=== 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: 843d9feafdfe50065c8ad5624e7456a67c33f1ab ################# 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 2022.11.001, LIBINT 2.6.0, LIBPEXSI 1.2.0, # LIBXC 6.1.0, LIBVORI 220621, LIBXSMM 1.17, PLUMED 2.8.1, # SIRIUS 7.4.3, 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 (03.02.2023) # # \ 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 -j${maxtasks} --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=${this_file##*.} libcp2k"; \ return # Set options DO_CHECKS := no USE_ACC := yes USE_COSMA := 2.6.2 USE_ELPA := 2022.11.001 USE_LIBINT := 2.6.0 USE_LIBPEXSI := 1.2.0 USE_LIBVORI := 220621 USE_LIBXC := 6.1.0 USE_LIBXSMM := 1.17 USE_PLUMED := 2.8.1 #USE_QUIP := 0.9.10 USE_SIRIUS := 7.4.3 USE_SPGLIB := 1.16.2 # Only needed for SIRIUS LIBVDWXC_VER := 0.4.0 SPFFT_VER := 1.0.6 SPLA_VER := 1.5.4 HDF5_VER := 1.12.0 # Only needed for LIBPEXSI SCOTCH_VER := 6.0.0 SUPERLU_VER := 6.1.0 LMAX := 5 MAX_CONTR := 4 GPUVER := P100 OFFLOAD_TARGET := cuda CC := cc CXX := CC OFFLOAD_CC := nvcc FC := ftn LD := ftn AR := ar -r # cc, CC, and ftn include already the proper -march flag CFLAGS := -O2 -fopenmp -fopenmp-simd -ftree-vectorize -funroll-loops -g DFLAGS := -D__parallel DFLAGS += -D__SCALAPACK DFLAGS += -D__FFTW3 DFLAGS += -D__MAX_CONTR=$(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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/01 job id: 44589109 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/02 job id: 44589111 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/03 job id: 44589113 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/04 job id: 44589115 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/05 job id: 44589117 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/06 job id: 44589118 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/07 job id: 44589121 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/08 job id: 44589123 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/09 job id: 44589124 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/10 job id: 44589125 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/11 job id: 44589126 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/12 job id: 44589127 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/13 job id: 44589128 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/14 job id: 44589129 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/15 job id: 44589130 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/16 job id: 44589131 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/17 job id: 44589132 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/18 job id: 44589133 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/19 job id: 44589134 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/20 job id: 44589135 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/21 job id: 44589136 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/22 