| MPI2007 license: | 0001 | Test date: | May-2007 |
|---|---|---|---|
| Test sponsor: | Hewlett-Packard Company | Hardware Availability: | May-2007 |
| Tested by: | HPCD | Software Availability: | May-2007 |
SPEC has determined that this result was not in compliance with the SPEC
MPI2007 run and reporting rules. Specifically, the result did not meet the
requirement for baseline optimization flags to not use assertion flags (the
flag -fno-alias is a violation of this rule). The result was found to be
performance neutral compared to runs without -fno-alias. Replacement
results could not be produced because of system access limitations.
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 | NC | NC | NC | NC | NC | NC | NC | |||||||
| 107.leslie3d | NC | NC | NC | NC | NC | NC | NC | |||||||
| 113.GemsFDTD | NC | NC | NC | NC | NC | NC | NC | |||||||
| 115.fds4 | NC | NC | NC | NC | NC | NC | NC | |||||||
| 121.pop2 | NC | NC | NC | NC | NC | NC | NC | |||||||
| 122.tachyon | NC | NC | NC | NC | NC | NC | NC | |||||||
| 126.lammps | NC | NC | NC | NC | NC | NC | NC | |||||||
| 127.wrf2 | NC | NC | NC | NC | NC | NC | NC | |||||||
| 128.GAPgeofem | NC | NC | NC | NC | NC | NC | NC | |||||||
| 129.tera_tf | NC | NC | NC | NC | NC | NC | NC | |||||||
| 130.socorro | NC | NC | NC | NC | NC | NC | NC | |||||||
| 132.zeusmp2 | NC | NC | NC | NC | NC | NC | NC | |||||||
| 137.lu | NC | NC | NC | NC | NC | NC | NC | |||||||
| Hardware Summary | |
|---|---|
| Type of System: | Homogenous |
| Compute Node: | HP XC Cluster |
| Interconnect: | DDR InfiniBand |
| File Server Node: | HP ProLiant DL380 G4 file server |
| Total Compute Nodes: | 16 |
| Total Chips: | 32 |
| Total Cores: | 64 |
| Total Threads: | 64 |
| Total Memory: | 128 GB |
| Base Ranks Run: | 64 |
| Minimum Peak Ranks: | -- |
| Maximum Peak Ranks: | -- |
| Software Summary | |
|---|---|
| C Compiler: | Intel C 9.1.045 |
| C++ Compiler: | Intel C++ 9.1.045 |
| Fortran Compiler: | Intel Fortran 9.1.040 |
| Base Pointers: | 64-bit |
| Peak Pointers: | 64-bit |
| MPI Library: | HP-MPI 2.2.5 |
| Other MPI Info: | None |
| Pre-processors: | No |
| Other Software: | None |
Node Description: HP XC Cluster
| Hardware | |
|---|---|
| Number of nodes: | 16 |
| Uses of the node: | compute |
| Vendor: | Hewlett-Packard Company |
| Model: | XC3000 |
| CPU Name: | Intel Xeon 5160 |
| CPU(s) orderable: | 1-2 chips |
| Chips enabled: | 2 |
| Cores enabled: | 4 |
| Cores per chip: | 2 |
| Threads per core: | 1 |
| CPU Characteristics: | 1333 MHz FSB |
| CPU MHz: | 3000 |
| Primary Cache: | 32 KB I + 32 KB D on chip per core |
| Secondary Cache: | 4 MB I+D on chip per chip |
| L3 Cache: | None |
| Other Cache: | None |
| Memory: | 8 GB |
| Disk Subsystem: | Ext 3 |
| Other Hardware: | None |
| Adapter: | PCI-Express DDR InfiniBand HCA |
| Number of Adapters: | 1 |
| Slot Type: | Mezzanine - PCIe x8 |
| Data Rate: | Infiniband 4x DDR |
| Ports Used: | 1 |
| Interconnect Type: | InfiniBand |
| Software | |
|---|---|
| Adapter: | PCI-Express DDR InfiniBand HCA |
| Adapter Driver: | Voltiare GridStack 3.5.5_25 |
| Adapter Firmware: | Hp 1.2 |
| Operating System: | Redhat 4 Update 4 |
| Local File System: | VXFS |
| Shared File System: | NFS |
| System State: | Multi User |
| Other Software: | None |
Node Description: HP ProLiant DL380 G4 file server
| Hardware | |
|---|---|
