# Invocation command line: # /autobench/sources/speccpu2006/bin/runspec --config speccpu_at.cfg --action run --rate 1 --copies 64 --reportable --nobuild --tune base --iterations 3 --define sles11sp2=1 --define mybasearchp7=1 --define libhuge=1 --define tcmalloc=1 --define lto=1 --define 16c730=1 --define numcores=16 --extension AT-601-sles11sp2 int # output_root was not used for this run ############################################################################ ####################################################################### # # SPEC CPU2006 Benchmark Config file IBM eServer pSeries Linux # ####################################################################### # THIS CONFIG FILE IS FOR BUILDING on P7 # # --define (-S) options that are passed to the runspec command # that dynamically affect this config file # # numcores=n The number of cores in the system # to determine how many copies to run in rate # # mybaseopt=x Allows override where x=3,4,5 (-O3, -O4, -O5) for base runs # # libhuge # tcmalloc # # debug Adds debug ('-g') compiler option # lto # libmass # # rhel # sles # noL3 # debug # # mymode64 # The following setting was inserted automatically as a result of # post-run basepeak application. basepeak = 1 ##################################################################### # System Under Test (SUT) Section # Replace the setting below with the ones that match your # machine. ##################################################################### ############ sysinfo_program = # System Description %ifdef %{48c760} %ifdef %{rhel64} hw_model = IBM Power 760 (3.4 GHz, 48 core, RHEL) %endif %ifdef %{sles11sp2} hw_model = IBM Power 760 (3.4 GHz, 48 core, SLES) %endif hw_avail = Mar-2013 hw_memory = 512 GB (64x8 GB) DDR3 1066 MHz hw_disk = 3 x 146.8 GB Raid0 SAS SFF 15K RPM hw_cpu_mhz = 3416 hw_nchips = 4 hw_ncores = 48 hw_ncoresperchip = 12 hw_nthreadspercore = 4 hw_fpu = Integrated hw_ncpuorder = 12, 24, 36, 48 cores hw_ocache = None hw_other = None hw_cpu_char000 = Intelligent Energy Optimization enabled, hw_cpu_char001 = up to 3.787 GHz %endif %ifdef %{32c750} %ifdef %{rhel64} hw_model = IBM Power 750 Express (4.0 GHz, 32 core, RHEL) %endif %ifdef %{sles11sp2} hw_model = IBM Power 750 Express (4.0 GHz, 32 core, SLES) %endif hw_avail = Mar-2013 hw_memory = 256 GB (64x4 GB) DDR3 1066 MHz hw_disk = 3 x 146.8 GB Raid0 SAS SFF 15K RPM hw_cpu_mhz = 4060 hw_nchips = 8 hw_ncores = 32 hw_ncoresperchip = 4 hw_nthreadspercore = 4 hw_fpu = Integrated hw_ncpuorder = 8, 16, 24, 32 cores hw_ocache = None hw_other = None hw_cpu_char000 = Intelligent Energy Optimization enabled, hw_cpu_char001 = up to 4.431 GHz %endif %ifdef %{16c730} %ifdef %{rhel64} hw_model = IBM Power 730 Express (4.2 GHz, 16 core, RHEL) %endif %ifdef %{sles11sp2} hw_model000= IBM Power 730 Express (4.2 GHz, 16 core, SLES, hw_model001 = GCC) %endif hw_avail = Feb-2013 hw_memory = 128 GB (16 x 8 GB) DDR3 1066 MHz hw_disk = 1 x 146.8 GB SAS SFF 15K RPM hw_cpu_mhz = 4228 hw_nchips = 2 hw_ncores = 16 hw_ncoresperchip = 8 hw_nthreadspercore = 4 hw_fpu = Integrated hw_ncpuorder = 8, 16 cores hw_ocache = None hw_other = None hw_cpu_char000 = Intelligent Energy Optimization enabled, hw_cpu_char001 = up to 4.540 GHz %endif %ifdef %{16c740} %ifdef %{rhel64} hw_model = IBM Power 740 (4.2 GHz, 16 core, RHEL) %endif %ifdef %{sles11sp2} hw_model = IBM Power 740 (4.2 GHz, 16 core, SLES) %endif hw_avail = Feb-2013 hw_memory = 128 GB (32x4 GB) DDR3 1066 MHz hw_disk = 5 x 146.8 GB Raid0 SAS SFF 15K RPM hw_cpu_mhz = 4228 hw_nchips = 2 hw_ncores = 16 hw_ncoresperchip = 8 hw_nthreadspercore = 4 hw_fpu = Integrated