=== This is the CP2K Performance-Test === Already up to date. Current branch master is up to date. Already up to date. Current branch master is up to date. GIT Revision: cea20fb160686b319917257b3a6652216704cc48 ################# ARCHITECTURE FILE ################## #!/bin/bash # # CP2K arch file for Cray-XC50 (Piz Daint, CSCS, GPU partition) # # Tested with: GNU 9.3.0, Cray-MPICH 7.7.18, Cray-libsci 20.09.1, Cray-FFTW 3.3.8.10, # COSMA 2.6.2, ELPA 2021.11.002, LIBINT 2.6.0, LIBPEXSI 1.2.0, # LIBXC 5.2.3, LIBVORI 220621, LIBXSMM 1.17, PLUMED 2.8.0, # SIRIUS 7.3.2, SPGLIB 1.16.2 # # Usage: Source this arch file and then run make as instructed. # A full toolchain installation is performed as default. # Replace or adapt the "module add" commands below if needed. # # Author: Matthias Krack (15.09.2022) # # \ if [ "${0}" = "${BASH_SOURCE}" ]; then \ echo "ERROR: Script ${0##*/} must be sourced"; \ echo "Usage: source ${0##*/}"; \ exit 1; \ fi; \ this_file=${BASH_SOURCE##*/}; \ if [ -n "${1}" ]; then \ gcc_version="${1}"; \ else \ gcc_version="9.3.0"; \ fi; \ module add daint-gpu; \ module rm PrgEnv-cray; \ module add PrgEnv-gnu; \ module rm gcc; \ module add gcc/${gcc_version}; \ module add cray-fftw/3.3.8.10; \ module add cudatoolkit; \ echo "Expected setup:"; \ echo " cray-mpich/7.7.18"; \ echo " craype-haswell"; \ echo " daint-gpu/21.09"; \ echo " craype/2.7.10"; \ echo " cray-libsci/20.09.1"; \ echo " PrgEnv-gnu/6.0.10"; \ echo " gcc/${gcc_version}"; \ echo " cray-fftw/3.3.8.10"; \ echo " cudatoolkit/11.0.2_3.38-8.1__g5b73779"; \ module list; \ module -f save cp2k_gpu_gnu_psmp; \ echo "To load the required modules in your batch job script, use:"; \ echo " module restore cp2k_gpu_gnu_psmp"; \ cd tools/toolchain; \ ./install_cp2k_toolchain.sh --enable-cuda=yes --gpu-ver=P100 --no-arch-files --with-gcc=system --with-libvdwxc --with-pexsi --with-plumed; \ cd ../..; \ printf "Sourcing ${PWD}/tools/toolchain/install/setup ... "; \ source ${PWD}/tools/toolchain/install/setup; \ printf "done\n"; \ echo "Check the output above for error messages and consistency!"; \ echo "If everything is OK, you can build a CP2K production binary with"; \ echo " make -j ARCH=${this_file%.*} VERSION=${this_file##*.}"; \ echo "Alternatively, you can add further checks, e.g. for regression testing, with"; \ echo " make -j ARCH=${this_file%.*} VERSION=${this_file##*.} DO_CHECKS=yes"; \ echo "or build CP2K as a library with"; \ echo " make -j ARCH=${this_file%.*} VERSION=${this_file##*.} libcp2k"; \ return # Set options DO_CHECKS := no USE_ACC := yes USE_COSMA := 2.6.2 USE_ELPA := 2021.11.002 USE_LIBINT := 2.6.0 USE_LIBPEXSI := 1.2.0 USE_LIBVORI := 220621 USE_LIBXC := 5.2.3 USE_LIBXSMM := 1.17 USE_PLUMED := 2.8.0 #USE_QUIP := b4336484fb65b0e73211a8f920ae4361c7c353fd USE_SIRIUS := 7.3.2 USE_SPGLIB := 1.16.2 # Only needed for SIRIUS LIBVDWXC_VER := 0.4.0 SPFFT_VER := 1.0.6 SPLA_VER := 1.5.4 HDF5_VER := 1.12.0 # Only needed for LIBPEXSI SCOTCH_VER := 6.0.0 SUPERLU_VER := 6.1.0 LMAX := 5 MAX_CONTR := 4 GPUVER := P100 OFFLOAD_TARGET := cuda CC := cc CXX := CC OFFLOAD_CC := nvcc FC := ftn LD := ftn AR := ar -r # cc, CC, and ftn include already the proper -march flag CFLAGS := -O2 -fopenmp -fopenmp-simd -ftree-vectorize -funroll-loops -g DFLAGS := -D__parallel DFLAGS += -D__SCALAPACK DFLAGS += -D__FFTW3 DFLAGS += -D__MPI_VERSION=3 DFLAGS += -D__MAX_CONTR=$(strip $(MAX_CONTR)) INSTALL_PATH := $(PWD)/tools/toolchain/install ifeq ($(DO_CHECKS), yes) DFLAGS += -D__CHECK_DIAG endif ifeq ($(USE_ACC), yes) DFLAGS += -D__DBCSR_ACC DFLAGS += -D__OFFLOAD_CUDA # Possibly no performance gain with PW_CUDA currently DFLAGS += -D__NO_OFFLOAD_PW endif ifneq ($(USE_PLUMED),) USE_PLUMED := $(strip $(USE_PLUMED)) PLUMED_LIB := $(INSTALL_PATH)/plumed-$(USE_PLUMED)/lib DFLAGS += -D__PLUMED2 USE_GSL := 2.7 LIBS += $(PLUMED_LIB)/libplumed.a endif ifneq ($(USE_ELPA),) USE_ELPA := $(strip $(USE_ELPA)) TARGET := nvidia ELPA_INC := $(INSTALL_PATH)/elpa-$(USE_ELPA)/$(TARGET)/include/elpa-$(USE_ELPA) ELPA_LIB := $(INSTALL_PATH)/elpa-$(USE_ELPA)/$(TARGET)/lib CFLAGS += -I$(ELPA_INC)/elpa -I$(ELPA_INC)/modules DFLAGS += -D__ELPA ifeq ($(TARGET), nvidia) DFLAGS += -D__ELPA_NVIDIA_GPU endif LIBS += $(ELPA_LIB)/libelpa.a endif ifneq ($(USE_QUIP),) USE_QUIP := $(strip $(USE_QUIP)) QUIP_INC := $(INSTALL_PATH)/quip-$(USE_QUIP)/include QUIP_LIB := $(INSTALL_PATH)/quip-$(USE_QUIP)/lib CFLAGS += -I$(QUIP_INC) DFLAGS += -D__QUIP LIBS += $(QUIP_LIB)/libquip_core.a LIBS += $(QUIP_LIB)/libatoms.a LIBS += $(QUIP_LIB)/libFoX_sax.a LIBS += $(QUIP_LIB)/libFoX_common.a LIBS += $(QUIP_LIB)/libFoX_utils.a LIBS += $(QUIP_LIB)/libFoX_fsys.a endif ifneq ($(USE_LIBPEXSI),) USE_LIBPEXSI := $(strip $(USE_LIBPEXSI)) SCOTCH_VER := $(strip $(SCOTCH_VER)) SUPERLU_VER := $(strip $(SUPERLU_VER)) LIBPEXSI_INC := $(INSTALL_PATH)/pexsi-$(USE_LIBPEXSI)/include LIBPEXSI_LIB := $(INSTALL_PATH)/pexsi-$(USE_LIBPEXSI)/lib SCOTCH_INC := $(INSTALL_PATH)/scotch-$(SCOTCH_VER)/include SCOTCH_LIB := $(INSTALL_PATH)/scotch-$(SCOTCH_VER)/lib SUPERLU_INC := $(INSTALL_PATH)/superlu_dist-$(SUPERLU_VER)/include SUPERLU_LIB := $(INSTALL_PATH)/superlu_dist-$(SUPERLU_VER)/lib CFLAGS += -I$(LIBPEXSI_INC) -I$(SCOTCH_INC) -I$(SUPERLU_INC) DFLAGS += -D__LIBPEXSI LIBS += $(LIBPEXSI_LIB)/libpexsi.a LIBS += $(SUPERLU_LIB)/libsuperlu_dist.a LIBS += $(SCOTCH_LIB)/libptscotchparmetis.a LIBS += $(SCOTCH_LIB)/libptscotch.a LIBS += $(SCOTCH_LIB)/libptscotcherr.a LIBS += $(SCOTCH_LIB)/libscotchmetis.a LIBS += $(SCOTCH_LIB)/libscotch.a endif ifneq ($(USE_LIBVORI),) USE_LIBVORI := $(strip $(USE_LIBVORI)) LIBVORI_LIB := $(INSTALL_PATH)/libvori-$(USE_LIBVORI)/lib DFLAGS += -D__LIBVORI LIBS += $(LIBVORI_LIB)/libvori.a endif ifneq ($(USE_LIBXC),) USE_LIBXC := $(strip $(USE_LIBXC)) LIBXC_INC := $(INSTALL_PATH)/libxc-$(USE_LIBXC)/include LIBXC_LIB := $(INSTALL_PATH)/libxc-$(USE_LIBXC)/lib CFLAGS += -I$(LIBXC_INC) DFLAGS += -D__LIBXC LIBS += $(LIBXC_LIB)/libxcf03.a LIBS += $(LIBXC_LIB)/libxc.a endif ifneq ($(USE_LIBINT),) USE_LIBINT := $(strip $(USE_LIBINT)) LMAX := $(strip $(LMAX)) LIBINT_INC := $(INSTALL_PATH)/libint-v$(USE_LIBINT)-cp2k-lmax-$(LMAX)/include LIBINT_LIB := $(INSTALL_PATH)/libint-v$(USE_LIBINT)-cp2k-lmax-$(LMAX)/lib CFLAGS += -I$(LIBINT_INC) DFLAGS += -D__LIBINT LIBS += $(LIBINT_LIB)/libint2.a endif ifneq ($(USE_SPGLIB),) USE_SPGLIB := $(strip $(USE_SPGLIB)) SPGLIB_INC := $(INSTALL_PATH)/spglib-$(USE_SPGLIB)/include SPGLIB_LIB := $(INSTALL_PATH)/spglib-$(USE_SPGLIB)/lib CFLAGS += -I$(SPGLIB_INC) DFLAGS += -D__SPGLIB LIBS += $(SPGLIB_LIB)/libsymspg.a endif ifneq ($(USE_LIBXSMM),) USE_LIBXSMM := $(strip $(USE_LIBXSMM)) LIBXSMM_INC := $(INSTALL_PATH)/libxsmm-$(USE_LIBXSMM)/include LIBXSMM_LIB := $(INSTALL_PATH)/libxsmm-$(USE_LIBXSMM)/lib CFLAGS += -I$(LIBXSMM_INC) DFLAGS += -D__LIBXSMM LIBS += $(LIBXSMM_LIB)/libxsmmf.a LIBS += $(LIBXSMM_LIB)/libxsmm.a endif ifneq ($(USE_SIRIUS),) USE_SIRIUS := $(strip $(USE_SIRIUS)) HDF5_VER := $(strip $(HDF5_VER)) HDF5_LIB := $(INSTALL_PATH)/hdf5-$(HDF5_VER)/lib LIBVDWXC_VER := $(strip $(LIBVDWXC_VER)) LIBVDWXC_INC := $(INSTALL_PATH)/libvdwxc-$(LIBVDWXC_VER)/include LIBVDWXC_LIB := $(INSTALL_PATH)/libvdwxc-$(LIBVDWXC_VER)/lib SPFFT_VER := $(strip $(SPFFT_VER)) SPFFT_INC := $(INSTALL_PATH)/SpFFT-$(SPFFT_VER)/include SPLA_VER := $(strip $(SPLA_VER)) SPLA_INC := $(INSTALL_PATH)/SpLA-$(SPLA_VER)/include/spla ifeq ($(USE_ACC), yes) DFLAGS += -D__OFFLOAD_GEMM SPFFT_LIB := $(INSTALL_PATH)/SpFFT-$(SPFFT_VER)/lib/cuda SPLA_LIB := $(INSTALL_PATH)/SpLA-$(SPLA_VER)/lib/cuda SIRIUS_INC := $(INSTALL_PATH)/sirius-$(USE_SIRIUS)/include/cuda SIRIUS_LIB := $(INSTALL_PATH)/sirius-$(USE_SIRIUS)/lib/cuda else SPFFT_LIB := $(INSTALL_PATH)/SpFFT-$(SPFFT_VER)/lib SPLA_LIB := $(INSTALL_PATH)/SpLA-$(SPLA_VER)/lib SIRIUS_INC := $(INSTALL_PATH)/sirius-$(USE_SIRIUS)/include SIRIUS_LIB := $(INSTALL_PATH)/sirius-$(USE_SIRIUS)/lib endif CFLAGS += -I$(LIBVDWXC_INC) CFLAGS += -I$(SPFFT_INC) CFLAGS += -I$(SPLA_INC) CFLAGS += -I$(SIRIUS_INC) DFLAGS += -D__HDF5 DFLAGS += -D__LIBVDWXC DFLAGS += -D__SPFFT DFLAGS += -D__SPLA DFLAGS += -D__SIRIUS LIBS += $(SIRIUS_LIB)/libsirius.a LIBS += $(SPLA_LIB)/libspla.a LIBS += $(SPFFT_LIB)/libspfft.a LIBS += $(LIBVDWXC_LIB)/libvdwxc.a LIBS += $(HDF5_LIB)/libhdf5.a endif ifneq ($(USE_COSMA),) USE_COSMA := $(strip $(USE_COSMA)) ifeq ($(USE_ACC), yes) USE_COSMA := $(USE_COSMA)-cuda endif COSMA_INC := $(INSTALL_PATH)/COSMA-$(USE_COSMA)/include COSMA_LIB := $(INSTALL_PATH)/COSMA-$(USE_COSMA)/lib CFLAGS += -I$(COSMA_INC) DFLAGS += -D__COSMA LIBS += $(COSMA_LIB)/libcosma_prefixed_pxgemm.a LIBS += $(COSMA_LIB)/libcosma.a LIBS += $(COSMA_LIB)/libcosta_prefixed_scalapack.a LIBS += $(COSMA_LIB)/libcosta.a LIBS += $(COSMA_LIB)/libTiled-MM.a endif ifneq ($(USE_GSL),) USE_GSL := $(strip $(USE_GSL)) GSL_INC := $(INSTALL_PATH)/gsl-$(USE_GSL)/include GSL_LIB := $(INSTALL_PATH)/gsl-$(USE_GSL)/lib CFLAGS += -I$(GSL_INC) DFLAGS += -D__GSL LIBS += $(GSL_LIB)/libgsl.a endif CFLAGS += $(DFLAGS) CXXFLAGS := $(CFLAGS) -std=c++11 OFFLOAD_FLAGS := $(DFLAGS) -O3 -Xcompiler="-fopenmp" -arch sm_60 --std=c++11 FCFLAGS := $(CFLAGS) ifeq ($(shell [ $(shell gcc -dumpversion | cut -d. -f1) -gt 9 ] && echo yes), yes) FCFLAGS += -fallow-argument-mismatch endif FCFLAGS += -fbacktrace FCFLAGS += -ffree-form FCFLAGS += -ffree-line-length-none FCFLAGS += -fno-omit-frame-pointer FCFLAGS += -std=f2008 ifneq ($(CUDA_HOME),) CUDA_LIB := $(CUDA_HOME)/lib64 LDFLAGS := $(FCFLAGS) -L$(CUDA_LIB) -Wl,-rpath=$(CUDA_LIB) else LDFLAGS := $(FCFLAGS) endif LIBS += -lcusolver -lcudart -lnvrtc -lcuda -lcufft -lcublas -lrt LIBS += -lz -ldl -lpthread -lstdc++ # End ############### END ARCHITECTURE FILE ################ ===== TESTS (description) ===== ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-32 RI-RPA/RI-MP2 correlation energy input file: benchmarks/QS_mp2_rpa/32-H2O/RI-RPA.inp required files: ['benchmarks/QS_mp2_rpa/32-H2O/BASIS_H2O', 'benchmarks/QS_mp2_rpa/32-H2O/H2O-32.xyz', 'benchmarks/QS_mp2_rpa/32-H2O/H2O-32-PBE-TZ.inp', 'benchmarks/QS_mp2_rpa/32-H2O/H2O-32-RI-dRPA-TZ.inp'] output file: result.log # nodes = 8 # ranks/node = 2 # threads/rank = 6 nrepeat = 1 time[min] = 15 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/01 job id: 42042251 --- Point --- name: 10 plot: h2o_32_ri_rpa_mp2 regex: Total RI-RPA Time= label: RI-RPA (8n/2r/6t) --- Point --- name: 11 plot: h2o_32_ri_rpa_mp2_mem regex: Estimated peak process memory label: RI-RPA (8n/2r/6t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-32 RI-RPA/RI-MP2 correlation energy input file: benchmarks/QS_mp2_rpa/32-H2O/RI-MP2.inp required files: ['benchmarks/QS_mp2_rpa/32-H2O/BASIS_H2O', 'benchmarks/QS_mp2_rpa/32-H2O/H2O-32.xyz', 'benchmarks/QS_mp2_rpa/32-H2O/H2O-32-PBE-TZ.inp', 'benchmarks/QS_mp2_rpa/32-H2O/H2O-32-HF-TZ.inp', 'benchmarks/QS_mp2_rpa/32-H2O/H2O-32-RI-MP2-TZ.inp'] output file: result.log # nodes = 8 # ranks/node = 6 # threads/rank = 2 nrepeat = 1 time[min] = 15 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/02 job id: 42042252 --- Point --- name: 20 plot: h2o_32_ri_rpa_mp2 regex: Total MP2 Time= label: RI-MP2 (8n/6r/2t) --- Point --- name: 21 plot: h2o_32_ri_rpa_mp2_mem regex: Estimated peak process memory label: RI-MP2 (8n/6r/2t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-64 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-64.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 12 # threads/rank = 1 nrepeat = 1 time[min] = 5 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/03 job id: 42042253 --- Point --- name: 100 plot: h2o_64_md regex: CP2K label: (8n/12r/1t) --- Point --- name: 101 plot: h2o_64_md_mem regex: Estimated peak process memory label: (8n/12r/1t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-64 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-64.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 6 # threads/rank = 2 nrepeat = 1 time[min] = 5 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/04 job id: 42042254 --- Point --- name: 102 plot: h2o_64_md regex: CP2K label: (8n/6r/2t) --- Point --- name: 103 plot: h2o_64_md_mem regex: Estimated peak process memory label: (8n/6r/2t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-64 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-64.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 4 # threads/rank = 3 nrepeat = 1 time[min] = 5 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/05 job id: 42042255 --- Point --- name: 104 plot: h2o_64_md regex: CP2K label: (8n/4r/3t) --- Point --- name: 105 plot: h2o_64_md_mem regex: Estimated peak process memory label: (8n/4r/3t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-64 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-64.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 3 # threads/rank = 4 nrepeat = 1 time[min] = 5 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/06 job id: 42042257 --- Point --- name: 106 plot: h2o_64_md regex: CP2K label: (8n/3r/4t) --- Point --- name: 107 plot: h2o_64_md_mem regex: Estimated peak process memory label: (8n/3r/4t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-64 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-64.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 2 # threads/rank = 6 nrepeat = 1 time[min] = 5 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/07 job id: 42042258 --- Point --- name: 108 plot: h2o_64_md regex: CP2K label: (8n/2r/6t) --- Point --- name: 109 plot: h2o_64_md_mem regex: Estimated peak process memory label: (8n/2r/6t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-64 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-64.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 1 # threads/rank = 12 nrepeat = 1 time[min] = 5 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/08 job id: 42042259 --- Point --- name: 110 plot: h2o_64_md regex: CP2K label: (8n/1r/12t) --- Point --- name: 111 plot: h2o_64_md_mem regex: Estimated peak process memory label: (8n/1r/12t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-128 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-128.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 12 # threads/rank = 1 nrepeat = 1 time[min] = 10 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/09 job id: 42042260 --- Point --- name: 200 plot: h2o_128_md regex: CP2K label: (8n/12r/1t) --- Point --- name: 201 plot: h2o_128_md_mem regex: Estimated peak process memory label: (8n/12r/1t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-128 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-128.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 6 # threads/rank = 2 nrepeat = 1 time[min] = 10 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/10 job id: 42042261 --- Point --- name: 202 plot: h2o_128_md regex: CP2K label: (8n/6r/2t) --- Point --- name: 203 plot: h2o_128_md_mem regex: Estimated peak process memory label: (8n/6r/2t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-128 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-128.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 4 # threads/rank = 3 nrepeat = 1 time[min] = 10 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/11 job id: 42042263 --- Point --- name: 204 plot: h2o_128_md regex: CP2K label: (8n/4r/3t) --- Point --- name: 205 plot: h2o_128_md_mem regex: Estimated peak process memory label: (8n/4r/3t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-128 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-128.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 3 # threads/rank = 4 nrepeat = 1 time[min] = 10 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/12 job id: 42042265 --- Point --- name: 206 plot: h2o_128_md regex: CP2K label: (8n/3r/4t) --- Point --- name: 207 plot: h2o_128_md_mem regex: Estimated peak process memory label: (8n/3r/4t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-128 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-128.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 2 # threads/rank = 6 nrepeat = 1 time[min] = 10 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/13 job id: 42042268 --- Point --- name: 208 plot: h2o_128_md regex: CP2K label: (8n/2r/6t) --- Point --- name: 209 plot: h2o_128_md_mem regex: Estimated peak process memory label: (8n/2r/6t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-128 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-128.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 1 # threads/rank = 12 nrepeat = 1 time[min] = 10 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/14 job id: 42042270 --- Point --- name: 210 plot: h2o_128_md regex: CP2K label: (8n/1r/12t) --- Point --- name: 211 plot: h2o_128_md_mem regex: Estimated peak process memory label: (8n/1r/12t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-256 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-256.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 12 # threads/rank = 1 nrepeat = 1 time[min] = 30 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/15 job id: 42042274 --- Point --- name: 400 plot: h2o_256_md regex: CP2K label: (8n/12r/1t) --- Point --- name: 401 plot: h2o_256_md_mem regex: Estimated peak process memory label: (8n/12r/1t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-256 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-256.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 6 # threads/rank = 2 nrepeat = 1 time[min] = 30 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/16 job id: 42042276 --- Point --- name: 402 plot: h2o_256_md regex: CP2K label: (8n/6r/2t) --- Point --- name: 403 plot: h2o_256_md_mem regex: Estimated peak process memory label: (8n/6r/2t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-256 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-256.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 4 # threads/rank = 3 nrepeat = 1 time[min] = 30 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/17 job id: 42042278 --- Point --- name: 404 plot: h2o_256_md regex: CP2K label: (8n/4r/3t) --- Point --- name: 405 plot: h2o_256_md_mem regex: Estimated peak process memory label: (8n/4r/3t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-256 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-256.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 3 # threads/rank = 4 nrepeat = 1 time[min] = 30 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/18 job id: 42042279 --- Point --- name: 406 plot: h2o_256_md regex: CP2K label: (8n/3r/4t) --- Point --- name: 407 plot: h2o_256_md_mem regex: Estimated peak process memory label: (8n/3r/4t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-256 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-256.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 2 # threads/rank = 6 nrepeat = 1 time[min] = 30 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/19 job id: 42042280 --- Point --- name: 408 plot: h2o_256_md regex: CP2K label: (8n/2r/6t) --- Point --- name: 409 plot: h2o_256_md_mem regex: Estimated peak process memory label: (8n/2r/6t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-256 test - DBCSR dominated (MPI/OMP) input file: benchmarks/QS/H2O-256.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 1 # threads/rank = 12 nrepeat = 1 time[min] = 30 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/20 job id: 42042281 --- Point --- name: 410 plot: h2o_256_md regex: CP2K label: (8n/1r/12t) --- Point --- name: 411 plot: h2o_256_md_mem regex: Estimated peak process memory label: (8n/1r/12t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-32 (NREP 3) linear scaling test (864 H2O) input file: benchmarks/QS_DM_LS/H2O-dft-ls.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 12 # threads/rank = 1 nrepeat = 1 time[min] = 10 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/21 job id: 42042282 --- Point --- name: 500 plot: h2o_32_nrep3_ls regex: CP2K label: (8n/12r/1t) --- Point --- name: 501 plot: h2o_32_nrep3_ls_mem regex: Estimated peak process memory label: (8n/12r/1t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-32 (NREP 3) linear scaling test (864 H2O) input file: benchmarks/QS_DM_LS/H2O-dft-ls.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 6 # threads/rank = 2 nrepeat = 1 time[min] = 10 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/22 job id: 42042283 --- Point --- name: 502 plot: h2o_32_nrep3_ls regex: CP2K label: (8n/6r/2t) --- Point --- name: 503 plot: h2o_32_nrep3_ls_mem regex: Estimated peak process memory label: (8n/6r/2t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-32 (NREP 3) linear scaling test (864 H2O) input file: benchmarks/QS_DM_LS/H2O-dft-ls.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 4 # threads/rank = 3 nrepeat = 1 time[min] = 10 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/23 job id: 42042284 --- Point --- name: 504 plot: h2o_32_nrep3_ls regex: CP2K label: (8n/4r/3t) --- Point --- name: 505 plot: h2o_32_nrep3_ls_mem regex: Estimated peak process memory label: (8n/4r/3t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-32 (NREP 3) linear scaling test (864 H2O) input file: benchmarks/QS_DM_LS/H2O-dft-ls.