# Invocation command line: # /export/bmk2/keeper/cpu2006.1.1/bin/runspec --config=jun03.dc1.cfg --reportable -r 127 --action=validate fp # output_root used was "/export/bmk/keeper/cpu2006.rep" ############################################################################ ############################################################################ ext = jun03 http_proxy = flagsurl000 = http://www.spec.org/cpu2006/flags/Sun-Solaris-Studio12-and-gccfss4.2.20090713.xml action = validate tune = all iterations = 1 # in a real run, would be 3 or more expand_notes = 1 ignore_errors = 1 log_line_width = 4090 # to avoid choking vi output_format = asc,cfgfile mean_anyway = 1 #minimize_rundirs = 1 #minimize_builddirs = 1 ############### Disk usage ############## # The output_root feature # www.spec.org/cpu2006/docs/runspec.html#output_root # nicely segregates your output directories from the original tree; # but if you don't like that feature, just comment out this whole # section. # # Recommended usage is: # # cd $SPEC # . ./shrc # cd /myoutputdirs # GO=`pwd` # runspec --define OUTDISK=$GO # # %if '%{OUTDISK}' eq '' # % error please say --define OUTDISK=/whatever # %endif # output_root=%{OUTDISK} output_root=/export/bmk/keeper/cpu2006.rep ####################### Processor usage ####################### # Note that the first instance of a "named section" occurs with # the copies line a little below. Therefore, any variables that # belong in the header section should be above this point. makeflags = -j16 # how many jobs specmake should spawn XJOBS = 16 # how many CPUs ipo can use (at link time) # Odds then Evens for 32 or 63 way on FF2 jupiter #bind = 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 0 bind = 8, 16, 24, 32, 40, 48, 56, 64, 72, 80, 88, 96, 104, 112, 120, 2, 10, 18, 26, 34, 42, 50, 58, 66, 74, 82, 90, 98, 106, 114, 122, 4, 12, 20, 28, 36, 44, 52, 60, 68, 76, 84, 92, 100, 108, 116, 124, 6, 14, 22, 30, 38, 46, 54, 62, 70, 78, 86, 94, 102, 110, 118, 126, 1, 9, 17, 25, 33, 41, 49, 57, 65, 73, 81, 89, 97, 105, 113, 121, 3, 11, 19, 27, 35, 43, 51, 59, 67, 75, 83, 91, 99, 107, 115, 123, 5, 13, 21, 29, 37, 45, 53, 61, 69, 77, 85, 93, 101, 109, 117, 125, 7, 15, 23, 31, 39, 47, 55, 63, 71, 79, 87, 95, 103, 111, 119, 127, 0 command_add_redirect = 1 # useful with submit submit = echo 'pbind -b $BIND \$\$ >> pbind.out; $command' > dobmk; sh dobmk %ifndef %{OverRideCopies} # # To cause the command line to win for both base and peak, say: # # runspec --define OverRideCopies --rate n # # If you _don't_ say that, then the order of precedence is: # # named section (highest), about 6 lines down, peak for several benchmarks # command line # the line marked "header section default" just below # # %endif parallel_setup = 16 parallel_setup_prefork=pbind -b $BIND $$ #################### Compiler invocation #################################### default=default=default=default: %define compiler_showversion 1 %define compiler_verbose 0 %ifdef %{compiler_showversion} CVERBOSE = -V CXXVERBOSE = -verbose=version FCVERBOSE = -V %endif %ifdef %{compiler_verbose} CVERBOSE = -V -\# CXXVERBOSE = -verbose=diags,version FCVERBOSE = -V -v %endif CC = $(SPECLANG)cc -xjobs=$(XJOBS) $(CVERBOSE) CXX = $(SPECLANG)CC -xjobs=$(XJOBS) $(CXXVERBOSE) -xdepend -library=stlport4 FC = $(SPECLANG)f90 -xjobs=$(XJOBS) $(FCVERBOSE) #################### Baseline Int Flags ######################### int=base=default=default: feedback = 0 OPTIMIZE = -fast -fma=fused -xipo=2 -xpagesize=4M $(XCACHE) COPTIMIZE = -xprefetch_level=1 -xalias_level=std CXXOPTIMIZE = -xprefetch_level=2 -xalias_level=compatible FOPTIMIZE = EXTRA_CXXLIBS = -lfast #################### Peak Int Optimization Flags ############################ int=peak: OPTIMIZE = -fast $(XCACHE) -xpagesize=4M CXXOPTIMIZE = -xalias_level=compatible 400.perlbench=peak=default=default: OPTIMIZE = -fast -fma=fused -xalias_level=std -xipo=2 -xprefetch_level=2 -xrestrict EXTRA_LIBS = -lfast 401.bzip2=peak=default=default: EXTRA_OPTIMIZE = -xalias_level=strong -fma=fused copies = 64 403.gcc=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -xalias_level=std -xprefetch=no -fma=fused EXTRA_LIBS = -ll2amm 429.mcf=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -xprefetch=no -xrestrict -xalias_level=std EXTRA_LIBS = -lfast feedback = 0 445.gobmk=peak=default=default: EXTRA_OPTIMIZE = -xalias_level=std -xrestrict -fma=fused 456.hmmer=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -fma=fused 458.sjeng=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -fma=fused -xchip=ultra3cu -xpagesize=4m 462.libquantum=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -xprefetch=latx:3 -fma=fused -xpagesize=64K EXTRA_CLIBS = -lbsdmalloc feedback=0 464.h264ref=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -xalias_level=std -xprefetch=no EXTRA_LIBS = -ll2amm 471.omnetpp=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -xprefetch_level=2 -Qoption cg -Qlp-av=0 -fma=fused EXTRA_LIBS = -lfast 473.astar=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -xprefetch_level=2 -fma=fused EXTRA_LIBS = -lfast feedback = 0 483.xalancbmk=peak: EXTRA_OPTIMIZE = -xipo=2 -xprefetch=no -fma=fused EXTRA_LIBS = -lfast feedback = 0 #################### Baseline Fp Flags ######################### fp=base=default=default: feedback = 0 OPTIMIZE = -fast -fma=fused $(XCACHE) -xipo=2 -xpagesize=4M -xprefetch_level=1 EXTRA_LIBS = COPTIMIZE = -xalias_level=std -xprefetch_auto_type=indirect_array_access CXXOPTIMIZE = -xalias_level=compatible FOPTIMIZE = #################### Peak FP Optimization Flags ############################ fp=peak: OPTIMIZE = -fast $(XCACHE) -xpagesize=4M CXXOPTIMIZE = -xalias_level=compatible 410.bwaves=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -xprefetch_level=2 -fma=fused feedback = 0 416.gamess=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -xprefetch_level=2 -fma=fused feedback = 1 433.milc=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -xprefetch_level=2 -fsimple=1 -xprefetch_auto_type=indirect_array_access -W2,-Ainline:rs=400 -xalias_level=std -fma=fused feedback = 0 434.zeusmp=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -fma=fused EXTRA_LIBS = -lmopt feedback = 0 435.gromacs=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -xarch=generic -xchip=generic -fsimple=0 -xunroll=5 -xprefetch=latx:0.5 436.cactusADM=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -fma=fused copies = 64 feedback = 0 437.leslie3d=peak=default=default: EXTRA_OPTIMIZE = -fma=fused -xipo=2 EXTRA_OPTIMIZE = -fma=fused -xipo=2 -xprefetch=latx:4 -xprefetch_level=2 feedback = 0 copies=64 444.namd=peak=default=default: EXTRA_OPTIMIZE = -fma=fused -xprefetch=latx:7 feedback = 0 447.dealII=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -xrestrict -xalias_level=compatible -fma=fused 450.soplex=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -xprefetch=no -fsimple=0 -xrestrict -xdepend -xalias_level=compatible 453.povray=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -xrestrict -fma=fused 454.calculix=peak=default=default: ###EXTRA_OPTIMIZE = -xipo=2 -xvector -xprefetch_level=3 -fma=fused -xprefetch=latx:3.0 ###EXTRA_COPTIMIZE = -xalias_level=std -xprefetch_auto_type=indirect_array_access EXTRA_OPTIMIZE = -xipo=2 -xprefetch_level=3 -fma=fused -xprefetch=latx:3.0 EXTRA_COPTIMIZE = -xalias_level=std feedback = 0 459.GemsFDTD=peak=default=default: EXTRA_OPTIMIZE = -fsimple=1 -xprefetch=no -fma=fused feedback = 0 465.tonto=peak=default=default: EXTRA_OPTIMIZE = -xipo=2 -xprefetch=no -xarch=generic EXTRA_LIBS = -lfast 470.lbm=peak=default=default: feedback = 0 copies=1 submit= OPTIMIZE = -fast -xipo=2 -xprefetch=latx:0.1 -m64 -xvector COPTIMIZE = -xalias_level=strong -xprefetch_level=3 -xprefetch_auto_type=indirect_array_access EXTRA_OPTIMIZE = -xarch=generic -xautopar -xreduction 481.wrf=peak: EXTRA_OPTIMIZE = -xipo=2 -xprefetch_level=3 -fma=fused -xunroll=8 copies=63 feedback=0 482.sphinx3=peak=default=default: #EXTRA_OPTIMIZE = -xinline= -xprefetch_level=2 -xrestrict -xalias_level=strong -Wc,-Qlp-ol=1 -fma=fused EXTRA_OPTIMIZE = -xinline= -xprefetch=no -xalias_level=strong -fma=fused EXTRA_LIBS = -lfast ###################### Portability ################################# 400.perlbench=default=default=default: CPORTABILITY = -DSPEC_CPU_SOLARIS_SPARC 403.gcc=default=default=default: CPORTABILITY = -DSPEC_CPU_SOLARIS 462.libquantum=default: CPORTABILITY=-DSPEC_CPU_SOLARIS 483.xalancbmk=default: CXXPORTABILITY=-DSPEC_CPU_SOLARIS ###################### Profile Feedback ################################# # # By default, turns on feedback for everything # ### default=default=default=default: fdo_pre0 = rm -rf ./feedback.profile ./SunWS_cache PASS1_CFLAGS = -xprofile=collect:./feedback PASS1_CXXFLAGS = -xprofile=collect:./feedback PASS1_FFLAGS = -xprofile=collect:./feedback PASS1_LDFLAGS = -xprofile=collect:./feedback PASS2_CFLAGS = -xprofile=use:./feedback PASS2_CXXFLAGS = -xprofile=use:./feedback PASS2_FFLAGS = -xprofile=use:./feedback PASS2_LDFLAGS = -xprofile=use:./feedback ################### Testbed ############################################## default=default=default=default: hw_avail = Jul-2008 hw_cpu_mhz = 2520 hw_cpu_name = SPARC64 VII hw_disk000 = 805 GB RAID 0 Solaris Volume hw_disk001 = 12 x Fujitsu 73 GB 10000 RPM SAS hw_disk002 = Stripe interlace size 512 Kbytes hw_fpu = Integrated hw_memory = 256 GB (128 x 2 GB) hw_model = Fujitsu SPARC Enterprise M8000 hw_nchips = 16 hw_ncores = 64 hw_ncoresperchip = 4 hw_ncpuorder = 1 to 4 CMUs; each CMU contains 2 or 4 chips hw_nthreadspercore = 2 hw_ocache = None hw_other = None hw_pcache = 64 KB I + 64 KB D on chip per core hw_scache = 6 MB I+D on chip per chip hw_tcache = None hw_vendor = Fujitsu Limited license_num = 19 sw_avail = Jul-2008 sw_base_ptrsize = 32-bit sw_compiler000 = Sun Studio 12 with patches sw_compiler001 = 124867-06, 124861-07, 124863-05, 127000-05 sw_compiler002 = (see patch information below) sw_file = ufs sw_os = Solaris 10 5/08 with Patch 137111-03 sw_other = None sw_peak_ptrsize = 32/64-bit sw_state = Default tester = Sun Microsystems notes_submit_000 = notes_submit_005 =Processes were assigned to specific processors using 'pbind' commands. The config notes_submit_010 =file option 'submit' was used, along with a list of processors in the 'BIND' notes_submit_015 =variable, to generate the pbind commands. (For details, please see the config file.) notes_submit_020 = notes_os_shell_000 = notes_os_shell_005 =Environment Variable Settings: notes_os_shell_010 = notes_os_shell_015 =The maximum number of threads a program can create was set with: notes_os_shell_020 = OMP_NUM_THREADS=127 notes_os_shell_025 = notes_os_shell_030 =Program threads were bound to processors with: notes_os_shell_035 = SUNW_MP_PROCBIND="1-127" notes_os_shell_040 = notes_os_shell_045 =Behavior of parallel threads was set with: notes_os_shell_050 = SUNW_MP_THR_IDLE=SPIN notes_os_shell_055 =SPIN specifies that an idle thread should spin while waiting at barrier notes_os_shell_060 =or waiting for new parallel regions to work on. notes_os_shell_065 = notes_os_sys_000 = notes_os_sys_005 =System Tunables (/etc/system parameters): notes_os_sys_010 = notes_os_sys_015 =tune_t_fsflushr=10 notes_os_sys_020 = Controls how many seconds elapse between runs of the notes_os_sys_025 = page flush daemon, fsflush. notes_os_sys_030 =autoup=300 notes_os_sys_035 = Causes pages older than the listed number of seconds to notes_os_sys_040 = be written by fsflush. notes_os_sys_045 =bufhwm=3000 notes_os_sys_050 = Memory byte limit for caching I/O buffers notes_os_sys_055 =segmap_percent=3 notes_os_sys_060 = Set maximum percent memory for file system cache notes_os_sys_065 =lpg_alloc_prefer=1 notes_os_sys_070 = Set lgroup page allocation to strongly prefer local pages notes_os_sys_075 = notes_os_sys_080 =Other System Settings: notes_os_sys_085 = notes_os_sys_090 =The webconsole service was turned off using notes_os_sys_095 = svcadm disable webconsole notes_os_sys_100 = # 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. 434.zeusmp: # The following setting was inserted automatically as a result of # post-run basepeak application. basepeak = 1 default: test_sponsor = Fujitsu Limited notes_plat_000 = notes_plat_005 =Memory is 8-way interleaved by filling all slots with notes_plat_010 =the same capacity DIMMs. notes_plat_015 = notes_plat_020 =This result is measured on a Sun SPARC Enterprise M8000 Server. notes_plat_025 =Note that the Sun SPARC Enterprise M8000 and Fujitsu SPARC Enterprise notes_plat_030 =M8000 are electrically equivalent. notes_plat_035 = notes_comp_000 =Sun Studio compiler patches are available at notes_comp_005 = http://developers.sun.com/sunstudio/downloads/patches/ss12_patches.jsp