job id: 44589137 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/23 job id: 44589138 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/24 job id: 44589139 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/25 job id: 44589140 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/26 job id: 44589141 --- 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/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 0.033 135.129 135.129 farming_run 1 2.0 133.749 133.751 135.101 135.105 ------------------------------------------------------------------------------- @@@@@@@@@@ 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 1.456845E+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 229 1108280. MP_Allreduce 485 2282278. MP_Sync 27 MP_Alltoall 38 9316958. MP_SendRecv 120 384007. MP_ISendRecv 45 235435. MP_Wait 191 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 0.011 0.026 115.374 115.374 qs_energies 1 2.0 0.000 0.000 115.155 115.157 mp2_main 1 3.0 0.000 0.000 113.247 113.248 mp2_gpw_main 1 4.0 0.019 0.026 112.430 112.432 mp2_ri_gpw_compute_in 1 5.0 0.171 0.173 93.464 93.937 mp2_ri_gpw_compute_in_loop 1 6.0 0.005 0.005 55.456 55.930 mp2_eri_3c_integrate_gpw 272 7.0 0.154 0.168 41.769 47.205 get_2c_integrals 1 6.0 0.000 0.001 36.996 37.836 integrate_v_rspace 273 8.0 0.436 0.452 25.181 30.415 pw_transfer 6555 10.6 0.377 0.384 27.392 27.926 grid_integrate_task_list 273 9.0 20.989 26.711 20.989 26.711 fft_wrap_pw1pw2 5465 11.4 0.045 0.048 26.061 26.624 fft_wrap_pw1pw2_100 2178 12.4 1.172 1.256 23.580 24.142 compute_2c_integrals 1 7.0 0.003 0.003 19.263 19.263 rpa_ri_compute_en 1 5.0 0.001 0.001 18.859 19.099 compute_2c_integrals_loop_lm 1 8.0 0.003 0.004 18.864 18.990 mp2_eri_2c_integrate_gpw 1 9.0 2.380 2.430 18.861 18.986 cp_fm_cholesky_decompose 12 8.2 17.686 18.530 17.686 18.530 cholesky_decomp 1 7.0 0.000 0.000 16.590 17.419 fft3d_s 5443 13.4 16.171 16.600 16.192 16.620 ao_to_mo_and_store_B_mult_1 272 7.0 10.858 15.607 10.858 15.607 calculate_wavefunction 272 8.0 5.397 5.577 12.512 13.113 rpa_num_int 1 6.0 0.000 0.001 10.681 10.681 rpa_num_int_RPA_matrix_operati 8 7.0 0.000 0.000 10.466 10.492 calc_mat_Q 8 8.0 0.000 0.000 9.323 9.408 contract_S_to_Q 8 9.0 0.000 0.000 8.745 8.831 calc_potential_gpw 544 9.5 0.005 0.005 8.224 8.609 mp2_eri_2c_integrate_gpw_pot_l 272 10.0 0.001 0.001 8.207 8.457 potential_pw2rs 545 10.0 0.107 0.110 7.712 8.414 parallel_gemm_fm 14 9.1 0.000 0.000 8.321 8.406 parallel_gemm_fm_cosma 14 10.1 8.321 8.406 8.321 8.406 collocate_single_gaussian 272 10.0 0.040 0.042 7.458 7.710 create_integ_mat 1 6.0 0.022 0.027 7.609 7.609 array2fm 1 7.0 0.000 0.000 6.714 7.038 pw_scatter_s 2720 13.7 4.436 4.616 4.436 4.616 pw_gather_s 2722 13.2 3.869 4.295 3.869 4.295 array2fm_buffer_send 1 8.0 2.998 3.111 2.998 3.111 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="10", plot="h2o_32_ri_rpa_mp2", label="RI-RPA (8n/2r/6t)", y=112.432136, yerr=0.000000 PlotPoint: name="11", plot="h2o_32_ri_rpa_mp2_mem", label="RI-RPA (8n/2r/6t)", y=2729.