hw_ncpuorder = 6, 8, 12, 16 cores hw_ocache = None hw_other = None hw_cpu_char000 = Intelligent Energy Optimization enabled, hw_cpu_char001 = up to 4.540 GHz %endif %ifdef %{16c7R2} %ifdef %{rhel64} hw_model = IBM PowerLinux 7R2 (4.2 GHz, 16 core, RHEL) %endif %ifdef %{sles11sp2} hw_model = IBM PowerLinux 7R2 (4.2 GHz, 16 core, SLES) %endif hw_avail = Feb-2013 hw_memory = 128 GB (16x8 GB) DDR3 1066 MHz hw_disk = 5 x 146.8 GB Raid0 SAS SFF 15K RPM hw_cpu_mhz = 4228 hw_nchips = 2 hw_ncores = 16 hw_ncoresperchip = 8 hw_nthreadspercore = 4 hw_fpu = Integrated hw_ncpuorder = 16 cores hw_ocache = None hw_other = None hw_cpu_char000 = Intelligent Energy Optimization enabled, hw_cpu_char001 = up to 4.540 GHz %endif hw_vendor = IBM Corporation # CPU description hw_cpu_name = POWER7+ hw_pcache = 32 KB I + 32 KB D on chip per core hw_scache = 256 KB I+D on chip per core hw_tcache = 10 MB I+D on chip per core # Tester description - edit the date, put your license number here! license_num = 11 prepared_by = -- test_sponsor = IBM Corporation tester = IBM Corporation # Operating system, file system sw_avail = Dec-2012 %ifdef %{rhel64} sw_os000 = Red Hat Enterprise Linux Server release 6.4 sw_os001 = (ppc64) kernel sw_file = ext4 %endif %ifdef %{sles11sp2} sw_os000 = SUSE Linux Enterprise Server 11 SP2 sw_os001 = (ppc64) kernel 3.0.42-0.7-ppc64 sw_file = ext3 %endif ############ hw_vendor = IBM Corporation # CPU description hw_cpu_name = POWER7+ hw_pcache = 32 KB I + 32 KB D on chip per core hw_scache = 256 KB I+D on chip per core hw_tcache = 10 MB I+D on chip per core # Tester description - edit the date, put your license number here! license_num = 11 prepared_by = -- test_sponsor = IBM Corporation tester = IBM Corporation # Operating system, file system sw_avail = Nov-2012 %ifdef %{rhel64} notes_comp_iler_invocation_000: For more information about IBM Advance Toolchain, including support, see notes_comp_iler_invocation_005: ftp://linuxpatch.ncsa.uiuc.edu/toolchain/at/redhat/RHEL6/at6.0/release_notes.at6.0-6.0-1.html %endif %ifdef %{sles11sp2} notes_comp_iler_invocation_000: For more information about IBM Advance Toolchain, including support, see notes_comp_iler_invocation_005: ftp://linuxpatch.ncsa.uiuc.edu/toolchain/at/suse/SLES_11/at6.0/release_notes.at6.0-6.0-1.html %endif sw_compiler000 = C/C++: Version 4.7.3 of IBM sw_compiler001 = Advance Toolchain 6.0-1 gcc/g++ compiler sw_other = -IBM Advance Toolchain 6.0-1 sw_base_ptrsize = 32-bit sw_peak_ptrsize = 32-bit sw_parallel_other = -- sw_state = Run level 3 (multi-user) notes_submit_000 = The config file option 'submit' was used notes_submit_005 = to assign benchmark copy to specific kernel thread using notes_submit_010 = the "numactl" command (see flags file for details). notes_os_000 = ulimit -s (stack) set to 1048576. notes_os_005 = notes_os_010 = Large pages reserved as follows by root user: %if defined(%{16c730}) || defined(%{16c740}) || defined(%{16c7R2}) notes_os_015 = echo 4224 > /proc/sys/vm/nr_hugepages %endif %ifdef %{32c750} notes_os_020 = echo 8448 > /proc/sys/vm/nr_hugepages %endif %ifdef %{48c760} notes_os_020 = echo 12672 > /proc/sys/vm/nr_hugepages %endif notes_os_020 = %if '%{benchmark}' eq 'fp' notes_os_030 = The Apache C++ Standard Library V4.2.1 was installed from notes_os_035 = http://stdcxx.apache.org/download.html using: notes_os_040 = gmake BUILDTYPE=8d CONFIG=gcc.config notes_os_045 = %endif notes_os_025 = The following environment varibles were set before the runspec command: notes_os_030 = export HUGETLB_VERBOSE=0 notes_os_035 = export HUGETLB_MORECORE=yes %if '%{benchmark}' eq 