# Invocation command line: # ./bin/runspec -I -c example-ibm-aix-p6-O4.cfg --define NOTES=node pump 4.7 GHz p6-IH - SMT mode off - *petaskbind.sh* 32 lcpus 64-way -a run --reportable --ranks 64 -T base medium # output_root was not used for this run ############################################################################ # # The following setting was inserted automatically as a result of # post-run basepeak application. basepeak = 1 #%define flagroot /specmpi/mpi2007-1.0 #%define flagroot /Perf/Benchmarks/spec/mpi2007 %define flagroot . flagsurl000 = http://www.spec.org/mpi2007/flags/MPI2007_flags.20080828.xml flagsurl001 = http://www.spec.org/mpi2007/flags/MPI2007_flags.0.20080828.xml flagsurl002 = http://www.spec.org/mpi2007/flags/MPI2007_flags.1.xml backup_config = 0 # No automatic backup. output_format = rsf,asc,html,csv,flags # Make sure we capture the config each time, instead. env_vars = 1 mean_anyway = 1 allow_extension_override = yes # Compilation is "extension neutral". makeflags = -j 16 ext = AIX64_p6_O4 #ENV_MP_ADAPTER_USE=shared #ENV_MP_EUILIB=us #ENV_MP_EUIDEVICE=sn_all #ENV_MP_EUILIBPATH=/u/lewars/iblibs/proto/ #ENV_OMP_DYNAMIC=FALSE #ENV_LAPI_DEBUG_SLOT_ATT_THRESH=500000 # # MPI process/processor mapping for a 32-way SMP. # # TO_DO: This needs to be changed for different systems. # # MPI Environment #ENV_MEMORY_AFFINITY =MCM #ENV_MP_SHARED_MEMORY =yes #ENV_MP_SINGLE_THREAD =yes #ENV_MP_WAIT_MODE =poll #ENV_MP_EAGER_LIMIT =65536 #ENV_MP_BUFFER_MEM =67108864 #ENV_MP_POLLING_INTERVAL =80000000 #ENV_MP_USE_BULK_XFER =yes #ENV_MP_BULK_MIN_MSG_SIZE=65536 #ENV_MP_RESD =yes #ENV_MP_STDINMODE =none #ENV_MP_PROCS = $ranks #ENV_XLFRTEOTPS =intrinthds=1 #ENV_LDR_CNTRL = DATAPSIZE=64K@TEXTPSIZE=64K@STACKPSIZE=64K #ENV_PEBND_PE_STRIDE = -1 #ENV_PEBND_PE_START = 0 BINDDIR = ../../../../.. notes_000=113.GemsFDTD (base): Applied maxprocandstop src.alt notes_005 =129.tera_tf (base): Applied fixbuffer src.alt notes_010 =127.wrf2 (base): Applied fixcalling src.alt notes_015 =all ulimits set to unlimited notes_020 ="petaskbind.sh" script used to bind each task to a unique processor notes_025 =POE Environment variables set before executing benchmarks: notes_030 = CWD =/specmpi/mpi2007-1.0 notes_035 = MP_ADAPTER_USE =shared notes_040 = MP_EUILIB =us notes_045 = MP_EUIDEVICE =sn_all notes_050 = MP_SHARED_MEMORY =yes notes_055 = MP_SINGLE_THREAD =yes notes_060 = MP_WAIT_MODE =poll notes_065 = MP_EAGER_LIMIT =65536 notes_070 = MP_BUFFER_MEM =67108864 notes_075 = MP_POLLING_INTERVAL =80000000 notes_080 = MP_USE_BULK_XFER =yes notes_085 = MP_BULK_MIN_MSG_SIZE=65536 notes_090 = MP_STDINMODE =none notes_095 = MP_LABELIO =no notes_100 = MP_HOSTFILE =$CWD/r35.64-2node notes_105 =Other Environment variables notes_110 = MEMORY_AFFINITY =MCM notes_115 = LDR_CNTRL =DATAPSIZE=64K@TEXTPSIZE=64K@STACKPSIZE=64K notes_120 = XLFRTEOTPS =intrinthds=1 notes_125 =submit command uses petaskbind.sh script to bind logical processors to ranks notes_130 = poe $CWD/petaskbind.sh $command -procs $ranks notes_135 =The Gigabit ethernet switch is shared among many nodes, not just the cluster used in this benchmark. # # Command to control the POE. # This can be overridden in the "peak" (or portability) rules. # submit = poe $BINDDIR/petaskbind.sh $command -procs $ranks # # Compiler invocations. # The