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 0.053 0.062 401.535 401.536 farming_run 1 2.0 400.388 400.430 401.474 401.494 ------------------------------------------------------------------------------- @@@@@@@@@@ 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 1.224729E+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 704 407793. MP_Allreduce 1821 23730. MP_Sync 38 MP_Alltoall 77 7592814. MP_SendRecv 2876 2171486. MP_ISendRecv 1034 172620. MP_Wait 1346 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 0.013 0.034 211.318 211.319 qs_energies 1 2.0 0.000 0.000 211.062 211.071 scf_env_do_scf 1 3.0 0.000 0.000 107.516 107.516 qs_ks_update_qs_env 5 5.0 0.000 0.000 106.618 106.627 rebuild_ks_matrix 4 6.0 0.000 0.000 106.617 106.626 qs_ks_build_kohn_sham_matrix 4 7.0 0.058 0.067 106.617 106.626 hfx_ks_matrix 4 8.0 0.001 0.001 106.229 106.233 integrate_four_center 4 9.0 0.144 0.448 106.228 106.232 mp2_main 1 3.0 0.000 0.000 103.246 103.255 mp2_gpw_main 1 4.0 0.035 0.062 102.359 102.369 integrate_four_center_main 4 10.0 0.096 0.520 96.804 100.828 integrate_four_center_bin 265 11.0 96.707 100.804 96.707 100.804 init_scf_loop 1 4.0 0.000 0.000 92.971 92.971 mp2_ri_gpw_compute_in 1 5.0 0.072 0.446 75.408 76.710 mp2_ri_gpw_compute_in_loop 1 6.0 0.002 0.002 54.664 55.963 mp2_eri_3c_integrate_gpw 91 7.0 0.143 0.164 41.899 47.371 integrate_v_rspace 95 8.0 0.397 0.573 28.268 33.575 pw_transfer 2240 10.6 0.144 0.166 29.940 30.367 fft_wrap_pw1pw2 1868 11.4 0.018 0.021 28.952 29.353 grid_integrate_task_list 95 9.0 23.549 29.063 23.549 29.063 mp2_ri_gpw_compute_en 1 5.0 0.056 0.066 26.800 28.776 fft_wrap_pw1pw2_100 730 12.4 1.267 1.469 26.667 27.052 ao_to_mo_and_store_B_mult_1 91 7.0 11.079 26.412 11.079 26.412 mp2_ri_gpw_compute_en_RI_loop 1 6.0 1.844 1.895 24.854 24.865 get_2c_integrals 1 6.0 0.000 0.016 20.277 20.302 compute_2c_integrals 1 7.0 0.003 0.007 19.252 19.259 compute_2c_integrals_loop_lm 1 8.0 0.002 0.011 18.854 19.116 mp2_eri_2c_integrate_gpw 1 9.0 1.742 1.844 18.851 19.115 fft3d_s 1823 13.4 18.432 18.860 18.445 18.873 scf_env_do_scf_inner_loop 4 4.0 0.000 0.000 14.543 14.543 calculate_wavefunction 91 8.0 2.023 2.054 9.748 9.945 potential_pw2rs 186 10.0 0.033 0.035 8.649 9.195 mp2_ri_gpw_compute_en_expansio 172 7.0 0.556 0.573 8.831 9.175 mp2_eri_2c_integrate_gpw_pot_l 91 10.0 0.001 0.001 8.234 8.613 local_gemm 172 8.0 8.275 8.605 8.275 8.605 mp2_ri_gpw_compute_en_comm 22 7.0 0.503 0.527 7.789 8.464 calc_potential_gpw 182 9.5 0.002 0.002 7.903 8.186 collocate_single_gaussian 91 10.0 0.016 0.021 7.879 8.120 mp_sync 38 10.4 4.430 7.353 4.430 7.353 mp_sendrecv_dm3 2068 8.0 5.815 6.472 5.815 6.472 mp2_ri_gpw_compute_en_ener 172 7.0 6.345 6.435 6.345 6.435 pw_gather_s 912 13.2 4.948 5.450 4.948 5.450 pw_scatter_s 910 13.7 3.910 4.271 3.910 4.271 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="20", plot="h2o_32_ri_rpa_mp2", label="RI-MP2 (8n/6r/2t)", y=102.360489, yerr=0.000000 PlotPoint: name="21", plot="h2o_32_ri_rpa_mp2_mem", label="RI-MP2 (8n/6r/2t)", y=1511.