# Invocation command line: # /home/opt/app/hpc2021/bin/harness/runhpc --rebuild --reportable --define EXPID=submission/7.12/small.omp.3node.rank_48.ppn_16.thread_10 -c xfusion.small.icx.base-peak.3node.cfg -T base,peak --define model=omp --define RANKS=48 --define PPN=16 --define THREADS=10 --size ref --iterations 2 --flagsurl ./config/flags/Intel-oneAPI-icx2021-official-linux64.xml --output_format=all small # output_root was not used for this run ############################################################################ #!/bin/bash ###################################################################### # Example configuration file for the Intel 2021 Compilers # # Defines: "acctype" => "mpi", "omp", default "mpi" # "label" => ext base label, default "xufsion" # # MPI-only Command: # runhpc --rebuild --reportable --define EXPID=submission/small.mpi.rank_240.ppn_80 --define model=mpi --iterations=3 -c xfusion.small.icx.base-peak.3node.cfg -T base --input ref --define RANKS=240 --define PPN=80 --flagsurl ./config/flags/Intel-oneAPI-icx2021-official-linux64.xml --output_format=all small # # OpenMP Command: # runhpc --rebuild --reportable --define EXPID=submission/small.omp.rank_48.ppn_16.thread_10 --define model=omp --iterations=3 -c xfusion.small.icx.base-peak.3node.cfg -T base,peak --input ref --define RANKS=48 --define PPN=16 --define THREADS=10 --flagsurl ./config/flags/Intel-oneAPI-icx2021-official-linux64.xml --output_format=all small # ####################################################################### expid= %ifdef %{EXPID} expid=%{EXPID} %endif allow_label_override = yes # label controls srcalt: simd - for simd build_in_build_dir=0 # build in run dir basepeak=0 %ifndef %{label} # IF label is not set use intel % define label xfusion %endif %ifndef %{model} # IF acctype is not set use mpi % define pmodel MPI %endif teeout = yes makeflags=-j # Tester description license_num = 6488 test_sponsor = xFusion tester = xFusion ###################################################### # SUT Section ###################################################### #include: Example_SUT.inc # ----- Begin inclusion of 'Example_SUT.inc' ############################################################################ ###################################################### # Example configuration information for a # system under test (SUT) Section ###################################################### # General SUT info system_vendor = xFusion system_name = FusionServer 2288H V6 (Intel Xeon Platinum 8380) hw_avail = Apr-2021 sw_avail = Oct-2021 prepared_by = xFusion # Computation node info # [Node_Description: Hardware] node_compute_syslbl = Fusionserver 2288H V6 node_compute_order = 1 node_compute_count = 3 node_compute_purpose = Compute node_compute_hw_vendor = xFusion node_compute_hw_model = FusionServer 2288H V6 node_compute_hw_cpu_name = Intel Xeon Platinum 8380 node_compute_hw_ncpuorder = 1, 2 chips node_compute_hw_nchips = 2 node_compute_hw_ncores = 80 node_compute_hw_ncoresperchip = 40 node_compute_hw_nthreadspercore = 2 node_compute_hw_cpu_char = Turbo Boost Technology up to 3.4 GHz node_compute_hw_cpu_mhz = 2300 node_compute_hw_pcache = 32 KB I + 48 KB D on chip per core node_compute_hw_scache = 1.25 MB I+D on chip per core node_compute_hw_tcache = 60 MB I+D on chip per chip node_compute_hw_ocache = None node_compute_hw_memory = 256 GB (16 x 16 GB 2Rx8 PC4-3200R) node_compute_hw_disk = 2 x 480 GB SATA 2.5" SSD (RAID 1) node_compute_hw_other = None #[Node_Description: Accelerator] node_compute_hw_accel_count = None node_compute_hw_accel_vendor= None node_compute_hw_accel_type = None node_compute_hw_accel_connect = None node_compute_hw_accel_ecc = None node_compute_hw_accel_desc = None #[Node_Description: Software] node_compute_hw_adapter_fs_model = MCX653105A-EFAT node_compute_hw_adapter_fs_count = 1 node_compute_hw_adapter_fs_slot_type = PCI-Express 4.0 x16 node_compute_hw_adapter_fs_data_rate = 100 Gb/s node_compute_hw_adapter_fs_ports_used = 1 node_compute_hw_adapter_fs_interconnect = Mellanox HDR node_compute_hw_adapter_fs_driver = 5.4-3.1.0 node_compute_hw_adapter_fs_firmware = 20.32.1010 node_compute_sw_os000 = CentOS Linux release 8.2.2004 node_compute_sw_os001 = 4.18.0-193.el8.x86_644 node_compute_sw_localfile = xfs node_compute_sw_sharedfile = NFS node_compute_sw_state = Multi-user, run level 3 node_compute_sw_other = None #[Fileserver] #[Interconnect] interconnect_fs_syslbl = Mellanox HDR interconnect_fs_order = 0 interconnect_fs_purpose = MPI Traffic