settings used here are for IBM C/C++ and XL Fortran compilers. # CC = /usr/bin/mpcc_r -w -qsuppress=1500-036 CXX = /usr/bin/mpCC_r -w -qsuppress=1500-036 FC = /usr/bin/mpxlf95_r -w -qsuppress=1500-036 -qsuppress=cmpmsg # # Base Level Optimizations. # default=base=default=default: FOPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads -qalias=nostd COPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads CXXOPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads -qstrict # # 64-bit compilation flags. # default=default=default=default: #EXTRA_LDFLAGS = -q64 -qmaxmem=-1 -qspillsize=32648 EXTRA_LDFLAGS = -q64 EXTRA_CFLAGS = -q64 EXTRA_CXXFLAGS = -q64 EXTRA_FFLAGS = -q64 # # # Per-benchmark Peak-Level optimizations. # 104.milc=peak=default=default: COPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads -qipa=level=1 107.leslie3d=peak=default=default: FOPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads 113.GemsFDTD=peak=default=default: FOPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads 115.fds4=peak=default=default: FOPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads -qalias=nostd 121.pop2=peak=default=default: FOPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads -qalias=nostd COPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads 122.tachyon=peak=default=default: COPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads 126.lammps=peak=default=default: CXXOPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads -qstrict 127.wrf2=peak=default=default: FOPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads COPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads 128.GAPgeofem=peak=default=default: FOPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads COPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads 129.tera_tf=peak=default=default: FOPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads 130.socorro=peak=default=default: FOPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads COPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads 132.zeusmp2=peak=default=default: FOPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads 137.lu=peak=default=default: FOPTIMIZE = -O4 -qarch=pwr6 -qtune=pwr6 -qipa=noobject -qipa=threads # # # Portability Flags. # # Only language-level flags, data-type selection, and data-space sizing are # allowed here. No optimizations are allowed. # # If there aren't any flags for the benchmark, we leave an empty entry as a # placeholder. # 104.milc=default=default=default: 107.leslie3d=default=default=default: FPORTABILITY = -qfixed 113.GemsFDTD=default=default=default: srcalt=maxprocandstop 115.fds4=default=default=default: FPORTABILITY = -qfixed CPORTABILITY = -DSPEC_MPI_LC_NO_TRAILING_UNDERSCORE 121.pop2=default=default=default: CPORTABILITY = -DSPEC_MPI_AIX 122.tachyon=default=default=default: 126.lammps=default=default=default: 127.wrf2=default=default=default: CPORTABILITY = -DNOUNDERSCORE -DSPEC_MPI_AIX srcalt=fixcalling 128.GAPgeofem=default=default=default: 129.tera_tf=default=default=default: srcalt=fixbuffer 130.socorro=default=default=default: CPORTABILITY = -DSPEC_NO_UNDERSCORE -qcpluscmt 132.zeusmp2=default=default=default: FPPPORTABILITY = -DSPEC_SINGLE_UNDERSCORE FPORTABILITY = -qfixed 137.lu=default=default=default: FPORTABILITY = -qfixed # 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 = May-2008 license_num = 0005 sw_auto_parallel = No sw_avail = May-2008 sw_base_ptrsize = 64-bit sw_other = None sw_peak_ptrsize = 64-bit system_name = IBM Power 575 node_POWER6IHb_sw_state = Multi-user node_POWER6IHb_sw_sharedfile = NFS over ethernet node_POWER6IHb_sw_localfile = AIX/JFS2 node_POWER6IHb_purpose = compute node_POWER6IHb_order = 1 node_POWER6IHb_label = IBM Power 575 node_POWER6IHb_hw_vendor = IBM Corporation node_POWER6IHb_hw_tcache = 32 MB I+D off chip per chip node_POWER6IHb_hw_scache = 4 MB I+D on chip per core node_POWER6IHb_hw_pcache = 64 KB I + 64 KB D on chip per core node_POWER6IHb_hw_other = None node_POWER6IHb_hw_ocache = None node_POWER6IHb_hw_nthreadspercore = 1 node_POWER6IHb_hw_ncpuorder = 32 cores node_POWER6IHb_hw_ncoresperchip = 2 node_POWER6IHb_hw_ncores = 32 node_POWER6IHb_hw_nchips = 16 node_POWER6IHb_hw_model = IBM Power 575 node_POWER6IHb_hw_memory = 128 GB (64x2 GB) DDR2 533 MHz node_POWER6IHb_hw_fpu = Integrated node_POWER6IHb_hw_disk = 1x146 GB SFF SAS, 10K RPM node_POWER6IHb_hw_cpu_name = POWER6 node_POWER6IHb_hw_cpu_mhz = 4700 node_POWER6IHb_hw_adapter_InfiniBand_slot_type = GX++ node_POWER6IHb_hw_adapter_InfiniBand_slot_ports_used = 4 node_POWER6IHb_hw_adapter_InfiniBand_ports_used = 4 node_POWER6IHb_hw_adapter_InfiniBand_model = IBM Dual 2-port 4x DDR Host Channel Adapter node_POWER6IHb_hw_adapter_InfiniBand_interconnect = DDR InfiniBand node_POWER6IHb_hw_adapter_InfiniBand_driver = fileset devices.common.IBM.ib.rte 5.3.8.2 node_POWER6IHb_hw_adapter_InfiniBand_data_rate = 4x DDR 20 Gbps node_POWER6IHb_hw_adapter_InfiniBand_count = 2 node_POWER6IHb_hw_adapter_Ethernet_slot_ports_used = integrated node_POWER6IHb_hw_adapter_Ethernet_ports_used = 1 node_POWER6IHb_hw_adapter_Ethernet_model = Integrated node_POWER6IHb_hw_adapter_Ethernet_interconnect = Gigabit Ethernet node_POWER6IHb_hw_adapter_Ethernet_driver = fileset devices.chrp.IBM.lhea.rte 5.3.8.2 node_POWER6IHb_hw_adapter_Ethernet_data_rate = 1 Gbps node_POWER6IHb_hw_adapter_Ethernet_count = 1 node_POWER6IHb_count = 1 system_vendor = IBM Corporation test_date = Jun-2008 test_sponsor = IBM Corporation tester = IBM Corporation system_class = homogeneous interconnect_Ethernet_hw_model000 = Cisco Systems WS-C6509-E interconnect_Ethernet_hw_model001 = Catalyst 6500 9-slot Chassis System interconnect_Ethernet_hw_switch_Cisco_count = 1 interconnect_Ethernet_hw_switch_Cisco_data_rate = 1 Gbps interconnect_Ethernet_hw_switch_Cisco_firmware = 01ES330_034_034 interconnect_Ethernet_hw_switch_Cisco_model000 = Cisco Systems WS-X6748-GE-TX interconnect_Ethernet_hw_switch_Cisco_model001 = CEF720 48 port 10/100/1000mb Ethernet card interconnect_Ethernet_hw_switch_Cisco_model002 = Cisco Systems WS-SUP720-3B interconnect_Ethernet_hw_switch_Cisco_model003 = 2 ports Supervisor Engine 720 Rev. 5.2 interconnect_Ethernet_hw_switch_Cisco_ports = 48 interconnect_Ethernet_hw_vendor = IBM Corporation interconnect_Ethernet_label = Gigabit Ethernet interconnect_Ethernet_order = 1 interconnect_Ethernet_purpose = File system interconnect_InfiniBand_hw_switch_QLogic_count = 2 interconnect_InfiniBand_hw_switch_QLogic_data_rate = InfiniBand 4x DDR 20 Gbps interconnect_InfiniBand_hw_switch_QLogic_firmware = 4.2.1.1.1 