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 452.091904E+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 578578. MP_SendRecv 22610 5520. MP_ISendRecv 22610 5520. MP_Wait 37876 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 0.021 0.048 53.507 53.510 qs_mol_dyn_low 1 2.0 0.003 0.006 53.163 53.171 qs_forces 11 3.9 0.002 0.002 53.037 53.039 qs_energies 11 4.9 0.001 0.001 51.574 51.588 scf_env_do_scf 11 5.9 0.000 0.001 44.438 44.438 scf_env_do_scf_inner_loop 108 6.5 0.002 0.007 42.377 42.377 dbcsr_multiply_generic 2286 12.5 0.094 0.096 33.838 34.415 qs_scf_new_mos 108 7.5 0.000 0.001 32.150 32.439 qs_scf_loop_do_ot 108 8.5 0.000 0.001 32.150 32.438 ot_scf_mini 108 9.5 0.002 0.002 30.507 30.708 multiply_cannon 2286 13.5 0.177 0.184 26.577 28.356 multiply_cannon_loop 2286 14.5 1.473 1.545 25.898 27.657 velocity_verlet 10 3.0 0.001 0.002 25.140 25.141 ot_mini 108 10.5 0.001 0.001 19.259 19.494 qs_ot_get_derivative 108 11.5 0.008 0.062 16.315 16.520 mp_waitall_1 245248 16.5 8.623 15.593 8.623 15.593 multiply_cannon_metrocomm3 54864 15.5 0.067 0.072 6.161 14.090 multiply_cannon_multrec 54864 15.5 4.226 6.509 7.751 11.269 rebuild_ks_matrix 119 8.3 0.000 0.000 7.933 8.067 qs_ks_build_kohn_sham_matrix 119 9.3 0.010 0.011 7.933 8.066 multiply_cannon_sync_h2d 54864 15.5 5.910 7.577 5.910 7.577 qs_ks_update_qs_env 119 7.6 0.001 0.001 7.002 7.127 mp_sum_l 7207 12.9 5.356 7.086 5.356 7.086 qs_ot_get_p 119 10.4 0.001 0.001 6.642 6.933 init_scf_run 11 5.9 0.000 0.001 5.925 5.925 scf_env_initial_rho_setup 11 6.9 0.001 0.001 5.925 5.925 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 5.361 5.807 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 5.640 5.748 dbcsr_mm_accdrv_process 76910 16.1 1.166 1.804 3.448 4.853 sum_up_and_integrate 119 10.3 0.012 0.015 4.537 4.544 integrate_v_rspace 119 11.3 0.002 0.002 4.524 4.533 qs_rho_update_rho_low 119 7.7 0.001 0.001 4.244 4.355 calculate_rho_elec 119 8.7 0.011 0.017 4.243 4.355 qs_ot_p2m_diag 50 11.0 0.004 0.006 3.772 3.828 calculate_first_density_matrix 1 7.0 0.000 0.001 3.809 3.818 multiply_cannon_metrocomm1 54864 15.5 0.051 0.058 1.781 3.421 cp_dbcsr_sm_fm_multiply 37 9.5 0.051 0.401 3.152 3.154 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 2.871 3.084 apply_single 119 13.6 0.000 0.000 2.871 3.084 calculate_dm_sparse 119 9.5 0.000 0.000 2.914 3.041 jit_kernel_multiply 13 15.8 2.221 2.962 2.221 2.962 rs_pw_transfer 974 11.9 0.011 0.012 2.755 2.881 cp_dbcsr_syevd 50 12.0 0.003 0.003 2.877 2.877 acc_transpose_blocks 54864 15.5 0.232 0.251 2.125 2.766 ot_diis_step 108 11.5 0.006 0.006 2.674 2.674 cp_fm_diag_elpa 50 13.0 0.000 0.000 2.665 2.665 cp_fm_redistribute_end 50 14.0 2.426 2.643 2.431 2.644 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 2.512 2.566 cp_fm_diag_elpa_base 