interconnect_fs_hw_vendor = Mellanox interconnect_fs_hw_model = Mellanox HDR interconnect_fs_hw_switch_fs_model000= Mellanox MQM8790-HS2F interconnect_fs_hw_switch_fs_model001 = InfiniBand Switch interconnect_fs_hw_switch_fs_count = 1 interconnect_fs_hw_switch_fs_ports = 40 interconnect_fs_hw_topo = Mesh interconnect_fs_hw_switch_fs_data_rate = 200 Gbit/s interconnect_fs_hw_switch_fs_firmware = -- ####################################################################### # End of SUT section # If this config file were to be applied to several SUTs, edits would # be needed only ABOVE this point. ###################################################################### # ---- End inclusion of '/panfs/projects/innl/emelnich/SPECMPIACCEL/SPECMPIACCEL_SUITE/kit39-Release-1.0.2/hpc2021/config/Example_SUT.inc' ###################################################################### # The header section of the config file. Must appear # before any instances of "section markers" (see below) # # ext = how the binaries you generated will be identified # tune = specify "base" or "peak" or "all" label = %{label}_%{model} tune = all output_format = all use_submit_for_speed = 1 # Compiler Settings default: AR = ar ARFLAGS = cr CC = mpiicc -cc=icx CXX = mpiicpc -cxx=icx FC = mpiifort -fc=ifx system_class = Homogeneous Cluster sw_compiler = Intel oneAPI Compiler 2021.4.0 # Compiler Version Flags CC_VERSION_OPTION = --version CXX_VERSION_OPTION = --version FC_VERSION_OPTION = --version # Optimization # Note that SPEC baseline rules require that all uses of a given compiler # use the same flags in the same order. See the SPEChpc Run Rules # for more details # http://www.spec.org/hpc2021/Docs/runrules.html # # OPTIMIZE = flags applicable to all compilers # COPTIMIZE = flags appliable to the C compiler # CXXOPTIMIZE = flags appliable to the C++ compiler # FOPTIMIZE = flags appliable to the Fortran compiler # # See your compiler manual for information on the flags available # for your compiler vec_novec=-no-vec vec_avx2=-xCORE-AVX2 vec_avx512=-xCORE-AVX512 vec_avx512_high=-xCORE-AVX512 -mprefer-vector-width=512 vec_avx512_streaming_stores=-xCORE-AVX512 -mllvm -hir-nontemporal-cacheline-count=0 vec_avx512_high_exp1=-xCORE-AVX512 -mprefer-vector-width=512 -ffast-math vecavx512_high_exp2=-xCORE-AVX512 -mprefer-vector-width=512 -flto vec_avx512_high_exp3=-xCORE-AVX512 -mprefer-vector-width=512 -funroll-loops vec_avx512_high_exp4=-xCORE-AVX512 -mprefer-vector-width=512 -ffast-math -flto -funroll-loops veci_avx512_high_exp5=-xCORE-AVX512 -mprefer-vector-width=512 -ffinite-math-only vec_avx512_high_exp6=-xCORE-AVX512 -mprefer-vector-width=512 -fimf-precision=low:sin,sqrt vec_avx512_high_exp7=-xCORE-AVX512 -mprefer-vector-width=512 -ffinite-math-only -fimf-precision=low:sin,sqrt -ffast-math -flto -funroll-loops vec_common512=-xCOMMON-AVX512 vec=-xCORE-AVX512 -mprefer-vector-width=512 default=base,peak: OPTIMIZE = -Ofast -ipo ${vec} -fiopenmp COPTIMIZE = -ansi-alias CXXOPTIMIZE = -ansi-alias PORTABILITY = -lstdc++ FOPTIMIZE = -nostandard-realloc-lhs -align array64byte mpicmd = mpiexec.hydra -bootstrap ssh -hostfile $[top]/3node -np $ranks -ppn $ppn -genv OMP_NUM_THREADS=$threads $command submit = $mpicmd default=base=default: ranks = %{RANKS} threads = %{THREADS} ppn = %{PPN} %if %{model} eq 'omp' pmodel=OMP %endif 605.lbm_s=peak: ranks=24 ppn=8 threads=20 vec=${vec_avx512_high} pmodel=OMP 613.soma_s=peak: ranks=6 ppn=2 threads=80 vec=${vec_avx512_high} pmodel=OMP 618.tealeaf_s=peak: basepeak=1 619.clvleaf_s=peak: ranks=24 ppn=8 threads=20 vec=${vec_avx512_streaming_stores} pmodel=OMP 621.miniswp_s=peak: ranks=24 ppn=8 threads=20 vec=${vec_avx512_high} pmodel=OMP 628.pot3d_s=peak: ranks=6 ppn=2 threads=80 vec=${vec_avx512_high} pmodel=OMP 632.sph_exa_s=peak: ranks=48 ppn=16 threads=10 vec=${vec_avx512_high_exp1} pmodel=OMP 634.hpgmgfv_s=peak: ranks=48 ppn=16 threads=10 vec=${vec_no_vec} pmodel=OMP 635.weather_s=peak: basepeak=1 613.soma_s=base,peak: PORTABILITY+=-DSPEC_NO_VAR_ARRAY_REDUCE # 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: flagsurl000 = http://www.spec.org/hpc2021/flags/Intel-oneAPI-icx2021-official-linux64.xml sw_mpi_library000 = Intel MPI Library for Linux* OS, Version 2021.4.0 sw_mpi_library001 = Build 20210831 notes_submit_000 =export LD_PRELOAD="/usr/lib64/libhugetlbfs.so $LD_PRELOAD" notes_submit_005 =export OMP_PROC_BIND=true notes_submit_010 =mpiexec.hydra -bootstrap ssh -hostfile $[top]/3node --bind-to core -np $ranks -ppn $ppn -genv OMP_NUM_THREADS=$threads $command