'fp' notes_os_065 = export HUGETLB_ELFMAP=RW notes_os_070 = export XLFRTEOPTS=intrinthds=1 %endif ###################################################################### # End of SUT section # If this config file were to be applied to several SUTs, edits would # be needed only ABOVE this point. ###################################################################### %define P7 -mcpu=power7 -mtune=power7 %ifdef %{libhuge} %define LHBDT -B/usr/share/libhugetlbfs/ -Wl,--hugetlbfs-align %define LHLIB -lhugetlbfs %else %define LHBDT %define LHLIB %endif %if %{tcmalloc} %if %{mymode64} %define TCMALLOCLIB -ltcmalloc %else %define TCMALLOCLIB -ltcmalloc %endif %else %define TCMALLOCLIB %endif %ifdef %{libmass} %define VEC_OPT -fvect-cost-model -mveclibabi=mass %define LIB_MATH -lm %if %{mymode64} %define LIB_MASS -L/opt/ibmcmp/xlmass/7.1/lib64 -lmassvp7_64 -lmass_simdp7_64 -lmass_64 %else %define LIB_MASS -L/opt/ibmcmp/xlmass/7.1/lib -lmassvp7 -lmass_simdp7 -lmass %endif %else %define VEC_OPT -fvect-cost-model %define LIB_MATH %define LIB_MASS %endif %ifdef %{lto} %define LTO -flto -fwhole-program -fuse-linker-plugin %else %define LTO %endif %define FDPRPRO /opt/ibm/fdprpro/bin/fdprpro %define FDPROPTS -O4 -omullX %define FDPRINT %ifdef %{numcores} %define make_j %{numcores} %else %define make_j 4 %endif #====================================================================== # Global Options apply to all runs #====================================================================== # # For just a report generated without a raw file or any runs required # Do a "runspec --reportonly" (also --action=report or --fakereport) # # Just set up the build or run directories, use --buildsetup or --setup # on the runspec command # Sets nobuild as the default - you will need -D to force a compile # nobuild = 1 # -D forces a build (recompile) # -N prohibits a build (recompile) #action = validate # -a #tune = base,peak # -T #iterations = 3 # -n output_format = raw,text,html,csv # -o #basepeak = 0 #reportable = 1 # -l turns off #ignore_errors = 0 # -I mean_anyway = 1 allow_extension_override = 1 #ext = LoP-ppc64 # -e #size = ref post_setup = sync parallel_test = 1 # This tells runspec to set enviornment variables of the form # ENV_xxx if found in here. Forced to 0 for reportable runs env_vars = 0 # Control if the embedded md5 string of compiler options is checked for # recompile if the options have changed check_md5 = 0 # Controls if the log file info is echo'd to the screen teeout = yes # If 1, make copy of the config file before modifing with MD5 # signitures of the compile options. backup_config = 0 # verbose -> -v #verbose = 1 to 99, 5 is default, 91 is good for debugging verbose = 1 %ifdef %{make_j} makeflags = -j%{make_j} %endif # We alway use submit to ensure there is a bindprocessor done. # use_submit_for_speed If set, use submit commands for speed as well as rate use_submit_for_speed = 1 # SPEED submit = echo "$command" > run.sh ; numactl --membind=\$SPECCOPYNUM --physcpubind=\$SPECCOPYNUM sh run.sh %ifdef %{48c760} bind = \ 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80 84 88 92 96 100 104 108 112 116 120 124 128 132 136 140 144 148 152 156 160 164 168 172 176 180 184 188 \ 1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97 101 105 109 113 117 121 125 129 133 137 141 145 149 153 157 161 165 169 173 177 181 185 189 \ 2 6 10 14 18 22 26 30 34 38 42 46 50 54 58 62 66 70 74 78 82 86 90 94 98 102 106 110 114 118 122 126 130 134 138 142 146 150 154 158 162 166 170 174 178 182 186 190 \ 3 7 11 15 19 23 27 31 35 39 43 47 51 55 59 63 67 71 75 79 83 87 91 95 99 103 107 