50 14.0 0.212 2.554 0.213 2.563 qs_ot_get_orbitals 108 10.5 0.000 0.000 2.421 2.499 density_rs2pw 119 9.7 0.004 0.004 2.256 2.424 grid_integrate_task_list 119 12.3 2.022 2.138 2.022 2.138 wfi_extrapolate 11 7.9 0.001 0.001 2.051 2.051 mp_sum_d 4129 12.0 1.350 2.045 1.350 2.045 init_scf_loop 11 6.9 0.000 0.000 2.044 2.044 potential_pw2rs 119 12.3 0.004 0.004 1.888 1.903 pw_transfer 1439 11.6 0.053 0.059 1.825 1.900 fft_wrap_pw1pw2 1201 12.6 0.007 0.008 1.748 1.826 make_m2s 4572 13.5 0.054 0.056 1.608 1.651 acc_transpose_blocks_kernels 54864 16.5 0.247 0.372 1.142 1.635 fft3d_ps 1201 14.6 0.370 0.473 1.519 1.589 make_images 4572 14.5 0.133 0.138 1.525 1.567 mp_alltoall_d11v 2130 13.8 1.274 1.522 1.274 1.522 mp_waitany 12084 13.8 1.240 1.455 1.240 1.455 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 1.431 1.451 fft_wrap_pw1pw2_140 487 13.2 0.082 0.095 1.346 1.426 grid_collocate_task_list 119 9.7 1.295 1.368 1.295 1.368 mp_alltoall_i22 627 13.8 0.975 1.266 0.975 1.266 jit_kernel_transpose 5 15.5 0.895 1.263 0.895 1.263 parallel_gemm_fm 81 9.0 0.000 0.000 1.255 1.259 parallel_gemm_fm_cosma 81 10.0 1.255 1.259 1.255 1.259 dbcsr_dot_sd 1205 11.9 0.051 0.063 0.731 1.130 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="100", plot="h2o_64_md", label="(8n/12r/1t)", y=53.510000, yerr=0.000000 PlotPoint: name="101", plot="h2o_64_md_mem", label="(8n/12r/1t)", y=431.090909, yerr=1.239835 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 488.030208E+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 1034343. MP_SendRecv 16779 37093. MP_ISendRecv 16779 37093. MP_Wait 23539 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 0.021 0.054 39.889 39.890 qs_mol_dyn_low 1 2.0 0.003 0.004 39.632 39.640 qs_forces 11 3.9 0.004 0.011 39.559 39.560 qs_energies 11 4.9 0.002 0.005 37.856 37.860 scf_env_do_scf 11 5.9 0.001 0.001 31.968 31.969 scf_env_do_scf_inner_loop 108 6.5 0.002 0.008 29.434 29.435 dbcsr_multiply_generic 2286 12.5 0.101 0.105 21.902 22.278 qs_scf_new_mos 108 7.5 0.001 0.001 20.108 20.344 qs_scf_loop_do_ot 108 8.5 0.001 0.001 20.108 20.344 ot_scf_mini 108 9.5 0.003 0.003 19.217 19.387 velocity_verlet 10 3.0 0.014 0.023 18.443 18.444 multiply_cannon 2286 13.5 0.206 0.214 16.907 18.413 multiply_cannon_loop 2286 14.5 0.902 0.974 15.812 17.251 ot_mini 108 10.5 0.001 0.001 11.956 12.189 mp_waitall_1 200699 16.5 5.920 11.530 5.920 11.530 multiply_cannon_metrocomm3 27432 15.5 0.068 0.071 4.431 10.270 qs_ot_get_derivative 108 11.5 0.001 0.001 9.501 9.674 multiply_cannon_multrec 27432 15.5 2.002 4.444 6.276 9.468 rebuild_ks_matrix 119 8.3 0.000 0.000 7.280 7.419 qs_ks_build_kohn_sham_matrix 119 9.3 0.012 0.015 7.279 7.418 qs_ks_update_qs_env 119 7.6 0.001 0.001 6.432 6.560 dbcsr_mm_accdrv_process 47894 16.0 3.076 5.562 4.206 6.426 qs_ot_get_p 119 10.4 0.001 0.001 4.367 4.610 init_scf_run 11 5.9 0.000 0.002 4.570 4.570 scf_env_initial_rho_setup 11 6.9 0.001 0.001 4.569 4.570 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 