# Invocation command line: # /store/mackey/specMPI/bin/runspec --ranks 320 -I --reportable -a validate -o asc,csv,html -n 3 -T base -c default.cfg --define ICC=14 --define PPN=40 -i mref --define SIZE=medium medium # output_root was not used for this run ############################################################################ # Invocation command line: # /store/hfeng/mpi2007-2014-07-09/bin/runspec --ranks 1536 -I --reportable -a validate -n 3 -o asc -i lref -T base -c sgi-linux-x86_64-sgimpi-14.0.3.174-avx2.cfg large # output_root was not used for this run ############################################################################ #################################################################### # # defaults # #################################################################### flagsurl000= http://www.spec.org/mpi2007/flags/SGI_x86_64_Intel14_flags.20140908.xml %if !defined(%{ICC}) || !defined(%{SIZE}) % error must define both SIZE and ICC %endif %if %{ICC} == 14 %if '%{SIZE}' eq 'medium' ext=sgimpi.intel.14.0.3.174.avx2.med %elif '%{SIZE}' eq 'large' ext=sgimpi.intel.14.0.3.174.avx2.large %else %error invalid SIZE '%{SIZE}' %endif %elif %{ICC} == 16 && '%{SIZE}' eq 'medium' ext=sgimpi.intel.16.0.1.150.msse2 %elif %{ICC} == 16 && '%{SIZE}' eq 'large' ext=sgimpi.intel.16.0.1.150.avx2 %else %error invalid combination of SIZE and ICC %endif action=validate tune=base input=ref teeout=no env_vars=1 no_input_handler=null mean_anyway=1 strict_rundir_verify = 1 FC = ifort CC = icc CXX = icpc #################################################################### # # Base flags # #################################################################### default=default=default=default: FOPTIMIZE = -O3 -xCORE-AVX2 -no-prec-div %if %{ICC} == 16 && '%{SIZE}' eq 'medium' FOPTIMIZE = -O3 -msse2 -no-prec-div %endif COPTIMIZE = -O3 -xCORE-AVX2 -no-prec-div CXXOPTIMIZE = -O3 -xCORE-AVX2 -no-prec-div -ansi-alias EXTRA_LIBS = -lmpi default=default=default=default: use_submit_for_speed=1 %ifndef %{PPN} submit=mpiexec_mpt -n $ranks $command %else submit=mpiexec_mpt -ppn %{PPN} -n $ranks $command %endif #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 130.socorro=default=default=default: srcalt=nullify_ptrs FPORTABILITY=-assume nostd_intent_in 104.milc=peak=default=default: basepeak=yes 107.leslie3d=peak=default=default: basepeak=yes 113.GemsFDTD=peak=default=default: ranks=96 115.fds4=peak=default=default: basepeak=yes #ranks=696 121.pop2=peak=default=default: basepeak=yes 122.tachyon=peak=default=default: basepeak=yes 126.lammps=peak=default=default: basepeak=yes #ranks=192 127.wrf2=peak=default=default: #ranks=744 basepeak=yes 128.GAPgeofem=peak=default=default: basepeak=yes 129.tera_tf=peak=default=default: basepeak=yes 129.tera_tf=default=default=default: %if '%{SIZE}' eq 'medium' srcalt=add_rank_support %endif 130.socorro=peak=default=default: basepeak=yes #RM_SOURCES=specblas.f90 #EXTRA_LIBS=-lacml -lmpi #EXTRA_LIBS=-Wl,--start-group -lmkl_intel_lp64 -lmkl_sequential -lmkl_core -Wl,--end-group -lmpi 132.zeusmp2=peak=default=default: basepeak=yes #ranks=528 137.lu=peak=default=default: basepeak=yes #ranks=504 143.dleslie=default=default=default: srcalt=integer_overflow notes_base_143=src.alt used: 143.dleslie->integer_overflow 145.lGemsFDTD=peak=default=default: #ranks=2048 #ENV_MPI_DSM_CPULIST=2-5,8-11:allhosts 147.l2wrf2=peak=default=default: #ranks=2048 #ENV_MPI_DSM_CPULIST=2-5,8-11:allhosts # 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: sw_base_ptrsize = 64-bit sw_peak_ptrsize = Not Applicable sw_other = None sw_mpi_library = SGI MPT 2.14 sw_mpi_other = MLNX_OFED_LINUX-3.1-1.0.3 sw_preprocessors = None test_sponsor = SGI license_num = 14 tester = SGI hw_avail = Mar-2016 sw_avail = May-2016 prepared_by = Mark Mackey system_vendor = SGI node_fileserver_sw_state = Multi-user, run level 5 node_fileserver_sw_localfile = xfs node_fileserver_purpose = fileserver node_fileserver_order = 3 node_fileserver_label = SGI MIS Server node_fileserver_hw_vendor = SGI node_fileserver_hw_tcache = 20 MB I+D on chip per chip node_fileserver_hw_scache = 256 KB I+D on chip per core node_fileserver_hw_pcache = 32 KB I + 32 KB D on chip per core node_fileserver_hw_other = None node_fileserver_hw_ocache = None node_fileserver_hw_nthreadspercore = 2 node_fileserver_hw_ncpuorder = 1-2 chips node_fileserver_hw_ncoresperchip = 8 node_fileserver_hw_ncores = 16 node_fileserver_hw_nchips = 2 node_fileserver_hw_memory = 128 GB (8 * 16 GB 2Rx4 PC3-10600R-9, ECC) node_fileserver_hw_cpu_name = Intel Xeon E5-2670 node_fileserver_hw_cpu_mhz = 2601 node_fileserver_hw_adapter_IB_slot_type = PCIe x8 Gen3 node_fileserver_hw_adapter_IB_ports_used = 2 node_fileserver_hw_adapter_IB_interconnect = InfiniBand node_fileserver_hw_adapter_IB_firmware = 2.35.5100 node_fileserver_hw_adapter_IB_driver = MLNX_OFED_LINUX-3.1-1.0.3 node_fileserver_hw_adapter_IB_data_rate = InfiniBand 4x FDR node_fileserver_hw_adapter_IB_count = 2 node_fileserver_count = 1 node_compute_sw_state = Multi-user, run level 3 node_compute_sw_sharedfile = NFSv3 IPoIB node_compute_sw_localfile = ext3 node_compute_purpose = compute node_compute_order = 2 node_compute_label = SGI Rackable C2112-4GP3 Compute Node node_compute_hw_vendor = SGI node_compute_hw_tcache = 55 MB I+D on chip per chip node_compute_hw_scache = 256 KB I+D on chip per core node_compute_hw_pcache = 32 KB I + 32 KB D on chip per core node_compute_hw_other = None node_compute_hw_ocache = None node_compute_hw_nthreadspercore = 1 node_compute_hw_ncpuorder = 1-2 chips node_compute_hw_ncoresperchip = 22 node_compute_hw_ncores = 44 node_compute_hw_nchips = 2 node_compute_hw_memory = 128 GB (8 x 16 GB 2Rx4 PC4-2400T-R) node_compute_hw_disk = None node_compute_hw_cpu_name = Intel Xeon E5-2699 v4 node_compute_hw_cpu_mhz = 2220 node_compute_hw_adapter_2_slot_type = PCIe x8 Gen3 node_compute_hw_adapter_2_ports_used = 1 node_compute_hw_adapter_2_interconnect = InfiniBand node_compute_hw_adapter_2_firmware = 2.35.5100 node_compute_hw_adapter_2_driver = OFED-3.1.1-0.0 node_compute_hw_adapter_2_data_rate = InfiniBand 4x FDR node_compute_hw_adapter_2_count = 1 node_compute_hw_adapter_1_slot_type = PCIe x16 Gen3 node_compute_hw_adapter_1_ports_used = 1 node_compute_hw_adapter_1_interconnect = InfiniBand node_compute_hw_adapter_1_firmware = 12.12.1240 node_compute_hw_adapter_1_driver = OFED-3.1.1-0.3 node_compute_hw_adapter_1_data_rate = InfiniBand 4x EDR node_compute_hw_adapter_1_count = 1 node_compute_count = 8 interconnect_MPI_purpose = MPI traffic interconnect_MPI_order = 1 interconnect_MPI_label = InfiniBand MPI interconnect_MPI_hw_vendor = Mellanox Technologies interconnect_MPI_hw_topo = Fat Tree interconnect_MPI_hw_switch_1_ports = 36 interconnect_MPI_hw_switch_1_model = Mellanox SB7790 interconnect_MPI_hw_switch_1_firmware = 11.1.102 interconnect_MPI_hw_switch_1_data_rate = InfiniBand 4x EDR interconnect_MPI_hw_switch_1_count = 6 interconnect_MPI_hw_model = None interconnect_IO_purpose = I/O traffic interconnect_IO_order = 2 interconnect_IO_label = InfiniBand I/O interconnect_IO_hw_vendor = Mellanox Technologies interconnect_IO_hw_topo = Fat Tree interconnect_IO_hw_switch_2_ports = 36 interconnect_IO_hw_switch_2_model = Mellanox MSX6025 interconnect_IO_hw_switch_2_firmware = 9.3.6000 interconnect_IO_hw_switch_2_data_rate = InfiniBand 4x FDR interconnect_IO_hw_switch_2_count = 4 interconnect_IO_hw_switch_1_ports = 36 interconnect_IO_hw_switch_1_model = Mellanox MSX6036F-1SFS interconnect_IO_hw_switch_1_firmware = 9.3.5080 interconnect_IO_hw_switch_1_data_rate = InfiniBand 4x FDR interconnect_IO_hw_switch_1_count = 2 interconnect_IO_hw_model = None notes_000 =130.socorro (base): "nullify_ptrs" src.alt was used. notes_005 = notes_010 =129.tera_tf (base): "add_rank_support" src.alt was used. notes_015 = notes_020 =Software environment: notes_025 = export MPI_REQUEST_MAX=65536 notes_030 = export MPI_TYPE_MAX=32768 notes_035 = export MPI_IB_DEVS=1 notes_040 = export MPI_CONNECTIONS_THRESHOLD=0 notes_045 = export MPI_IB_UPGRADE_SENDS=50 notes_050 = export MPI_IB_IMM_UPGRADE=false notes_055 = export MPI_IB_HYPER_LAZY=false notes_060 = ulimit -s unlimited notes_065 = notes_070 =BIOS settings: notes_075 = AMI BIOS version T20151001184140 notes_080 = Hyper-Threading Technology disabled notes_085 = Transparent HugePages enabled notes_090 = Intel Turbo Boost Technology enabled (default) notes_095 = Intel Turbo Boost Technology activated with notes_100 = modprobe acpi_cpufreq notes_105 = cpupower frequency-set -u 2601MHz -d 2601MHz -g performance notes_110 = notes_115 =Job Placement: notes_120 = Twenty ranks were assigned to each CPU chip, leaving 2 notes_125 = cores per chip idle. Each MPI job was assigned to a notes_130 = topologically compact set of nodes, i.e. the minimal notes_135 = number of leaf switches was used for each job: 1 switch notes_140 = for up to 32 sockets, and 2 switches for up to 64 sockets. notes_145 = notes_150 =Additional notes regarding interconnect: notes_155 = The Infiniband network consists of two independent planes, notes_160 = with half the switches in the system allocated to each plane. notes_165 = I/O traffic is restricted to one plane, while MPI traffic is notes_170 = restricted to the other plane. node_compute_hw_adapter_1_model000 = Mellanox MT27620 with ConnectX-4 node_compute_hw_adapter_1_model001 = (PCIe x16 Gen3 8 GT/s) node_compute_hw_adapter_2_model000 = Mellanox MT27500 with ConnectX-3 node_compute_hw_adapter_2_model001 = (PCIe x8 Gen3 8 GT/s) node_compute_hw_cpu_char000 = 22 Core, 2.20 GHz, 9.6 GT/s QPI node_compute_hw_cpu_char001 = Intel Turbo Boost Technology up to 3.60 GHz node_compute_hw_cpu_char002 = Hyper-Threading Technology disabled node_compute_hw_model000 = SGI Rackable C2112-4GP3 (Intel Xeon E5-2699 v4, node_compute_hw_model001 = 2.20 GHz) node_compute_sw_os000 = SUSE Linux Enterprise Server 12 (x86_64), node_compute_sw_os001 = Kernel 3.12.44-52.10-default node_compute_sw_other000 = SGI Tempo Service Node 3.2.0, node_compute_sw_other001 = Build 713r26.sles12-1510192000 node_fileserver_hw_adapter_IB_model000 = Mellanox MT27500 with ConnectX-3 ASIC node_fileserver_hw_cpu_char000 = Intel Turbo Boost Technology up to 3.30 GHz node_fileserver_hw_cpu_char001 = Hyper-Threading Technology enabled node_fileserver_hw_disk000 = 45 TB RAID 6 node_fileserver_hw_disk001 = 12 x 1 TB SATA (Seagate Constellation, 7200RPM) node_fileserver_hw_model000 = SGI MIS Server (Intel Xeon X2670, 2.60 GHz) node_fileserver_sw_os000 = SUSE Linux Enterprise Server 11 SP3 (x86_64), node_fileserver_sw_os001 = Kernel 3.0.101-0.46-default node_fileserver_sw_other000 = SGI Foundation Software 2.10 node_fileserver_sw_other001 = Build 710r16.sles11sp3-1404092103 sw_c_compiler000 = Intel C++ Composer XE 2013 for Linux, sw_c_compiler001 = Version 14.0.3.174 Build 20140422 sw_cxx_compiler000 = Intel C++ Composer XE 2013 for Linux, sw_cxx_compiler001 = Version 14.0.3.174 Build 20140422 sw_f_compiler000 = Intel Fortran Composer XE 2013 for Linux, sw_f_compiler001 = Version 14.0.3.174 Build 20140422 system_class = Homogeneous system_name000 = SGI Rackable C2112-4GP3 system_name001 = (Intel Xeon E5-2699 v4, 2.20 GHz)