# Invocation command line: # /nfsshare/hpc2021/bin/harness/runhpc --reportable --define EXPID=submission/8490H/tiny.omp.4.32.8.30 --define model=omp --iterations=2 -c xfusion.omp.cfg.4 --tune=base --define HOSTFILE=node4 --define RANKS=32 --define PPN=8 --define THREADS=30 --size=ref --flagsurl ./config/flags/Intel_compiler_flags.2023-01-19.xml --output_format=all tiny # output_root was not used for this run ############################################################################ # Invocation command line: # runhpc --reportable --define EXPID=submission/8490H/tiny.omp.1.8.30 --define model=omp --iterations=2 -c xfusion.omp.cfg --tune=base,peak --define HOSTFILE=node1 --define RANKS=8 --define PPN=8 --define THREADS=30 --size=ref --flagsurl ./config/flags/Intel_compiler_flags.2023-01-19.xml --output_format=all tiny # output_root was not used for this run ############################################################################ #!/bin/sh ###################################################################### # Configuration file for the Intel Compilers # # Defines: "acctype" => "mpi", "omp", default "mpi" # "label" => ext base label, default "hpc" # # ####################################################################### #strict_rundir_verify=0 expid= %ifdef %{EXPID} expid=%{EXPID} %endif basepeak=0 allow_label_override=yes %ifndef %{label} # IF label is not set use hpc % define label xfusion %endif %ifndef %{model} # IF model is not set use mpi % define model mpi %endif teeout = yes makeflags=-j # License and tester information license_num = 6488 showtimer = 0 test_sponsor = xFusion tester = xFusion #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 = xFusion FusionServer 2288H V7 (Intel Xeon Platinum 8490H) node_compute_sw_accel_driver = N/A node_compute_hw_accel_vendor = N/A node_compute_hw_accel_type = N/A node_compute_hw_accel_model = N/A node_compute_hw_accel_ecc = N/A node_compute_hw_accel_desc = N/A node_compute_hw_accel_count = N/A node_compute_hw_accel_connect = N/A hw_avail = Jan-2023 sw_avail = Nov-2022 prepared_by = xFusion # Computation node info # [Node_Description: Hardware] node_compute_syslbl = xFusion FusionServer 2288H V7 node_compute_order = 1 node_compute_count = 4 node_compute_purpose = Compute Node node_compute_hw_vendor = xFusion node_compute_hw_model = xFusion FusionServer 2288H V7 node_compute_hw_cpu_name = Intel Xeon Platinum 8490H node_compute_hw_ncpuorder = 2 chips node_compute_hw_nchips = 2 node_compute_hw_ncores = 120 node_compute_hw_ncoresperchip = 60 node_compute_hw_nthreadspercore = 2 node_compute_hw_cpu_char = Turbo Boost Technology up to 3.5 GHz node_compute_hw_cpu_mhz = 1900 node_compute_hw_pcache = 32 KB I + 48 KB D on chip per core node_compute_hw_scache = 2 MB I+D on chip per core node_compute_hw_tcache = 112.5 MB I+D on chip per chip node_compute_hw_ocache = None node_compute_hw_memory = 512 GB (16 x 32 GB 2Rx4 PC5-4800B-R) node_compute_hw_disk = 1 x 7.68 TB NVMe SSD node_compute_hw_other = None #[Node_Description: Accelerator] #[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 = N/A #[Fileserver] #[Interconnect] interconnect_fs_syslbl = Mellanox HDR interconnect_fs_order = 0 interconnect_fs_purpose = MPI 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 = 27.2010.1202 ####################################################################### # 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 '/home/hpc2021/config/Example_SUT.inc' sw_compiler000 = C/C++/Fortran: Version 2022.2.1 of sw_compiler001 = Intel oneAPI Compiler sw_mpi_library000 = Intel MPI Library for Linux* OS, Version 2022.2.1 sw_mpi_library001 = Build 20221020 sw_mpi_other = N/A sw_other = N/A label = %{label}_%{model} tune = base output_format = text use_submit_for_speed = 1 # Compiler Settings default: AR = ar ARFLAGS = cr CXX = mpiicpc -cxx=icpx CC = mpiicc -cc=icx FC = mpiifort -fc=ifx system_class = Homogenous Cluster # 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=default: OPTIMIZE = -Ofast -ipo ${vec} COPTIMIZE = -ansi-alias CXXOPTIMIZE = -ansi-alias FOPTIMIZE = -nostandard-realloc-lhs -align array64byte PORTABILITY = -lstdc++ %if %{model} eq 'mpi' submit = mpiexec.hydra -bootstrap ssh --bind-to core -hostfile $[top]/$hostfile -np $ranks -ppn $ppn $command %else %{model} eq 'omp' submit = mpiexec.hydra -bootstrap ssh --bind-to core -hostfile $[top]/$hostfile -np $ranks -ppn $ppn -genv OMP_NUM_THREADS=$threads -genv OMP_PROC_BIND=true $command %endif default=base=default: ranks = %{RANKS} threads = %{THREADS} ppn = %{PPN} hostfile = %{HOSTFILE} %if %{model} eq 'mpi' pmodel=MPI %endif # OpenACC flags %if %{model} eq 'acc' pmodel=ACC OPTIMIZE += -fopenacc -foffload=-lm %endif # OpenMP (CPU) flags %if %{model} eq 'omp' pmodel=OMP OPTIMIZE += -fiopenmp %endif 513.soma_t=base,peak: PORTABILITY+=-DSPEC_NO_VAR_ARRAY_REDUCE 528.pot3d_t=base,peak: OPTIMIZE+=-Wno-incompatible-function-pointer-types # omp peak 505.lbm_t=peak: ranks=240 threads=4 vec=${vec_avx512_high} pmodel=OMP 513.soma_t=peak: ranks=8 threads=120 vec=${vec_avx512_high} pmodel=OMP 518.tealeaf_t=peak: basepeak=1 519.clvleaf_t=peak: ranks=240 threads=4 vec=${vec_avx512_streaming_stores} pmodel=OMP 521.miniswp_t=peak: ranks=16 threads=60 vec=${vec_avx512_high} pmodel=OMP 528.pot3d_t=peak: ranks=480 threads=2 vec=${vec_avx512_high} pmodel=OMP 532.sph_exa_t=peak: ranks=120 threads=8 vec=${vec_avx512_high_exp1} pmodel=OMP 534.hpgmgfv_t=peak: basepeak=1 535.weather_t=peak: ranks=160 threads=6 vec=${vec_avx512_high} pmodel=OMP # 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_compiler_flags.2023-02-09.xml 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 --bind-to core -hostfile $[top]/$hostfile -np $ranks -ppn $ppn -genv OMP_NUM_THREADS=$threads $command