3.580 4.427 sum_up_and_integrate 119 10.3 0.024 0.026 4.257 4.262 integrate_v_rspace 119 11.3 0.002 0.003 4.232 4.239 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 3.048 4.116 apply_single 119 13.6 0.000 0.000 3.047 4.115 mp_sum_l 7207 12.9 2.060 4.010 2.060 4.010 qs_rho_update_rho_low 119 7.7 0.001 0.001 3.880 3.908 calculate_rho_elec 119 8.7 0.021 0.024 3.880 3.907 rs_pw_transfer 974 11.9 0.010 0.011 2.732 3.173 calculate_first_density_matrix 1 7.0 0.001 0.004 3.038 3.041 qs_ot_p2m_diag 50 11.0 0.008 0.013 2.925 2.944 multiply_cannon_sync_h2d 27432 15.5 2.185 2.843 2.185 2.843 make_m2s 4572 13.5 0.052 0.054 2.489 2.724 density_rs2pw 119 9.7 0.004 0.004 2.207 2.637 make_images 4572 14.5 0.199 0.237 2.399 2.633 init_scf_loop 11 6.9 0.001 0.003 2.513 2.513 cp_dbcsr_syevd 50 12.0 0.003 0.003 2.495 2.495 ot_diis_step 108 11.5 0.011 0.011 2.403 2.404 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 2.342 2.345 jit_kernel_multiply 8 16.0 1.078 2.253 1.078 2.253 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 2.138 2.227 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 2.134 2.174 cp_fm_diag_elpa 50 13.0 0.000 0.000 2.152 2.153 cp_fm_redistribute_end 50 14.0 1.780 2.126 1.784 2.128 calculate_dm_sparse 119 9.5 0.000 0.000 2.026 2.102 cp_fm_diag_elpa_base 50 14.0 0.329 2.009 0.342 2.073 pw_transfer 1439 11.6 0.065 0.073 2.028 2.068 potential_pw2rs 119 12.3 0.006 0.006 1.983 1.992 fft_wrap_pw1pw2 1201 12.6 0.008 0.009 1.936 1.980 grid_integrate_task_list 119 12.3 1.827 1.944 1.827 1.944 fft3d_ps 1201 14.6 0.512 0.566 1.638 1.677 prepare_preconditioner 11 7.9 0.000 0.001 1.594 1.620 make_preconditioner 11 8.9 0.000 0.001 1.594 1.620 fft_wrap_pw1pw2_140 487 13.2 0.079 0.086 1.510 1.552 make_full_inverse_cholesky 11 9.9 0.000 0.000 1.493 1.549 make_images_data 4572 15.5 0.045 0.051 1.139 1.546 wfi_extrapolate 11 7.9 0.001 0.001 1.479 1.479 acc_transpose_blocks 27432 15.5 0.109 0.116 1.183 1.474 hybrid_alltoall_any 4725 16.4 0.050 0.112 0.995 1.422 mp_alltoall_d11v 2130 13.8 1.199 1.411 1.199 1.411 grid_collocate_task_list 119 9.7 1.228 1.364 1.228 1.364 mp_allgather_i34 2286 14.5 0.539 1.303 0.539 1.303 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 1.246 1.253 qs_ot_get_orbitals 108 10.5 0.000 0.000 1.169 1.216 mp_sum_d 4129 12.0 0.613 1.066 0.613 1.066 rs_pw_transfer_RS2PW_140 130 11.5 0.139 0.150 0.610 1.059 mp_waitany 5720 13.7 0.585 1.043 0.585 1.043 qs_energies_init_hamiltonians 11 5.9 0.001 0.002 1.040 1.040 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 0.906 0.919 acc_transpose_blocks_kernels 27432 16.5 0.180 0.267 0.650 0.854 mp_alltoall_z22v 1201 16.6 0.722 0.821 0.722 0.821 rs_pw_transfer_PW2RS_50 119 14.3 0.587 0.606 0.753 0.806 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="102", plot="h2o_64_md", label="(8n/6r/2t)", y=39.890000, yerr=0.000000 PlotPoint: name="103", plot="h2o_64_md_mem", label="(8n/6r/2t)", y=464.727273, yerr=2.004128 