interconnect_InfiniBand_hw_switch_QLogic_interconnect = InfiniBand interconnect_InfiniBand_hw_switch_QLogic_model = QLogic SilverStorm 9024 interconnect_InfiniBand_hw_switch_QLogic_ports = 24 interconnect_InfiniBand_hw_switch_QLogic_ports_used = 16 interconnect_InfiniBand_hw_topo = linear interconnect_InfiniBand_hw_vendor = QLogic interconnect_InfiniBand_label = InfiniBand interconnect_InfiniBand_order = 0 interconnect_InfiniBand_purpose = MPI Communication node_POWER6IH_count = 1 node_POWER6IH_hw_adapter_Ethernet_count = 1 node_POWER6IH_hw_adapter_Ethernet_data_rate = 1 Gbps node_POWER6IH_hw_adapter_Ethernet_driver = fileset devices.chrp.IBM.lhea.rte 5.3.8.2 node_POWER6IH_hw_adapter_Ethernet_interconnect = Gigabit Ethernet node_POWER6IH_hw_adapter_Ethernet_model = Integrated node_POWER6IH_hw_adapter_Ethernet_ports_used = 1 node_POWER6IH_hw_adapter_Ethernet_slot_ports_used = integrated node_POWER6IH_hw_adapter_InfiniBand_count = 2 node_POWER6IH_hw_adapter_InfiniBand_data_rate = 4x DDR 20 Gbps node_POWER6IH_hw_adapter_InfiniBand_driver = fileset devices.common.IBM.ib.rte 5.3.8.2 node_POWER6IH_hw_adapter_InfiniBand_interconnect = DDR InfiniBand node_POWER6IH_hw_adapter_InfiniBand_model = IBM Dual 2-port 4x DDR Host Channel Adapter node_POWER6IH_hw_adapter_InfiniBand_ports_used = 4 node_POWER6IH_hw_adapter_InfiniBand_slot_ports_used = 4 node_POWER6IH_hw_adapter_InfiniBand_slot_type = GX++ node_POWER6IH_hw_cpu_mhz = 4700 node_POWER6IH_hw_cpu_name = POWER6 node_POWER6IH_hw_disk = 1x146 GB SFF SAS, 10K RPM node_POWER6IH_hw_fpu = Integrated node_POWER6IH_hw_memory = 128 GB (64x2 GB) DDR2 533 MHz node_POWER6IH_hw_model = IBM Power 575 node_POWER6IH_hw_nchips = 16 node_POWER6IH_hw_ncores = 32 node_POWER6IH_hw_ncoresperchip = 2 node_POWER6IH_hw_ncpuorder = 32 cores node_POWER6IH_hw_nthreadspercore = 1 node_POWER6IH_hw_ocache = None node_POWER6IH_hw_other = None node_POWER6IH_hw_pcache = 64 KB I + 64 KB D on chip per core node_POWER6IH_hw_scache = 4 MB I+D on chip per core node_POWER6IH_hw_tcache = 32 MB I+D off chip per chip node_POWER6IH_hw_vendor = IBM Corporation node_POWER6IH_label = IBM Power 575 node_POWER6IH_order = 0 node_POWER6IH_purpose = compute, head, fileserver node_POWER6IH_sw_localfile = AIX/JFS2 node_POWER6IH_sw_os000 = IBM AIX V5.3 node_POWER6IH_sw_os001 = with the 5300-08-02 Technology Level node_POWER6IH_sw_other000 = APAR IZ26983 node_POWER6IH_sw_other001 = software update for InfiniBand adapter drivers node_POWER6IH_sw_other004 = IBM LoadLeveler for AIX node_POWER6IH_sw_other005 = V3.4.3.2 node_POWER6IH_sw_sharedfile = NFS over ethernet node_POWER6IH_sw_state = Multi-user sw_c_compiler000 = IBM XL C/C++ Enterprise Edition V9.0 sw_c_compiler001 = Updated with the Oct2007 PTF sw_cxx_compiler000 = IBM XL C/C++ Enterprise Edition V9.0 sw_cxx_compiler001 = Updated with the Oct2007 PTF sw_f_compiler000 = IBM XL Fortran Enterprise Edition V11.1 sw_f_compiler001 = Updated with the Oct2007 PTF sw_mpi_library000 = IBM Parallel Environment for AIX sw_mpi_library001 = V4.3.2.2 node_POWER6IHb_sw_os000 = IBM AIX V5.3 node_POWER6IHb_sw_os001 = with the 5300-08-02 Technology Level node_POWER6IHb_sw_other000 = APAR IZ26983 node_POWER6IHb_sw_other001 = software update for InfiniBand adapter drivers node_POWER6IHb_sw_other004 = IBM LoadLeveler for AIX node_POWER6IHb_sw_other005 = V3.4.3.2