# Invocation command line: # /store/hfeng/mpi2007/v2-2/bin/runspec --ranks 3072 -I --reportable -a validate -n 3 -o asc -i lref -T base,peak -c sgi-linux-x86_64-intel-sgimpi-campeak-12.cfg large # output_root was not used for this run ############################################################################ # Invocation command line: # /nas/store/hfeng/mpi2007/v1.1/bin/runspec --ranks 512 --reportable -a validate -n 2 -o asc -i mref -T base -c sgi-linux-x86_64-intel-sgimpi.cfg medium # output_root was not used for this run ############################################################################ #################################################################### # # defaults # #################################################################### #flagsurl = http://www.spec.org/mpi2007/flags/EM64T_Intel101_flags.20080618.xml flagsurl000= http://www.spec.org/mpi2007/flags/SGI_x86_64_Intel12_flags.xml ext=sgimpi.intel.12.0.3 action=validate tune=base input=ref teeout=no env_vars=1 no_input_handler=null mean_anyway=1 FC = ifort CC = icc CXX = icpc #################################################################### # # Base flags # #################################################################### default=default=default=default: FOPTIMIZE = -O3 -xSSE4.2 -no-prec-div COPTIMIZE = -O3 -xSSE4.2 -no-prec-div CXXOPTIMIZE = -O3 -xSSE4.2 -no-prec-div -ansi-alias EXTRA_LIBS = -lmpi default=default=default=default: use_submit_for_speed=1 #submit=mpirun -np $ranks MPInside $command submit=mpiexec_mpt -np $ranks $command #submit=mpiexec_mpt -n $ranks dplace -s1 -c2-5,8-11 $command #################################################################### # # Portability flags # #################################################################### 121.pop2=default=default=default: CPORTABILITY = -DSPEC_MPI_CASE_FLAG 127.wrf2=default=default=default: CPORTABILITY = -DSPEC_MPI_CASE_FLAG -DSPEC_MPI_LINUX #default=peak=default=default: #basepeak=yes 121.pop2=peak=default=default: ranks=2048 ENV_MPI_DSM_CPULIST=2-5,8-11:allhosts submit=mpirun `uniq \$PBS_NODEFILE | awk '{printf ",%s", \$0}' | sed 's/^,/ /'` 8 $command 122.tachyon=peak=default=default: basepeak=yes 125.RAxML=peak=default=default: basepeak=yes 126.lammps=peak=default=default: basepeak=yes 128.GAPgeofem=peak=default=default: basepeak=yes #ranks=2048 #ENV_MPI_DSM_CPULIST=2-5,8-11:allhosts #submit=mpirun `uniq \$PBS_NODEFILE | awk '{printf ",%s", \$0}' | sed 's/^,/ /'` 8 $command 129.tera_tf=peak=default=default: basepeak=yes 132.zeusmp2=peak=default=default: ranks=2048 ENV_MPI_DSM_CPULIST=2-5,8-11:allhosts submit=mpirun `uniq \$PBS_NODEFILE | awk '{printf ",%s", \$0}' | sed 's/^,/ /'` 8 $command 137.lu=peak=default=default: ranks=2048 ENV_MPI_DSM_CPULIST=2-5,8-11:allhosts submit=mpirun `uniq \$PBS_NODEFILE | awk '{printf ",%s", \$0}' | sed 's/^,/ /'` 8 $command 142.dmilc=peak=default=default: basepeak=yes 143.dleslie=peak=default=default: ranks=2048 ENV_MPI_DSM_CPULIST=2-5,8-11:allhosts submit=mpirun `uniq \$PBS_NODEFILE | awk '{printf ",%s", \$0}' | sed 's/^,/ /'` 8 $command 145.lGemsFDTD=peak=default=default: ranks=2048 ENV_MPI_DSM_CPULIST=2-5,8-11:allhosts submit=mpirun `uniq \$PBS_NODEFILE | awk '{printf ",%s", \$0}' | sed 's/^,/ /'` 8 $command 147.l2wrf2=peak=default=default: basepeak=yes #ranks=2048 #ENV_MPI_DSM_CPULIST=2-5,8-11:allhosts #submit=mpirun `uniq \$PBS_NODEFILE | awk '{printf ",%s", \$0}' | sed 's/^,/ /'` 8 $command # The following section was added