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 521.654272E+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 1607838. MP_SendRecv 11067 57667. MP_ISendRecv 11067 57667. MP_Wait 21987 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 0.089 0.226 34.587 34.589 qs_mol_dyn_low 1 2.0 0.003 0.007 34.100 34.107 qs_forces 11 3.9 0.034 0.079 34.035 34.040 qs_energies 11 4.9 0.006 0.020 32.336 32.358 scf_env_do_scf 11 5.9 0.029 0.077 26.112 26.112 scf_env_do_scf_inner_loop 108 6.5 0.007 0.032 23.408 23.413 dbcsr_multiply_generic 2286 12.5 0.098 0.105 17.963 18.099 velocity_verlet 10 3.0 0.006 0.012 15.490 15.491 multiply_cannon 2286 13.5 0.195 0.201 14.586 15.351 qs_scf_new_mos 108 7.5 0.001 0.001 15.084 15.109 qs_scf_loop_do_ot 108 8.5 0.001 0.002 15.083 15.108 multiply_cannon_loop 2286 14.5 0.639 0.665 13.778 14.540 ot_scf_mini 108 9.5 0.003 0.004 14.344 14.362 ot_mini 108 10.5 0.001 0.001 8.936 8.959 multiply_cannon_multrec 18288 15.5 2.028 2.817 7.520 7.860 qs_ot_get_derivative 108 11.5 0.001 0.001 7.430 7.451 rebuild_ks_matrix 119 8.3 0.000 0.000 6.488 6.505 qs_ks_build_kohn_sham_matrix 119 9.3 0.013 0.015 6.487 6.504 dbcsr_mm_accdrv_process 38222 16.0 4.558 6.151 5.409 6.355 qs_ks_update_qs_env 119 7.6 0.001 0.001 5.718 5.734 init_scf_run 11 5.9 0.000 0.001 4.972 4.972 scf_env_initial_rho_setup 11 6.9 0.001 0.001 4.972 4.972 mp_waitall_1 158411 16.6 2.946 4.254 2.946 4.254 sum_up_and_integrate 119 10.3 0.029 0.031 4.042 4.049 integrate_v_rspace 119 11.3 0.003 0.003 4.012 4.019 calculate_first_density_matrix 1 7.0 0.002 0.003 3.742 3.743 qs_rho_update_rho_low 119 7.7 0.001 0.001 3.521 3.532 calculate_rho_elec 119 8.7 0.030 0.032 3.520 3.532 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 2.874 3.489 qs_ot_get_p 119 10.4 0.001 0.002 3.247 3.273 multiply_cannon_metrocomm3 18288 15.5 0.045 0.046 1.676 2.953 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 2.635 2.637 init_scf_loop 11 6.9 0.001 0.003 2.610 2.614 rs_pw_transfer 974 11.9 0.009 0.010 2.376 2.612 jit_kernel_multiply 10 15.8 0.800 2.510 0.800 2.510 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 2.470 2.478 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 2.037 2.366 apply_single 119 13.6 0.000 0.000 2.036 2.365 calculate_dm_sparse 119 9.5 0.000 0.000 2.313 2.322 density_rs2pw 119 9.7 0.004 0.004 2.045 2.316 qs_ot_p2m_diag 50 11.0 0.012 0.012 2.140 2.146 pw_transfer 1439 11.6 0.066 0.071 2.001 2.011 make_m2s 4572 13.5 0.044 0.046 1.823 1.974 fft_wrap_pw1pw2 1201 12.6 0.008 0.009 1.907 1.922 grid_integrate_task_list 119 12.3 1.807 1.902 1.807 1.902 make_images 4572 14.5 0.190 0.201 1.737 1.887 cp_dbcsr_syevd 50 12.0 0.003 0.003 1.877 1.878 potential_pw2rs 119 12.3 0.007 0.008 1.819 1.827 prepare_preconditioner 11 7.9 0.000 0.000 1.785 1.787 make_preconditioner 11 8.9 0.001 0.001 1.785 1.787 acc_transpose_blocks 18288 15.5 0.077 0.078 1.679 1.756 make_full_inverse_cholesky 11 9.9 0.000 0.000 1.637 1.719 