| Number of nodes: | 1 |
| Uses of the node: | file server |
| Vendor: | Hewlett-Packard Company |
| Model: | DL380 G4 |
| CPU Name: | Intel Xeon |
| CPU(s) orderable: | 1-2 chips |
| Chips enabled: | 2 |
| Cores enabled: | 2 |
| Cores per chip: | 1 |
| Threads per core: | 1 |
| CPU Characteristics: | None |
| CPU MHz: | 3800 |
| Primary Cache: | 32 KB I + 32 KB D on chip per core |
| Secondary Cache: | 2 MB I+D on chip per chip |
| L3 Cache: | None |
| Other Cache: | None |
| Memory: | 6 GB |
| Disk Subsystem: | SCSI |
| Other Hardware: | None |
| Adapter: | PCI-Express DDR InfiniBand HCA |
| Number of Adapters: | 1 |
| Slot Type: | PCIe x4 |
| Data Rate: | Infiniband 4x DDR |
| Ports Used: | 1 |
| Interconnect Type: | InfiniBand |
| Software | |
|---|---|
| Adapter: | PCI-Express DDR InfiniBand HCA |
| Adapter Driver: | Voltiare GridStack 3.5.5_25 |
| Adapter Firmware: | Hp 1.2 |
| Operating System: | Redhat 4 Update 4 |
| Local File System: | 2 x 146GB Ultra320 SCSI 10k RPM Disks RAID 0+1 |
| Shared File System: | N/A |
| System State: | Multi User |
| Other Software: | None |
Interconnect Description: DDR InfiniBand
| Hardware | |
|---|---|
| Vendor: | Voltaire |
| Model: | ISR |
| Switch Model: | HP 4X DDR IB Switch Module for HP BladeSystem C-Class |
| Number of Switches: | 1 |
| Number of Ports: | 16 |
| Data Rate: | Infiniband 4x DDR |
| Firmware: | 1.30 |
| Switch Model: | Voltaire 24 port External switch for subnet management ISR 9024 |
| Number of Switches: | 1 |
| Number of Ports: | 24 |
| Data Rate: | Infiniband 4x DDR |
| Firmware: | 0.8.6 |
| Topology: | two level, 1 for subnet manager, and 1 for leaf. |
| Primary Use: | MPI and file system traffic |
General Notes
To run spec mpi2007 on XC3000 cluster. module load mpi . ./shrc . /opt/intel/fc/9.1.040/bin/ifortvars.sh . /opt/intel/cc/9.1.045/bin/iccvars.sh runspec -c linux_amd_intel medium --reportable --ranks=NN --define CMD_PREFIX= --define MPIRUN_OPTIONS= --flagsurl=hpmpi-intel.xml -o cfg,txt,html,check cluster launch performed via slurm's srun submit= \$MPI_ROOT/bin/mpirun -srun $command
Base Compiler Invocation
C benchmarks:
| mpicc -mpicc icc |
C++ benchmarks:
| 126.lammps: | mpiCC -mpicxx icpc |
Fortran benchmarks:
| 107.leslie3d: | mpif90 -mpif90 ifort |
| 113.GemsFDTD: | mpif90 -mpif90 ifort |
| 115.fds4: | mpif90 -mpif90 ifort |
| 129.tera_tf: | mpif90 -mpif90 ifort |
| 132.zeusmp2: | mpif90 -mpif90 ifort |
| 137.lu: | mpif90 -mpif90 ifort |
Benchmarks using both Fortran and C (except as noted below):
| mpicc -mpicc icc mpif90 -mpif90 ifort |
Base Portability Flags
| 121.pop2: | -DSPEC_MPI_CASE_FLAG |
| 127.wrf2: | -DSPEC_MPI_LINUX -DSPEC_MPI_CASE_FLAG |
Base Optimization Flags
C benchmarks:
| -O3 -no-prec-div -ftz -fno-alias -xT |
C++ benchmarks:
| 126.lammps: | -O3 -no-prec-div -ftz -fno-alias -xT |
Fortran benchmarks:
| 107.leslie3d: | -O3 -no-prec-div -ftz -fno-alias -xT |
| 113.GemsFDTD: | -O3 -no-prec-div -ftz -fno-alias -xT |
| 115.fds4: | -O3 -no-prec-div -ftz -fno-alias -xT |
| 129.tera_tf: | -O3 -no-prec-div -ftz -fno-alias -xT |
| 132.zeusmp2: | -O3 -no-prec-div -ftz -fno-alias -xT |
| 137.lu: | -O3 -no-prec-div -ftz -fno-alias -xT |
Benchmarks using both Fortran and C:
| 121.pop2: | -O3 -no-prec-div -ftz -fno-alias -xT |
| 127.wrf2: | Same as 121.pop2 |
| 128.GAPgeofem: | Same as 121.pop2 |
| 130.socorro: | Same as 121.pop2 |