111 115 119 123 127 131 135 139 143 147 151 155 159 163 167 171 175 179 183 187 191 copies = 192 submit = let "MYCPU=\$BIND"; MYMEM=`printf '%d' \$((\$MYCPU/24))` ; numactl --physcpubind=\$MYCPU --membind=\$MYMEM $command %endif %ifdef %{32c750} bind = \ 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80 84 88 92 96 100 104 108 112 116 120 124 \ 1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97 101 105 109 113 117 121 125 \ 2 6 10 14 18 22 26 30 34 38 42 46 50 54 58 62 66 70 74 78 82 86 90 94 98 102 106 110 114 118 122 126 \ 3 7 11 15 19 23 27 31 35 39 43 47 51 55 59 63 67 71 75 79 83 87 91 95 99 103 107 111 115 119 123 127 copies = 128 submit = let "MYCPU=\$BIND"; MYMEM=`printf '%d' \$((\$MYCPU/16))` ; numactl --physcpubind=\$MYCPU --membind=\$MYMEM $command %endif %if defined(%{16c730}) || defined(%{16c740}) || defined(%{16c7R2}) bind = \ 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 \ 1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 \ 2 6 10 14 18 22 26 30 34 38 42 46 50 54 58 62 \ 3 7 11 15 19 23 27 31 35 39 43 47 51 55 59 63 copies = 64 submit = echo "$command" > run.sh; let "MYCPU=\$BIND"; MYMEM=`printf '%d' \$((\$MYCPU/32))` ; numactl --physcpubind=\$MYCPU --membind=\$MYMEM sh run.sh %endif #--prevent log file lines that are too long for vi log_line_width = 130 # Point to where the usr/compiler/OS Flags file is flagsurl000= http://www.spec.org/cpu2006/flags/IBM-Linux-AT.xml flagsurl001= http://www.spec.org/cpu2006/flags/IBM-Power.20130226.xml #=============================================================================== # How the compilers are called # The settings used here are for IBM Advance Toolchain C/C++ and Fortran compilers #=============================================================================== default=default=default=default: %define MY_PATH /opt/at6.0/bin CC = %{MY_PATH}/gcc CXX = %{MY_PATH}/g++ FC = %{MY_PATH}/gfortran #====================================================================== # Base Level Optimizations # No limit to number of flags, must be same for all of each Language. #====================================================================== default=base=default=default: %ifndef %{debug} %define DEBUG %else %define DEBUG -g %endif # Check for BASEARCH overrides # Default BASEARCH is P7 %define BASEARCH %{P7} %define MODE -m32 %if %{mymode64} %define MODE -m64 %endif # Note if set -mcpu=power7, then don't need to specify -mvsx ,-mpopcntd, -mcmpb, or -mhard-dfp. # But -maltivec is not set automatically. # Note when using -mcpu=power7, # - do not disable Altivec without also disabling VSX # this is illegal -mcpu=power7 -mno-altivec. %define GCCOPT -ffast-math -fpeel-loops -funroll-loops -mvsx -maltivec -ftree-vectorize -mpopcntd -mrecip=rsqrt %define GCCOPT_MATH -fno-associative-math #%define GCCOPT_LOOP -ftree-loop-linear %define DPFOPT -mno-hard-dfp -D__STDC_WANT_DEC_FP__ -ldfp # Default mode is -O3 with p7+ %define BASEOPT -O3 %{BASEARCH} %{MODE} %{GCCOPT} %{DEBUG} #Check for BASEOPT overrides %if %{mybaseopt} == 2 %define BASEOPT -O2 %{BASEARCH} %{MODE} %{GCCOPT} %{DEBUG} %endif %if %{mybaseopt} == 1 %define BASEOPT -O1 %{BASEARCH} %{MODE} %{GCCOPT} %{DEBUG} %endif int=base=default=default: COPTIMIZE = %{BASEOPT} -fno-strict-aliasing %if %{mybasearchp7} COPTIMIZE = %{BASEOPT} %{LTO} %endif EXTRA_CLIBS = %{LHLIB} CXXOPTIMIZE = %{BASEOPT} %{LTO} EXTRA_CXXLIBS = %{TCMALLOCLIB} fp=base=default=default: COPTIMIZE = %{BASEOPT} %{VEC_OPT} CXXOPTIMIZE = %{BASEOPT} %{VEC_OPT} %{GCCOPT_MATH} FOPTIMIZE = %{BASEOPT} %{VEC_OPT} EXTRA_LIBS = %{LHLIB} %{LIB_MASS} %{LIB_MATH} default=base=default=default: %if %{mymode64} PORTABILITY = -DSPEC_CPU_LP64 %endif #====================================================================== # INT Portability Flags # no optimizations allowed here #====================================================================== 400.perlbench=default=default=default: %if %{mymode64} CPORTABILITY = -DSPEC_CPU_LINUX_PPC -DSPEC_CPU_LP64 %else CPORTABILITY = -DSPEC_CPU_LINUX_PPC %endif 403.gcc=default=default=default: %if %{mymode64} CPORTABILITY = -DSPEC_CPU_LP64 %endif 462.libquantum=default=default=default: CPORTABILITY = -DSPEC_CPU_LINUX 464.h264ref=default=default=default: CPORTABILITY = -fsigned-char 483.xalancbmk=default=default=default: CXXPORTABILITY = -DSPEC_CPU_LINUX #====================================================================== # FP Portability Flags # no optimizations allowed here #====================================================================== 436.cactusADM=peak=default=default: %if %{mymode64} CPORTABILITY = -DSPEC_CPU_LP64 %endif 447.dealII=default=default=default: %if %{mymode64} PORTABILITY = -DSPEC_CPU_LP64 %endif CXXPORTABILITY = -DSPEC_CPU_LINUX 481.wrf=default=default=default: CPORTABILITY = -DSPEC_CPU_CASE_FLAG -DSPEC_CPU_LINUX_PPC 482.sphinx3=default=default=default: CPORTABILITY = -fsigned-char #====================================================================== # # Default peak flags. # #====================================================================== default=peak=default=default: %define O2 -O2 %{BASEARCH} %{MODE} %{GCCOPT} %{DEBUG} %define O3 -O3 %{BASEARCH} %{MODE} %{GCCOPT} %{DEBUG} #====================================================================== # INT peak Flags #====================================================================== int=peak=default=default: 400.perlbench=peak=default=default: # C only Benchmark OPTIMIZE = %{O3} %{LTO} EXTRA_LIBS = %{TCMALLOCLIB} PASS1_CFLAGS = -fprofile-generate PASS1_LDFLAGS = -fprofile-generate PASS2_CFLAGS = -fprofile-use PASS2_LDFLAGS = -fprofile-use %ifdef %{64} CPORTABILITY = -DSPEC_CPU_LINUX_PPC -DSPEC_CPU_LP64 COPTIMIZE = -m64 %else CPORTABILITY = -DSPEC_CPU_LINUX_PPC %endif 401.bzip2=peak=default=default: # C only Benchmark %define FDPROPTS_BZIP2 -O4 -vrox OPTIMIZE = %{O3} %{LTO} -fno-branch-count-reg -fprefetch-loop-arrays EXTRA_LIBS = -lhugetlbfs LDCFLAGS = -Wl,-q fdo_pre1 = cp ${baseexe} ${baseexe}.orig ; %{FDPRPRO} -a instr %{FDPRINT} ${baseexe}; cp ${baseexe}.instr ${baseexe} fdo_run1 = $command fdo_post1 = cp ${baseexe}.orig ${baseexe} ; %{FDPRPRO} -a opt -f ${baseexe}.nprof %{FDPROPTS_BZIP2} ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} COPTIMIZE = -m64 %endif 403.gcc=peak=default=default: # C only Benchmark %define FDPROPTS_GCC -O4 -nodp -rtb -m power7 -A 1 OPTIMIZE = %{O3} %{LTO} EXTRA_LIBS = -lhugetlbfs LDCFLAGS = -Wl,-q #PASS1_CFLAGS = -fprofile-generate #PASS1_LDFLAGS = -fprofile-generate -fwhole-program -fuse-linker-plugin #PASS2_CFLAGS = -fprofile-use #PASS2_LDFLAGS = -fprofile-use -fwhole-program -fuse-linker-plugin fdo_pre2 = cp ${baseexe} ${baseexe}.orig ; %{FDPRPRO} -a instr %{FDPRINT} ${baseexe}; cp ${baseexe}.instr ${baseexe} fdo_run2 = $command fdo_post2 = cp ${baseexe}.orig ${baseexe} ; %{FDPRPRO} -a opt -f ${baseexe}.nprof %{FDPROPTS_GCC} ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} CPORTABILITY = -DSPEC_CPU_LP64 COPTIMIZE = -m64 %endif 429.mcf=peak=default=default: # C only Benchmark %define FDPROPTS_MCF -O3 -m power7 -A 1 -omullX OPTIMIZE = %{O3} %{LTO} --param max-goto-duplication-insns=16 EXTRA_LIBS = -lhugetlbfs LDCFLAGS = -Wl,-q EXTRA_LDFLAGS = -fwhole-program fdo_pre1 = cp ${baseexe} ${baseexe}.orig ; %{FDPRPRO} -a instr %{FDPRINT} ${baseexe}; cp ${baseexe}.instr ${baseexe} fdo_run1 = $command fdo_post1 = cp ${baseexe}.orig ${baseexe} ; %{FDPRPRO} -a opt -f ${baseexe}.nprof %{FDPROPTS_MCF} ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} COPTIMIZE = -m64 %endif 445.gobmk=peak=default=default: # C only Benchmark %define FDPROPTS_GOBMK -m power7 -A 1 -q -O3 -omullX OPTIMIZE = %{O3} %{LTO} EXTRA_LIBS = -lhugetlbfs LDCFLAGS = -Wl,-q #PASS1_CFLAGS = -fprofile-generate #PASS1_LDFLAGS = -fprofile-generate -fwhole-program -fuse-linker-plugin #PASS2_CFLAGS = -fprofile-use #PASS2_LDFLAGS = -fprofile-use -fwhole-program -fuse-linker-plugin fdo_pre2 = cp ${baseexe} ${baseexe}.orig ; %{FDPRPRO} -a instr %{FDPRINT} ${baseexe}; cp ${baseexe}.instr ${baseexe} fdo_run2 = $command fdo_post2 = cp ${baseexe}.orig ${baseexe} ; %{FDPRPRO} -a opt -f ${baseexe}.nprof %{FDPROPTS_GOBMK} ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} COPTIMIZE = -m64 %endif 456.hmmer=peak=default=default: # C only Benchmark %define FDPROPTS_HMMER -O4 -omullX #OPTIMIZE = %{O3} %{LTO} -finline-limit=11880 OPTIMIZE = %{O3} -finline-limit=11880 EXTRA_LIBS = -lhugetlbfs LDCFLAGS = -Wl,-q #PASS1_CFLAGS = -fprofile-generate #PASS1_LDFLAGS = -fprofile-generate -fwhole-program -fuse-linker-plugin #PASS2_CFLAGS = -fprofile-use #PASS2_LDFLAGS = -fprofile-use -fwhole-program -fuse-linker-plugin fdo_pre2 = cp ${baseexe} ${baseexe}.orig ; %{FDPRPRO} -a instr %{FDPRINT} ${baseexe}; cp ${baseexe}.instr ${baseexe} fdo_run2 = $command fdo_post2 = cp ${baseexe}.orig ${baseexe} ; %{FDPRPRO} -a opt -f ${baseexe}.nprof %{FDPROPTS_HMMER} ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} COPTIMIZE = -m64 %endif 458.sjeng=peak=default=default: # C only Benchmark %define FDPROPTS_SJENG -O4 -omullX OPTIMIZE = %{O3} %{LTO} -misel EXTRA_LIBS = -lhugetlbfs LDCFLAGS = -Wl,-q PASS1_CFLAGS = -fprofile-generate PASS1_LDFLAGS = -fprofile-generate PASS2_CFLAGS = -fprofile-use PASS2_LDFLAGS = -fprofile-use fdo_pre2 = cp ${baseexe} ${baseexe}.orig ; %{FDPRPRO} -a instr %{FDPRINT} ${baseexe}; cp ${baseexe}.instr ${baseexe} fdo_run2 = $command fdo_post2 = cp ${baseexe}.orig ${baseexe} ; %{FDPRPRO} -a opt -f ${baseexe}.nprof %{FDPROPTS_SJENG} ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} COPTIMIZE = -m64 %endif 462.libquantum=peak=default=default: # C only Benchmark # We want just 1 copy per core, not SMT, when we have L3cache %define FDPROPTS_LIBQUANTUM -m power7 -A 1 -bf -dp -lro -nop -RC -RD -tb -tlo -vro -isf 88 -lu 8 -hrf 0.10 -sdp 4 -lun 27 -omullX %if %{noL3} %define FDPROPTS_LIBQUANTUM -ece -pca -RC -RD -tb -tlo -lu 5 -rt 0.75 -hrf 0.75 -sdpla 4 -rcctf 0.55 -rccrf 0.10 -rcaf 2 -omullX %endif OPTIMIZE = %{O3} %{LTO} EXTRA_LIBS = -lhugetlbfs LDCFLAGS = -Wl,-q PASS1_CFLAGS = -fprofile-generate PASS1_LDFLAGS = -fprofile-generate PASS2_CFLAGS = -fprofile-use PASS2_LDFLAGS = -fprofile-use fdo_pre2 = cp ${baseexe} ${baseexe}.orig ; %{FDPRPRO} -a instr %{FDPRINT} ${baseexe}; cp ${baseexe}.instr ${baseexe} fdo_run2 = $command fdo_post2 = cp ${baseexe}.orig ${baseexe} ; %{FDPRPRO} -a opt -f ${baseexe}.nprof %{FDPROPTS_LIBQUANTUM} ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} COPTIMIZE = -m64 %endif 464.h264ref=peak=default=default: # C only Benchmark #OPTIMIZE = %{O3} %{LTO} -fno-branch-count-reg -misel -fprefetch-loop-arrays OPTIMIZE = %{O3} -fno-branch-count-reg -misel -fprefetch-loop-arrays EXTRA_LIBS = -lhugetlbfs LDCFLAGS = -Wl,-q PASS1_CFLAGS = -fprofile-generate PASS1_LDFLAGS = -fprofile-generate PASS2_CFLAGS = -fprofile-use PASS2_LDFLAGS = -fprofile-use #fdo_pre1 = cp ${baseexe} ${baseexe}.orig ; %{FDPRPRO} -a instr ${baseexe}; cp ${baseexe}.instr ${baseexe} #fdo_run1 = $command #fdo_post1 = cp ${baseexe}.orig ${baseexe} ; %{FDPRPRO} -a opt -f ${baseexe}.nprof -O4 -vrox -nodp ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} COPTIMIZE = -m64 %endif 471.omnetpp=peak=default=default: # C++ Benchmark #%define FDPROPTS_OMNETPP -O3 -lu -1 -nodp -sdp 9 OPTIMIZE = %{O3} %{LTO} EXTRA_LIBS = %{TCMALLOCLIB} LDCXXFLAGS = -Wl,-q #EXTRA_LDFLAGS = -fwhole-program -fuse-linker-plugin #fdo_pre1 = cp ${baseexe} ${baseexe}.orig ; %{FDPRPRO} -a instr %{FDPRINT} ${baseexe}; cp ${baseexe}.instr ${baseexe} #fdo_run1 = $command #fdo_post1 = cp ${baseexe}.orig ${baseexe} ; %{FDPRPRO} -a opt -f ${baseexe}.nprof %{FDPROPTS_OMNETPP} ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} CXXOPTIMIZE = -m64 %endif 473.astar=peak=default=default: # C++ Benchmark %define FDPROPTS_ASTAR -O4 -m power7 -A 1 OPTIMIZE = %{O3} %{LTO} EXTRA_LIBS = %{TCMALLOCLIB} LDCXXFLAGS = -Wl,-q #PASS1_CFLAGS = -fprofile-generate #PASS1_LDFLAGS = -fprofile-generate #PASS2_CFLAGS = -fprofile-use #PASS2_LDFLAGS = -fprofile-use fdo_pre1 = cp ${baseexe} ${baseexe}.orig ; %{FDPRPRO} -a instr %{FDPRINT} ${baseexe}; cp ${baseexe}.instr ${baseexe} fdo_run1 = $command fdo_post1 = cp ${baseexe}.orig ${baseexe} ; %{FDPRPRO} -a opt -f ${baseexe}.nprof %{FDPROPTS_ASTAR} ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} CXXOPTIMIZE = -m64 %endif 483.xalancbmk=peak=default=default: # C++ Benchmark #%define FDPROPTS_XALAN -O3 -m power7 OPTIMIZE = %{O3} %{LTO} EXTRA_LIBS = %{TCMALLOCLIB} LDCXXFLAGS = -Wl,-q #fdo_pre1 = cp ${baseexe} ${baseexe}.orig ; %{FDPRPRO} -a instr %{FDPRINT} ${baseexe}; cp ${baseexe}.instr ${baseexe} #fdo_run1 = $command #fdo_post1 = cp ${baseexe}.orig ${baseexe} ; %{FDPRPRO} -a opt -f ${baseexe}.nprof %{FDPROPTS_XALAN} ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} CXXOPTIMIZE = -m64 %endif 999.specrand=peak=default=default: # C only Benchmark - Must run, but not reported unless fails OPTIMIZE = %{O3} #============================================================ # FP peak Flags #============================================================ fp=peak=default=default: 410.bwaves=peak=default=default: # Fortran Benchmark %define 64 OPTIMIZE = %{O3} EXTRA_LIBS = -lhugetlbfs %ifdef %{64} OPTIMIZE = %{O3} FOPTIMIZE = -m64 %endif %undef 64 416.gamess=peak=default=default: # Fortran Benchmark OPTIMIZE = %{O3} LDFFLAGS = -Wl,-q fdo_pre1 = cp ${baseexe} ${baseexe}.orig ; /opt/bin/fdprpro -a instr %{FDPRINT} ${baseexe}; cp ${baseexe}.instr ${baseexe} fdo_run1 = $command fdo_post1 = cp ${baseexe}.orig ${baseexe} ; /opt/bin/fdprpro -a opt -f ${baseexe}.nprof %{FDPROPTS} ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} FOPTIMIZE = -m64 %endif 433.milc=peak=default=default: # C only Benchmark %define 64 OPTIMIZE = %{O3} EXTRA_LIBS = -lhugetlbfs LDCFLAGS = -Wl,-q fdo_pre1 = cp ${baseexe} ${baseexe}.orig ; /opt/bin/fdprpro -a instr %{FDPRINT} ${baseexe}; cp ${baseexe}.instr ${baseexe} fdo_run1 = $command fdo_post1 = cp ${baseexe}.orig ${baseexe} ; /opt/bin/fdprpro -a opt -f ${baseexe}.nprof %{FDPROPTS} ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} COPTIMIZE = -m64 %endif %undef 64 434.zeusmp=peak=default=default: # Fortran Benchmark basepeak = 1 435.gromacs=peak=default=default: # Fortran & C Benchmark OPTIMIZE = %{O3} EXTRA_LIBS = -lhugetlbfs LDCFLAGS = -Wl,-q LDFFLAGS = -Wl,-q fdo_pre2 = cp ${baseexe} ${baseexe}.orig ; /opt/bin/fdprpro -a instr %{FDPRINT} ${baseexe}; cp ${baseexe}.instr ${baseexe} fdo_run2 = $command fdo_post2 = cp ${baseexe}.orig ${baseexe} ; /opt/bin/fdprpro -a opt -f ${baseexe}.nprof %{FDPROPTS} ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} COPTIMIZE = -m64 FOPTIMIZE = -m64 %endif 436.cactusADM=peak=default=default: # Fortran & C Benchmark %define 64 OPTIMIZE = %{O3} %{LHBDT} %ifdef %{64} #FPPPORTABILITY = -WF,-DSPEC_CPU_LP64 FPPPORTABILITY = -DSPEC_CPU_LP64 CPORTABILITY = -DSPEC_CPU_LP64 FOPTIMIZE = -m64 COPTIMIZE = -m64 %endif %undef 64 437.leslie3d=peak=default=default: # Fortran Benchmark # We want just 1 copy per core, not SMT %define 64 OPTIMIZE = %{O3} %{LHBDT} %ifdef %{64} FOPTIMIZE = -m64 %endif %undef 64 444.namd=peak=default=default: # C++ Benchmark basepeak = 1 447.dealII=peak=default=default: # C++ Benchmark OPTIMIZE = %{O3} -static EXTRA_LIBS = -Wl,--whole-archive /usr/lib/libhugetlbfs.a -Wl,--no-whole-archive %ifdef %{64} CXXOPTIMIZE = -m64 %endif 450.soplex=peak=default=default: # C++ Benchmark basepeak = 1 %ifdef %{64} CXXOPTIMIZE = -m64 %endif 453.povray=peak=default=default: # C++ Benchmark %define FDPROPTS_POVRAY -O4 -omullX -see 1 -ihf -1 OPTIMIZE = %{O3} EXTRA_LIBS = %{SMARTHEAPLIB} LDCXXFLAGS = -Wl,-q PASS1_CXXFLAGS = -fprofile-generate PASS1_LDFLAGS = -fprofile-generate PASS2_CXXFLAGS = -fprofile-use PASS2_LDFLAGS = -fprofile-use fdo_pre2 = cp ${baseexe} ${baseexe}.orig ; /opt/bin/fdprpro -a instr %{FDPRINT} ${baseexe}; cp ${baseexe}.instr ${baseexe} fdo_run2 = $command fdo_post2 = cp ${baseexe}.orig ${baseexe} ; /opt/bin/fdprpro -a opt -f ${baseexe}.nprof %{FDPROPTS_POVRAY} ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} CXXPORTABILITY = -DSPEC_CPU_LP64 CXXOPTIMIZE = -m64 %endif 454.calculix=peak=default=default: # Fortran & C Benchmark basepeak = 1 %ifdef %{64} FOPTIMIZE = -m64 COPTIMIZE = -m64 %endif 459.GemsFDTD=peak=default=default: # Fortran Benchmark basepeak = 1 %ifdef %{64} FOPTIMIZE = -m64 %endif 465.tonto=peak=default=default: # Fortran Benchmark basepeak = 1 %ifdef %{64} FOPTIMIZE = -m64 %endif 470.lbm=peak=default=default: # C Benchmark #%define 64 OPTIMIZE = %{O3} %{LHBDT} LDCFLAGS = -Wl,-q PASS1_CFLAGS = -fprofile-generate PASS1_LDFLAGS = -fprofile-generate PASS2_CFLAGS = -fprofile-use PASS2_LDFLAGS = -fprofile-use fdo_pre2 = cp ${baseexe} ${baseexe}.orig ; /opt/bin/fdprpro -a instr %{FDPRINT} ${baseexe}; cp ${baseexe}.instr ${baseexe} fdo_run2 = $command fdo_post2 = cp ${baseexe}.orig ${baseexe} ; /opt/bin/fdprpro -a opt -f ${baseexe}.nprof %{FDPROPTS} ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} COPTIMIZE = -m64 %endif #%undef 64 481.wrf=peak=default=default: # Fortran & C-for netCDF Benchmark basepeak = 1 %ifdef %{64} COPTIMIZE = -m64 FOPTIMIZE = -m64 %endif 482.sphinx3=peak=default=default: # C only Benchmark OPTIMIZE = %{O3} EXTRA_LIBS = -lhugetlbfs LDCFLAGS = -Wl,-q PASS1_CFLAGS = -fprofile-generate PASS1_LDFLAGS = -fprofile-generate PASS2_CFLAGS = -fprofile-use PASS2_LDFLAGS = -fprofile-use fdo_pre2 = cp ${baseexe} ${baseexe}.orig ; /opt/bin/fdprpro -a instr %{FDPRINT} ${baseexe}; cp ${baseexe}.instr ${baseexe} fdo_run2 = $command fdo_post2 = cp ${baseexe}.orig ${baseexe} ; /opt/bin/fdprpro -a opt -f ${baseexe}.nprof %{FDPROPTS} ${baseexe} ; cp ${baseexe}.fdpr ${baseexe} %ifdef %{64} COPTIMIZE = -m64 %endif 998.specrand=peak=default=default: # C only Benchmark - Must run, but not reported unless fails OPTIMIZE = %{O3} # 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: notes_comp_iler_invocation_000 =For more information about IBM Advance Toolchain, including support, see notes_comp_iler_invocation_005 =ftp://linuxpatch.ncsa.uiuc.edu/toolchain/at/suse/SLES_11/at6.0/release_notes.at6.0-6.0-1.html