

C++ benchmarks:

126.lammps:  -O2   -LANG:copyinout=OFF 

Fortran benchmarks:

 -O2   -LANG:copyinout=OFF 

Benchmarks using both Fortran and C:

 -O2   -LANG:copyinout=OFF 

Base Other Flags

C benchmarks:

122.tachyon:  -I. 

Fortran benchmarks:

 -I. 

Benchmarks using both Fortran and C:

115.fds4:  -I. 

Peak Optimization Flags

C benchmarks:

104.milc:  basepeak = yes 
122.tachyon:  basepeak = yes 

C++ benchmarks:

126.lammps:  basepeak = yes 

Fortran benchmarks:

107.leslie3d:  basepeak = yes 
113.GemsFDTD:  basepeak = yes 
129.tera_tf:  basepeak = yes 
137.lu:  basepeak = yes 

Benchmarks using both Fortran and C:

115.fds4:  basepeak = yes 
121.pop2:  basepeak = yes 
127.wrf2:  basepeak = yes 
128.GAPgeofem:  basepeak = yes 
130.socorro:  basepeak = yes 
132.zeusmp2:  basepeak = yes 

The flags file that was used to format this result can be browsed at
http://www.spec.org/mpi2007/results/flags/MPI2007_flags.20100413.09.html.

You can also download the XML flags source by saving the following link:
http://www.spec.org/mpi2007/results/flags/MPI2007_flags.20100413.09.xml.