automatically, and contains settings that # did not appear in the original configuration file, but were added to the # raw file after the run. default: hw_avail = Feb-2011 license_num = 4 prepared_by = Huiyu Feng sw_avail = Aug-2011 sw_base_ptrsize = 64-bit sw_other = None sw_peak_ptrsize = 64-bit system_vendor = SGI interconnect_IOMPI_purpose = MPI and I/O traffic interconnect_IOMPI_order = 2 interconnect_IOMPI_label = InfiniBand (MPI and I/O) interconnect_IOMPI_hw_vendor = Mellanox Technologies and SGI interconnect_IOMPI_hw_switch_1_ports = 36 interconnect_IOMPI_hw_switch_1_firmware = 5040005 interconnect_IOMPI_hw_switch_1_data_rate = InfiniBand 4x QDR interconnect_IOMPI_hw_switch_1_count = 64 interconnect_IOMPI_hw_model = None test_sponsor = SGI tester = SGI system_class = Homogeneous sw_preprocessors = None sw_mpi_other = OFED 1.4.2 sw_mpi_library = SGI MPT 2.04 Patch 10789 node_compute_count = 256 node_compute_hw_adapter_IB_count = 2 node_compute_hw_adapter_IB_data_rate = InfiniBand 4x QDR node_compute_hw_adapter_IB_driver = OFED-1.4.2 node_compute_hw_adapter_IB_firmware = 2.7.8200 node_compute_hw_adapter_IB_interconnect = InfiniBand node_compute_hw_adapter_IB_model000 = Mellanox MT26428 ConnectX IB QDR node_compute_hw_adapter_IB_model001 = (PCIe x8 Gen2 5 GT/s) node_compute_hw_adapter_IB_ports_used = 1 node_compute_hw_adapter_IB_slot_type = PCIe x8 Gen2 node_compute_hw_cpu_char000= Six Core, 3.46 GHz, 6.4 GT/s QPI node_compute_hw_cpu_char001 = Intel Turbo Boost Technology up to 3.73 GHz node_compute_hw_cpu_char002 = Hyper-Threading Technology enabled node_compute_hw_cpu_mhz = 3467 node_compute_hw_cpu_name = Intel Xeon X5690 node_compute_hw_disk = None node_compute_hw_memory = 24 GB (6 x 4 GB 2Rx4 PC3-10600R-9, ECC) node_compute_hw_model = SGI Altix ICE 8400EX (Intel Xeon X5690, 3.46 GHz) node_compute_hw_nchips = 2 node_compute_hw_ncores = 12 node_compute_hw_ncoresperchip = 6 node_compute_hw_ncpuorder = 1-2 chips node_compute_hw_nthreadspercore = 2 node_compute_hw_ocache = None node_compute_hw_other = None node_compute_hw_pcache = 32 KB I + 32 KB D on chip per core node_compute_hw_scache = 256 KB I+D on chip per core node_compute_hw_tcache = 12 MB I+D on chip per chip node_compute_hw_vendor = SGI node_compute_label = SGI Altix ICE 8400EX Compute Node node_compute_order = 2 node_compute_purpose = compute node_compute_sw_localfile = NFSv3 node_compute_sw_os000 = SUSE Linux Enterprise Server 11 SP1, node_compute_sw_os001 = Kernel 2.6.32.13-0.4-default node_compute_sw_other000= SGI ProPack 7SP1 for Linux, node_compute_sw_other001 = Build 701r3.sles11-1005252113 node_compute_sw_other002 = SGI Tempo Compute Node 2.1, node_compute_sw_other003 = Build 701r3.sles11-1005252113 node_compute_sw_sharedfile = NFSv3 IPoIB node_compute_sw_state = Multi-user, run level 3 node_fileserver_count = 1 node_fileserver_hw_adapter_IB_count = 2 node_fileserver_hw_adapter_IB_data_rate = InfiniBand 4x QDR node_fileserver_hw_adapter_IB_driver = OFED-1.4.0 node_fileserver_hw_adapter_IB_firmware = 2.7.0 node_fileserver_hw_adapter_IB_interconnect = InfiniBand node_fileserver_hw_adapter_IB_model000 = Mellanox MT26428 ConnectX IB QDR node_fileserver_hw_adapter_IB_ports_used = 2 node_fileserver_hw_adapter_IB_slot_type = PCIe x8 Gen2 node_fileserver_hw_cpu_char000= Intel Turbo Boost Technology up to 3.33 GHz node_fileserver_hw_cpu_char001 = Hyper-Threading Technology enabled node_fileserver_hw_cpu_mhz = 2933 node_fileserver_hw_cpu_name = Intel Xeon X5670 