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 3 # threads/rank = 4 nrepeat = 1 time[min] = 10 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/24 job id: 42042285 --- Point --- name: 506 plot: h2o_32_nrep3_ls regex: CP2K label: (8n/3r/4t) --- Point --- name: 507 plot: h2o_32_nrep3_ls_mem regex: Estimated peak process memory label: (8n/3r/4t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-32 (NREP 3) linear scaling test (864 H2O) input file: benchmarks/QS_DM_LS/H2O-dft-ls.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 2 # threads/rank = 6 nrepeat = 1 time[min] = 10 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/25 job id: 42042286 --- Point --- name: 508 plot: h2o_32_nrep3_ls regex: CP2K label: (8n/2r/6t) --- Point --- name: 509 plot: h2o_32_nrep3_ls_mem regex: Estimated peak process memory label: (8n/2r/6t) ~~~~~~~ END TEST ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: H2O-32 (NREP 3) linear scaling test (864 H2O) input file: benchmarks/QS_DM_LS/H2O-dft-ls.inp required files: [] output file: result.log # nodes = 8 # ranks/node = 1 # threads/rank = 12 nrepeat = 1 time[min] = 10 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/26 job id: 42042287 --- Point --- name: 510 plot: h2o_32_nrep3_ls regex: CP2K label: (8n/1r/12t) --- Point --- name: 511 plot: h2o_32_nrep3_ls_mem regex: Estimated peak process memory label: (8n/1r/12t) ~~~~~~~ END TEST ~~~~~~~ === END TESTS (description) === ===== PLOTS (description) ===== ~~~~~~~~~ PLOT ~~~~~~~~~ Plot: name="h2o_32_ri_rpa_mp2", title="32 H2O molecules (RI-MP2, RI-RPA)", xlabel="Revision", ylabel="Time [s]" ~~~~~~~~~ PLOT ~~~~~~~~~ Plot: name="h2o_32_ri_rpa_mp2_mem", title="32 H2O molecules (RI-MP2, RI-RPA)", xlabel="Revision", ylabel="Est. peak process memory [MiB]" ~~~~~~~~~ PLOT ~~~~~~~~~ Plot: name="h2o_64_md", title="64 H2O molecules (10 MD steps)", xlabel="Revision", ylabel="Time [s]" ~~~~~~~~~ PLOT ~~~~~~~~~ Plot: name="h2o_64_md_mem", title="64 H2O molecules (10 MD steps)", xlabel="Revision", ylabel="Est. peak process memory [MiB]" ~~~~~~~~~ PLOT ~~~~~~~~~ Plot: name="h2o_128_md", title="128 H2O molecules (10 MD steps)", xlabel="Revision", ylabel="Time [s]" ~~~~~~~~~ PLOT ~~~~~~~~~ Plot: name="h2o_128_md_mem", title="128 H2O molecules (10 MD steps)", xlabel="Revision", ylabel="Est. peak process memory [MiB]" ~~~~~~~~~ PLOT ~~~~~~~~~ Plot: name="h2o_256_md", title="256 H2O molecules (10 MD steps)", xlabel="Revision", ylabel="Time [s]" ~~~~~~~~~ PLOT ~~~~~~~~~ Plot: name="h2o_256_md_mem", title="256 H2O molecules (10 MD steps)", xlabel="Revision", ylabel="Est. peak process memory [MiB]" ~~~~~~~~~ PLOT ~~~~~~~~~ Plot: name="h2o_32_nrep3_ls", title="864 H2O molecules (LS SCF)", xlabel="Revision", ylabel="Time [s]" ~~~~~~~~~ PLOT ~~~~~~~~~ Plot: name="h2o_32_nrep3_ls_mem", title="864 H2O molecules (LS SCF)", xlabel="Revision", ylabel="Est. peak process memory [MiB]" === END PLOTS (description) === ============ RESULTS ============ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/01/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 0.000000E+00 0.0% 0.0% 0.0% flops max/rank 0.000000E+00 0.0% 0.0% 0.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 0 0.0% 0.0% 0.0% number of processed stacks 0 0.0% 0.0% 0.0% average stack size 0.0 0.0 0.0 marketing flops 0.000000E+00 ------------------------------------------------------------------------------- - - - DBCSR MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Bcast 1 12. MP_Allreduce 19 21. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Bcast 15 177869. MP_Allreduce 344 9. MP_Sync 3 MP_comm_split 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.021 0.033 133.810 133.810 farming_run 1 2.0 133.225 133.232 133.781 133.785 ------------------------------------------------------------------------------- @@@@@@@@@@ Run number: 2 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 32 x 32 x 32 4194304 0.0% 0.0% 100.0% flops 14 x 32 x 32 154140672 0.0% 0.0% 100.0% flops 29 x 32 x 32 159645696 0.0% 0.0% 100.0% flops 14 x 14 x 32 208732160 0.0% 0.0% 100.0% flops 29 x 14 x 32 212860928 0.0% 0.0% 100.0% flops 14 x 29 x 32 212860928 0.0% 0.0% 100.0% flops 29 x 29 x 32 227352576 0.0% 0.0% 100.0% flops 14 x 32 x 14 896801644032 0.0% 0.0% 100.0% flops 29 x 32 x 14 928925089792 0.0% 0.0% 100.0% flops 14 x 32 x 29 928925089792 0.0% 0.0% 100.0% flops 29 x 32 x 29 962100985856 0.0% 0.0% 100.0% flops 32 x 32 x 14 1693169221632 0.0% 0.0% 100.0% flops 32 x 32 x 29 1753639550976 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 7.164741E+12 0.0% 0.0% 100.0% flops max/rank 447.801317E+09 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 249492158 0.0% 0.0% 100.0% number of processed stacks 164328 0.0% 0.0% 100.0% average stack size 0.0 0.0 1518.3 marketing flops 7.165779E+12 ------------------------------------------------------------------------------- # multiplications 1160 max memory usage/rank 1.454768E+09 # max total images/rank 1 # max 3D layers 1 # MPI messages exchanged 2592 MPI messages size (bytes): total size 1.140326E+09 min size 0.000000E+00 max size 1.663488E+06 average size 439.940750E+03 MPI breakdown and total messages size (bytes): size <= 128 132 0 128 < size <= 8192 348 2850816 8192 < size <= 32768 0 0 32768 < size <= 131072 1536 179306496 131072 < size <= 4194304 576 958169088 4194304 < size <= 16777216 0 0 16777216 < size 0 0 ------------------------------------------------------------------------------- - - - DBCSR MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Bcast 14 12. MP_Allreduce 2308 54. MP_Alltoall 4670 822215. MP_ISend 2604 90577. MP_IRecv 2604 90574. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 12 MP_Bcast 228 1113141. MP_Allreduce 489 2263609. MP_Sync 27 MP_Alltoall 38 9316958. MP_SendRecv 30 829726. MP_ISendRecv 135 235435. MP_Wait 281 MP_comm_split 8 MP_ISend 127 3867574. MP_IRecv 127 3866554. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.019 0.035 115.494 115.494 qs_energies 1 2.0 0.000 0.000 115.246 115.249 mp2_main 1 3.0 0.000 0.000 113.240 113.243 mp2_gpw_main 1 4.0 0.020 0.027 112.276 112.279 mp2_ri_gpw_compute_in 1 5.0 0.172 0.186 93.448 93.637 mp2_ri_gpw_compute_in_loop 1 6.0 0.005 0.006 55.307 55.495 mp2_eri_3c_integrate_gpw 272 7.0 0.153 0.168 41.619 46.652 get_2c_integrals 1 6.0 0.000 0.000 36.962 37.969 integrate_v_rspace 273 8.0 0.436 0.450 25.015 29.690 pw_transfer 6555 10.6 0.380 0.396 27.271 27.685 fft_wrap_pw1pw2 5465 11.4 0.046 0.049 25.919 26.337 grid_integrate_task_list 273 9.0 20.847 25.952 20.847 25.952 fft_wrap_pw1pw2_100 2178 12.4 1.185 1.319 23.478 23.933 compute_2c_integrals 1 7.0 0.002 0.003 19.548 19.549 compute_2c_integrals_loop_lm 1 8.0 0.004 0.004 18.866 19.212 mp2_eri_2c_integrate_gpw 1 9.0 2.391 2.435 18.862 19.209 rpa_ri_compute_en 1 5.0 0.000 0.004 18.715 18.858 cp_fm_cholesky_decompose 12 8.2 17.404 18.467 17.404 18.467 cholesky_decomp 1 7.0 0.000 0.000 16.254 17.294 fft3d_s 5443 13.4 16.068 16.430 16.090 16.450 ao_to_mo_and_store_B_mult_1 272 7.0 10.852 15.567 10.852 15.567 calculate_wavefunction 272 8.0 5.423 5.495 12.523 13.170 rpa_num_int 1 6.0 0.000 0.001 10.614 10.615 rpa_num_int_RPA_matrix_operati 8 7.0 0.000 0.000 10.564 10.583 calc_mat_Q 8 8.0 0.000 0.000 9.367 9.445 contract_S_to_Q 8 9.0 0.000 0.000 8.790 8.868 calc_potential_gpw 544 9.5 0.006 0.007 8.248 8.670 mp2_eri_2c_integrate_gpw_pot_l 272 10.0 0.002 0.002 8.215 8.514 parallel_gemm_fm 14 9.1 0.000 0.000 8.368 8.444 parallel_gemm_fm_cosma 14 10.1 8.368 8.444 8.368 8.444 potential_pw2rs 545 10.0 0.107 0.110 7.673 8.391 create_integ_mat 1 6.0 0.022 0.028 7.817 7.817 collocate_single_gaussian 272 10.0 0.040 0.042 7.441 7.712 array2fm 1 7.0 0.000 0.000 6.704 7.275 pw_scatter_s 2720 13.7 4.427 4.609 4.427 4.609 pw_gather_s 2722 13.2 3.870 4.212 3.870 4.212 array2fm_buffer_send 1 8.0 2.942 3.155 2.942 3.155 pw_poisson_solve 545 10.5 1.138 1.197 2.192 2.319 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="10", plot="h2o_32_ri_rpa_mp2", label="RI-RPA (8n/2r/6t)", y=112.274845, yerr=0.000000 PlotPoint: name="11", plot="h2o_32_ri_rpa_mp2_mem", label="RI-RPA (8n/2r/6t)", y=2725.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/02/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 0.000000E+00 0.0% 0.0% 0.0% flops max/rank 0.000000E+00 0.0% 0.0% 0.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 0 0.0% 0.0% 0.0% number of processed stacks 0 0.0% 0.0% 0.0% average stack size 0.0 0.0 0.0 marketing flops 0.000000E+00 ------------------------------------------------------------------------------- - - - DBCSR MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Bcast 1 12. MP_Allreduce 19 21. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Bcast 22 205321. MP_Allreduce 344 10. MP_Sync 4 MP_comm_split 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.029 0.038 396.290 396.291 farming_run 1 2.0 395.473 395.483 396.243 396.245 ------------------------------------------------------------------------------- @@@@@@@@@@ Run number: 2 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 32 x 32 x 32 16777216 0.0% 0.0% 100.0% flops 14 x 32 x 32 565182464 0.0% 0.0% 100.0% flops 29 x 32 x 32 585367552 0.0% 0.0% 100.0% flops 14 x 14 x 32 626196480 0.0% 0.0% 100.0% flops 29 x 14 x 32 638582784 0.0% 0.0% 100.0% flops 14 x 29 x 32 638582784 0.0% 0.0% 100.0% flops 29 x 29 x 32 682057728 0.0% 0.0% 100.0% flops 14 x 32 x 14 897827141120 0.0% 0.0% 100.0% flops 29 x 32 x 14 929989394432 0.0% 0.0% 100.0% flops 14 x 32 x 29 929989394432 0.0% 0.0% 100.0% flops 29 x 32 x 29 963203301376 0.0% 0.0% 100.0% flops 32 x 32 x 14 1693481172992 0.0% 0.0% 100.0% flops 32 x 32 x 29 1753962643456 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 7.172206E+12 0.0% 0.0% 100.0% flops max/rank 150.696064E+09 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 249788822 0.0% 0.0% 100.0% number of processed stacks 98736 0.0% 0.0% 100.0% average stack size 0.0 0.0 2529.9 marketing flops 7.174951E+12 ------------------------------------------------------------------------------- # multiplications 1140 max memory usage/rank 1.214640E+09 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 61440 MPI messages size (bytes): total size 6.073508E+09 min size 0.000000E+00 max size 642.960000E+03 average size 98.852664E+03 MPI breakdown and total messages size (bytes): size <= 128 32004 0 128 < size <= 8192 1820 14909440 8192 < size <= 32768 0 0 32768 < size <= 131072 18640 1081442304 131072 < size <= 4194304 8976 4977156096 4194304 < size <= 16777216 0 0 16777216 < size 0 0 ------------------------------------------------------------------------------- - - - DBCSR MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Bcast 14 12. MP_Allreduce 1003 44. MP_Alltoall 1797 713538. MP_ISend 3686 54943. MP_IRecv 3622 54292. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 12 MP_Bcast 703 408373. MP_Allreduce 1825 23678. MP_Sync 38 MP_Alltoall 77 MP_SendRecv 2171 2843495. MP_ISendRecv 1739 144022. MP_Wait 2051 MP_comm_split 7 MP_ISend 264 362227. MP_IRecv 264 362718. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.026 0.052 209.132 209.133 qs_energies 1 2.0 0.000 0.000 208.886 208.893 scf_env_do_scf 1 3.0 0.000 0.000 106.271 106.271 qs_ks_update_qs_env 5 5.0 0.000 0.000 105.355 105.363 rebuild_ks_matrix 4 6.0 0.000 0.000 105.354 105.362 qs_ks_build_kohn_sham_matrix 4 7.0 0.056 0.064 105.354 105.362 hfx_ks_matrix 4 8.0 0.001 0.001 104.973 104.976 integrate_four_center 4 9.0 0.145 0.460 104.972 104.975 mp2_main 1 3.0 0.000 0.000 102.318 102.324 mp2_gpw_main 1 4.0 0.034 0.053 101.455 101.463 integrate_four_center_main 4 10.0 0.130 0.574 96.582 99.149 integrate_four_center_bin 267 11.0 96.452 98.902 96.452 98.902 init_scf_loop 1 4.0 0.000 0.000 92.116 92.116 mp2_ri_gpw_compute_in 1 5.0 0.066 0.078 74.798 75.891 mp2_ri_gpw_compute_in_loop 1 6.0 0.002 0.003 54.436 55.523 mp2_eri_3c_integrate_gpw 91 7.0 0.143 0.159 42.064 47.336 integrate_v_rspace 95 8.0 0.400 0.570 28.433 33.495 pw_transfer 2240 10.6 0.146 0.163 29.888 30.234 fft_wrap_pw1pw2 1868 11.4 0.018 0.020 28.910 29.251 ao_to_mo_and_store_B_mult_1 91 7.0 10.692 29.044 10.692 29.044 grid_integrate_task_list 95 9.0 23.717 29.007 23.717 29.007 mp2_ri_gpw_compute_en 1 5.0 0.055 0.062 26.484 28.168 fft_wrap_pw1pw2_100 730 12.4 1.256 1.429 26.648 27.043 mp2_ri_gpw_compute_en_RI_loop 1 6.0 1.839 1.896 24.756 24.765 get_2c_integrals 1 6.0 0.002 0.014 20.264 20.301 compute_2c_integrals 1 7.0 0.003 0.004 19.243 19.257 compute_2c_integrals_loop_lm 1 8.0 0.002 0.006 18.828 19.120 mp2_eri_2c_integrate_gpw 1 9.0 1.745 1.854 18.826 19.119 fft3d_s 1823 13.4 18.430 18.871 18.443 18.884 scf_env_do_scf_inner_loop 4 4.0 0.000 0.000 14.154 14.154 calculate_wavefunction 91 8.0 2.025 2.065 9.751 9.991 mp2_ri_gpw_compute_en_expansio 172 7.0 0.560 0.588 8.756 9.224 potential_pw2rs 186 10.0 0.033 0.035 8.615 9.068 local_gemm 172 8.0 8.196 8.665 8.196 8.665 mp2_eri_2c_integrate_gpw_pot_l 91 10.0 0.001 0.001 8.216 8.491 mp2_ri_gpw_compute_en_comm 22 7.0 0.508 0.531 7.775 8.209 calc_potential_gpw 182 9.5 0.002 0.002 7.909 8.136 collocate_single_gaussian 91 10.0 0.020 0.080 7.864 8.086 mp2_ri_gpw_compute_en_ener 172 7.0 6.341 6.439 6.341 6.439 mp_sendrecv_dm3 2068 8.0 5.787 6.274 5.787 6.274 mp_sync 38 10.4 3.052 5.851 3.052 5.851 pw_gather_s 912 13.2 4.926 5.454 4.926 5.454 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="20", plot="h2o_32_ri_rpa_mp2", label="RI-MP2 (8n/6r/2t)", y=101.449059, yerr=0.000000 PlotPoint: name="21", plot="h2o_32_ri_rpa_mp2_mem", label="RI-MP2 (8n/6r/2t)", y=1503.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/03/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 32 x 32 x 32 26877100032 0.0% 0.0% 100.0% flops 9 x 9 x 32 44168260608 0.0% 0.0% 100.0% flops 22 x 9 x 32 53835724800 0.0% 0.0% 100.0% flops 9 x 22 x 32 53885500416 0.0% 0.0% 100.0% flops 32 x 32 x 9 63568871424 0.0% 0.0% 100.0% flops 22 x 22 x 32 67007283200 0.0% 0.0% 100.0% flops 32 x 32 x 22 77695287296 0.0% 0.0% 100.0% flops 9 x 32 x 32 78422999040 0.0% 0.0% 100.0% flops 22 x 32 x 32 95850332160 0.0% 0.0% 100.0% flops 9 x 32 x 9 266263676928 0.0% 0.0% 100.0% flops 22 x 32 x 9 326697440256 0.0% 0.0% 100.0% flops 9 x 32 x 22 326697440256 0.0% 0.0% 100.0% flops 22 x 32 x 22 399918497792 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 1.880888E+12 0.0% 0.0% 100.0% flops max/rank 29.277748E+09 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 146984760 0.0% 0.0% 100.0% number of processed stacks 5055360 0.0% 0.0% 100.0% average stack size 0.0 0.0 29.1 marketing flops 2.107592E+12 ------------------------------------------------------------------------------- # multiplications 2286 max memory usage/rank 449.826816E+06 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 9436608 MPI messages size (bytes): total size 333.233553E+09 min size 0.000000E+00 max size 315.840000E+03 average size 35.312852E+03 MPI breakdown and total messages size (bytes): size <= 128 4913240 0 128 < size <= 8192 1155432 9465298944 8192 < size <= 32768 1984512 54190407680 32768 < size <= 131072 551296 42776657920 131072 < size <= 4194304 832128 226802306368 4194304 < size <= 16777216 0 0 16777216 < size 0 0 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 3683 62385. MP_Allreduce 10249 271. MP_Sync 580 MP_Alltoall 2083 MP_ISendRecv 45220 5520. MP_Wait 60486 MP_comm_split 50 MP_ISend 20771 42672. MP_IRecv 20771 42672. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.016 0.034 48.293 48.295 qs_mol_dyn_low 1 2.0 0.003 0.005 47.965 47.973 qs_forces 11 3.9 0.002 0.004 47.902 47.904 qs_energies 11 4.9 0.001 0.002 46.449 46.459 scf_env_do_scf 11 5.9 0.000 0.001 40.504 40.505 scf_env_do_scf_inner_loop 108 6.5 0.002 0.006 38.428 38.428 dbcsr_multiply_generic 2286 12.5 0.092 0.097 30.369 30.723 qs_scf_new_mos 108 7.5 0.000 0.000 28.413 28.685 qs_scf_loop_do_ot 108 8.5 0.000 0.001 28.413 28.685 ot_scf_mini 108 9.5 0.002 0.002 26.861 27.054 multiply_cannon 2286 13.5 0.179 0.187 24.472 25.768 multiply_cannon_loop 2286 14.5 1.479 1.576 23.749 25.065 velocity_verlet 10 3.0 0.001 0.001 23.237 23.237 ot_mini 108 10.5 0.001 0.001 16.749 16.988 qs_ot_get_derivative 108 11.5 0.001 0.001 14.052 14.221 mp_waitall_1 267858 16.1 7.097 12.819 7.097 12.819 multiply_cannon_metrocomm3 54864 15.5 0.068 0.072 4.910 11.264 multiply_cannon_multrec 54864 15.5 4.403 6.731 7.630 11.074 rebuild_ks_matrix 119 8.3 0.000 0.000 7.968 8.093 qs_ks_build_kohn_sham_matrix 119 9.3 0.010 0.011 7.968 8.092 qs_ks_update_qs_env 119 7.6 0.001 0.001 7.038 7.151 multiply_cannon_sync_h2d 54864 15.5 6.092 6.733 6.092 6.733 qs_ot_get_p 119 10.4 0.001 0.001 5.804 6.055 mp_sum_l 7207 12.9 3.837 5.447 3.837 5.447 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 5.021 5.112 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 4.394 4.846 sum_up_and_integrate 119 10.3 0.012 0.015 4.601 4.609 integrate_v_rspace 119 11.3 0.002 0.002 4.589 4.598 init_scf_run 11 5.9 0.000 0.001 4.588 4.588 scf_env_initial_rho_setup 11 6.9 0.000 0.002 4.587 4.588 dbcsr_mm_accdrv_process 76910 16.1 1.110 1.829 3.150 4.481 qs_rho_update_rho_low 119 7.7 0.000 0.001 4.026 4.105 calculate_rho_elec 119 8.7 0.011 0.016 4.026 4.105 qs_ot_p2m_diag 50 11.0 0.004 0.006 3.392 3.434 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 2.852 3.030 apply_single 119 13.6 0.000 0.000 2.852 3.030 rs_pw_transfer 974 11.9 0.012 0.013 2.874 2.994 calculate_dm_sparse 119 9.5 0.000 0.000 2.804 2.931 cp_dbcsr_syevd 50 12.0 0.002 0.003 2.596 2.596 jit_kernel_multiply 13 15.8 1.981 2.564 1.981 2.564 ot_diis_step 108 11.5 0.006 0.006 2.563 2.563 calculate_first_density_matrix 1 7.0 0.000 0.001 2.434 2.437 cp_fm_diag_elpa 50 13.0 0.000 0.000 2.357 2.358 density_rs2pw 119 9.7 0.004 0.004 2.244 2.347 cp_fm_redistribute_end 50 14.0 2.138 2.332 2.143 2.334 cp_fm_diag_elpa_base 50 14.0 0.189 2.280 0.190 2.292 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 2.191 2.193 acc_transpose_blocks 54864 15.5 0.217 0.244 1.684 2.165 qs_ot_get_orbitals 108 10.5 0.000 0.000 2.079 2.128 grid_integrate_task_list 119 12.3 2.018 2.117 2.018 2.117 wfi_extrapolate 11 7.9 0.001 0.001 2.096 2.097 init_scf_loop 11 6.9 0.000 0.000 2.056 2.057 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 1.924 1.988 potential_pw2rs 119 12.3 0.004 0.004 1.947 1.966 multiply_cannon_metrocomm1 54864 15.5 0.055 0.061 1.231 1.862 pw_transfer 1439 11.6 0.051 0.057 1.773 1.851 make_m2s 4572 13.5 0.053 0.056 1.759 1.809 fft_wrap_pw1pw2 1201 12.6 0.007 0.007 1.698 1.778 make_images 4572 14.5 0.133 0.138 1.677 1.727 mp_sum_d 4129 12.0 1.009 1.592 1.009 1.592 fft3d_ps 1201 14.6 0.360 0.464 1.477 1.552 mp_waitany 12084 13.8 1.285 1.538 1.285 1.538 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 1.424 1.439 mp_alltoall_d11v 2130 13.8 1.284 1.426 1.284 1.426 fft_wrap_pw1pw2_140 487 13.2 0.079 0.094 1.296 1.377 grid_collocate_task_list 119 9.7 1.281 1.345 1.281 1.345 prepare_preconditioner 11 7.9 0.000 0.000 0.996 1.027 make_preconditioner 11 8.9 0.000 0.000 0.996 1.027 acc_transpose_blocks_kernels 54864 16.5 0.240 0.385 0.732 1.000 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 0.974 0.988 make_full_inverse_cholesky 11 9.9 0.000 0.000 0.941 0.978 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="100", plot="h2o_64_md", label="(8n/12r/1t)", y=48.295000, yerr=0.000000 PlotPoint: name="101", plot="h2o_64_md_mem", label="(8n/12r/1t)", y=428.545455, yerr=0.655555 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/04/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 32 x 32 x 32 26877100032 0.0% 0.0% 100.0% flops 9 x 9 x 32 44168260608 0.0% 0.0% 100.0% flops 22 x 9 x 32 53835724800 0.0% 0.0% 100.0% flops 9 x 22 x 32 53885500416 0.0% 0.0% 100.0% flops 32 x 32 x 9 63568871424 0.0% 0.0% 100.0% flops 22 x 22 x 32 67007283200 0.0% 0.0% 100.0% flops 32 x 32 x 22 77695287296 0.0% 0.0% 100.0% flops 9 x 32 x 32 78422999040 0.0% 0.0% 100.0% flops 22 x 32 x 32 95850332160 0.0% 0.0% 100.0% flops 9 x 32 x 9 266263676928 0.0% 0.0% 100.0% flops 22 x 32 x 9 326697440256 0.0% 0.0% 100.0% flops 9 x 32 x 22 326697440256 0.0% 0.0% 100.0% flops 22 x 32 x 22 399918497792 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 1.880888E+12 0.0% 0.0% 100.0% flops max/rank 57.173320E+09 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 146984760 0.0% 0.0% 100.0% number of processed stacks 3066240 0.0% 0.0% 100.0% average stack size 0.0 0.0 47.9 marketing flops 2.107592E+12 ------------------------------------------------------------------------------- # multiplications 2286 max memory usage/rank 487.014400E+06 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 2194560 MPI messages size (bytes): total size 310.646604E+09 min size 0.000000E+00 max size 1.145520E+06 average size 141.553031E+03 MPI breakdown and total messages size (bytes): size <= 128 724648 0 128 < size <= 8192 253512 2076770304 8192 < size <= 32768 281952 4619501568 32768 < size <= 131072 494448 39143342080 131072 < size <= 4194304 440000 264807943488 4194304 < size <= 16777216 0 0 16777216 < size 0 0 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 3672 62664. MP_Allreduce 10226 305. MP_Sync 104 MP_Alltoall 2060 1146865. MP_ISendRecv 33558 37093. MP_Wait 40318 MP_comm_split 50 MP_ISend 5720 128509. MP_IRecv 5720 128509. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.012 0.027 38.792 38.793 qs_mol_dyn_low 1 2.0 0.003 0.003 38.622 38.630 qs_forces 11 3.9 0.002 0.002 38.565 38.566 qs_energies 11 4.9 0.001 0.001 36.881 36.883 scf_env_do_scf 11 5.9 0.000 0.001 31.818 31.819 scf_env_do_scf_inner_loop 108 6.5 0.002 0.007 29.249 29.249 dbcsr_multiply_generic 2286 12.5 0.101 0.104 21.524 21.897 qs_scf_new_mos 108 7.5 0.001 0.001 20.132 20.373 qs_scf_loop_do_ot 108 8.5 0.001 0.001 20.131 20.373 ot_scf_mini 108 9.5 0.002 0.003 19.214 19.388 velocity_verlet 10 3.0 0.001 0.001 18.453 18.454 multiply_cannon 2286 13.5 0.208 0.219 16.370 17.987 multiply_cannon_loop 2286 14.5 0.907 0.976 15.212 16.742 ot_mini 108 10.5 0.001 0.001 11.957 12.196 mp_waitall_1 217478 16.2 6.220 11.398 6.220 11.398 qs_ot_get_derivative 108 11.5 0.001 0.001 9.477 9.654 multiply_cannon_metrocomm3 27432 15.5 0.068 0.071 4.197 9.555 multiply_cannon_multrec 27432 15.5 1.960 4.550 5.837 8.652 rebuild_ks_matrix 119 8.3 0.000 0.000 7.218 7.359 qs_ks_build_kohn_sham_matrix 119 9.3 0.012 0.013 7.217 7.359 qs_ks_update_qs_env 119 7.6 0.001 0.001 6.386 6.515 dbcsr_mm_accdrv_process 47894 16.0 2.937 4.953 3.808 5.625 qs_ot_get_p 119 10.4 0.001 0.001 4.329 4.567 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 3.612 4.453 sum_up_and_integrate 119 10.3 0.024 0.027 4.248 4.254 integrate_v_rspace 119 11.3 0.002 0.002 4.223 4.230 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 3.058 4.134 apply_single 119 13.6 0.000 0.000 3.057 4.134 mp_sum_l 7207 12.9 2.082 4.020 2.082 4.020 init_scf_run 11 5.9 0.000 0.001 3.835 3.836 scf_env_initial_rho_setup 11 6.9 0.001 0.001 3.835 3.835 qs_rho_update_rho_low 119 7.7 0.001 0.001 3.681 3.720 calculate_rho_elec 119 8.7 0.021 0.024 3.680 3.720 multiply_cannon_sync_h2d 27432 15.5 2.167 2.870 2.167 2.870 make_m2s 4572 13.5 0.052 0.054 2.618 2.844 qs_ot_p2m_diag 50 11.0 0.008 0.012 2.816 2.835 make_images 4572 14.5 0.200 0.238 2.529 2.753 rs_pw_transfer 974 11.9 0.010 0.011 2.624 2.701 init_scf_loop 11 6.9 0.000 0.000 2.545 2.545 ot_diis_step 108 11.5 0.011 0.011 2.426 2.427 cp_dbcsr_syevd 50 12.0 0.003 0.003 2.371 2.371 calculate_first_density_matrix 1 7.0 0.000 0.000 2.306 2.307 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 2.207 2.296 calculate_dm_sparse 119 9.5 0.000 0.000 2.057 2.131 density_rs2pw 119 9.7 0.004 0.004 2.038 2.120 cp_fm_diag_elpa 50 13.0 0.000 0.000 1.997 1.998 potential_pw2rs 119 12.3 0.006 0.006 1.972 1.986 cp_fm_redistribute_end 50 14.0 1.644 1.969 1.649 1.970 cp_fm_diag_elpa_base 50 14.0 0.312 1.900 0.320 1.931 grid_integrate_task_list 119 12.3 1.839 1.929 1.839 1.929 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 1.900 1.903 pw_transfer 1439 11.6 0.063 0.067 1.836 1.873 fft_wrap_pw1pw2 1201 12.6 0.007 0.008 1.747 1.785 jit_kernel_multiply 9 16.1 0.819 1.762 0.819 1.762 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 1.680 1.725 prepare_preconditioner 11 7.9 0.000 0.000 1.637 1.665 make_preconditioner 11 8.9 0.000 0.000 1.637 1.665 make_images_data 4572 15.5 0.045 0.051 1.179 1.625 make_full_inverse_cholesky 11 9.9 0.000 0.000 1.531 1.590 hybrid_alltoall_any 4725 16.4 0.050 0.111 1.030 1.505 acc_transpose_blocks 27432 15.5 0.108 0.111 1.198 1.494 fft3d_ps 1201 14.6 0.498 0.551 1.454 1.491 wfi_extrapolate 11 7.9 0.001 0.001 1.477 1.477 mp_alltoall_d11v 2130 13.8 1.270 1.406 1.270 1.406 fft_wrap_pw1pw2_140 487 13.2 0.076 0.084 1.331 1.368 mp_allgather_i34 2286 14.5 0.603 1.357 0.603 1.357 grid_collocate_task_list 119 9.7 1.239 1.322 1.239 1.322 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 1.252 1.260 qs_ot_get_orbitals 108 10.5 0.000 0.000 1.194 1.240 mp_sum_d 4129 12.0 0.593 1.046 0.593 1.046 qs_energies_init_hamiltonians 11 5.9 0.000 0.001 0.955 0.956 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 0.888 0.902 make_images_sizes 4572 15.5 0.005 0.005 0.626 0.885 mp_alltoall_i44 4572 16.5 0.621 0.881 0.621 0.881 acc_transpose_blocks_kernels 27432 16.5 0.182 0.269 0.653 0.861 rs_pw_transfer_PW2RS_50 119 14.3 0.589 0.607 0.764 0.820 mp_alltoall_z22v 1201 16.6 0.720 0.800 0.720 0.800 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="102", plot="h2o_64_md", label="(8n/6r/2t)", y=38.793000, yerr=0.000000 PlotPoint: name="103", plot="h2o_64_md_mem", label="(8n/6r/2t)", y=463.272727, yerr=1.762793 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/05/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 32 x 32 x 32 26877100032 0.0% 0.0% 100.0% flops 9 x 9 x 32 44168260608 0.0% 0.0% 100.0% flops 22 x 9 x 32 53835724800 0.0% 0.0% 100.0% flops 9 x 22 x 32 53885500416 0.0% 0.0% 100.0% flops 32 x 32 x 9 63568871424 0.0% 0.0% 100.0% flops 22 x 22 x 32 67007283200 0.0% 0.0% 100.0% flops 32 x 32 x 22 77695287296 0.0% 0.0% 100.0% flops 9 x 32 x 32 78422999040 0.0% 0.0% 100.0% flops 22 x 32 x 32 95850332160 0.0% 0.0% 100.0% flops 9 x 32 x 9 266263676928 0.0% 0.0% 100.0% flops 22 x 32 x 9 326697440256 0.0% 0.0% 100.0% flops 9 x 32 x 22 326697440256 0.0% 0.0% 100.0% flops 22 x 32 x 22 399918497792 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 1.880888E+12 0.0% 0.0% 100.0% flops max/rank 59.051995E+09 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 146984760 0.0% 0.0% 100.0% number of processed stacks 3143552 0.0% 0.0% 100.0% average stack size 0.0 0.0 46.8 marketing flops 2.107587E+12 ------------------------------------------------------------------------------- # multiplications 2286 max memory usage/rank 519.032832E+06 # max total images/rank 2 # max 3D layers 1 # MPI messages exchanged 950976 MPI messages size (bytes): total size 203.844256E+09 min size 0.000000E+00 max size 1.638400E+06 average size 214.352688E+03 MPI breakdown and total messages size (bytes): size <= 128 6424 0 128 < size <= 8192 253512 2076770304 8192 < size <= 32768 179424 2939682816 32768 < size <= 131072 181440 14863564800 131072 < size <= 4194304 330176 183964913216 4194304 < size <= 16777216 0 0 16777216 < size 0 0 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 3672 62660. MP_Allreduce 10225 303. MP_Sync 104 MP_Alltoall 1821 1607811. MP_ISendRecv 22134 57667. MP_Wait 33054 MP_comm_split 50 MP_ISend 9880 92618. MP_IRecv 9880 92618. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.013 0.028 32.228 32.229 qs_mol_dyn_low 1 2.0 0.003 0.003 32.026 32.034 qs_forces 11 3.9 0.002 0.002 31.969 31.970 qs_energies 11 4.9 0.001 0.002 30.391 30.394 scf_env_do_scf 11 5.9 0.000 0.001 25.689 25.689 scf_env_do_scf_inner_loop 108 6.5 0.002 0.006 23.115 23.115 dbcsr_multiply_generic 2286 12.5 0.094 0.095 16.490 16.567 velocity_verlet 10 3.0 0.001 0.001 15.424 15.425 qs_scf_new_mos 108 7.5 0.001 0.001 14.937 14.955 qs_scf_loop_do_ot 108 8.5 0.001 0.001 14.936 14.954 ot_scf_mini 108 9.5 0.002 0.002 14.201 14.217 multiply_cannon 2286 13.5 0.194 0.203 13.031 13.837 multiply_cannon_loop 2286 14.5 0.642 0.669 12.223 13.053 ot_mini 108 10.5 0.001 0.001 8.823 8.843 qs_ot_get_derivative 108 11.5 0.001 0.001 7.319 7.334 multiply_cannon_multrec 18288 15.5 1.911 2.867 6.735 7.046 rebuild_ks_matrix 119 8.3 0.000 0.000 6.448 6.465 qs_ks_build_kohn_sham_matrix 119 9.3 0.012 0.013 6.448 6.465 qs_ks_update_qs_env 119 7.6 0.001 0.001 5.700 5.716 dbcsr_mm_accdrv_process 38222 16.0 4.017 5.425 4.740 5.512 mp_waitall_1 169478 16.3 3.148 4.270 3.148 4.270 sum_up_and_integrate 119 10.3 0.030 0.031 4.065 4.071 integrate_v_rspace 119 11.3 0.002 0.003 4.035 4.045 init_scf_run 11 5.9 0.000 0.001 3.479 3.479 scf_env_initial_rho_setup 11 6.9 0.001 0.001 3.479 3.479 qs_rho_update_rho_low 119 7.7 0.001 0.001 3.356 3.366 calculate_rho_elec 119 8.7 0.030 0.031 3.355 3.366 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 2.727 3.345 qs_ot_get_p 119 10.4 0.001 0.001 3.179 3.209 init_scf_loop 11 6.9 0.000 0.000 2.553 2.554 rs_pw_transfer 974 11.9 0.009 0.010 2.421 2.552 multiply_cannon_metrocomm3 18288 15.5 0.045 0.046 1.381 2.365 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 2.040 2.352 apply_single 119 13.6 0.000 0.000 2.040 2.352 calculate_first_density_matrix 1 7.0 0.000 0.000 2.230 2.231 make_m2s 4572 13.5 0.044 0.045 1.956 2.105 density_rs2pw 119 9.7 0.004 0.004 1.944 2.058 qs_ot_p2m_diag 50 11.0 0.012 0.012 2.042 2.048 make_images 4572 14.5 0.190 0.201 1.871 2.018 grid_integrate_task_list 119 12.3 1.796 1.899 1.796 1.899 jit_kernel_multiply 10 15.9 0.671 1.859 0.671 1.859 potential_pw2rs 119 12.3 0.007 0.008 1.846 1.854 calculate_dm_sparse 119 9.5 0.000 0.000 1.820 1.833 pw_transfer 1439 11.6 0.063 0.066 1.810 1.821 prepare_preconditioner 11 7.9 0.000 0.000 1.767 1.769 make_preconditioner 11 8.9 0.000 0.000 1.767 1.769 cp_dbcsr_syevd 50 12.0 0.003 0.003 1.767 1.767 fft_wrap_pw1pw2 1201 12.6 0.008 0.008 1.720 1.733 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 1.716 1.718 make_full_inverse_cholesky 11 9.9 0.000 0.000 1.623 1.706 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 1.626 1.635 multiply_cannon_sync_h2d 18288 15.5 1.385 1.588 1.385 1.588 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 1.555 1.568 cp_fm_diag_elpa 50 13.0 0.000 0.000 1.504 1.505 ot_diis_step 108 11.5 0.011 0.011 1.484 1.484 cp_fm_redistribute_end 50 14.0 1.116 1.480 1.117 1.481 mp_sum_l 7207 12.9 1.114 1.461 1.114 1.461 cp_fm_diag_elpa_base 50 14.0 0.347 1.411 0.362 1.450 fft3d_ps 1201 14.6 0.507 0.524 1.408 1.420 fft_wrap_pw1pw2_140 487 13.2 0.085 0.092 1.312 1.325 grid_collocate_task_list 119 9.7 1.204 1.300 1.204 1.300 acc_transpose_blocks 18288 15.5 0.075 0.076 1.238 1.260 wfi_extrapolate 11 7.9 0.001 0.001 1.204 1.204 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 1.180 1.185 make_images_data 4572 15.5 0.045 0.049 0.860 1.038 qs_energies_init_hamiltonians 11 5.9 0.000 0.001 0.947 0.948 hybrid_alltoall_any 4725 16.4 0.054 0.113 0.735 0.943 qs_ot_get_orbitals 108 10.5 0.000 0.000 0.896 0.917 mp_alltoall_d11v 2130 13.8 0.728 0.856 0.728 0.856 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 0.818 0.820 acc_transpose_blocks_kernels 18288 16.5 0.211 0.220 0.787 0.796 mp_alltoall_z22v 1201 16.6 0.734 0.793 0.734 0.793 cp_fm_cholesky_invert 11 10.9 0.746 0.750 0.746 0.750 build_core_hamiltonian_matrix_ 11 4.9 0.000 0.000 0.655 0.732 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="104", plot="h2o_64_md", label="(8n/4r/3t)", y=32.229000, yerr=0.000000 PlotPoint: name="105", plot="h2o_64_md_mem", label="(8n/4r/3t)", y=494.454545, yerr=1.233151 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/06/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 32 x 32 x 32 26877100032 0.0% 0.0% 100.0% flops 9 x 9 x 32 44168260608 0.0% 0.0% 100.0% flops 22 x 9 x 32 53835724800 0.0% 0.0% 100.0% flops 9 x 22 x 32 53885500416 0.0% 0.0% 100.0% flops 32 x 32 x 9 63568871424 0.0% 0.0% 100.0% flops 22 x 22 x 32 67007283200 0.0% 0.0% 100.0% flops 32 x 32 x 22 77695287296 0.0% 0.0% 100.0% flops 9 x 32 x 32 78422999040 0.0% 0.0% 100.0% flops 22 x 32 x 32 95850332160 0.0% 0.0% 100.0% flops 9 x 32 x 9 266263676928 0.0% 0.0% 100.0% flops 22 x 32 x 9 326697440256 0.0% 0.0% 100.0% flops 9 x 32 x 22 326697440256 0.0% 0.0% 100.0% flops 22 x 32 x 22 399918497792 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 1.880888E+12 0.0% 0.0% 100.0% flops max/rank 114.044384E+09 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 146984760 0.0% 0.0% 100.0% number of processed stacks 3805952 0.0% 0.0% 100.0% average stack size 0.0 0.0 38.6 marketing flops 2.107592E+12 ------------------------------------------------------------------------------- # multiplications 2286 max memory usage/rank 557.989888E+06 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 1042416 MPI messages size (bytes): total size 150.443262E+09 min size 0.000000E+00 max size 1.188816E+06 average size 144.321719E+03 MPI breakdown and total messages size (bytes): size <= 128 228256 0 128 < size <= 8192 126888 1039466496 8192 < size <= 32768 191472 3137077248 32768 < size <= 131072 295800 25899827200 131072 < size <= 4194304 200000 120367247040 4194304 < size <= 16777216 0 0 16777216 < size 0 0 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 3672 62659. MP_Allreduce 10224 344. MP_Sync 104 MP_Alltoall 1582 2412273. MP_ISendRecv 16422 74133. MP_Wait 24482 MP_comm_split 50 MP_ISend 7280 135929. MP_IRecv 7280 135929. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.047 0.110 35.857 35.859 qs_mol_dyn_low 1 2.0 0.003 0.003 35.538 35.545 qs_forces 11 3.9 0.002 0.003 35.480 35.482 qs_energies 11 4.9 0.003 0.013 33.679 33.686 scf_env_do_scf 11 5.9 0.001 0.001 28.539 28.540 scf_env_do_scf_inner_loop 108 6.5 0.002 0.006 24.850 24.854 dbcsr_multiply_generic 2286 12.5 0.102 0.107 18.336 18.472 velocity_verlet 10 3.0 0.001 0.001 18.184 18.186 qs_scf_new_mos 108 7.5 0.001 0.001 16.491 16.545 qs_scf_loop_do_ot 108 8.5 0.001 0.001 16.491 16.544 ot_scf_mini 108 9.5 0.002 0.003 15.543 15.596 multiply_cannon 2286 13.5 0.227 0.260 14.487 14.854 multiply_cannon_loop 2286 14.5 0.939 0.972 13.473 13.813 ot_mini 108 10.5 0.001 0.001 9.430 9.499 multiply_cannon_multrec 27432 15.5 2.330 3.023 8.573 8.948 qs_ot_get_derivative 108 11.5 0.001 0.001 7.612 7.666 dbcsr_mm_accdrv_process 47916 15.9 5.249 6.963 6.149 7.358 rebuild_ks_matrix 119 8.3 0.000 0.000 6.686 6.730 qs_ks_build_kohn_sham_matrix 119 9.3 0.012 0.013 6.686 6.729 qs_ks_update_qs_env 119 7.6 0.001 0.001 5.891 5.929 sum_up_and_integrate 119 10.3 0.035 0.039 3.922 3.937 integrate_v_rspace 119 11.3 0.002 0.003 3.887 3.903 init_scf_run 11 5.9 0.000 0.001 3.711 3.712 scf_env_initial_rho_setup 11 6.9 0.001 0.002 3.711 3.712 init_scf_loop 11 6.9 0.000 0.002 3.663 3.672 qs_ot_get_p 119 10.4 0.001 0.001 3.473 3.549 qs_rho_update_rho_low 119 7.7 0.001 0.001 3.425 3.460 calculate_rho_elec 119 8.7 0.040 0.046 3.424 3.459 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 2.858 3.255 mp_waitall_1 145218 16.4 2.363 2.871 2.363 2.871 prepare_preconditioner 11 7.9 0.000 0.000 2.745 2.753 make_preconditioner 11 8.9 0.000 0.000 2.745 2.753 make_full_inverse_cholesky 11 9.9 0.000 0.000 2.337 2.675 make_m2s 4572 13.5 0.054 0.056 2.379 2.483 make_images 4572 14.5 0.273 0.333 2.271 2.373 calculate_first_density_matrix 1 7.0 0.000 0.002 2.294 2.296 rs_pw_transfer 974 11.9 0.009 0.009 2.200 2.279 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 2.047 2.163 apply_single 119 13.6 0.000 0.000 2.047 2.162 calculate_dm_sparse 119 9.5 0.000 0.000 2.042 2.102 qs_ot_p2m_diag 50 11.0 0.015 0.023 2.074 2.085 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 2.005 2.029 density_rs2pw 119 9.7 0.003 0.004 1.904 1.982 pw_transfer 1439 11.6 0.063 0.066 1.885 1.920 grid_integrate_task_list 119 12.3 1.832 1.897 1.832 1.897 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 1.856 1.858 fft_wrap_pw1pw2 1201 12.6 0.008 0.008 1.796 1.834 jit_kernel_multiply 10 15.9 0.840 1.818 0.840 1.818 ot_diis_step 108 11.5 0.012 0.012 1.778 1.778 cp_dbcsr_syevd 50 12.0 0.003 0.003 1.745 1.745 potential_pw2rs 119 12.3 0.009 0.009 1.705 1.715 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 1.655 1.671 fft3d_ps 1201 14.6 0.533 0.586 1.478 1.509 acc_transpose_blocks 27432 15.5 0.112 0.114 1.475 1.504 cp_fm_diag_elpa 50 13.0 0.000 0.000 1.475 1.476 fft_wrap_pw1pw2_140 487 13.2 0.084 0.091 1.425 1.464 cp_fm_redistribute_end 50 14.0 0.971 1.447 0.973 1.447 cp_fm_diag_elpa_base 50 14.0 0.451 1.379 0.472 1.419 wfi_extrapolate 11 7.9 0.001 0.001 1.367 1.367 mp_sum_l 7207 12.9 1.056 1.354 1.056 1.354 grid_collocate_task_list 119 9.7 1.221 1.328 1.221 1.328 multiply_cannon_metrocomm3 27432 15.5 0.038 0.039 0.775 1.320 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 1.219 1.229 cp_fm_upper_to_full 72 13.5 0.833 1.155 0.833 1.155 qs_ot_get_orbitals 108 10.5 0.000 0.000 1.139 1.155 dbcsr_complete_redistribute 329 12.2 0.120 0.143 0.850 1.134 qs_energies_init_hamiltonians 11 5.9 0.001 0.002 1.113 1.114 make_images_data 4572 15.5 0.045 0.049 0.925 1.061 multiply_cannon_sync_h2d 27432 15.5 0.979 1.025 0.979 1.025 hybrid_alltoall_any 4725 16.4 0.061 0.150 0.797 0.970 mp_alltoall_d11v 2130 13.8 0.803 0.895 0.803 0.895 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 0.887 0.892 copy_fm_to_dbcsr 176 11.2 0.001 0.001 0.615 0.889 build_core_hamiltonian_matrix_ 11 4.9 0.000 0.000 0.798 0.874 cp_fm_cholesky_invert 11 10.9 0.858 0.862 0.858 0.862 mp_alltoall_z22v 1201 16.6 0.828 0.857 0.828 0.857 acc_transpose_blocks_kernels 27432 16.5 0.268 0.276 0.832 0.850 mp_alltoall_i22 627 13.8 0.454 0.754 0.454 0.754 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="106", plot="h2o_64_md", label="(8n/3r/4t)", y=35.859000, yerr=0.000000 PlotPoint: name="107", plot="h2o_64_md_mem", label="(8n/3r/4t)", y=529.454545, yerr=2.934548 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/07/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 32 x 32 x 32 26877100032 0.0% 0.0% 100.0% flops 9 x 9 x 32 44168260608 0.0% 0.0% 100.0% flops 22 x 9 x 32 53835724800 0.0% 0.0% 100.0% flops 9 x 22 x 32 53885500416 0.0% 0.0% 100.0% flops 32 x 32 x 9 63568871424 0.0% 0.0% 100.0% flops 22 x 22 x 32 67007283200 0.0% 0.0% 100.0% flops 32 x 32 x 22 77695287296 0.0% 0.0% 100.0% flops 9 x 32 x 32 78422999040 0.0% 0.0% 100.0% flops 22 x 32 x 32 95850332160 0.0% 0.0% 100.0% flops 9 x 32 x 9 266263676928 0.0% 0.0% 100.0% flops 22 x 32 x 9 326697440256 0.0% 0.0% 100.0% flops 9 x 32 x 22 326697440256 0.0% 0.0% 100.0% flops 22 x 32 x 22 399918497792 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 1.880888E+12 0.0% 0.0% 100.0% flops max/rank 117.977176E+09 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 146984760 0.0% 0.0% 100.0% number of processed stacks 1384136 0.0% 0.0% 100.0% average stack size 0.0 0.0 106.2 marketing flops 2.107587E+12 ------------------------------------------------------------------------------- # multiplications 2286 max memory usage/rank 608.079872E+06 # max total images/rank 1 # max 3D layers 1 # MPI messages exchanged 219456 MPI messages size (bytes): total size 97.042514E+09 min size 0.000000E+00 max size 3.276800E+06 average size 442.195750E+03 MPI breakdown and total messages size (bytes): size <= 128 1452 0 128 < size <= 8192 0 0 8192 < size <= 32768 101892 3336634368 32768 < size <= 131072 0 0 131072 < size <= 4194304 116112 93705670464 4194304 < size <= 16777216 0 0 16777216 < size 0 0 ------------------------------------------------------------------------------- - - - DBCSR MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Bcast 14 12. MP_Allreduce 8156 20. MP_Alltoall 8655 64935. MP_ISend 36532 168375. MP_IRecv 36532 168349. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 3672 62658. MP_Allreduce 10224 344. MP_Sync 104 MP_Alltoall 1582 3682667. MP_ISendRecv 10710 94533. MP_Wait 16690 MP_comm_split 50 MP_ISend 5200 225425. MP_IRecv 5200 225425. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.015 0.032 28.135 28.136 qs_mol_dyn_low 1 2.0 0.003 0.003 27.962 27.969 qs_forces 11 3.9 0.002 0.002 27.904 27.905 qs_energies 11 4.9 0.001 0.001 26.207 26.209 scf_env_do_scf 11 5.9 0.000 0.001 21.520 21.520 scf_env_do_scf_inner_loop 108 6.5 0.002 0.006 18.933 18.934 velocity_verlet 10 3.0 0.001 0.001 14.445 14.448 dbcsr_multiply_generic 2286 12.5 0.092 0.095 12.403 12.478 qs_scf_new_mos 108 7.5 0.001 0.001 11.317 11.340 qs_scf_loop_do_ot 108 8.5 0.001 0.001 11.316 11.339 ot_scf_mini 108 9.5 0.002 0.002 10.633 10.656 multiply_cannon 2286 13.5 0.230 0.235 9.723 10.201 multiply_cannon_loop 2286 14.5 0.331 0.342 8.776 8.959 multiply_cannon_multrec 9144 15.5 1.663 1.919 5.948 6.174 ot_mini 108 10.5 0.001 0.001 5.980 6.006 rebuild_ks_matrix 119 8.3 0.000 0.000 5.806 5.828 qs_ks_build_kohn_sham_matrix 119 9.3 0.012 0.013 5.806 5.828 qs_ks_update_qs_env 119 7.6 0.001 0.001 5.176 5.196 qs_ot_get_derivative 108 11.5 0.001 0.001 4.675 4.698 dbcsr_mm_accdrv_process 12550 15.8 3.161 4.051 4.183 4.255 sum_up_and_integrate 119 10.3 0.037 0.041 3.571 3.576 integrate_v_rspace 119 11.3 0.003 0.003 3.533 3.539 qs_rho_update_rho_low 119 7.7 0.001 0.001 3.270 3.276 calculate_rho_elec 119 8.7 0.059 0.061 3.269 3.275 init_scf_run 11 5.9 0.000 0.001 3.230 3.230 scf_env_initial_rho_setup 11 6.9 0.001 0.001 3.230 3.230 qs_ot_get_p 119 10.4 0.001 0.001 2.852 2.885 init_scf_loop 11 6.9 0.000 0.000 2.563 2.565 mp_waitall_1 121218 16.5 1.738 2.207 1.738 2.207 calculate_first_density_matrix 1 7.0 0.000 0.000 2.081 2.082 make_m2s 4572 13.5 0.034 0.035 1.844 1.975 grid_integrate_task_list 119 12.3 1.839 1.920 1.839 1.920 make_images 4572 14.5 0.267 0.301 1.754 1.883 qs_ot_p2m_diag 50 11.0 0.022 0.023 1.854 1.858 prepare_preconditioner 11 7.9 0.000 0.000 1.831 1.835 make_preconditioner 11 8.9 0.000 0.000 1.831 1.835 rs_pw_transfer 974 11.9 0.008 0.008 1.713 1.790 density_rs2pw 119 9.7 0.004 0.004 1.672 1.755 pw_transfer 1439 11.6 0.063 0.065 1.742 1.752 make_full_inverse_cholesky 11 9.9 0.000 0.000 1.717 1.743 calculate_dm_sparse 119 9.5 0.000 0.000 1.713 1.733 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 1.694 1.707 fft_wrap_pw1pw2 1201 12.6 0.008 0.008 1.652 1.662 cp_dbcsr_syevd 50 12.0 0.003 0.003 1.647 1.648 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 1.553 1.555 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 1.377 1.386 potential_pw2rs 119 12.3 0.010 0.011 1.367 1.372 grid_collocate_task_list 119 9.7 1.271 1.370 1.271 1.370 cp_fm_diag_elpa 50 13.0 0.000 0.000 1.339 1.339 fft_wrap_pw1pw2_140 487 13.2 0.082 0.085 1.315 1.329 fft3d_ps 1201 14.6 0.539 0.550 1.319 1.328 cp_fm_redistribute_end 50 14.0 0.666 1.314 0.667 1.315 cp_fm_diag_elpa_base 50 14.0 0.603 1.240 0.647 1.296 ot_diis_step 108 11.5 0.012 0.012 1.290 1.290 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 1.237 1.238 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 1.203 1.217 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 1.160 1.175 apply_single 119 13.6 0.000 0.000 1.159 1.175 jit_kernel_multiply 6 16.0 0.985 1.167 0.985 1.167 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 1.126 1.133 wfi_extrapolate 11 7.9 0.001 0.001 1.102 1.102 hybrid_alltoall_any 4725 16.4 0.062 0.175 0.813 1.033 make_images_data 4572 15.5 0.038 0.042 0.836 1.019 mp_alltoall_d11v 2130 13.8 0.843 0.974 0.843 0.974 cp_fm_cholesky_invert 11 10.9 0.941 0.944 0.941 0.944 build_core_hamiltonian_matrix_ 11 4.9 0.000 0.000 0.871 0.923 acc_transpose_blocks 9144 15.5 0.039 0.040 0.897 0.903 multiply_cannon_sync_h2d 9144 15.5 0.705 0.779 0.705 0.779 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 0.753 0.755 qs_ot_get_orbitals 108 10.5 0.000 0.000 0.742 0.750 qs_env_update_s_mstruct 11 6.9 0.001 0.001 0.692 0.741 mp_allgather_i34 2286 14.5 0.279 0.701 0.279 0.701 mp_alltoall_z22v 1201 16.6 0.654 0.688 0.654 0.688 multiply_cannon_metrocomm3 9144 15.5 0.018 0.019 0.319 0.677 acc_transpose_blocks_kernels 9144 16.5 0.117 0.120 0.649 0.652 qs_create_task_list 11 7.9 0.027 0.027 0.571 0.596 dbcsr_complete_redistribute 329 12.2 0.166 0.174 0.558 0.586 generate_qs_task_list 11 8.9 0.189 0.211 0.545 0.570 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="108", plot="h2o_64_md", label="(8n/2r/6t)", y=28.136000, yerr=0.000000 PlotPoint: name="109", plot="h2o_64_md_mem", label="(8n/2r/6t)", y=577.636364, yerr=4.558164 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/08/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 32 x 32 x 32 26877100032 0.0% 0.0% 100.0% flops 9 x 9 x 32 44168260608 0.0% 0.0% 100.0% flops 22 x 9 x 32 53835724800 0.0% 0.0% 100.0% flops 9 x 22 x 32 53885500416 0.0% 0.0% 100.0% flops 32 x 32 x 9 63568871424 0.0% 0.0% 100.0% flops 22 x 22 x 32 67007283200 0.0% 0.0% 100.0% flops 32 x 32 x 22 77695287296 0.0% 0.0% 100.0% flops 9 x 32 x 32 78422999040 0.0% 0.0% 100.0% flops 22 x 32 x 32 95850332160 0.0% 0.0% 100.0% flops 9 x 32 x 9 266263676928 0.0% 0.0% 100.0% flops 22 x 32 x 9 326697440256 0.0% 0.0% 100.0% flops 9 x 32 x 22 326697440256 0.0% 0.0% 100.0% flops 22 x 32 x 22 399918497792 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 1.880888E+12 0.0% 0.0% 100.0% flops max/rank 235.585836E+09 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 146984760 0.0% 0.0% 100.0% number of processed stacks 1388964 0.0% 0.0% 100.0% average stack size 0.0 0.0 105.8 marketing flops 2.107587E+12 ------------------------------------------------------------------------------- # multiplications 2286 max memory usage/rank 743.170048E+06 # max total images/rank 2 # max 3D layers 1 # MPI messages exchanged 91440 MPI messages size (bytes): total size 85.748679E+09 min size 0.000000E+00 max size 6.553600E+06 average size 937.758938E+03 MPI breakdown and total messages size (bytes): size <= 128 572 0 128 < size <= 8192 0 0 8192 < size <= 32768 21148 692256768 32768 < size <= 131072 19224 1259864064 131072 < size <= 4194304 41040 21941452800 4194304 < size <= 16777216 9456 61855174464 16777216 < size 0 0 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 3622 63729. MP_Allreduce 10074 433. MP_Sync 54 MP_Alltoall 1582 7383731. MP_ISendRecv 4998 189067. MP_Wait 8898 MP_ISend 3120 546875. MP_IRecv 3120 546875. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.063 0.101 41.512 41.513 qs_mol_dyn_low 1 2.0 0.003 0.003 41.153 41.160 qs_forces 11 3.9 0.001 0.002 41.089 41.090 qs_energies 11 4.9 0.001 0.001 39.127 39.130 scf_env_do_scf 11 5.9 0.000 0.001 33.181 33.181 scf_env_do_scf_inner_loop 108 6.5 0.003 0.006 25.191 25.191 velocity_verlet 10 3.0 0.001 0.001 23.305 23.310 dbcsr_multiply_generic 2286 12.5 0.101 0.102 17.812 18.004 qs_scf_new_mos 108 7.5 0.001 0.001 16.118 16.224 qs_scf_loop_do_ot 108 8.5 0.001 0.001 16.117 16.223 ot_scf_mini 108 9.5 0.002 0.002 15.023 15.133 multiply_cannon 2286 13.5 0.306 0.316 13.783 14.711 multiply_cannon_loop 2286 14.5 0.343 0.348 12.465 13.380 ot_mini 108 10.5 0.001 0.001 8.889 9.014 multiply_cannon_multrec 9144 15.5 3.373 4.703 8.643 8.751 init_scf_loop 11 6.9 0.000 0.000 7.958 7.959 rebuild_ks_matrix 119 8.3 0.000 0.000 6.929 7.080 qs_ks_build_kohn_sham_matrix 119 9.3 0.013 0.013 6.928 7.080 prepare_preconditioner 11 7.9 0.000 0.000 7.020 7.035 make_preconditioner 11 8.9 0.000 0.000 7.020 7.035 qs_ot_get_derivative 108 11.5 0.001 0.001 6.866 6.977 make_full_inverse_cholesky 11 9.9 0.000 0.000 5.566 6.911 dbcsr_mm_accdrv_process 12550 15.8 4.151 5.375 5.146 6.491 qs_ks_update_qs_env 119 7.6 0.001 0.001 6.261 6.399 cp_fm_upper_to_full 72 14.2 3.175 4.568 3.175 4.568 init_scf_run 11 5.9 0.000 0.001 3.916 3.916 scf_env_initial_rho_setup 11 6.9 0.001 0.001 3.915 3.915 qs_rho_update_rho_low 119 7.7 0.001 0.001 3.797 3.801 calculate_rho_elec 119 8.7 0.118 0.121 3.797 3.801 sum_up_and_integrate 119 10.3 0.064 0.066 3.757 3.763 integrate_v_rspace 119 11.3 0.003 0.003 3.692 3.701 mp_waitall_1 97218 16.6 2.648 3.596 2.648 3.596 qs_ot_get_p 119 10.4 0.001 0.001 3.378 3.523 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 2.514 2.918 dbcsr_complete_redistribute 329 12.2 0.286 0.291 2.031 2.846 make_m2s 4572 13.5 0.038 0.038 2.404 2.607 copy_fm_to_dbcsr 176 11.2 0.001 0.001 1.705 2.508 make_images 4572 14.5 0.352 0.386 2.284 2.486 calculate_first_density_matrix 1 7.0 0.000 0.000 2.460 2.461 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 2.155 2.425 apply_single 119 13.6 0.000 0.000 2.155 2.425 mp_alltoall_i22 627 13.8 1.503 2.328 1.503 2.328 calculate_dm_sparse 119 9.5 0.000 0.000 2.288 2.306 transfer_fm_to_dbcsr 11 9.9 0.000 0.000 1.449 2.246 multiply_cannon_metrocomm3 9144 15.5 0.019 0.019 1.252 2.132 pw_transfer 1439 11.6 0.066 0.067 2.108 2.111 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 2.087 2.088 grid_integrate_task_list 119 12.3 2.024 2.056 2.024 2.056 qs_ot_p2m_diag 50 11.0 0.042 0.043 2.031 2.032 fft_wrap_pw1pw2 1201 12.6 0.008 0.008 2.012 2.015 ot_diis_step 108 11.5 0.014 0.014 1.995 1.995 density_rs2pw 119 9.7 0.003 0.003 1.880 1.893 mp_sum_l 7207 12.9 1.085 1.825 1.085 1.825 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 1.743 1.800 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 1.744 1.773 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 1.752 1.752 cp_dbcsr_syevd 50 12.0 0.003 0.003 1.748 1.749 fft_wrap_pw1pw2_140 487 13.2 0.087 0.089 1.660 1.666 fft3d_ps 1201 14.6 0.568 0.576 1.647 1.650 cp_fm_cholesky_invert 11 10.9 1.575 1.578 1.575 1.578 rs_pw_transfer 974 11.9 0.009 0.009 1.480 1.507 grid_collocate_task_list 119 9.7 1.461 1.474 1.461 1.474 hybrid_alltoall_any 4725 16.4 0.087 0.147 1.193 1.446 cp_fm_diag_elpa 50 13.0 0.000 0.000 1.423 1.424 cp_fm_diag_elpa_base 50 14.0 1.279 1.332 1.421 1.422 make_images_data 4572 15.5 0.042 0.045 1.139 1.403 wfi_extrapolate 11 7.9 0.001 0.001 1.396 1.396 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 1.346 1.354 potential_pw2rs 119 12.3 0.014 0.014 1.264 1.268 mp_alltoall_d11v 2130 13.8 1.216 1.248 1.216 1.248 qs_ot_get_orbitals 108 10.5 0.000 0.000 1.154 1.172 qs_env_update_s_mstruct 11 6.9 0.001 0.001 1.091 1.113 jit_kernel_multiply 6 15.8 0.968 1.086 0.968 1.086 build_core_hamiltonian_matrix_ 11 4.9 0.000 0.000 0.989 1.044 multiply_cannon_sync_h2d 9144 15.5 1.037 1.042 1.037 1.042 acc_transpose_blocks 9144 15.5 0.038 0.039 1.030 1.031 mp_alltoall_z22v 1201 16.6 0.944 0.960 0.944 0.960 qs_create_task_list 11 7.9 0.001 0.001 0.943 0.957 generate_qs_task_list 11 8.9 0.371 0.390 0.942 0.956 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 0.879 0.893 copy_dbcsr_to_fm 153 11.3 0.002 0.002 0.806 0.861 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="110", plot="h2o_64_md", label="(8n/1r/12t)", y=41.513000, yerr=0.000000 PlotPoint: name="111", plot="h2o_64_md_mem", label="(8n/1r/12t)", y=698.181818, yerr=14.843537 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/09/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 32 x 32 x 32 184415158272 0.0% 0.0% 100.0% flops 9 x 9 x 32 269180485632 0.0% 0.0% 100.0% flops 9 x 22 x 32 349395425280 0.0% 0.0% 100.0% flops 22 x 9 x 32 350042406912 0.0% 0.0% 100.0% flops 22 x 22 x 32 453581815808 0.0% 0.0% 100.0% flops 32 x 32 x 9 465064427520 0.0% 0.0% 100.0% flops 32 x 32 x 22 568412078080 0.0% 0.0% 100.0% flops 9 x 32 x 32 572195340288 0.0% 0.0% 100.0% flops 22 x 32 x 32 699349860352 0.0% 0.0% 100.0% flops 9 x 32 x 9 1735942275072 0.0% 0.0% 100.0% flops 22 x 32 x 9 2216407818240 0.0% 0.0% 100.0% flops 9 x 32 x 22 2216407818240 0.0% 0.0% 100.0% flops 22 x 32 x 22 2803661053952 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 12.884056E+12 0.0% 0.0% 100.0% flops max/rank 198.287135E+09 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 984178160 0.0% 0.0% 100.0% number of processed stacks 8410880 0.0% 0.0% 100.0% average stack size 0.0 0.0 117.0 marketing flops 15.646302E+12 ------------------------------------------------------------------------------- # multiplications 2055 max memory usage/rank 501.575680E+06 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 8483040 MPI messages size (bytes): total size 1.160510E+12 min size 0.000000E+00 max size 1.161504E+06 average size 136.803609E+03 MPI breakdown and total messages size (bytes): size <= 128 1836752 0 128 < size <= 8192 1040592 8524529664 8192 < size <= 32768 1486976 24362614784 32768 < size <= 131072 2491776 216971345920 131072 < size <= 4194304 1626944 910632720448 4194304 < size <= 16777216 0 0 16777216 < size 0 0 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 3521 65372. MP_Allreduce 9840 486. MP_Sync 100 MP_Alltoall 1938 590705. MP_ISendRecv 41800 9096. MP_Wait 58168 MP_comm_split 48 MP_ISend 14300 82312. MP_IRecv 14300 82312. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.013 0.030 82.235 82.236 qs_mol_dyn_low 1 2.0 0.003 0.003 81.968 82.015 qs_forces 11 3.9 0.002 0.003 81.900 81.900 qs_energies 11 4.9 0.001 0.002 79.062 79.074 scf_env_do_scf 11 5.9 0.000 0.001 70.229 70.231 scf_env_do_scf_inner_loop 99 6.5 0.002 0.006 64.457 64.458 dbcsr_multiply_generic 2055 12.4 0.106 0.108 51.593 51.842 qs_scf_new_mos 99 7.5 0.000 0.001 47.539 47.653 qs_scf_loop_do_ot 99 8.5 0.001 0.001 47.539 47.653 ot_scf_mini 99 9.5 0.002 0.002 45.146 45.224 multiply_cannon 2055 13.4 0.174 0.180 42.200 43.352 multiply_cannon_loop 2055 14.4 1.557 1.611 41.187 42.362 velocity_verlet 10 3.0 0.001 0.001 41.737 41.737 ot_mini 99 10.5 0.001 0.001 27.443 27.541 qs_ot_get_derivative 99 11.5 0.001 0.001 20.632 20.751 multiply_cannon_multrec 49320 15.4 12.237 13.073 17.216 17.976 rebuild_ks_matrix 110 8.3 0.000 0.001 14.444 14.574 qs_ks_build_kohn_sham_matrix 110 9.3 0.011 0.012 14.444 14.573 qs_ks_update_qs_env 110 7.6 0.001 0.001 12.706 12.820 mp_waitall_1 241148 16.1 11.258 12.384 11.258 12.384 multiply_cannon_sync_h2d 49320 15.4 10.076 10.801 10.076 10.801 qs_ot_get_p 110 10.4 0.001 0.001 9.306 9.442 multiply_cannon_metrocomm3 49320 15.4 0.077 0.081 6.521 7.691 apply_preconditioner_dbcsr 110 12.6 0.000 0.000 7.161 7.638 apply_single 110 13.6 0.000 0.000 7.161 7.638 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 6.967 7.540 sum_up_and_integrate 110 10.3 0.037 0.043 7.042 7.055 integrate_v_rspace 110 11.3 0.003 0.003 7.006 7.026 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 6.785 6.842 init_scf_run 11 5.9 0.000 0.001 6.794 6.794 scf_env_initial_rho_setup 11 6.9 0.001 0.001 6.794 6.794 ot_diis_step 99 11.5 0.006 0.006 6.527 6.528 qs_rho_update_rho_low 110 7.6 0.001 0.001 6.173 6.285 calculate_rho_elec 110 8.6 0.020 0.024 6.173 6.284 qs_ot_p2m_diag 48 11.0 0.012 0.018 6.081 6.105 mp_sum_l 6514 12.8 4.864 5.777 4.864 5.777 init_scf_loop 11 6.9 0.000 0.000 5.742 5.742 dbcsr_mm_accdrv_process 87628 16.1 2.042 2.243 4.859 5.267 cp_dbcsr_syevd 48 12.0 0.002 0.002 5.077 5.078 cp_fm_diag_elpa 48 13.0 0.000 0.000 4.539 4.540 cp_fm_redistribute_end 48 14.0 3.947 4.504 3.951 4.506 cp_fm_diag_elpa_base 48 14.0 0.548 4.395 0.552 4.428 make_m2s 4110 13.4 0.061 0.066 4.066 4.179 make_images 4110 14.4 0.178 0.193 3.970 4.085 wfi_extrapolate 11 7.9 0.001 0.001 4.078 4.078 rs_pw_transfer 902 11.9 0.012 0.013 3.794 3.930 calculate_dm_sparse 110 9.5 0.000 0.001 3.758 3.845 prepare_preconditioner 11 7.9 0.000 0.000 3.587 3.609 make_preconditioner 11 8.9 0.000 0.000 3.587 3.609 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 3.553 3.558 multiply_cannon_metrocomm1 49320 15.4 0.063 0.068 2.550 3.536 density_rs2pw 110 9.6 0.004 0.004 3.317 3.446 make_full_inverse_cholesky 11 9.9 0.000 0.000 3.379 3.422 grid_integrate_task_list 110 12.3 3.247 3.386 3.247 3.386 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 3.220 3.268 qs_ot_get_orbitals 99 10.5 0.000 0.001 3.154 3.198 pw_transfer 1331 11.6 0.053 0.065 2.946 3.007 fft_wrap_pw1pw2 1111 12.6 0.007 0.008 2.859 2.922 potential_pw2rs 110 12.3 0.006 0.007 2.684 2.706 calculate_first_density_matrix 1 7.0 0.000 0.000 2.627 2.633 jit_kernel_multiply 13 15.9 2.537 2.552 2.537 2.552 fft3d_ps 1111 14.6 0.751 0.842 2.397 2.451 fft_wrap_pw1pw2_140 451 13.1 0.169 0.187 2.336 2.403 mp_alltoall_d11v 2046 13.8 2.007 2.371 2.007 2.371 acc_transpose_blocks 49320 15.4 0.210 0.225 2.168 2.309 grid_collocate_task_list 110 9.6 2.083 2.188 2.083 2.188 cp_fm_cholesky_invert 11 10.9 1.988 1.992 1.988 1.992 mp_waitany 14300 13.8 1.761 1.951 1.761 1.951 mp_sum_d 3883 11.9 1.448 1.936 1.448 1.936 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.881 1.896 make_images_data 4110 15.4 0.043 0.046 1.742 1.873 hybrid_alltoall_any 4261 16.3 0.081 0.479 1.522 1.747 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.001 0.001 1.636 1.663 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="200", plot="h2o_128_md", label="(8n/12r/1t)", y=82.236000, yerr=0.000000 PlotPoint: name="201", plot="h2o_128_md_mem", label="(8n/12r/1t)", y=475.909091, yerr=2.842999 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/10/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 32 x 32 x 32 184415158272 0.0% 0.0% 100.0% flops 9 x 9 x 32 269180485632 0.0% 0.0% 100.0% flops 9 x 22 x 32 349395425280 0.0% 0.0% 100.0% flops 22 x 9 x 32 350042406912 0.0% 0.0% 100.0% flops 22 x 22 x 32 453581815808 0.0% 0.0% 100.0% flops 32 x 32 x 9 465064427520 0.0% 0.0% 100.0% flops 32 x 32 x 22 568412078080 0.0% 0.0% 100.0% flops 9 x 32 x 32 572195340288 0.0% 0.0% 100.0% flops 22 x 32 x 32 699349860352 0.0% 0.0% 100.0% flops 9 x 32 x 9 1735942275072 0.0% 0.0% 100.0% flops 22 x 32 x 9 2216407818240 0.0% 0.0% 100.0% flops 9 x 32 x 22 2216407818240 0.0% 0.0% 100.0% flops 22 x 32 x 22 2803661053952 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 12.884056E+12 0.0% 0.0% 100.0% flops max/rank 390.715586E+09 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 984178160 0.0% 0.0% 100.0% number of processed stacks 5019072 0.0% 0.0% 100.0% average stack size 0.0 0.0 196.1 marketing flops 15.646302E+12 ------------------------------------------------------------------------------- # multiplications 2055 max memory usage/rank 586.219520E+06 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 1972800 MPI messages size (bytes): total size 1.077520E+12 min size 0.000000E+00 max size 4.537280E+06 average size 546.188250E+03 MPI breakdown and total messages size (bytes): size <= 128 14916 0 128 < size <= 8192 222984 1826684928 8192 < size <= 32768 520356 13399818240 32768 < size <= 131072 372336 35386294272 131072 < size <= 4194304 787758 788321309808 4194304 < size <= 16777216 54450 238588003280 16777216 < size 0 0 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 3521 65587. MP_Allreduce 9839 562. MP_Sync 100 MP_Alltoall 1717 3485785. MP_ISendRecv 20680 26400. MP_Wait 32692 MP_comm_split 48 MP_ISend 10164 155761. MP_IRecv 10164 155761. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.022 0.042 69.955 69.958 qs_mol_dyn_low 1 2.0 0.003 0.005 69.594 69.604 qs_forces 11 3.9 0.002 0.002 69.525 69.527 qs_energies 11 4.9 0.003 0.010 66.214 66.219 scf_env_do_scf 11 5.9 0.001 0.001 57.338 57.342 scf_env_do_scf_inner_loop 99 6.5 0.003 0.006 49.538 49.538 dbcsr_multiply_generic 2055 12.4 0.114 0.119 38.423 38.648 velocity_verlet 10 3.0 0.001 0.001 36.314 36.316 qs_scf_new_mos 99 7.5 0.001 0.001 33.510 33.632 qs_scf_loop_do_ot 99 8.5 0.001 0.001 33.509 33.631 multiply_cannon 2055 13.4 0.221 0.247 31.546 32.934 ot_scf_mini 99 9.5 0.003 0.003 31.830 31.934 multiply_cannon_loop 2055 14.4 0.932 0.959 30.212 31.203 ot_mini 99 10.5 0.001 0.001 18.669 18.777 multiply_cannon_multrec 24660 15.4 7.571 9.163 14.033 15.643 rebuild_ks_matrix 110 8.3 0.000 0.001 13.720 13.855 qs_ks_build_kohn_sham_matrix 110 9.3 0.012 0.014 13.719 13.855 qs_ot_get_derivative 99 11.5 0.001 0.001 12.791 12.901 qs_ks_update_qs_env 110 7.6 0.001 0.001 12.108 12.223 mp_waitall_1 186928 16.3 8.198 11.043 8.198 11.043 multiply_cannon_metrocomm3 24660 15.4 0.070 0.074 5.400 8.824 multiply_cannon_sync_h2d 24660 15.4 7.031 8.469 7.031 8.469 init_scf_loop 11 6.9 0.001 0.003 7.762 7.763 apply_preconditioner_dbcsr 110 12.6 0.000 0.000 6.577 7.318 apply_single 110 13.6 0.000 0.001 6.577 7.317 qs_ot_get_p 110 10.4 0.001 0.001 6.471 6.622 sum_up_and_integrate 110 10.3 0.052 0.060 6.603 6.614 dbcsr_mm_accdrv_process 52282 16.1 4.877 5.339 6.304 6.607 integrate_v_rspace 110 11.3 0.002 0.003 6.550 6.564 init_scf_run 11 5.9 0.000 0.001 6.411 6.412 scf_env_initial_rho_setup 11 6.9 0.001 0.001 6.411 6.412 qs_rho_update_rho_low 110 7.6 0.001 0.001 5.852 5.863 calculate_rho_elec 110 8.6 0.039 0.047 5.851 5.863 ot_diis_step 99 11.5 0.010 0.010 5.819 5.820 prepare_preconditioner 11 7.9 0.000 0.000 5.737 5.756 make_preconditioner 11 8.9 0.000 0.001 5.737 5.756 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 4.784 5.496 make_full_inverse_cholesky 11 9.9 0.000 0.000 5.321 5.483 make_m2s 4110 13.4 0.056 0.059 4.326 4.760 make_images 4110 14.4 0.397 0.444 4.218 4.648 qs_ot_p2m_diag 48 11.0 0.028 0.043 4.517 4.540 cp_dbcsr_syevd 48 12.0 0.003 0.003 4.056 4.056 wfi_extrapolate 11 7.9 0.001 0.001 3.536 3.536 pw_transfer 1331 11.6 0.065 0.072 3.395 3.531 cp_fm_diag_elpa 48 13.0 0.000 0.000 3.462 3.467 cp_fm_redistribute_end 48 14.0 2.578 3.429 2.582 3.430 fft_wrap_pw1pw2 1111 12.6 0.008 0.008 3.289 3.425 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 3.361 3.422 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.003 3.399 3.401 cp_fm_diag_elpa_base 48 14.0 0.809 3.288 0.842 3.375 density_rs2pw 110 9.6 0.004 0.004 3.143 3.355 grid_integrate_task_list 110 12.3 3.141 3.342 3.141 3.342 rs_pw_transfer 902 11.9 0.012 0.014 3.072 3.322 calculate_dm_sparse 110 9.5 0.001 0.001 3.131 3.163 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 3.044 3.095 fft_wrap_pw1pw2_140 451 13.1 0.199 0.216 2.718 2.854 make_images_data 4110 15.4 0.047 0.051 2.392 2.841 calculate_first_density_matrix 1 7.0 0.000 0.002 2.787 2.789 fft3d_ps 1111 14.6 1.072 1.263 2.647 2.768 hybrid_alltoall_any 4261 16.3 0.101 0.441 2.116 2.709 cp_fm_cholesky_invert 11 10.9 2.686 2.693 2.686 2.693 mp_sum_l 6514 12.8 1.944 2.560 1.944 2.560 potential_pw2rs 110 12.3 0.008 0.009 2.450 2.469 grid_collocate_task_list 110 9.6 2.118 2.285 2.118 2.285 mp_alltoall_d11v 2046 13.8 1.856 2.274 1.856 2.274 jit_kernel_multiply 10 16.4 1.074 2.161 1.074 2.161 qs_ot_get_orbitals 99 10.5 0.000 0.001 2.019 2.050 qs_energies_init_hamiltonians 11 5.9 0.000 0.003 1.919 1.921 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.776 1.794 mp_allgather_i34 2055 14.4 0.665 1.705 0.665 1.705 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.001 0.001 1.687 1.700 multiply_cannon_metrocomm4 22605 15.4 0.074 0.079 0.778 1.646 acc_transpose_blocks 24660 15.4 0.106 0.109 1.598 1.633 build_core_hamiltonian_matrix_ 11 4.9 0.000 0.001 1.373 1.480 mp_irecv_dv 57340 16.2 0.653 1.468 0.653 1.468 dbcsr_complete_redistribute 325 12.2 0.236 0.299 1.180 1.449 mp_waitany 10164 13.8 1.237 1.433 1.237 1.433 cp_fm_cholesky_decompose 22 10.9 1.396 1.404 1.396 1.404 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="202", plot="h2o_128_md", label="(8n/6r/2t)", y=69.958000, yerr=0.000000 PlotPoint: name="203", plot="h2o_128_md_mem", label="(8n/6r/2t)", y=553.090909, yerr=6.894722 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/11/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 32 x 32 x 32 184415158272 0.0% 0.0% 100.0% flops 9 x 9 x 32 269180485632 0.0% 0.0% 100.0% flops 9 x 22 x 32 349395425280 0.0% 0.0% 100.0% flops 22 x 9 x 32 350042406912 0.0% 0.0% 100.0% flops 22 x 22 x 32 453581815808 0.0% 0.0% 100.0% flops 32 x 32 x 9 465064427520 0.0% 0.0% 100.0% flops 32 x 32 x 22 568412078080 0.0% 0.0% 100.0% flops 9 x 32 x 32 572195340288 0.0% 0.0% 100.0% flops 22 x 32 x 32 699349860352 0.0% 0.0% 100.0% flops 9 x 32 x 9 1735942275072 0.0% 0.0% 100.0% flops 22 x 32 x 9 2216407818240 0.0% 0.0% 100.0% flops 9 x 32 x 22 2216407818240 0.0% 0.0% 100.0% flops 22 x 32 x 22 2803661053952 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 12.884056E+12 0.0% 0.0% 100.0% flops max/rank 404.681598E+09 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 984178160 0.0% 0.0% 100.0% number of processed stacks 3346752 0.0% 0.0% 100.0% average stack size 0.0 0.0 294.1 marketing flops 15.646297E+12 ------------------------------------------------------------------------------- # multiplications 2055 max memory usage/rank 659.501056E+06 # max total images/rank 2 # max 3D layers 1 # MPI messages exchanged 854880 MPI messages size (bytes): total size 708.322787E+09 min size 0.000000E+00 max size 6.553600E+06 average size 828.564000E+03 MPI breakdown and total messages size (bytes): size <= 128 6424 0 128 < size <= 8192 0 0 8192 < size <= 32768 222984 7302414336 32768 < size <= 131072 153888 10085203968 131072 < size <= 4194304 389376 200257044480 4194304 < size <= 16777216 82208 490679162176 16777216 < size 0 0 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 3521 65578. MP_Allreduce 9838 559. MP_Sync 100 MP_Alltoall 1496 4511006. MP_ISendRecv 13640 27424. MP_Wait 32318 MP_comm_split 48 MP_ISend 17072 115022. MP_IRecv 17072 115022. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.013 0.029 59.897 59.897 qs_mol_dyn_low 1 2.0 0.003 0.005 59.637 59.646 qs_forces 11 3.9 0.002 0.003 59.569 59.570 qs_energies 11 4.9 0.001 0.002 56.423 56.427 scf_env_do_scf 11 5.9 0.000 0.001 48.597 48.597 scf_env_do_scf_inner_loop 99 6.5 0.002 0.007 39.835 39.836 velocity_verlet 10 3.0 0.001 0.001 32.536 32.538 dbcsr_multiply_generic 2055 12.4 0.108 0.112 28.490 28.795 qs_scf_new_mos 99 7.5 0.001 0.001 25.346 25.455 qs_scf_loop_do_ot 99 8.5 0.001 0.001 25.346 25.455 ot_scf_mini 99 9.5 0.002 0.003 24.114 24.252 multiply_cannon 2055 13.4 0.210 0.222 22.048 23.226 multiply_cannon_loop 2055 14.4 0.621 0.635 20.886 21.826 ot_mini 99 10.5 0.001 0.001 13.741 13.887 rebuild_ks_matrix 110 8.3 0.000 0.000 12.133 12.281 qs_ks_build_kohn_sham_matrix 110 9.3 0.013 0.016 12.132 12.281 qs_ks_update_qs_env 110 7.6 0.001 0.001 10.707 10.834 multiply_cannon_multrec 16440 15.4 3.797 4.833 9.530 10.489 mp_waitall_1 146766 16.3 7.370 10.324 7.370 10.324 qs_ot_get_derivative 99 11.5 0.001 0.001 9.238 9.376 init_scf_loop 11 6.9 0.000 0.000 8.723 8.724 multiply_cannon_metrocomm3 16440 15.4 0.042 0.044 4.462 7.274 prepare_preconditioner 11 7.9 0.000 0.000 6.959 6.977 make_preconditioner 11 8.9 0.000 0.000 6.959 6.977 make_full_inverse_cholesky 11 9.9 0.000 0.000 6.304 6.638 sum_up_and_integrate 110 10.3 0.060 0.060 6.420 6.435 integrate_v_rspace 110 11.3 0.003 0.003 6.360 6.376 dbcsr_mm_accdrv_process 34862 16.1 4.731 5.316 5.588 5.747 qs_ot_get_p 110 10.4 0.001 0.001 5.557 5.727 qs_rho_update_rho_low 110 7.6 0.001 0.001 5.547 5.556 calculate_rho_elec 110 8.6 0.058 0.058 5.546 5.556 apply_preconditioner_dbcsr 110 12.6 0.000 0.000 5.022 5.467 apply_single 110 13.6 0.000 0.000 5.021 5.467 init_scf_run 11 5.9 0.000 0.001 5.422 5.423 scf_env_initial_rho_setup 11 6.9 0.001 0.002 5.422 5.422 make_m2s 4110 13.4 0.050 0.052 4.236 4.588 make_images 4110 14.4 0.391 0.512 4.121 4.472 ot_diis_step 99 11.5 0.010 0.011 4.471 4.471 multiply_cannon_sync_h2d 16440 15.4 3.646 4.198 3.646 4.198 qs_ot_p2m_diag 48 11.0 0.042 0.044 3.920 3.924 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 3.105 3.746 cp_dbcsr_syevd 48 12.0 0.003 0.003 3.572 3.572 grid_integrate_task_list 110 12.3 3.192 3.366 3.192 3.366 pw_transfer 1331 11.6 0.064 0.071 3.199 3.207 fft_wrap_pw1pw2 1111 12.6 0.008 0.008 3.093 3.104 cp_fm_diag_elpa 48 13.0 0.000 0.000 3.031 3.033 cp_fm_redistribute_end 48 14.0 1.886 3.002 1.888 3.002 density_rs2pw 110 9.6 0.004 0.004 2.823 2.974 cp_fm_diag_elpa_base 48 14.0 1.050 2.853 1.109 2.962 wfi_extrapolate 11 7.9 0.001 0.001 2.944 2.944 make_images_data 4110 15.4 0.044 0.048 2.449 2.896 rs_pw_transfer 902 11.9 0.010 0.011 2.705 2.849 cp_fm_cholesky_invert 11 10.9 2.780 2.787 2.780 2.787 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.003 2.778 2.779 hybrid_alltoall_any 4261 16.3 0.105 0.375 2.158 2.756 fft_wrap_pw1pw2_140 451 13.1 0.210 0.215 2.628 2.641 calculate_dm_sparse 110 9.5 0.001 0.001 2.507 2.534 multiply_cannon_metrocomm4 14385 15.4 0.046 0.050 0.897 2.515 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 2.435 2.507 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 2.412 2.463 fft3d_ps 1111 14.6 1.066 1.076 2.407 2.419 calculate_first_density_matrix 1 7.0 0.000 0.000 2.396 2.397 mp_irecv_dv 48980 15.7 0.825 2.389 0.825 2.389 grid_collocate_task_list 110 9.6 2.175 2.367 2.175 2.367 potential_pw2rs 110 12.3 0.011 0.011 2.248 2.258 mp_sum_l 6514 12.8 1.586 2.160 1.586 2.160 mp_alltoall_d11v 2046 13.8 1.815 2.096 1.815 2.096 qs_energies_init_hamiltonians 11 5.9 0.000 0.001 1.971 1.972 dbcsr_complete_redistribute 325 12.2 0.329 0.382 1.509 1.970 cp_fm_upper_to_full 70 13.6 1.369 1.827 1.369 1.827 cp_fm_cholesky_decompose 22 10.9 1.655 1.676 1.655 1.676 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.625 1.647 mp_allgather_i34 2055 14.4 0.489 1.552 0.489 1.552 jit_kernel_multiply 8 16.8 0.468 1.471 0.468 1.471 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.001 0.001 1.457 1.469 build_core_hamiltonian_matrix_ 11 4.9 0.000 0.001 1.356 1.455 copy_fm_to_dbcsr 174 11.2 0.001 0.001 1.007 1.451 mp_waitany 17072 13.8 1.152 1.321 1.152 1.321 qs_ot_get_orbitals 99 10.5 0.001 0.001 1.282 1.289 acc_transpose_blocks 16440 15.4 0.073 0.075 1.222 1.235 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="204", plot="h2o_128_md", label="(8n/4r/3t)", y=59.897000, yerr=0.000000 PlotPoint: name="205", plot="h2o_128_md_mem", label="(8n/4r/3t)", y=625.454545, yerr=8.348801 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/12/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 32 x 32 x 32 184415158272 0.0% 0.0% 100.0% flops 9 x 9 x 32 269180485632 0.0% 0.0% 100.0% flops 9 x 22 x 32 349395425280 0.0% 0.0% 100.0% flops 22 x 9 x 32 350042406912 0.0% 0.0% 100.0% flops 22 x 22 x 32 453581815808 0.0% 0.0% 100.0% flops 32 x 32 x 9 465064427520 0.0% 0.0% 100.0% flops 32 x 32 x 22 568412078080 0.0% 0.0% 100.0% flops 9 x 32 x 32 572195340288 0.0% 0.0% 100.0% flops 22 x 32 x 32 699349860352 0.0% 0.0% 100.0% flops 9 x 32 x 9 1735942275072 0.0% 0.0% 100.0% flops 22 x 32 x 9 2216407818240 0.0% 0.0% 100.0% flops 9 x 32 x 22 2216407818240 0.0% 0.0% 100.0% flops 22 x 32 x 22 2803661053952 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 12.884056E+12 0.0% 0.0% 100.0% flops max/rank 601.317074E+09 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 984178160 0.0% 0.0% 100.0% number of processed stacks 4916280 0.0% 0.0% 100.0% average stack size 0.0 0.0 200.2 marketing flops 15.646302E+12 ------------------------------------------------------------------------------- # multiplications 2055 max memory usage/rank 735.399936E+06 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 937080 MPI messages size (bytes): total size 523.723932E+09 min size 0.000000E+00 max size 4.537280E+06 average size 558.889250E+03 MPI breakdown and total messages size (bytes): size <= 128 6996 0 128 < size <= 8192 264 2162688 8192 < size <= 32768 304932 8165326848 32768 < size <= 131072 110640 6338641920 131072 < size <= 4194304 489498 400769458320 4194304 < size <= 16777216 24750 108449092400 16777216 < size 0 0 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 3521 65576. MP_Allreduce 9838 600. MP_Sync 100 MP_Alltoall 1496 5863162. MP_ISendRecv 10120 43184. MP_Wait 25102 MP_comm_split 48 MP_ISend 13376 163145. MP_IRecv 13376 163145. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.042 0.134 65.089 65.091 qs_mol_dyn_low 1 2.0 0.003 0.003 64.689 64.698 qs_forces 11 3.9 0.002 0.003 64.621 64.622 qs_energies 11 4.9 0.002 0.005 61.261 61.267 scf_env_do_scf 11 5.9 0.001 0.002 52.927 52.930 scf_env_do_scf_inner_loop 99 6.5 0.002 0.006 41.059 41.060 velocity_verlet 10 3.0 0.001 0.001 36.842 36.844 dbcsr_multiply_generic 2055 12.4 0.117 0.121 29.632 29.845 qs_scf_new_mos 99 7.5 0.001 0.001 26.677 26.782 qs_scf_loop_do_ot 99 8.5 0.001 0.002 26.676 26.781 ot_scf_mini 99 9.5 0.003 0.003 25.047 25.148 multiply_cannon 2055 13.4 0.245 0.263 22.489 23.358 multiply_cannon_loop 2055 14.4 0.883 0.908 21.018 21.637 ot_mini 99 10.5 0.001 0.001 14.181 14.296 multiply_cannon_multrec 24660 15.4 4.214 6.764 12.697 13.837 rebuild_ks_matrix 110 8.3 0.000 0.000 11.935 12.041 qs_ks_build_kohn_sham_matrix 110 9.3 0.012 0.013 11.935 12.040 init_scf_loop 11 6.9 0.000 0.001 11.826 11.827 qs_ks_update_qs_env 110 7.6 0.001 0.001 10.575 10.668 qs_ot_get_derivative 99 11.5 0.001 0.001 10.036 10.129 prepare_preconditioner 11 7.9 0.000 0.000 10.091 10.107 make_preconditioner 11 8.9 0.000 0.003 10.091 10.107 make_full_inverse_cholesky 11 9.9 0.000 0.000 8.324 9.779 dbcsr_mm_accdrv_process 52304 16.0 6.910 8.319 8.337 9.267 mp_waitall_1 126806 16.4 4.573 6.469 4.573 6.469 sum_up_and_integrate 110 10.3 0.068 0.071 6.338 6.353 integrate_v_rspace 110 11.3 0.003 0.003 6.270 6.284 qs_ot_get_p 110 10.4 0.001 0.001 5.816 5.959 make_m2s 4110 13.4 0.059 0.062 5.495 5.812 make_images 4110 14.4 0.576 0.700 5.355 5.669 qs_rho_update_rho_low 110 7.6 0.001 0.001 5.577 5.587 calculate_rho_elec 110 8.6 0.077 0.081 5.577 5.587 init_scf_run 11 5.9 0.000 0.001 5.550 5.551 scf_env_initial_rho_setup 11 6.9 0.001 0.002 5.550 5.551 cp_fm_upper_to_full 70 13.8 3.307 4.690 3.307 4.690 ot_diis_step 99 11.5 0.011 0.011 4.105 4.106 apply_preconditioner_dbcsr 110 12.6 0.000 0.000 4.004 4.103 apply_single 110 13.6 0.000 0.000 4.003 4.103 qs_ot_p2m_diag 48 11.0 0.054 0.063 3.952 3.966 dbcsr_complete_redistribute 325 12.2 0.415 0.458 2.684 3.829 cp_dbcsr_syevd 48 12.0 0.003 0.003 3.470 3.470 grid_integrate_task_list 110 12.3 3.287 3.466 3.287 3.466 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 3.317 3.365 multiply_cannon_sync_h2d 24660 15.4 3.171 3.335 3.171 3.335 copy_fm_to_dbcsr 174 11.2 0.001 0.002 2.149 3.276 pw_transfer 1331 11.6 0.064 0.072 3.215 3.244 multiply_cannon_metrocomm3 24660 15.4 0.035 0.036 1.426 3.201 fft_wrap_pw1pw2 1111 12.6 0.008 0.008 3.110 3.143 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 3.086 3.136 make_images_data 4110 15.4 0.046 0.050 2.766 3.064 hybrid_alltoall_any 4261 16.3 0.119 0.457 2.306 3.030 calculate_dm_sparse 110 9.5 0.001 0.001 2.950 2.985 wfi_extrapolate 11 7.9 0.001 0.001 2.967 2.967 cp_fm_diag_elpa 48 13.0 0.000 0.000 2.896 2.898 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 2.866 2.867 cp_fm_redistribute_end 48 14.0 1.443 2.862 1.444 2.863 transfer_fm_to_dbcsr 11 9.9 0.000 0.000 1.758 2.862 cp_fm_cholesky_invert 11 10.9 2.849 2.858 2.849 2.858 density_rs2pw 110 9.6 0.004 0.004 2.767 2.852 cp_fm_diag_elpa_base 48 14.0 1.335 2.725 1.416 2.837 mp_alltoall_i22 605 13.7 1.655 2.829 1.655 2.829 fft_wrap_pw1pw2_140 451 13.1 0.202 0.212 2.616 2.653 rs_pw_transfer 902 11.9 0.010 0.011 2.502 2.598 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 2.478 2.519 calculate_first_density_matrix 1 7.0 0.000 0.002 2.495 2.498 fft3d_ps 1111 14.6 1.062 1.094 2.412 2.430 grid_collocate_task_list 110 9.6 2.224 2.405 2.224 2.405 qs_energies_init_hamiltonians 11 5.9 0.001 0.003 2.274 2.275 jit_kernel_multiply 12 15.8 1.098 2.245 1.098 2.245 potential_pw2rs 110 12.3 0.012 0.013 2.132 2.144 mp_alltoall_d11v 2046 13.8 1.740 2.046 1.740 2.046 cp_fm_cholesky_decompose 22 10.9 1.743 1.791 1.743 1.791 qs_ot_get_orbitals 99 10.5 0.001 0.001 1.735 1.769 build_core_hamiltonian_matrix_ 11 4.9 0.000 0.001 1.606 1.701 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.605 1.617 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.001 0.001 1.561 1.576 multiply_cannon_metrocomm4 20550 15.4 0.057 0.059 0.838 1.564 mp_sum_l 6514 12.8 0.986 1.563 0.986 1.563 acc_transpose_blocks 24660 15.4 0.104 0.106 1.511 1.529 mp_irecv_dv 62702 16.1 0.738 1.483 0.738 1.483 mp_allgather_i34 2055 14.4 0.573 1.368 0.573 1.368 qs_env_update_s_mstruct 11 6.9 0.000 0.001 1.221 1.318 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="206", plot="h2o_128_md", label="(8n/3r/4t)", y=65.091000, yerr=0.000000 PlotPoint: name="207", plot="h2o_128_md_mem", label="(8n/3r/4t)", y=694.818182, yerr=10.969530 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/13/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 32 x 32 x 32 184415158272 0.0% 0.0% 100.0% flops 9 x 9 x 32 269180485632 0.0% 0.0% 100.0% flops 9 x 22 x 32 349395425280 0.0% 0.0% 100.0% flops 22 x 9 x 32 350042406912 0.0% 0.0% 100.0% flops 22 x 22 x 32 453581815808 0.0% 0.0% 100.0% flops 32 x 32 x 9 465064427520 0.0% 0.0% 100.0% flops 32 x 32 x 22 568412078080 0.0% 0.0% 100.0% flops 9 x 32 x 32 572195340288 0.0% 0.0% 100.0% flops 22 x 32 x 32 699349860352 0.0% 0.0% 100.0% flops 9 x 32 x 9 1735942275072 0.0% 0.0% 100.0% flops 22 x 32 x 9 2216407818240 0.0% 0.0% 100.0% flops 9 x 32 x 22 2216407818240 0.0% 0.0% 100.0% flops 22 x 32 x 22 2803661053952 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 12.884056E+12 0.0% 0.0% 100.0% flops max/rank 807.299199E+09 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 984178160 0.0% 0.0% 100.0% number of processed stacks 1438408 0.0% 0.0% 100.0% average stack size 0.0 0.0 684.2 marketing flops 15.646297E+12 ------------------------------------------------------------------------------- # multiplications 2055 max memory usage/rank 824.635392E+06 # max total images/rank 1 # max 3D layers 1 # MPI messages exchanged 197280 MPI messages size (bytes): total size 339.125567E+09 min size 0.000000E+00 max size 13.107200E+06 average size 1.719006E+06 MPI breakdown and total messages size (bytes): size <= 128 1452 0 128 < size <= 8192 0 0 8192 < size <= 32768 132 4325376 32768 < size <= 131072 88656 11620319232 131072 < size <= 4194304 89424 117209825280 4194304 < size <= 16777216 17616 210291069504 16777216 < size 0 0 ------------------------------------------------------------------------------- - - - DBCSR MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Bcast 14 12. MP_Allreduce 7346 33. MP_Alltoall 8043 263767. MP_ISend 32836 654203. MP_IRecv 32836 654587. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 3521 65574. MP_Allreduce 9838 640. MP_Sync 100 MP_Alltoall 1496 8504061. MP_ISendRecv 6600 54848. MP_Wait 17226 MP_comm_split 48 MP_ISend 9240 278857. MP_IRecv 9240 278857. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.021 0.037 54.870 54.872 qs_mol_dyn_low 1 2.0 0.003 0.003 54.624 54.633 qs_forces 11 3.9 0.002 0.002 54.560 54.561 qs_energies 11 4.9 0.001 0.002 50.955 50.960 scf_env_do_scf 11 5.9 0.000 0.001 42.809 42.810 scf_env_do_scf_inner_loop 99 6.5 0.002 0.006 34.872 34.874 velocity_verlet 10 3.0 0.001 0.001 30.895 30.898 dbcsr_multiply_generic 2055 12.4 0.105 0.106 22.616 22.741 qs_scf_new_mos 99 7.5 0.001 0.001 20.721 20.764 qs_scf_loop_do_ot 99 8.5 0.001 0.001 20.721 20.764 ot_scf_mini 99 9.5 0.002 0.002 19.464 19.485 multiply_cannon 2055 13.4 0.246 0.267 17.103 18.328 multiply_cannon_loop 2055 14.4 0.323 0.336 15.736 16.034 rebuild_ks_matrix 110 8.3 0.000 0.000 11.432 11.457 qs_ks_build_kohn_sham_matrix 110 9.3 0.012 0.013 11.432 11.457 ot_mini 99 10.5 0.001 0.001 10.355 10.371 qs_ks_update_qs_env 110 7.6 0.001 0.001 10.186 10.205 multiply_cannon_multrec 8220 15.4 3.213 4.506 7.443 8.425 init_scf_loop 11 6.9 0.000 0.000 7.884 7.885 mp_waitall_1 106626 16.5 5.978 7.615 5.978 7.615 qs_ot_get_derivative 99 11.5 0.001 0.001 6.570 6.590 prepare_preconditioner 11 7.9 0.000 0.000 6.261 6.264 make_preconditioner 11 8.9 0.000 0.000 6.261 6.263 sum_up_and_integrate 110 10.3 0.079 0.081 6.210 6.222 integrate_v_rspace 110 11.3 0.003 0.003 6.130 6.141 make_full_inverse_cholesky 11 9.9 0.000 0.000 5.847 5.922 qs_rho_update_rho_low 110 7.6 0.001 0.001 5.683 5.698 calculate_rho_elec 110 8.6 0.115 0.116 5.682 5.698 qs_ot_get_p 110 10.4 0.001 0.001 5.218 5.238 dbcsr_mm_accdrv_process 17442 15.9 2.891 4.040 4.099 5.073 init_scf_run 11 5.9 0.000 0.001 5.028 5.028 scf_env_initial_rho_setup 11 6.9 0.001 0.001 5.028 5.028 make_m2s 4110 13.4 0.038 0.040 4.182 4.386 multiply_cannon_metrocomm3 8220 15.4 0.017 0.017 2.902 4.280 make_images 4110 14.4 0.633 0.687 4.053 4.255 qs_ot_p2m_diag 48 11.0 0.081 0.084 3.796 3.800 ot_diis_step 99 11.5 0.012 0.012 3.760 3.760 apply_preconditioner_dbcsr 110 12.6 0.000 0.000 3.673 3.701 apply_single 110 13.6 0.000 0.000 3.672 3.700 grid_integrate_task_list 110 12.3 3.344 3.504 3.344 3.504 cp_dbcsr_syevd 48 12.0 0.003 0.003 3.484 3.485 pw_transfer 1331 11.6 0.065 0.071 3.319 3.341 fft_wrap_pw1pw2 1111 12.6 0.008 0.008 3.211 3.237 cp_fm_cholesky_invert 11 10.9 3.083 3.086 3.083 3.086 multiply_cannon_sync_h2d 8220 15.4 2.912 3.030 2.912 3.030 cp_fm_diag_elpa 48 13.0 0.000 0.000 2.928 2.930 cp_fm_redistribute_end 48 14.0 0.743 2.896 0.748 2.897 cp_fm_diag_elpa_base 48 14.0 1.961 2.700 2.141 2.861 density_rs2pw 110 9.6 0.004 0.004 2.714 2.830 fft_wrap_pw1pw2_140 451 13.1 0.211 0.215 2.732 2.767 wfi_extrapolate 11 7.9 0.001 0.001 2.705 2.705 qs_energies_init_hamiltonians 11 5.9 0.001 0.002 2.654 2.655 make_images_data 4110 15.4 0.038 0.044 2.323 2.650 hybrid_alltoall_any 4261 16.3 0.199 0.858 2.252 2.628 calculate_dm_sparse 110 9.5 0.001 0.001 2.485 2.524 grid_collocate_task_list 110 9.6 2.326 2.516 2.326 2.516 fft3d_ps 1111 14.6 1.119 1.169 2.471 2.494 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 2.482 2.483 rs_pw_transfer 902 11.9 0.010 0.011 2.231 2.368 calculate_first_density_matrix 1 7.0 0.000 0.000 2.233 2.234 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 2.115 2.132 build_core_hamiltonian_matrix_ 11 4.9 0.000 0.001 1.777 1.985 potential_pw2rs 110 12.3 0.015 0.016 1.969 1.978 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 1.945 1.961 mp_alltoall_d11v 2046 13.8 1.706 1.895 1.706 1.895 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 1.844 1.856 cp_fm_cholesky_decompose 22 10.9 1.731 1.761 1.731 1.761 qs_env_update_s_mstruct 11 6.9 0.001 0.002 1.496 1.616 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.579 1.584 mp_allgather_i34 2055 14.4 0.552 1.578 0.552 1.578 dbcsr_complete_redistribute 325 12.2 0.552 0.576 1.481 1.577 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.001 0.001 1.436 1.445 qs_create_task_list 11 7.9 0.001 0.001 1.226 1.326 generate_qs_task_list 11 8.9 0.378 0.446 1.226 1.325 jit_kernel_multiply 7 15.7 0.898 1.276 0.898 1.276 copy_dbcsr_to_fm 151 11.3 0.003 0.003 1.164 1.186 mp_waitany 9240 13.8 1.021 1.154 1.021 1.154 mp_alltoall_z22v 1111 16.6 1.119 1.143 1.119 1.143 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="208", plot="h2o_128_md", label="(8n/2r/6t)", y=54.872000, yerr=0.000000 PlotPoint: name="209", plot="h2o_128_md_mem", label="(8n/2r/6t)", y=777.545455, yerr=12.412337 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/14/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 32 x 32 x 32 184415158272 0.0% 0.0% 100.0% flops 9 x 9 x 32 269180485632 0.0% 0.0% 100.0% flops 9 x 22 x 32 349395425280 0.0% 0.0% 100.0% flops 22 x 9 x 32 350042406912 0.0% 0.0% 100.0% flops 22 x 22 x 32 453581815808 0.0% 0.0% 100.0% flops 32 x 32 x 9 465064427520 0.0% 0.0% 100.0% flops 32 x 32 x 22 568412078080 0.0% 0.0% 100.0% flops 9 x 32 x 32 572195340288 0.0% 0.0% 100.0% flops 22 x 32 x 32 699349860352 0.0% 0.0% 100.0% flops 9 x 32 x 9 1735942275072 0.0% 0.0% 100.0% flops 22 x 32 x 9 2216407818240 0.0% 0.0% 100.0% flops 9 x 32 x 22 2216407818240 0.0% 0.0% 100.0% flops 22 x 32 x 22 2803661053952 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 12.884056E+12 0.0% 0.0% 100.0% flops max/rank 1.612391E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 984178160 0.0% 0.0% 100.0% number of processed stacks 1464624 0.0% 0.0% 100.0% average stack size 0.0 0.0 672.0 marketing flops 15.646297E+12 ------------------------------------------------------------------------------- # multiplications 2055 max memory usage/rank 1.316987E+09 # max total images/rank 2 # max 3D layers 1 # MPI messages exchanged 82200 MPI messages size (bytes): total size 297.640985E+09 min size 0.000000E+00 max size 26.214400E+06 average size 3.620936E+06 MPI breakdown and total messages size (bytes): size <= 128 572 0 128 < size <= 8192 0 0 8192 < size <= 32768 44 1441792 32768 < size <= 131072 18560 2432696320 131072 < size <= 4194304 54216 84915781632 4194304 < size <= 16777216 0 0 16777216 < size 8808 210291069504 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 3462 67104. MP_Allreduce 9672 819. MP_Sync 52 MP_Alltoall 1474 16505187. MP_ISendRecv 4620 360267. MP_Wait 7524 MP_ISend 2420 1187840. MP_IRecv 2420 1187840. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.020 0.034 87.104 87.105 qs_mol_dyn_low 1 2.0 0.003 0.003 86.804 86.812 qs_forces 11 3.9 0.002 0.002 86.720 86.722 qs_energies 11 4.9 0.002 0.002 82.686 82.689 scf_env_do_scf 11 5.9 0.000 0.001 72.796 72.797 velocity_verlet 10 3.0 0.001 0.001 55.791 55.800 scf_env_do_scf_inner_loop 99 6.5 0.002 0.006 43.765 43.768 dbcsr_multiply_generic 2055 12.4 0.121 0.123 28.838 28.966 init_scf_loop 11 6.9 0.000 0.000 28.957 28.961 qs_scf_new_mos 99 7.5 0.001 0.001 27.094 27.160 qs_scf_loop_do_ot 99 8.5 0.001 0.001 27.093 27.159 prepare_preconditioner 11 7.9 0.000 0.000 27.003 27.012 make_preconditioner 11 8.9 0.000 0.000 27.003 27.012 make_full_inverse_cholesky 11 9.9 0.000 0.000 21.050 26.464 ot_scf_mini 99 9.5 0.002 0.002 25.307 25.348 multiply_cannon 2055 13.4 0.346 0.370 21.557 22.259 multiply_cannon_loop 2055 14.4 0.342 0.346 19.689 20.054 cp_fm_upper_to_full 70 14.2 12.819 18.485 12.819 18.485 ot_mini 99 10.5 0.001 0.001 13.888 13.944 rebuild_ks_matrix 110 8.3 0.000 0.001 13.269 13.351 qs_ks_build_kohn_sham_matrix 110 9.3 0.013 0.013 13.269 13.350 qs_ks_update_qs_env 110 7.6 0.001 0.001 12.067 12.143 dbcsr_complete_redistribute 325 12.2 1.023 1.050 7.619 10.882 copy_fm_to_dbcsr 174 11.2 0.001 0.001 6.540 9.803 multiply_cannon_multrec 8220 15.4 4.338 4.536 9.544 9.636 qs_ot_get_derivative 99 11.5 0.001 0.001 9.263 9.306 transfer_fm_to_dbcsr 11 9.9 0.000 0.000 5.937 9.174 mp_waitall_1 87304 16.6 7.941 9.036 7.941 9.036 mp_alltoall_i22 605 13.7 5.598 8.874 5.598 8.874 sum_up_and_integrate 110 10.3 0.150 0.150 6.761 6.773 qs_rho_update_rho_low 110 7.6 0.001 0.001 6.646 6.683 calculate_rho_elec 110 8.6 0.227 0.227 6.646 6.683 integrate_v_rspace 110 11.3 0.004 0.004 6.610 6.623 qs_ot_get_p 110 10.4 0.001 0.001 6.192 6.258 make_m2s 4110 13.4 0.043 0.044 5.340 5.895 make_images 4110 14.4 0.881 0.930 5.153 5.707 init_scf_run 11 5.9 0.000 0.001 5.652 5.652 scf_env_initial_rho_setup 11 6.9 0.001 0.001 5.652 5.652 cp_fm_cholesky_invert 11 10.9 5.578 5.583 5.578 5.583 dbcsr_mm_accdrv_process 11614 15.7 3.259 3.587 5.063 5.328 apply_preconditioner_dbcsr 110 12.6 0.000 0.000 4.682 5.188 apply_single 110 13.6 0.000 0.000 4.681 5.188 multiply_cannon_metrocomm3 8220 15.4 0.018 0.018 4.796 5.104 ot_diis_step 99 11.5 0.015 0.016 4.591 4.591 qs_ot_p2m_diag 48 11.0 0.151 0.155 4.531 4.538 cp_dbcsr_syevd 48 12.0 0.003 0.003 4.071 4.072 multiply_cannon_sync_h2d 8220 15.4 3.945 3.952 3.945 3.952 pw_transfer 1331 11.6 0.074 0.074 3.911 3.915 fft_wrap_pw1pw2 1111 12.6 0.009 0.009 3.795 3.799 grid_integrate_task_list 110 12.3 3.669 3.742 3.669 3.742 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 3.628 3.630 make_images_data 4110 15.4 0.041 0.044 2.892 3.589 hybrid_alltoall_any 4261 16.3 0.255 0.550 2.883 3.582 cp_fm_diag_elpa 48 13.0 0.000 0.000 3.414 3.415 cp_fm_diag_elpa_base 48 14.0 2.872 3.079 3.412 3.412 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 2.937 3.402 wfi_extrapolate 11 7.9 0.001 0.001 3.271 3.271 fft_wrap_pw1pw2_140 451 13.1 0.215 0.219 3.260 3.265 calculate_dm_sparse 110 9.5 0.001 0.001 3.200 3.223 density_rs2pw 110 9.6 0.004 0.004 3.122 3.139 fft3d_ps 1111 14.6 1.271 1.280 3.004 3.010 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 2.858 2.860 grid_collocate_task_list 110 9.6 2.635 2.653 2.635 2.653 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 2.493 2.517 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 2.336 2.362 calculate_first_density_matrix 1 7.0 0.000 0.000 2.276 2.277 rs_pw_transfer 902 11.9 0.011 0.011 2.243 2.268 qs_env_update_s_mstruct 11 6.9 0.001 0.002 2.172 2.233 build_core_hamiltonian_matrix_ 11 4.9 0.001 0.001 2.087 2.177 mp_alltoall_d11v 2046 13.8 2.105 2.166 2.105 2.166 potential_pw2rs 110 12.3 0.021 0.021 2.155 2.159 cp_fm_cholesky_decompose 22 10.9 2.124 2.140 2.124 2.140 qs_create_task_list 11 7.9 0.001 0.001 1.899 1.945 generate_qs_task_list 11 8.9 0.737 0.791 1.899 1.944 copy_dbcsr_to_fm 151 11.3 0.003 0.003 1.818 1.864 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.808 1.815 jit_kernel_multiply 10 15.3 1.603 1.779 1.603 1.779 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="210", plot="h2o_128_md", label="(8n/1r/12t)", y=87.105000, yerr=0.000000 PlotPoint: name="211", plot="h2o_128_md_mem", label="(8n/1r/12t)", y=1184.363636, yerr=49.306597 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/15/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1420241154048 0.0% 0.0% 100.0% flops 32 x 32 x 32 1943472701440 0.0% 0.0% 100.0% flops 22 x 9 x 32 1972057190400 0.0% 0.0% 100.0% flops 9 x 22 x 32 1977770336256 0.0% 0.0% 100.0% flops 22 x 22 x 32 2734287699968 0.0% 0.0% 100.0% flops 32 x 32 x 9 4416300122112 0.0% 0.0% 100.0% flops 32 x 32 x 22 5397700149248 0.0% 0.0% 100.0% flops 9 x 32 x 32 5443971710976 0.0% 0.0% 100.0% flops 22 x 32 x 32 6653743202304 0.0% 0.0% 100.0% flops 9 x 32 x 9 11528896499712 0.0% 0.0% 100.0% flops 22 x 32 x 9 15129160814592 0.0% 0.0% 100.0% flops 9 x 32 x 22 15129160814592 0.0% 0.0% 100.0% flops 22 x 32 x 22 19767995056128 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 93.514757E+12 0.0% 0.0% 100.0% flops max/rank 1.094965E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 6755939872 0.0% 0.0% 100.0% number of processed stacks 11950464 0.0% 0.0% 100.0% average stack size 0.0 0.0 565.3 marketing flops 144.580175E+12 ------------------------------------------------------------------------------- # multiplications 2507 max memory usage/rank 628.576256E+06 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 10348896 MPI messages size (bytes): total size 4.491514E+12 min size 0.000000E+00 max size 4.537280E+06 average size 434.009000E+03 MPI breakdown and total messages size (bytes): size <= 128 65736 0 128 < size <= 8192 1232 10092544 8192 < size <= 32768 3576680 95640223744 32768 < size <= 131072 1294784 74079797248 131072 < size <= 4194304 5148576 3175955098256 4194304 < size <= 16777216 261888 1145794321408 16777216 < size 0 0 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 4075 56898. MP_Allreduce 11228 786. MP_Sync 170 MP_Alltoall 2226 2368995. MP_ISendRecv 48640 18752. MP_Wait 66796 MP_comm_split 83 MP_ISend 16020 108028. MP_IRecv 16020 108028. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.016 0.034 205.610 205.612 qs_mol_dyn_low 1 2.0 0.008 0.042 205.107 205.121 qs_forces 11 3.9 0.003 0.004 204.976 204.977 qs_energies 11 4.9 0.003 0.004 199.461 199.475 scf_env_do_scf 11 5.9 0.001 0.001 181.921 181.925 scf_env_do_scf_inner_loop 117 6.6 0.003 0.008 160.868 160.870 dbcsr_multiply_generic 2507 12.6 0.176 0.179 125.563 126.432 velocity_verlet 10 3.0 0.001 0.001 123.378 123.379 qs_scf_new_mos 117 7.6 0.001 0.001 121.835 122.089 qs_scf_loop_do_ot 117 8.6 0.001 0.001 121.834 122.088 ot_scf_mini 117 9.6 0.003 0.003 115.256 115.483 multiply_cannon 2507 13.6 0.238 0.246 101.630 104.101 multiply_cannon_loop 2507 14.6 2.119 2.199 99.378 101.928 ot_mini 117 10.6 0.001 0.001 66.285 66.510 multiply_cannon_multrec 60168 15.6 33.060 36.625 41.481 44.078 qs_ot_get_derivative 117 11.6 0.001 0.001 41.328 41.548 rebuild_ks_matrix 128 8.3 0.001 0.001 33.478 34.025 qs_ks_build_kohn_sham_matrix 128 9.3 0.015 0.017 33.478 34.025 mp_waitall_1 291448 16.2 29.952 32.374 29.952 32.374 qs_ks_update_qs_env 128 7.6 0.001 0.001 30.131 30.619 multiply_cannon_sync_h2d 60168 15.6 27.220 30.200 27.220 30.200 qs_ot_get_p 128 10.4 0.001 0.001 26.865 27.207 apply_preconditioner_dbcsr 128 12.6 0.000 0.001 24.326 25.346 apply_single 128 13.6 0.001 0.001 24.326 25.346 ot_diis_step 117 11.6 0.007 0.008 24.572 24.573 init_scf_loop 11 6.9 0.000 0.000 20.975 20.976 qs_ot_p2m_diag 83 11.4 0.077 0.091 20.095 20.181 qs_ot_get_derivative_diag 77 12.4 0.002 0.002 19.313 19.478 multiply_cannon_metrocomm3 60168 15.6 0.110 0.116 16.166 19.365 cp_dbcsr_syevd 83 12.4 0.004 0.005 17.520 17.522 prepare_preconditioner 11 7.9 0.000 0.000 16.362 16.417 make_preconditioner 11 8.9 0.000 0.000 16.362 16.417 make_full_inverse_cholesky 11 9.9 0.000 0.000 15.588 15.743 make_m2s 5014 13.6 0.103 0.111 14.155 14.671 cp_fm_diag_elpa 83 13.4 0.000 0.001 14.614 14.620 cp_fm_redistribute_end 83 14.4 11.530 14.527 11.542 14.530 make_images 5014 14.6 0.403 0.422 13.974 14.499 cp_fm_diag_elpa_base 83 14.4 2.946 14.217 2.977 14.339 sum_up_and_integrate 128 10.3 0.090 0.108 14.147 14.164 integrate_v_rspace 128 11.3 0.003 0.004 14.057 14.077 init_scf_run 11 5.9 0.000 0.001 13.272 13.272 scf_env_initial_rho_setup 11 6.9 0.012 0.028 13.271 13.272 qs_rho_update_rho_low 128 7.7 0.001 0.001 12.774 12.849 calculate_rho_elec 128 8.7 0.045 0.064 12.773 12.848 mp_sum_l 7870 13.0 8.517 9.912 8.517 9.912 cp_fm_cholesky_invert 11 10.9 9.372 9.381 9.372 9.381 wfi_extrapolate 11 7.9 0.001 0.001 9.090 9.090 dbcsr_mm_accdrv_process 124484 16.2 3.352 3.515 7.984 8.957 multiply_cannon_metrocomm1 60168 15.6 0.091 0.097 6.431 8.886 calculate_dm_sparse 128 9.5 0.001 0.001 8.392 8.502 qs_ot_get_derivative_taylor 40 13.0 0.001 0.001 8.118 8.269 qs_ot_get_orbitals 117 10.6 0.001 0.001 7.953 8.074 make_images_data 5014 15.6 0.070 0.077 6.988 7.914 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.003 7.584 7.593 grid_integrate_task_list 128 12.3 7.072 7.586 7.072 7.586 hybrid_alltoall_any 5200 16.5 0.289 2.248 6.080 7.545 pw_transfer 1547 11.6 0.074 0.107 6.921 7.190 density_rs2pw 128 9.7 0.006 0.007 6.588 7.076 fft_wrap_pw1pw2 1291 12.7 0.010 0.013 6.716 6.959 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 6.567 6.695 rs_pw_transfer 1046 11.9 0.017 0.019 5.876 6.377 fft_wrap_pw1pw2_140 523 13.2 0.443 0.514 5.716 5.897 mp_alltoall_d11v 2415 14.1 4.331 5.706 4.331 5.706 fft3d_ps 1291 14.7 2.092 2.563 5.510 5.700 grid_collocate_task_list 128 9.7 4.722 5.070 4.722 5.070 cp_fm_cholesky_decompose 22 10.9 4.788 4.802 4.788 4.802 potential_pw2rs 128 12.3 0.010 0.011 4.579 4.614 mp_sum_d 4459 12.1 3.673 4.435 3.673 4.435 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="400", plot="h2o_256_md", label="(8n/12r/1t)", y=205.612000, yerr=0.000000 PlotPoint: name="401", plot="h2o_256_md_mem", label="(8n/12r/1t)", y=594.818182, yerr=7.309000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/16/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1420242647040 0.0% 0.0% 100.0% flops 32 x 32 x 32 1943472701440 0.0% 0.0% 100.0% flops 22 x 9 x 32 1972057190400 0.0% 0.0% 100.0% flops 9 x 22 x 32 1977770336256 0.0% 0.0% 100.0% flops 22 x 22 x 32 2734287699968 0.0% 0.0% 100.0% flops 32 x 32 x 9 4416300122112 0.0% 0.0% 100.0% flops 32 x 32 x 22 5397700149248 0.0% 0.0% 100.0% flops 9 x 32 x 32 5443971710976 0.0% 0.0% 100.0% flops 22 x 32 x 32 6653743202304 0.0% 0.0% 100.0% flops 9 x 32 x 9 11528903135232 0.0% 0.0% 100.0% flops 22 x 32 x 9 15129160814592 0.0% 0.0% 100.0% flops 9 x 32 x 22 15129160814592 0.0% 0.0% 100.0% flops 22 x 32 x 22 19767995056128 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 93.514766E+12 0.0% 0.0% 100.0% flops max/rank 2.183246E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 6755941440 0.0% 0.0% 100.0% number of processed stacks 5975232 0.0% 0.0% 100.0% average stack size 0.0 0.0 1130.7 marketing flops 144.580175E+12 ------------------------------------------------------------------------------- # multiplications 2507 max memory usage/rank 827.416576E+06 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 2406720 MPI messages size (bytes): total size 4.100943E+12 min size 0.000000E+00 max size 17.653760E+06 average size 1.703955E+06 MPI breakdown and total messages size (bytes): size <= 128 14916 0 128 < size <= 8192 0 0 8192 < size <= 32768 70860 2317615104 32768 < size <= 131072 722992 55511613440 131072 < size <= 4194304 1375664 1398181724160 4194304 < size <= 16777216 154704 1463834845264 16777216 < size 67584 1181116006400 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 4085 57203. MP_Allreduce 11252 946. MP_Sync 170 MP_Alltoall 1969 5063619. MP_ISendRecv 24064 47072. MP_Wait 37948 MP_comm_split 83 MP_ISend 11748 212467. MP_IRecv 11748 212467. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.023 0.053 193.211 193.213 qs_mol_dyn_low 1 2.0 0.003 0.005 192.756 192.769 qs_forces 11 3.9 0.003 0.005 192.665 192.667 qs_energies 11 4.9 0.003 0.008 185.952 185.963 scf_env_do_scf 11 5.9 0.001 0.001 169.337 169.347 scf_env_do_scf_inner_loop 117 6.6 0.003 0.008 135.745 135.748 velocity_verlet 10 3.0 0.001 0.001 121.152 121.154 dbcsr_multiply_generic 2507 12.6 0.186 0.192 97.861 98.943 qs_scf_new_mos 117 7.6 0.001 0.001 97.214 97.516 qs_scf_loop_do_ot 117 8.6 0.001 0.002 97.214 97.516 ot_scf_mini 117 9.6 0.004 0.005 92.393 92.746 multiply_cannon 2507 13.6 0.474 0.531 77.781 82.189 multiply_cannon_loop 2507 14.6 1.262 1.297 74.428 77.071 ot_mini 117 10.6 0.001 0.001 50.025 50.396 mp_waitall_1 226760 16.4 24.788 38.588 24.788 38.588 multiply_cannon_multrec 30084 15.6 22.104 26.959 31.976 37.005 init_scf_loop 11 6.9 0.001 0.005 33.495 33.496 rebuild_ks_matrix 128 8.3 0.001 0.001 32.782 33.250 qs_ks_build_kohn_sham_matrix 128 9.3 0.017 0.021 32.782 33.249 qs_ks_update_qs_env 128 7.6 0.001 0.001 29.566 29.994 prepare_preconditioner 11 7.9 0.000 0.000 29.069 29.115 make_preconditioner 11 8.9 0.000 0.002 29.069 29.115 qs_ot_get_derivative 117 11.6 0.001 0.002 28.062 28.392 make_full_inverse_cholesky 11 9.9 0.000 0.000 27.773 28.317 multiply_cannon_metrocomm3 30084 15.6 0.095 0.100 15.579 27.867 qs_ot_get_p 128 10.4 0.001 0.001 24.041 24.561 apply_preconditioner_dbcsr 128 12.6 0.000 0.000 22.019 23.048 apply_single 128 13.6 0.001 0.001 22.019 23.047 ot_diis_step 117 11.6 0.014 0.015 21.781 21.784 multiply_cannon_sync_h2d 30084 15.6 19.419 21.628 19.419 21.628 qs_ot_p2m_diag 83 11.4 0.188 0.216 19.155 19.194 cp_dbcsr_syevd 83 12.4 0.005 0.006 17.958 17.959 cp_fm_cholesky_invert 11 10.9 17.008 17.021 17.008 17.021 make_m2s 5014 13.6 0.088 0.095 14.271 15.846 make_images 5014 14.6 1.147 1.346 14.064 15.639 cp_fm_diag_elpa 83 13.4 0.000 0.001 14.703 14.712 cp_fm_redistribute_end 83 14.4 8.597 14.615 8.611 14.619 cp_fm_diag_elpa_base 83 14.4 5.767 14.090 5.988 14.495 sum_up_and_integrate 128 10.3 0.116 0.133 14.399 14.426 integrate_v_rspace 128 11.3 0.003 0.004 14.283 14.314 qs_rho_update_rho_low 128 7.7 0.001 0.001 12.866 12.897 calculate_rho_elec 128 8.7 0.088 0.105 12.865 12.896 init_scf_run 11 5.9 0.000 0.001 11.770 11.771 scf_env_initial_rho_setup 11 6.9 0.001 0.002 11.770 11.771 qs_ot_get_derivative_diag 77 12.4 0.002 0.003 11.053 11.308 multiply_cannon_metrocomm4 27577 15.6 0.097 0.112 3.733 10.475 make_images_data 5014 15.6 0.066 0.073 8.542 10.346 mp_irecv_dv 69486 16.3 3.538 10.098 3.538 10.098 dbcsr_mm_accdrv_process 62242 16.2 4.727 5.355 9.332 9.920 hybrid_alltoall_any 5200 16.5 0.341 1.508 7.261 9.592 wfi_extrapolate 11 7.9 0.001 0.001 8.462 8.462 pw_transfer 1547 11.6 0.085 0.106 7.650 7.714 grid_integrate_task_list 128 12.3 7.189 7.533 7.189 7.533 fft_wrap_pw1pw2 1291 12.7 0.010 0.011 7.427 7.489 cp_fm_cholesky_decompose 22 10.9 7.204 7.302 7.204 7.302 density_rs2pw 128 9.7 0.006 0.006 6.725 7.154 qs_ot_get_derivative_taylor 40 13.0 0.001 0.001 6.264 7.017 calculate_dm_sparse 128 9.5 0.001 0.001 6.625 6.752 fft_wrap_pw1pw2_140 523 13.2 0.468 0.514 6.362 6.427 rs_pw_transfer 1046 11.9 0.015 0.017 5.769 6.225 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 6.216 6.224 fft3d_ps 1291 14.7 2.772 2.914 5.803 5.835 mp_sum_l 7870 13.0 4.035 5.724 4.035 5.724 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 5.360 5.505 qs_ot_get_orbitals 117 10.6 0.001 0.001 5.406 5.476 grid_collocate_task_list 128 9.7 4.912 5.275 4.912 5.275 potential_pw2rs 128 12.3 0.015 0.018 4.908 4.929 mp_allgather_i34 2507 14.6 1.927 4.905 1.927 4.905 mp_alltoall_d11v 2415 14.1 4.013 4.725 4.013 4.725 mp_sum_d 4464 12.1 2.704 4.062 2.704 4.062 dbcsr_complete_redistribute 395 12.7 0.765 0.859 3.209 4.057 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="402", plot="h2o_256_md", label="(8n/6r/2t)", y=193.213000, yerr=0.000000 PlotPoint: name="403", plot="h2o_256_md_mem", label="(8n/6r/2t)", y=788.363636, yerr=3.082877 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/17/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1420242647040 0.0% 0.0% 100.0% flops 32 x 32 x 32 1943472701440 0.0% 0.0% 100.0% flops 22 x 9 x 32 1972057190400 0.0% 0.0% 100.0% flops 9 x 22 x 32 1977770336256 0.0% 0.0% 100.0% flops 22 x 22 x 32 2734287699968 0.0% 0.0% 100.0% flops 32 x 32 x 9 4416300122112 0.0% 0.0% 100.0% flops 32 x 32 x 22 5397700149248 0.0% 0.0% 100.0% flops 9 x 32 x 32 5443971710976 0.0% 0.0% 100.0% flops 22 x 32 x 32 6653743202304 0.0% 0.0% 100.0% flops 9 x 32 x 9 11528903135232 0.0% 0.0% 100.0% flops 22 x 32 x 9 15129160814592 0.0% 0.0% 100.0% flops 9 x 32 x 22 15129160814592 0.0% 0.0% 100.0% flops 22 x 32 x 22 19767995056128 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 93.514766E+12 0.0% 0.0% 100.0% flops max/rank 2.928533E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 6755941440 0.0% 0.0% 100.0% number of processed stacks 3984192 0.0% 0.0% 100.0% average stack size 0.0 0.0 1695.7 marketing flops 144.579337E+12 ------------------------------------------------------------------------------- # multiplications 2507 max memory usage/rank 942.841856E+06 # max total images/rank 2 # max 3D layers 1 # MPI messages exchanged 1042912 MPI messages size (bytes): total size 2.716210E+12 min size 0.000000E+00 max size 26.214400E+06 average size 2.604448E+06 MPI breakdown and total messages size (bytes): size <= 128 6424 0 128 < size <= 8192 0 0 8192 < size <= 32768 264 8650752 32768 < size <= 131072 281856 36943429632 131072 < size <= 4194304 660064 996105256960 4194304 < size <= 16777216 65632 931531265168 16777216 < size 28672 751619276800 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 4103 56951. MP_Allreduce 11296 984. MP_Sync 170 MP_Alltoall 1712 9388896. MP_ISendRecv 15872 75008. MP_Wait 29756 MP_comm_split 83 MP_ISend 11748 275205. MP_IRecv 11748 275205. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.024 0.043 176.414 176.418 qs_mol_dyn_low 1 2.0 0.003 0.003 175.944 175.958 qs_forces 11 3.9 0.003 0.004 175.845 175.853 qs_energies 11 4.9 0.004 0.014 169.353 169.364 scf_env_do_scf 11 5.9 0.001 0.002 153.851 153.851 scf_env_do_scf_inner_loop 117 6.6 0.003 0.010 118.312 118.314 velocity_verlet 10 3.0 0.001 0.001 113.022 113.024 dbcsr_multiply_generic 2507 12.6 0.183 0.188 81.848 82.902 qs_scf_new_mos 117 7.6 0.001 0.001 81.841 82.218 qs_scf_loop_do_ot 117 8.6 0.001 0.001 81.841 82.217 ot_scf_mini 117 9.6 0.005 0.008 77.677 78.066 multiply_cannon 2507 13.6 0.497 0.540 62.028 66.165 multiply_cannon_loop 2507 14.6 0.863 0.891 58.898 61.330 ot_mini 117 10.6 0.001 0.001 42.796 43.191 init_scf_loop 11 6.9 0.001 0.008 35.438 35.439 mp_waitall_1 178456 16.5 25.655 34.942 25.655 34.942 prepare_preconditioner 11 7.9 0.000 0.000 31.442 31.485 make_preconditioner 11 8.9 0.000 0.001 31.442 31.485 rebuild_ks_matrix 128 8.3 0.001 0.001 29.984 30.469 qs_ks_build_kohn_sham_matrix 128 9.3 0.016 0.018 29.984 30.468 make_full_inverse_cholesky 11 9.9 0.000 0.000 29.035 30.443 qs_ks_update_qs_env 128 7.6 0.001 0.001 27.088 27.535 multiply_cannon_metrocomm3 20056 15.6 0.058 0.061 15.645 25.244 multiply_cannon_multrec 20056 15.6 13.394 16.490 22.126 25.110 qs_ot_get_derivative 117 11.6 0.001 0.002 22.891 23.289 apply_preconditioner_dbcsr 128 12.6 0.000 0.000 20.021 20.933 apply_single 128 13.6 0.001 0.001 20.021 20.933 qs_ot_get_p 128 10.4 0.001 0.001 20.194 20.672 ot_diis_step 117 11.6 0.018 0.018 19.795 19.796 qs_ot_p2m_diag 83 11.4 0.266 0.272 16.014 16.023 multiply_cannon_sync_h2d 20056 15.6 14.184 16.022 14.184 16.022 make_m2s 5014 13.6 0.081 0.085 14.753 15.538 make_images 5014 14.6 1.175 1.266 14.520 15.309 cp_dbcsr_syevd 83 12.4 0.005 0.005 14.959 14.960 cp_fm_cholesky_invert 11 10.9 14.812 14.821 14.812 14.821 sum_up_and_integrate 128 10.3 0.134 0.145 14.147 14.171 integrate_v_rspace 128 11.3 0.003 0.004 14.013 14.039 qs_rho_update_rho_low 128 7.7 0.001 0.001 13.023 13.060 calculate_rho_elec 128 8.7 0.132 0.146 13.023 13.059 cp_fm_diag_elpa 83 13.4 0.000 0.001 11.752 11.754 cp_fm_redistribute_end 83 14.4 4.451 11.689 4.465 11.691 cp_fm_diag_elpa_base 83 14.4 6.801 11.103 7.207 11.595 init_scf_run 11 5.9 0.000 0.001 10.463 10.464 scf_env_initial_rho_setup 11 6.9 0.001 0.002 10.463 10.464 make_images_data 5014 15.6 0.061 0.067 9.058 10.328 hybrid_alltoall_any 5200 16.5 0.431 1.977 7.889 9.370 multiply_cannon_metrocomm4 17549 15.6 0.063 0.073 3.501 9.335 qs_ot_get_derivative_diag 77 12.4 0.002 0.002 9.008 9.289 mp_irecv_dv 50230 16.2 3.376 9.088 3.376 9.088 dbcsr_mm_accdrv_process 41502 16.2 4.487 5.336 8.191 8.315 grid_integrate_task_list 128 12.3 7.240 7.911 7.240 7.911 pw_transfer 1547 11.6 0.084 0.104 7.743 7.846 fft_wrap_pw1pw2 1291 12.7 0.010 0.011 7.520 7.629 cp_fm_cholesky_decompose 22 10.9 7.580 7.613 7.580 7.613 wfi_extrapolate 11 7.9 0.001 0.001 7.370 7.370 cp_fm_upper_to_full 105 14.5 5.583 7.344 5.583 7.344 density_rs2pw 128 9.7 0.006 0.006 6.522 6.956 fft_wrap_pw1pw2_140 523 13.2 0.478 0.530 6.585 6.701 dbcsr_complete_redistribute 395 12.7 1.165 1.189 4.709 6.423 calculate_dm_sparse 128 9.5 0.001 0.001 5.780 5.854 fft3d_ps 1291 14.7 2.694 2.901 5.789 5.847 rs_pw_transfer 1046 11.9 0.014 0.014 5.148 5.602 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 5.523 5.530 mp_alltoall_d11v 2415 14.1 4.358 5.477 4.358 5.477 grid_collocate_task_list 128 9.7 5.074 5.443 5.074 5.443 qs_ot_get_derivative_taylor 40 13.0 0.001 0.001 4.587 5.242 copy_fm_to_dbcsr 209 11.7 0.002 0.002 3.491 5.208 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 4.581 4.736 mp_sum_l 7870 13.0 3.261 4.736 3.261 4.736 mp_allgather_i34 2507 14.6 1.653 4.572 1.653 4.572 potential_pw2rs 128 12.3 0.020 0.022 4.530 4.552 qs_ot_get_orbitals 117 10.6 0.001 0.001 4.084 4.125 transfer_fm_to_dbcsr 11 9.9 0.000 0.000 2.387 4.083 mp_alltoall_i22 716 14.1 1.981 3.850 1.981 3.850 qs_energies_init_hamiltonians 11 5.9 0.001 0.002 3.751 3.752 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="404", plot="h2o_256_md", label="(8n/4r/3t)", y=176.418000, yerr=0.000000 PlotPoint: name="405", plot="h2o_256_md_mem", label="(8n/4r/3t)", y=888.090909, yerr=12.071413 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/18/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1430456039424 0.0% 0.0% 100.0% flops 32 x 32 x 32 1962800054272 0.0% 0.0% 100.0% flops 22 x 9 x 32 1986255912960 0.0% 0.0% 100.0% flops 9 x 22 x 32 1992003932160 0.0% 0.0% 100.0% flops 22 x 22 x 32 2753958699008 0.0% 0.0% 100.0% flops 32 x 32 x 9 4454954827776 0.0% 0.0% 100.0% flops 32 x 32 x 22 5444944789504 0.0% 0.0% 100.0% flops 9 x 32 x 32 5492290093056 0.0% 0.0% 100.0% flops 22 x 32 x 32 6712799002624 0.0% 0.0% 100.0% flops 9 x 32 x 9 11613072052224 0.0% 0.0% 100.0% flops 22 x 32 x 9 15239176077312 0.0% 0.0% 100.0% flops 9 x 32 x 22 15239176077312 0.0% 0.0% 100.0% flops 22 x 32 x 22 19911132921856 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 94.233020E+12 0.0% 0.0% 100.0% flops max/rank 4.387242E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 6806383904 0.0% 0.0% 100.0% number of processed stacks 6026880 0.0% 0.0% 100.0% average stack size 0.0 0.0 1129.3 marketing flops 145.651870E+12 ------------------------------------------------------------------------------- # multiplications 2529 max memory usage/rank 1.139020E+09 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 1153224 MPI messages size (bytes): total size 2.039489E+12 min size 0.000000E+00 max size 17.653760E+06 average size 1.768511E+06 MPI breakdown and total messages size (bytes): size <= 128 6996 0 128 < size <= 8192 0 0 8192 < size <= 32768 396 8650752 32768 < size <= 131072 322096 36390305792 131072 < size <= 4194304 721976 792118951936 4194304 < size <= 16777216 70800 669922227920 16777216 < size 30960 541065216000 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 4127 56626. MP_Allreduce 11356 1063. MP_Sync 172 MP_Alltoall 1724 12509604. MP_ISendRecv 11868 75008. MP_Wait 28546 MP_comm_split 84 MP_ISend 15064 244788. MP_IRecv 15064 244788. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.027 0.044 188.503 188.505 qs_mol_dyn_low 1 2.0 0.003 0.004 188.070 188.083 qs_forces 11 3.9 0.003 0.005 187.958 187.966 qs_energies 11 4.9 0.003 0.007 180.936 180.945 scf_env_do_scf 11 5.9 0.001 0.001 164.406 164.417 velocity_verlet 10 3.0 0.001 0.001 125.185 125.189 scf_env_do_scf_inner_loop 118 6.6 0.004 0.010 117.338 117.340 qs_scf_new_mos 118 7.6 0.001 0.001 81.529 81.822 qs_scf_loop_do_ot 118 8.6 0.001 0.002 81.528 81.821 dbcsr_multiply_generic 2529 12.6 0.194 0.199 79.758 80.417 ot_scf_mini 118 9.6 0.004 0.004 77.015 77.319 multiply_cannon 2529 13.6 0.562 0.596 55.135 58.644 multiply_cannon_loop 2529 14.6 1.190 1.224 51.330 52.856 init_scf_loop 11 6.9 0.002 0.005 46.946 46.954 ot_mini 118 10.6 0.001 0.001 42.631 42.932 prepare_preconditioner 11 7.9 0.000 0.000 42.882 42.905 make_preconditioner 11 8.9 0.000 0.001 42.882 42.905 make_full_inverse_cholesky 11 9.9 0.000 0.000 36.451 41.509 multiply_cannon_multrec 30348 15.6 14.037 18.911 26.044 30.709 rebuild_ks_matrix 129 8.3 0.001 0.001 29.206 29.375 qs_ks_build_kohn_sham_matrix 129 9.3 0.017 0.022 29.205 29.374 mp_waitall_1 155106 16.5 17.806 28.031 17.806 28.031 qs_ks_update_qs_env 129 7.6 0.001 0.001 26.416 26.569 qs_ot_get_derivative 118 11.6 0.001 0.002 22.860 23.177 make_m2s 5058 13.6 0.097 0.102 20.224 21.336 make_images 5058 14.6 1.937 2.123 19.916 21.028 qs_ot_get_p 129 10.4 0.001 0.002 20.560 20.900 apply_preconditioner_dbcsr 129 12.6 0.000 0.001 19.219 19.670 apply_single 129 13.6 0.001 0.001 19.219 19.669 ot_diis_step 118 11.6 0.018 0.019 19.640 19.642 cp_fm_upper_to_full 106 14.7 11.276 16.664 11.276 16.664 cp_fm_cholesky_invert 11 10.9 16.489 16.498 16.489 16.498 qs_ot_p2m_diag 84 11.4 0.347 0.394 16.431 16.483 multiply_cannon_metrocomm3 30348 15.6 0.047 0.050 6.668 15.716 cp_dbcsr_syevd 84 12.4 0.005 0.005 15.137 15.138 sum_up_and_integrate 129 10.3 0.142 0.156 14.151 14.181 integrate_v_rspace 129 11.3 0.003 0.004 14.008 14.044 qs_rho_update_rho_low 129 7.7 0.001 0.001 13.204 13.237 calculate_rho_elec 129 8.7 0.176 0.192 13.204 13.237 dbcsr_complete_redistribute 397 12.7 1.506 1.624 9.042 12.851 multiply_cannon_sync_h2d 30348 15.6 11.832 12.839 11.832 12.839 make_images_data 5058 15.6 0.064 0.068 10.907 12.657 dbcsr_mm_accdrv_process 62780 16.2 7.510 8.405 11.580 12.074 cp_fm_diag_elpa 84 13.4 0.000 0.000 11.970 11.973 cp_fm_redistribute_end 84 14.4 2.062 11.892 2.079 11.897 cp_fm_diag_elpa_base 84 14.4 9.190 11.271 9.785 11.755 hybrid_alltoall_any 5245 16.5 0.521 2.177 9.763 11.641 copy_fm_to_dbcsr 210 11.7 0.002 0.002 7.635 11.418 init_scf_run 11 5.9 0.000 0.001 10.574 10.575 scf_env_initial_rho_setup 11 6.9 0.001 0.002 10.574 10.575 transfer_fm_to_dbcsr 11 9.9 0.000 0.000 6.407 10.087 qs_ot_get_derivative_diag 78 12.4 0.002 0.002 9.499 9.725 mp_alltoall_i22 720 14.1 5.657 9.496 5.657 9.496 pw_transfer 1559 11.6 0.085 0.102 7.966 8.064 grid_integrate_task_list 129 12.3 7.575 8.056 7.575 8.056 fft_wrap_pw1pw2 1301 12.7 0.010 0.011 7.739 7.841 cp_fm_cholesky_decompose 22 10.9 7.611 7.693 7.611 7.693 wfi_extrapolate 11 7.9 0.001 0.001 7.483 7.483 multiply_cannon_metrocomm4 25290 15.6 0.076 0.086 2.823 7.103 fft_wrap_pw1pw2_140 527 13.2 0.485 0.496 6.829 6.958 mp_irecv_dv 76751 16.2 2.675 6.839 2.675 6.839 density_rs2pw 129 9.7 0.005 0.006 6.293 6.755 calculate_dm_sparse 129 9.5 0.001 0.001 6.186 6.256 fft3d_ps 1301 14.7 2.828 2.924 5.954 6.014 grid_collocate_task_list 129 9.7 5.249 5.561 5.249 5.561 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 5.423 5.466 mp_alltoall_d11v 2429 14.1 4.804 5.290 4.804 5.290 rs_pw_transfer 1054 12.0 0.013 0.015 4.603 5.068 qs_ot_get_derivative_taylor 40 13.0 0.001 0.001 4.380 4.464 qs_energies_init_hamiltonians 11 5.9 0.005 0.024 4.444 4.463 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 4.345 4.439 potential_pw2rs 129 12.3 0.023 0.023 4.309 4.329 qs_ot_get_orbitals 118 10.6 0.001 0.001 4.234 4.288 mp_allgather_i34 2529 14.6 1.895 4.079 1.895 4.079 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="406", plot="h2o_256_md", label="(8n/3r/4t)", y=188.505000, yerr=0.000000 PlotPoint: name="407", plot="h2o_256_md_mem", label="(8n/3r/4t)", y=1078.090909, yerr=16.549949 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/19/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1410023282688 0.0% 0.0% 100.0% flops 32 x 32 x 32 1924145348608 0.0% 0.0% 100.0% flops 22 x 9 x 32 1957871443968 0.0% 0.0% 100.0% flops 9 x 22 x 32 1963544850432 0.0% 0.0% 100.0% flops 22 x 22 x 32 2714615709696 0.0% 0.0% 100.0% flops 32 x 32 x 9 4377645416448 0.0% 0.0% 100.0% flops 32 x 32 x 22 5350455508992 0.0% 0.0% 100.0% flops 9 x 32 x 32 5395653328896 0.0% 0.0% 100.0% flops 22 x 32 x 32 6594687401984 0.0% 0.0% 100.0% flops 9 x 32 x 9 11444707676160 0.0% 0.0% 100.0% flops 22 x 32 x 9 15019188129792 0.0% 0.0% 100.0% flops 9 x 32 x 22 15019188129792 0.0% 0.0% 100.0% flops 22 x 32 x 22 19624853225472 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 92.796579E+12 0.0% 0.0% 100.0% flops max/rank 5.820059E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 6705500928 0.0% 0.0% 100.0% number of processed stacks 1944496 0.0% 0.0% 100.0% average stack size 0.0 0.0 3448.5 marketing flops 143.507742E+12 ------------------------------------------------------------------------------- # multiplications 2485 max memory usage/rank 1.494901E+09 # max total images/rank 1 # max 3D layers 1 # MPI messages exchanged 238560 MPI messages size (bytes): total size 1.321104E+12 min size 0.000000E+00 max size 52.428800E+06 average size 5.537828E+06 MPI breakdown and total messages size (bytes): size <= 128 1452 0 128 < size <= 8192 0 0 8192 < size <= 32768 0 0 32768 < size <= 131072 132 8650752 131072 < size <= 4194304 112800 59139686400 4194304 < size <= 16777216 104112 545846722560 16777216 < size 20064 716108638608 ------------------------------------------------------------------------------- - - - DBCSR MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Bcast 14 12. MP_Allreduce 8852 52. MP_Alltoall 9584 804353. MP_ISend 39716 2104723. MP_IRecv 39716 2103824. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 4003 58193. MP_Allreduce 11005 1175. MP_Sync 86 MP_Alltoall 1700 18828160. MP_ISendRecv 7620 122880. MP_Wait 19810 MP_ISend 10600 423612. MP_IRecv 10600 423612. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.019 0.035 172.581 172.582 qs_mol_dyn_low 1 2.0 0.003 0.003 172.104 172.117 qs_forces 11 3.9 0.003 0.003 171.992 171.998 qs_energies 11 4.9 0.002 0.002 164.629 164.637 scf_env_do_scf 11 5.9 0.001 0.001 147.503 147.516 velocity_verlet 10 3.0 0.001 0.001 112.318 112.322 scf_env_do_scf_inner_loop 116 6.6 0.003 0.008 111.156 111.158 qs_scf_new_mos 116 7.6 0.001 0.001 75.835 75.964 qs_scf_loop_do_ot 116 8.6 0.001 0.001 75.834 75.963 dbcsr_multiply_generic 2485 12.5 0.180 0.185 73.816 74.633 ot_scf_mini 116 9.6 0.003 0.004 71.451 71.561 multiply_cannon 2485 13.5 0.581 0.610 54.004 58.045 multiply_cannon_loop 2485 14.5 0.445 0.453 49.392 50.847 ot_mini 116 10.6 0.001 0.001 39.477 39.586 init_scf_loop 11 6.9 0.000 0.000 36.194 36.196 mp_waitall_1 128490 16.6 25.815 33.424 25.815 33.424 prepare_preconditioner 11 7.9 0.000 0.000 32.357 32.384 make_preconditioner 11 8.9 0.000 0.000 32.357 32.384 make_full_inverse_cholesky 11 9.9 0.000 0.000 30.250 30.534 rebuild_ks_matrix 127 8.3 0.001 0.001 28.394 28.570 qs_ks_build_kohn_sham_matrix 127 9.3 0.017 0.017 28.393 28.570 qs_ks_update_qs_env 127 7.6 0.001 0.001 25.867 26.026 multiply_cannon_multrec 9940 15.5 10.353 14.002 17.785 20.900 qs_ot_get_derivative 116 11.6 0.002 0.002 19.730 19.838 apply_preconditioner_dbcsr 127 12.6 0.000 0.000 19.169 19.704 apply_single 127 13.6 0.001 0.001 19.169 19.704 ot_diis_step 116 11.6 0.020 0.020 19.677 19.677 qs_ot_get_p 127 10.4 0.001 0.001 19.162 19.292 multiply_cannon_metrocomm3 9940 15.5 0.022 0.023 12.035 18.985 cp_fm_cholesky_invert 11 10.9 18.616 18.622 18.616 18.622 make_m2s 4970 13.5 0.066 0.070 16.022 18.346 make_images 4970 14.5 2.285 2.762 15.720 18.048 qs_ot_p2m_diag 82 11.4 0.489 0.495 15.278 15.299 cp_dbcsr_syevd 82 12.4 0.005 0.005 14.137 14.140 sum_up_and_integrate 127 10.3 0.179 0.188 14.037 14.085 integrate_v_rspace 127 11.3 0.004 0.004 13.857 13.914 qs_rho_update_rho_low 127 7.7 0.001 0.001 13.056 13.094 calculate_rho_elec 127 8.7 0.256 0.267 13.056 13.093 multiply_cannon_sync_h2d 9940 15.5 11.508 12.400 11.508 12.400 make_images_data 4970 15.5 0.052 0.060 9.744 12.345 hybrid_alltoall_any 5155 16.4 0.828 3.744 9.511 11.696 cp_fm_diag_elpa 82 13.4 0.000 0.000 11.025 11.028 cp_fm_diag_elpa_base 82 14.4 10.780 10.857 11.018 11.021 init_scf_run 11 5.9 0.000 0.001 10.329 10.329 scf_env_initial_rho_setup 11 6.9 0.001 0.001 10.328 10.329 cp_fm_cholesky_decompose 22 10.9 8.128 8.223 8.128 8.223 grid_integrate_task_list 127 12.3 7.716 8.026 7.716 8.026 qs_ot_get_derivative_diag 76 12.4 0.002 0.002 7.919 7.999 dbcsr_mm_accdrv_process 20590 16.1 2.640 3.535 7.066 7.743 multiply_cannon_metrocomm1 9940 15.5 0.029 0.030 4.454 7.714 pw_transfer 1535 11.6 0.082 0.090 7.669 7.691 fft_wrap_pw1pw2 1281 12.7 0.010 0.010 7.451 7.479 wfi_extrapolate 11 7.9 0.001 0.001 7.282 7.282 mp_allgather_i34 2485 14.5 2.831 7.055 2.831 7.055 fft_wrap_pw1pw2_140 519 13.2 0.499 0.519 6.530 6.556 calculate_dm_sparse 127 9.5 0.001 0.001 6.133 6.248 density_rs2pw 127 9.7 0.005 0.006 5.939 6.239 grid_collocate_task_list 127 9.7 5.442 5.707 5.442 5.707 fft3d_ps 1281 14.7 2.704 2.770 5.623 5.647 dbcsr_complete_redistribute 393 12.7 2.116 2.155 5.216 5.569 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 5.231 5.232 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 5.177 5.189 mp_alltoall_d11v 2401 14.1 4.658 5.159 4.658 5.159 rs_pw_transfer 1038 11.9 0.012 0.013 4.209 4.535 multiply_cannon_metrocomm4 7455 15.5 0.023 0.026 1.841 4.181 potential_pw2rs 127 12.3 0.026 0.027 4.130 4.149 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 4.086 4.128 mp_irecv_dv 28618 15.9 1.804 4.103 1.804 4.103 build_core_hamiltonian_matrix_ 11 4.9 0.001 0.001 3.570 3.868 copy_fm_to_dbcsr 208 11.6 0.002 0.002 3.400 3.717 qs_ot_get_orbitals 116 10.6 0.001 0.001 3.678 3.709 copy_dbcsr_to_fm 185 11.7 0.004 0.004 3.547 3.626 qs_ot_get_derivative_taylor 40 13.0 0.001 0.001 3.462 3.493 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="408", plot="h2o_256_md", label="(8n/2r/6t)", y=172.582000, yerr=0.000000 PlotPoint: name="409", plot="h2o_256_md_mem", label="(8n/2r/6t)", y=1389.727273, yerr=40.233615 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/20/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1410023282688 0.0% 0.0% 100.0% flops 32 x 32 x 32 1924145348608 0.0% 0.0% 100.0% flops 22 x 9 x 32 1957871443968 0.0% 0.0% 100.0% flops 9 x 22 x 32 1963544850432 0.0% 0.0% 100.0% flops 22 x 22 x 32 2714615709696 0.0% 0.0% 100.0% flops 32 x 32 x 9 4377645416448 0.0% 0.0% 100.0% flops 32 x 32 x 22 5350455508992 0.0% 0.0% 100.0% flops 9 x 32 x 32 5395653328896 0.0% 0.0% 100.0% flops 22 x 32 x 32 6594687401984 0.0% 0.0% 100.0% flops 9 x 32 x 9 11444707676160 0.0% 0.0% 100.0% flops 22 x 32 x 9 15019188129792 0.0% 0.0% 100.0% flops 9 x 32 x 22 15019188129792 0.0% 0.0% 100.0% flops 22 x 32 x 22 19624853225472 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 92.796579E+12 0.0% 0.0% 100.0% flops max/rank 11.606413E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 6705500928 0.0% 0.0% 100.0% number of processed stacks 1947808 0.0% 0.0% 100.0% average stack size 0.0 0.0 3442.6 marketing flops 143.507742E+12 ------------------------------------------------------------------------------- # multiplications 2485 max memory usage/rank 2.999669E+09 # max total images/rank 2 # max 3D layers 1 # MPI messages exchanged 99400 MPI messages size (bytes): total size 1.127422E+12 min size 0.000000E+00 max size 104.857600E+06 average size 11.342275E+06 MPI breakdown and total messages size (bytes): size <= 128 572 0 128 < size <= 8192 0 0 8192 < size <= 32768 0 0 32768 < size <= 131072 44 2883584 131072 < size <= 4194304 44768 34745614336 4194304 < size <= 16777216 43984 376564613120 16777216 < size 10032 716108638608 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 3997 59213. MP_Allreduce 10990 1515. MP_Sync 86 MP_Alltoall 1700 36954383. MP_ISendRecv 3556 218624. MP_Wait 11506 MP_ISend 6360 1080477. MP_IRecv 6360 1080477. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.026 0.045 291.874 291.876 qs_mol_dyn_low 1 2.0 0.003 0.003 291.329 291.341 qs_forces 11 3.9 0.003 0.005 291.237 291.240 qs_energies 11 4.9 0.001 0.002 282.558 282.573 scf_env_do_scf 11 5.9 0.001 0.001 260.781 260.793 velocity_verlet 10 3.0 0.001 0.001 210.898 210.906 scf_env_do_scf_inner_loop 116 6.6 0.003 0.008 133.326 133.328 init_scf_loop 11 6.9 0.000 0.000 127.187 127.190 prepare_preconditioner 11 7.9 0.000 0.000 122.507 122.525 make_preconditioner 11 8.9 0.000 0.000 122.507 122.525 make_full_inverse_cholesky 11 9.9 0.000 0.000 97.610 119.640 qs_scf_new_mos 116 7.6 0.001 0.001 91.640 91.701 qs_scf_loop_do_ot 116 8.6 0.001 0.001 91.639 91.700 ot_scf_mini 116 9.6 0.004 0.004 86.929 87.002 dbcsr_multiply_generic 2485 12.5 0.212 0.220 81.896 82.104 cp_fm_upper_to_full 104 14.8 54.107 78.035 54.107 78.035 multiply_cannon 2485 13.5 0.693 0.719 57.663 58.790 multiply_cannon_loop 2485 14.5 0.470 0.475 54.014 54.603 dbcsr_complete_redistribute 393 12.7 3.991 4.039 30.622 44.236 ot_mini 116 10.6 0.001 0.001 43.921 44.009 copy_fm_to_dbcsr 208 11.6 0.001 0.002 27.201 40.867 transfer_fm_to_dbcsr 11 9.9 0.000 0.000 24.849 38.344 mp_alltoall_i22 712 14.1 22.626 36.418 22.626 36.418 cp_fm_cholesky_invert 11 10.9 34.251 34.258 34.251 34.258 rebuild_ks_matrix 127 8.3 0.001 0.001 33.234 33.313 qs_ks_build_kohn_sham_matrix 127 9.3 0.017 0.018 33.233 33.312 qs_ks_update_qs_env 127 7.6 0.001 0.001 30.932 31.014 mp_waitall_1 104546 16.7 27.465 30.623 27.465 30.623 qs_ot_get_p 127 10.4 0.001 0.001 27.830 27.980 qs_ot_get_derivative 116 11.6 0.001 0.002 24.171 24.238 qs_ot_p2m_diag 82 11.4 0.868 0.875 23.728 23.756 cp_dbcsr_syevd 82 12.4 0.005 0.006 21.990 21.991 make_m2s 4970 13.5 0.076 0.078 19.524 20.417 make_images 4970 14.5 3.746 3.855 19.050 19.946 ot_diis_step 116 11.6 0.022 0.022 19.712 19.712 apply_preconditioner_dbcsr 127 12.6 0.000 0.000 18.951 19.041 apply_single 127 13.6 0.001 0.001 18.951 19.041 cp_fm_diag_elpa 82 13.4 0.000 0.000 18.752 18.756 cp_fm_diag_elpa_base 82 14.4 14.245 15.935 18.748 18.752 multiply_cannon_metrocomm3 9940 15.5 0.023 0.023 18.203 18.537 multiply_cannon_multrec 9940 15.5 10.344 12.086 17.714 17.773 sum_up_and_integrate 127 10.3 0.319 0.320 15.746 15.843 multiply_cannon_sync_h2d 9940 15.5 15.520 15.529 15.520 15.529 integrate_v_rspace 127 11.3 0.004 0.004 15.427 15.524 qs_rho_update_rho_low 127 7.7 0.001 0.001 15.000 15.029 calculate_rho_elec 127 8.7 0.479 0.479 14.999 15.029 hybrid_alltoall_any 5155 16.4 1.291 3.011 10.327 11.903 init_scf_run 11 5.9 0.000 0.001 11.868 11.869 scf_env_initial_rho_setup 11 6.9 0.001 0.001 11.868 11.868 make_images_data 4970 15.5 0.059 0.063 10.253 11.850 qs_ot_get_derivative_diag 76 12.4 0.002 0.002 9.381 9.436 cp_fm_cholesky_decompose 22 10.9 9.147 9.179 9.147 9.179 dbcsr_mm_accdrv_process 20590 16.0 3.840 5.619 7.134 8.953 wfi_extrapolate 11 7.9 0.001 0.001 8.778 8.778 grid_integrate_task_list 127 12.3 8.477 8.647 8.477 8.647 pw_transfer 1535 11.6 0.090 0.091 8.410 8.420 fft_wrap_pw1pw2 1281 12.7 0.011 0.011 8.177 8.186 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 7.945 7.945 mp_alltoall_d11v 2401 14.1 7.100 7.245 7.100 7.245 fft_wrap_pw1pw2_140 519 13.2 0.535 0.538 7.216 7.227 calculate_dm_sparse 127 9.5 0.001 0.001 6.551 6.694 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 6.330 6.389 grid_collocate_task_list 127 9.7 6.289 6.322 6.289 6.322 fft3d_ps 1281 14.7 2.733 2.738 6.238 6.251 copy_dbcsr_to_fm 185 11.7 0.004 0.004 6.075 6.144 density_rs2pw 127 9.7 0.005 0.005 6.080 6.113 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="410", plot="h2o_256_md", label="(8n/1r/12t)", y=291.876000, yerr=0.000000 PlotPoint: name="411", plot="h2o_256_md_mem", label="(8n/1r/12t)", y=2675.909091, yerr=179.297850 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/21/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 23 x 23 x 23 234439235724792 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 234.439236E+12 0.0% 0.0% 100.0% flops max/rank 2.766000E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 9634225188 0.0% 0.0% 100.0% number of processed stacks 419739 0.0% 0.0% 100.0% average stack size 0.0 0.0 22952.9 marketing flops 1.742116E+15 ------------------------------------------------------------------------------- # multiplications 111 max memory usage/rank 1.260073E+09 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 458208 MPI messages size (bytes): total size 3.456111E+12 min size 0.000000E+00 max size 18.735064E+06 average size 7.542668E+06 MPI breakdown and total messages size (bytes): size <= 128 112896 0 128 < size <= 8192 0 0 8192 < size <= 32768 224 5687808 32768 < size <= 131072 10528 813356544 131072 < size <= 4194304 36422 76284728544 4194304 < size <= 16777216 294266 3312457683808 16777216 < size 3872 66548597808 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 1026 255669. MP_Allreduce 3059 6274. MP_Sync 4 MP_Alltoall 54 MP_ISendRecv 570 19200. MP_Wait 1302 MP_ISend 642 197829. MP_IRecv 642 197607. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.044 0.076 85.146 85.155 qs_energies 1 2.0 0.000 0.000 84.669 84.683 ls_scf 1 3.0 0.000 0.000 83.764 83.778 dbcsr_multiply_generic 111 6.7 0.014 0.016 72.771 72.950 multiply_cannon 111 7.7 0.017 0.020 55.946 57.083 multiply_cannon_loop 111 8.7 0.211 0.227 52.510 53.805 ls_scf_main 1 4.0 0.000 0.000 52.471 52.472 density_matrix_trs4 2 5.0 0.002 0.003 46.972 47.057 ls_scf_init_scf 1 4.0 0.000 0.000 28.242 28.243 ls_scf_init_matrix_S 1 5.0 0.000 0.000 27.170 27.228 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.001 25.065 25.084 mp_waitall_1 11316 10.9 22.627 24.832 22.627 24.832 multiply_cannon_multrec 2664 9.7 8.178 8.930 15.478 17.192 multiply_cannon_sync_h2d 2664 9.7 13.736 15.187 13.736 15.187 make_m2s 222 7.7 0.008 0.011 13.231 13.711 make_images 222 8.7 0.099 0.109 13.209 13.692 multiply_cannon_metrocomm1 2664 9.7 0.010 0.011 9.660 13.529 make_images_data 222 9.7 0.004 0.005 7.823 8.432 hybrid_alltoall_any 227 10.6 0.215 1.833 6.601 8.241 multiply_cannon_metrocomm3 2664 9.7 0.009 0.010 5.459 7.921 dbcsr_mm_accdrv_process 4760 10.4 0.513 0.605 6.919 7.875 dbcsr_mm_accdrv_process_sort 4760 11.4 6.207 7.094 6.207 7.094 calculate_norms 4752 9.8 5.513 6.222 5.513 6.222 apply_matrix_preconditioner 6 5.3 0.000 0.000 5.077 5.199 mp_sum_l 807 5.4 3.062 4.674 3.062 4.674 make_images_sizes 222 9.7 0.000 0.000 0.789 3.526 mp_alltoall_i44 222 10.7 0.788 3.525 0.788 3.525 dbcsr_multiply_generic_mpsum_f 86 7.8 0.000 0.000 2.254 3.507 multiply_cannon_metrocomm4 2442 9.7 0.011 0.014 2.053 3.431 mp_irecv_dv 6231 10.9 2.035 3.408 2.035 3.408 arnoldi_extremal 4 6.8 0.000 0.000 3.364 3.393 arnoldi_normal_ev 4 7.8 0.001 0.002 3.364 3.393 build_subspace 16 8.4 0.009 0.012 3.264 3.266 ls_scf_post 1 4.0 0.000 0.000 3.051 3.065 ls_scf_store_result 1 5.0 0.000 0.000 2.873 2.912 dbcsr_special_finalize 555 9.7 0.005 0.006 2.271 2.669 dbcsr_merge_single_wm 555 10.7 0.456 0.580 2.263 2.660 make_images_pack 222 9.7 2.203 2.629 2.205 2.631 dbcsr_matrix_vector_mult 304 9.0 0.003 0.010 2.351 2.609 dbcsr_matrix_vector_mult_local 304 10.0 2.064 2.498 2.066 2.500 dbcsr_sort_data 658 11.4 2.065 2.442 2.065 2.442 ls_scf_dm_to_ks 2 5.0 0.000 0.000 2.223 2.332 buffer_matrices_ensure_size 222 8.7 1.752 2.037 1.752 2.037 qs_ks_update_qs_env 3 6.3 0.000 0.000 1.735 1.736 rebuild_ks_matrix 3 7.3 0.000 0.000 1.725 1.727 qs_ks_build_kohn_sham_matrix 3 8.3 0.000 0.001 1.725 1.727 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="500", plot="h2o_32_nrep3_ls", label="(8n/12r/1t)", y=85.155000, yerr=0.000000 PlotPoint: name="501", plot="h2o_32_nrep3_ls_mem", label="(8n/12r/1t)", y=1140.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/22/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 23 x 23 x 23 234439235724792 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 234.439236E+12 0.0% 0.0% 100.0% flops max/rank 5.588524E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 9634225188 0.0% 0.0% 100.0% number of processed stacks 368848 0.0% 0.0% 100.0% average stack size 0.0 0.0 26119.8 marketing flops 1.742116E+15 ------------------------------------------------------------------------------- # multiplications 111 max memory usage/rank 2.120851E+09 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 106560 MPI messages size (bytes): total size 2.699093E+12 min size 0.000000E+00 max size 72.286792E+06 average size 25.329324E+06 MPI breakdown and total messages size (bytes): size <= 128 23040 0 128 < size <= 8192 0 0 8192 < size <= 32768 0 0 32768 < size <= 131072 3264 325830144 131072 < size <= 4194304 5280 3328561104 4194304 < size <= 16777216 12709 156766962056 16777216 < size 62267 2538670978840 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 1026 266696. MP_Allreduce 3058 10339. MP_Sync 4 MP_Alltoall 47 15335933. MP_ISendRecv 282 57600. MP_Wait 828 MP_ISend 462 414589. MP_IRecv 462 413870. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.032 0.045 90.487 90.488 qs_energies 1 2.0 0.000 0.000 89.960 89.964 ls_scf 1 3.0 0.000 0.001 88.647 88.651 dbcsr_multiply_generic 111 6.7 0.015 0.016 74.583 74.937 multiply_cannon 111 7.7 0.028 0.039 52.899 56.379 ls_scf_main 1 4.0 0.000 0.004 54.596 54.598 multiply_cannon_loop 111 8.7 0.117 0.125 49.589 53.173 density_matrix_trs4 2 5.0 0.002 0.003 48.989 49.175 ls_scf_init_scf 1 4.0 0.000 0.002 30.545 30.546 ls_scf_init_matrix_S 1 5.0 0.000 0.000 29.379 29.504 mp_waitall_1 9246 10.9 20.677 28.837 20.677 28.837 multiply_cannon_multrec 1332 9.7 13.115 17.370 22.310 27.746 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.001 26.998 27.015 multiply_cannon_metrocomm3 1332 9.7 0.006 0.008 11.392 21.023 make_m2s 222 7.7 0.006 0.008 15.126 15.789 make_images 222 8.7 1.571 1.930 15.095 15.759 dbcsr_mm_accdrv_process 4041 10.4 0.254 0.429 8.797 10.310 dbcsr_mm_accdrv_process_sort 4041 11.4 8.391 9.878 8.391 9.878 make_images_data 222 9.7 0.004 0.004 8.655 9.502 hybrid_alltoall_any 227 10.6 0.520 2.444 8.062 9.262 mp_sum_l 807 5.4 5.253 8.395 5.253 8.395 multiply_cannon_metrocomm4 1221 9.7 0.006 0.008 3.210 7.605 mp_irecv_dv 3311 11.0 3.190 7.542 3.190 7.542 calculate_norms 2376 9.8 6.021 6.847 6.021 6.847 dbcsr_multiply_generic_mpsum_f 86 7.8 0.000 0.000 4.036 6.532 multiply_cannon_sync_h2d 1332 9.7 4.817 5.891 4.817 5.891 apply_matrix_preconditioner 6 5.3 0.000 0.000 4.952 5.167 arnoldi_extremal 4 6.8 0.000 0.000 4.792 4.817 arnoldi_normal_ev 4 7.8 0.001 0.005 4.792 4.817 build_subspace 16 8.4 0.014 0.021 4.525 4.528 ls_scf_post 1 4.0 0.000 0.001 3.507 3.512 dbcsr_matrix_vector_mult 304 9.0 0.005 0.017 3.190 3.434 ls_scf_store_result 1 5.0 0.000 0.000 3.210 3.332 dbcsr_matrix_vector_mult_local 304 10.0 2.764 3.264 2.766 3.266 multiply_cannon_metrocomm1 1332 9.7 0.003 0.003 1.254 2.881 ls_scf_dm_to_ks 2 5.0 0.000 0.000 2.524 2.636 make_images_pack 222 9.7 2.031 2.392 2.033 2.394 mp_allgather_i34 111 8.7 1.016 2.335 1.016 2.335 dbcsr_sort_data 436 11.2 1.790 1.985 1.790 1.985 qs_ks_update_qs_env 3 6.3 0.000 0.000 1.828 1.830 dbcsr_data_new 4174 10.1 1.608 1.828 1.608 1.828 rebuild_ks_matrix 3 7.3 0.000 0.000 1.815 1.817 qs_ks_build_kohn_sham_matrix 3 8.3 0.001 0.004 1.815 1.817 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="502", plot="h2o_32_nrep3_ls", label="(8n/6r/2t)", y=90.488000, yerr=0.000000 PlotPoint: name="503", plot="h2o_32_nrep3_ls_mem", label="(8n/6r/2t)", y=1766.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/23/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 23 x 23 x 23 234439235724792 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 234.439236E+12 0.0% 0.0% 100.0% flops max/rank 8.404608E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 9634225188 0.0% 0.0% 100.0% number of processed stacks 353133 0.0% 0.0% 100.0% average stack size 0.0 0.0 27282.1 marketing flops 1.742118E+15 ------------------------------------------------------------------------------- # multiplications 111 max memory usage/rank 2.703270E+09 # max total images/rank 2 # max 3D layers 1 # MPI messages exchanged 46176 MPI messages size (bytes): total size 1.924064E+12 min size 0.000000E+00 max size 108.059888E+06 average size 41.668048E+06 MPI breakdown and total messages size (bytes): size <= 128 9984 0 128 < size <= 8192 0 0 8192 < size <= 32768 0 0 32768 < size <= 131072 0 0 131072 < size <= 4194304 3328 1170063360 4194304 < size <= 16777216 1870 19378539600 16777216 < size 30994 1903514987232 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 1026 265470. MP_Allreduce 3058 11181. MP_Sync 4 MP_Alltoall 47 23526250. MP_ISendRecv 186 57600. MP_Wait 732 MP_ISend 462 560046. MP_IRecv 462 560662. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.030 0.045 93.648 93.649 qs_energies 1 2.0 0.000 0.000 93.101 93.107 ls_scf 1 3.0 0.000 0.000 91.711 91.716 dbcsr_multiply_generic 111 6.7 0.015 0.016 76.366 76.598 ls_scf_main 1 4.0 0.000 0.000 57.482 57.489 multiply_cannon 111 7.7 0.043 0.109 52.846 56.785 multiply_cannon_loop 111 8.7 0.100 0.105 49.264 52.243 density_matrix_trs4 2 5.0 0.002 0.003 51.597 51.762 mp_waitall_1 7374 11.0 23.941 33.301 23.941 33.301 ls_scf_init_scf 1 4.0 0.000 0.000 30.673 30.674 ls_scf_init_matrix_S 1 5.0 0.000 0.000 29.557 29.627 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.001 27.150 27.164 multiply_cannon_multrec 888 9.7 12.492 14.981 21.025 24.132 multiply_cannon_metrocomm3 888 9.7 0.004 0.004 11.152 22.434 make_m2s 222 7.7 0.006 0.007 17.022 18.227 make_images 222 8.7 1.963 2.276 16.984 18.189 make_images_data 222 9.7 0.003 0.004 9.775 10.741 hybrid_alltoall_any 227 10.6 0.618 2.862 9.455 10.590 dbcsr_mm_accdrv_process 3754 10.4 0.241 0.421 8.035 9.257 dbcsr_mm_accdrv_process_sort 3754 11.4 7.664 8.836 7.664 8.836 mp_sum_l 807 5.4 4.928 7.862 4.928 7.862 multiply_cannon_sync_h2d 888 9.7 6.092 7.823 6.092 7.823 multiply_cannon_metrocomm4 777 9.7 0.004 0.005 2.490 6.949 mp_irecv_dv 2335 11.1 2.474 6.894 2.474 6.894 multiply_cannon_metrocomm1 888 9.7 0.003 0.003 3.631 6.287 dbcsr_multiply_generic_mpsum_f 86 7.8 0.000 0.000 3.787 6.089 arnoldi_extremal 4 6.8 0.000 0.000 5.292 5.309 arnoldi_normal_ev 4 7.8 0.001 0.005 5.292 5.309 apply_matrix_preconditioner 6 5.3 0.000 0.000 4.905 5.138 build_subspace 16 8.4 0.014 0.020 4.984 4.993 calculate_norms 1584 9.8 4.259 4.595 4.259 4.595 dbcsr_matrix_vector_mult 304 9.0 0.005 0.016 3.508 3.865 mp_allgather_i34 111 8.7 1.411 3.789 1.411 3.789 dbcsr_matrix_vector_mult_local 304 10.0 3.038 3.606 3.040 3.608 ls_scf_post 1 4.0 0.000 0.000 3.556 3.563 ls_scf_store_result 1 5.0 0.000 0.000 3.281 3.382 ls_scf_dm_to_ks 2 5.0 0.000 0.000 2.757 2.855 make_images_sizes 222 9.7 0.000 0.000 0.919 2.149 mp_alltoall_i44 222 10.7 0.919 2.148 0.919 2.148 make_images_pack 222 9.7 1.811 2.117 1.813 2.119 dbcsr_sort_data 325 11.1 1.898 2.092 1.898 2.092 dbcsr_data_release 9322 10.9 1.339 2.021 1.339 2.021 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="504", plot="h2o_32_nrep3_ls", label="(8n/4r/3t)", y=93.649000, yerr=0.000000 PlotPoint: name="505", plot="h2o_32_nrep3_ls_mem", label="(8n/4r/3t)", y=2182.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/24/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 23 x 23 x 23 234439235724792 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 234.439236E+12 0.0% 0.0% 100.0% flops max/rank 10.747127E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 9634225188 0.0% 0.0% 100.0% number of processed stacks 369794 0.0% 0.0% 100.0% average stack size 0.0 0.0 26053.0 marketing flops 1.742116E+15 ------------------------------------------------------------------------------- # multiplications 111 max memory usage/rank 3.309785E+09 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 50616 MPI messages size (bytes): total size 1.536549E+12 min size 0.000000E+00 max size 72.286792E+06 average size 30.356986E+06 MPI breakdown and total messages size (bytes): size <= 128 10368 0 128 < size <= 8192 0 0 8192 < size <= 32768 0 0 32768 < size <= 131072 1056 104411904 131072 < size <= 4194304 3168 831638784 4194304 < size <= 16777216 3103 33613273640 16777216 < size 32921 1501999894888 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 1026 266696. MP_Allreduce 3058 13371. MP_Sync 4 MP_Alltoall 47 30278988. MP_ISendRecv 138 86400. MP_Wait 600 MP_ISend 378 823502. MP_IRecv 378 823753. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.045 0.058 97.113 97.115 qs_energies 1 2.0 0.000 0.000 96.438 96.445 ls_scf 1 3.0 0.000 0.001 94.781 94.790 dbcsr_multiply_generic 111 6.7 0.017 0.023 78.673 78.928 ls_scf_main 1 4.0 0.000 0.001 58.956 58.957 multiply_cannon 111 7.7 0.053 0.130 51.637 56.229 density_matrix_trs4 2 5.0 0.002 0.003 52.896 53.048 multiply_cannon_loop 111 8.7 0.115 0.128 46.584 49.959 ls_scf_init_scf 1 4.0 0.001 0.006 32.524 32.527 ls_scf_init_matrix_S 1 5.0 0.000 0.000 31.362 31.440 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.001 28.875 28.887 mp_waitall_1 6438 11.0 22.754 28.585 22.754 28.585 multiply_cannon_multrec 1332 9.7 14.165 17.350 21.986 25.062 make_m2s 222 7.7 0.007 0.009 21.166 22.535 make_images 222 8.7 3.134 3.605 21.116 22.487 multiply_cannon_metrocomm3 1332 9.7 0.003 0.003 9.089 17.408 make_images_data 222 9.7 0.004 0.005 11.736 13.313 hybrid_alltoall_any 227 10.6 0.797 3.830 11.180 12.816 dbcsr_mm_accdrv_process 3641 10.4 0.195 0.377 7.457 8.986 dbcsr_mm_accdrv_process_sort 3641 11.4 7.085 8.566 7.085 8.566 mp_sum_l 807 5.4 4.288 8.362 4.288 8.362 dbcsr_multiply_generic_mpsum_f 86 7.8 0.000 0.000 3.298 6.530 multiply_cannon_sync_h2d 1332 9.7 5.549 6.226 5.549 6.226 multiply_cannon_metrocomm4 1110 9.7 0.004 0.006 2.082 6.185 mp_irecv_dv 3229 10.9 2.061 6.104 2.061 6.104 multiply_cannon_metrocomm1 1332 9.7 0.003 0.003 2.756 5.802 arnoldi_extremal 4 6.8 0.000 0.000 5.215 5.227 arnoldi_normal_ev 4 7.8 0.001 0.005 5.215 5.227 build_subspace 16 8.4 0.015 0.021 4.863 4.870 apply_matrix_preconditioner 6 5.3 0.000 0.000 4.594 4.809 calculate_norms 2376 9.8 4.180 4.557 4.180 4.557 mp_allgather_i34 111 8.7 2.178 4.390 2.178 4.390 dbcsr_matrix_vector_mult 304 9.0 0.006 0.016 3.583 3.891 dbcsr_matrix_vector_mult_local 304 10.0 3.183 3.697 3.185 3.699 dbcsr_sort_data 658 11.4 3.140 3.581 3.140 3.581 ls_scf_post 1 4.0 0.000 0.001 3.301 3.308 dbcsr_special_finalize 555 9.7 0.006 0.006 2.887 3.263 dbcsr_merge_single_wm 555 10.7 0.541 0.668 2.879 3.255 ls_scf_store_result 1 5.0 0.000 0.000 3.013 3.106 ls_scf_dm_to_ks 2 5.0 0.000 0.000 2.888 2.934 dbcsr_data_release 10477 10.7 1.572 2.434 1.572 2.434 dbcsr_finalize 304 7.8 0.049 0.062 1.801 2.045 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="506", plot="h2o_32_nrep3_ls", label="(8n/3r/4t)", y=97.115000, yerr=0.000000 PlotPoint: name="507", plot="h2o_32_nrep3_ls_mem", label="(8n/3r/4t)", y=2722.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/25/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 23 x 23 x 23 234439235724792 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 234.439236E+12 0.0% 0.0% 100.0% flops max/rank 15.383312E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 9634225188 0.0% 0.0% 100.0% number of processed stacks 336818 0.0% 0.0% 100.0% average stack size 0.0 0.0 28603.7 marketing flops 1.742118E+15 ------------------------------------------------------------------------------- # multiplications 111 max memory usage/rank 4.607013E+09 # max total images/rank 1 # max 3D layers 1 # MPI messages exchanged 10656 MPI messages size (bytes): total size 1.149035E+12 min size 0.000000E+00 max size 203.538048E+06 average size 107.829832E+06 MPI breakdown and total messages size (bytes): size <= 128 2304 0 128 < size <= 8192 0 0 8192 < size <= 32768 0 0 32768 < size <= 131072 0 0 131072 < size <= 4194304 768 702038016 4194304 < size <= 16777216 0 0 16777216 < size 7584 1148332810224 ------------------------------------------------------------------------------- - - - DBCSR MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Bcast 2 12. MP_Allreduce 705 128. MP_Alltoall 310 12920694. MP_ISend 1776 40180424. MP_IRecv 1776 40465030. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 1026 265558. MP_Allreduce 3049 15663. MP_Sync 4 MP_Alltoall 47 46208988. MP_ISendRecv 90 115200. MP_Wait 573 MP_ISend 420 924980. MP_IRecv 420 924528. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.068 0.094 97.324 97.334 qs_energies 1 2.0 0.000 0.000 96.510 96.517 ls_scf 1 3.0 0.000 0.000 94.565 94.572 dbcsr_multiply_generic 111 6.7 0.017 0.019 76.458 76.728 ls_scf_main 1 4.0 0.000 0.000 60.943 60.944 multiply_cannon 111 7.7 0.101 0.185 54.321 59.425 density_matrix_trs4 2 5.0 0.002 0.003 54.037 54.111 multiply_cannon_loop 111 8.7 0.070 0.075 49.735 51.812 mp_waitall_1 5481 11.0 25.137 31.011 25.137 31.011 ls_scf_init_scf 1 4.0 0.000 0.000 30.040 30.045 ls_scf_init_matrix_S 1 5.0 0.000 0.000 28.883 28.924 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.001 26.752 26.763 multiply_cannon_multrec 444 9.7 14.010 16.188 21.023 23.029 make_m2s 222 7.7 0.004 0.005 17.356 19.961 make_images 222 8.7 3.715 4.387 17.294 19.901 multiply_cannon_metrocomm1 444 9.7 0.002 0.002 10.234 14.748 multiply_cannon_metrocomm3 444 9.7 0.001 0.001 6.060 14.214 make_images_data 222 9.7 0.003 0.004 9.642 12.142 hybrid_alltoall_any 227 10.6 0.789 3.769 9.344 11.881 dbcsr_mm_accdrv_process 3003 10.4 0.165 0.343 6.718 7.853 dbcsr_mm_accdrv_process_sort 3003 11.4 6.402 7.502 6.402 7.502 multiply_cannon_sync_h2d 444 9.7 6.532 7.364 6.532 7.364 mp_allgather_i34 111 8.7 2.813 6.954 2.813 6.954 arnoldi_extremal 4 6.8 0.000 0.000 5.809 5.820 arnoldi_normal_ev 4 7.8 0.001 0.004 5.809 5.820 build_subspace 16 8.4 0.015 0.020 5.408 5.418 mp_sum_l 807 5.4 2.835 4.999 2.835 4.999 apply_matrix_preconditioner 6 5.3 0.000 0.000 4.579 4.740 dbcsr_matrix_vector_mult 304 9.0 0.007 0.016 4.137 4.328 multiply_cannon_metrocomm4 333 9.7 0.001 0.002 1.688 4.210 mp_irecv_dv 1241 11.2 1.669 4.187 1.669 4.187 dbcsr_matrix_vector_mult_local 304 10.0 3.655 4.109 3.657 4.111 dbcsr_multiply_generic_mpsum_f 86 7.8 0.000 0.000 1.965 3.844 calculate_norms 792 9.8 3.563 3.727 3.563 3.727 ls_scf_post 1 4.0 0.000 0.000 3.583 3.590 ls_scf_store_result 1 5.0 0.000 0.000 3.354 3.403 ls_scf_dm_to_ks 2 5.0 0.000 0.000 3.317 3.388 make_images_sizes 222 9.7 0.000 0.000 1.072 3.229 mp_alltoall_i44 222 10.7 1.072 3.229 1.072 3.229 dbcsr_finalize 304 7.8 0.062 0.078 2.200 2.278 dbcsr_merge_all 275 8.9 0.472 0.517 2.044 2.093 dbcsr_data_release 10123 10.8 1.335 2.002 1.335 2.002 qs_ks_update_qs_env 3 6.3 0.000 0.000 1.950 1.951 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="508", plot="h2o_32_nrep3_ls", label="(8n/2r/6t)", y=97.334000, yerr=0.000000 PlotPoint: name="509", plot="h2o_32_nrep3_ls_mem", label="(8n/2r/6t)", y=3609.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/cea20fb160686b319917257b3a6652216704cc48_performance_tests/26/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 23 x 23 x 23 234439235724792 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 234.439236E+12 0.0% 0.0% 100.0% flops max/rank 30.358840E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 9634225188 0.0% 0.0% 100.0% number of processed stacks 339931 0.0% 0.0% 100.0% average stack size 0.0 0.0 28341.7 marketing flops 1.742118E+15 ------------------------------------------------------------------------------- # multiplications 111 max memory usage/rank 8.706064E+09 # max total images/rank 2 # max 3D layers 1 # MPI messages exchanged 4440 MPI messages size (bytes): total size 770.525954E+09 min size 0.000000E+00 max size 399.069120E+06 average size 173.541888E+06 MPI breakdown and total messages size (bytes): size <= 128 640 0 128 < size <= 8192 0 0 8192 < size <= 32768 0 0 32768 < size <= 131072 0 0 131072 < size <= 4194304 640 468025344 4194304 < size <= 16777216 0 0 16777216 < size 3160 770057961712 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 1026 284111. MP_Allreduce 3043 21950. MP_Sync 4 MP_Alltoall 47 88727262. MP_ISendRecv 84 732600. MP_Wait 309 MP_ISend 180 3337386. MP_IRecv 180 3339494. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.080 0.097 107.413 107.414 qs_energies 1 2.0 0.000 0.000 105.974 105.988 ls_scf 1 3.0 0.000 0.000 103.102 103.116 dbcsr_multiply_generic 111 6.7 0.023 0.027 77.343 77.460 ls_scf_main 1 4.0 0.000 0.000 65.479 65.479 density_matrix_trs4 2 5.0 0.002 0.003 56.583 56.637 multiply_cannon 111 7.7 0.213 0.398 49.899 52.108 multiply_cannon_loop 111 8.7 0.067 0.069 46.371 47.163 ls_scf_init_scf 1 4.0 0.000 0.000 33.955 33.955 ls_scf_init_matrix_S 1 5.0 0.000 0.000 32.628 32.636 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.001 29.802 29.808 mp_waitall_1 4569 11.1 22.136 25.634 22.136 25.634 make_m2s 222 7.7 0.005 0.005 23.867 24.919 make_images 222 8.7 4.585 5.013 23.759 24.810 multiply_cannon_multrec 444 9.7 17.881 18.654 22.468 23.151 hybrid_alltoall_any 227 10.6 1.660 3.619 12.941 15.743 make_images_data 222 9.7 0.003 0.003 13.116 15.624 multiply_cannon_metrocomm3 444 9.7 0.001 0.001 10.389 10.858 multiply_cannon_sync_h2d 444 9.7 8.850 8.885 8.850 8.885 arnoldi_extremal 4 6.8 0.000 0.000 7.353 7.362 arnoldi_normal_ev 4 7.8 0.002 0.008 7.353 7.362 build_subspace 16 8.4 0.026 0.036 6.783 6.792 dbcsr_matrix_vector_mult 304 9.0 0.009 0.025 5.406 5.571 apply_matrix_preconditioner 6 5.3 0.000 0.000 5.079 5.351 dbcsr_matrix_vector_mult_local 304 10.0 4.982 5.288 4.984 5.290 ls_scf_dm_to_ks 2 5.0 0.000 0.000 4.802 4.893 dbcsr_mm_accdrv_process 1814 10.4 0.206 0.318 4.417 4.543 dbcsr_mm_accdrv_process_sort 1814 11.4 4.120 4.255 4.120 4.255 ls_scf_post 1 4.0 0.000 0.000 3.668 3.682 mp_allgather_i34 111 8.7 1.174 3.614 1.174 3.614 make_images_sizes 222 9.7 0.000 0.000 1.430 3.534 mp_alltoall_i44 222 10.7 1.430 3.533 1.430 3.533 ls_scf_store_result 1 5.0 0.000 0.000 3.423 3.434 calculate_norms 792 9.8 3.241 3.277 3.241 3.277 dbcsr_finalize 304 7.8 0.082 0.089 3.089 3.143 dbcsr_merge_all 275 8.9 0.896 0.920 2.874 2.921 dbcsr_complete_redistribute 5 7.6 1.427 1.468 2.772 2.883 qs_energies_init_hamiltonians 1 3.0 0.001 0.001 2.842 2.842 dbcsr_data_release 12724 10.6 2.328 2.835 2.328 2.835 matrix_ls_to_qs 2 6.0 0.000 0.000 2.423 2.544 dbcsr_sort_data 325 11.1 2.442 2.518 2.442 2.518 dbcsr_new_transposed 4 7.5 0.246 0.257 2.303 2.332 dbcsr_frobenius_norm 74 6.6 2.055 2.131 2.188 2.230 dbcsr_add_d 103 6.2 0.000 0.000 2.138 2.205 dbcsr_add_anytype 103 7.2 0.859 0.893 2.137 2.205 qs_ks_update_qs_env 3 6.3 0.000 0.000 2.198 2.200 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="510", plot="h2o_32_nrep3_ls", label="(8n/1r/12t)", y=107.414000, yerr=0.000000 PlotPoint: name="511", plot="h2o_32_nrep3_ls_mem", label="(8n/1r/12t)", y=6803.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ========= END RESULTS =========== CommitSHA: cea20fb160686b319917257b3a6652216704cc48 Summary: empty Status: OK