node_fileserver_hw_disk000 = 8.8 TB RAID 5 node_fileserver_hw_disk001 = 60 x 146 GB SAS (Seagate Cheetah 15K.5) node_fileserver_hw_memory000= 96 GB (12*8 GB DDR3-1333 CL9 DIMMs) node_fileserver_hw_model = SGI Altix XE 270 (Intel Xeon X5670, 2.93 GHz) node_fileserver_hw_nchips = 2 node_fileserver_hw_ncores = 12 node_fileserver_hw_ncoresperchip = 6 node_fileserver_hw_ncpuorder = 1-2 chips node_fileserver_hw_nthreadspercore = 2 node_fileserver_hw_ocache = None node_fileserver_hw_other = None node_fileserver_hw_pcache = 32 KB I + 32 KB D on chip per core node_fileserver_hw_scache = 256 KB I+D on chip per chip node_fileserver_hw_tcache = 12 MB I+D on chip per chip node_fileserver_hw_vendor = SGI node_fileserver_label = SGI InfiniteStorage Nexis 2000 NAS node_fileserver_order = 3 node_fileserver_purpose = fileserver node_fileserver_sw_localfile = xfs node_fileserver_sw_os000 = SUSE Linux Enterprise Server 11 (x86_64) node_fileserver_sw_os001 = Kernel 2.6.27.19-5-default node_fileserver_sw_other000= SGI Foundation Software 2, Build node_fileserver_sw_other001 = 700r3.sles11-1004061553 node_fileserver_sw_state = Multi-user, run level 3 sw_c_compiler000 = Intel C++ Composer XE 2011 for Linux, sw_c_compiler001 = Version 12.0.3.174 Build 20110309 sw_cxx_compiler000 = Intel C++ Composer XE 2011 for Linux, sw_cxx_compiler001 = Version 12.0.3.174 Build 20110309 sw_f_compiler000 = Intel Fortran Composer XE 2011 for Linux, sw_f_compiler001 = Version 12.0.3.174 Build 20110309 system_name000 = SGI Altix ICE 8400EX system_name001 = (Intel Xeon X5690, 3.46 GHz) notes_000 = Software environment: notes_005 = export MPI_REQUEST_MAX=65536 notes_010 = export MPI_TYPE_MAX=32768 notes_015 = export MPI_BUFS_THRESHOLD=1 notes_020 = export MPI_IB_RAILS=2 notes_025 = ulimit -s unlimited notes_030 = notes_035 = BIOS settings: notes_040 = AMI BIOS version 080016 notes_045 = Hyper-Threading Technology enabled (default) notes_050 = Intel Turbo Boost Technology enabled (default) notes_055 = Intel Turbo Boost Technology activated in the OS via notes_060 = /etc/init.d/acpid start notes_065 = /etc/init.d/powersaved start notes_070 = powersave -f notes_075 = notes_080 = Job Placement: notes_085 = In the base run, each MPI job was assigned to a topologically compact notes_090 = set of nodes, i.e. the minimal needed number of switches was notes_095 = used for each job: 2 switches for 96 ranks, notes_100 = 4 switches for 192 ranks, 8 switches for 384 ranks, notes_105 = 16 switches for 768 ranks, 32 switches for 1536 ranks, notes_110 = and 64 switches for 3072 ranks. notes_115 = notes_120 = Additional notes regarding interconnect: notes_125 = The Infiniband network consists of two independent planes, notes_130 = with half the switches in the system allocated to each plane. notes_135 = I/O traffic is restricted to one plane, while MPI traffic can notes_140 = use both planes. # The following section was added automatically, and contains settings that # did not appear in the original configuration file, but were added to the # raw file after the run. default: interconnect_IOMPI_hw_switch_1_model000 = SGI QDR_1.5_HYPR_2454 with Mellanox Device 48438 interconnect_IOMPI_hw_switch_1_model001 = (Infiniscale IV) interconnect_IOMPI_hw_topo000 = Enhanced Hypercube node_fileserver_hw_adapter_IB_model003 = (PCIe x8 Gen2 5 GT/s) notes_submit_000 =For peak benchmarks that used 2048 MPI ranks, four ranks notes_submit_005 =were assigned to each CPU chip, leaving 2 cores per chip idle.