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 1.667 1.671 cp_fm_diag_elpa 50 13.0 0.000 0.000 1.636 1.637 cp_fm_redistribute_end 50 14.0 1.220 1.616 1.221 1.616 fft3d_ps 1201 14.6 0.524 0.541 1.586 1.600 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 1.585 1.594 cp_fm_diag_elpa_base 50 14.0 0.379 1.536 0.393 1.582 multiply_cannon_sync_h2d 18288 15.5 1.416 1.555 1.416 1.555 fft_wrap_pw1pw2_140 487 13.2 0.090 0.094 1.513 1.527 mp_sum_l 7207 12.9 1.164 1.513 1.164 1.513 ot_diis_step 108 11.5 0.011 0.011 1.483 1.483 grid_collocate_task_list 119 9.7 1.215 1.371 1.215 1.371 acc_transpose_blocks_kernels 18288 16.5 0.208 0.216 1.240 1.314 wfi_extrapolate 11 7.9 0.001 0.001 1.179 1.179 jit_kernel_transpose 5 15.6 1.032 1.100 1.032 1.100 qs_energies_init_hamiltonians 11 5.9 0.001 0.002 0.983 0.986 make_images_data 4572 15.5 0.045 0.048 0.794 0.961 qs_ot_get_orbitals 108 10.5 0.000 0.000 0.889 0.915 hybrid_alltoall_any 4725 16.4 0.055 0.113 0.685 0.874 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 0.841 0.843 multiply_cannon_metrocomm1 18288 15.5 0.028 0.029 0.450 0.806 mp_waitany 9880 13.7 0.543 0.803 0.543 0.803 mp_alltoall_z22v 1201 16.6 0.660 0.773 0.660 0.773 rs_pw_transfer_RS2PW_140 130 11.5 0.119 0.122 0.525 0.768 cp_fm_cholesky_invert 11 10.9 0.758 0.762 0.758 0.762 build_core_hamiltonian_matrix_ 11 4.9 0.000 0.000 0.662 0.742 mp_alltoall_d11v 2130 13.8 0.651 0.739 0.651 0.739 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="104", plot="h2o_64_md", label="(8n/4r/3t)", y=34.589000, yerr=0.000000 PlotPoint: name="105", plot="h2o_64_md_mem", label="(8n/4r/3t)", y=496.181818, yerr=1.799908 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 552.038400E+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_SendRecv 8211 74133. MP_ISendRecv 8211 74133. MP_Wait 16271 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 0.015 0.035 36.212 36.214 qs_mol_dyn_low 1 2.0 0.003 0.004 35.896 35.903 qs_forces 11 3.9 0.004 0.018 35.778 35.779 qs_energies 11 4.9 0.003 0.023 34.038 34.043 scf_env_do_scf 11 5.9 0.001 0.001 28.308 28.309 scf_env_do_scf_inner_loop 108 6.5 0.002 0.006 24.646 24.647 dbcsr_multiply_generic 2286 12.5 0.101 0.102 18.427 18.525 velocity_verlet 10 3.0 0.008 0.035 18.035 18.040 qs_scf_new_mos 108 7.5 0.001 0.001 15.980 16.039 qs_scf_loop_do_ot 108 8.5 0.001 0.001 15.979 16.039 multiply_cannon 2286 13.5 0.221 0.262 14.839 15.248 ot_scf_mini 108 9.5 0.003 0.004 15.045 15.096 multiply_cannon_loop 2286 14.5 0.929 0.965 13.891 14.202 ot_mini 108 10.5 0.001 0.001 9.165 9.229 multiply_cannon_multrec 27432 15.5 2.327 3.012 8.704 9.119 dbcsr_mm_accdrv_process 47916 15.9 5.407 7.193 6.285 7.499 qs_ot_get_derivative 108 11.5 0.001 0.001 7.363 7.418 rebuild_ks_matrix 119 8.3 0.000 0.000 6.883 6.937 qs_ks_build_kohn_sham_matrix 119 9.3 0.012 0.013 6.882 6.936 qs_ks_update_qs_env 119 7.6 0.001 0.001 6.140 6.189 init_scf_run 11 5.9 0.000 0.001 4.286 4.289 scf_env_initial_rho_setup 11 6.9 0.001 0.001 4.286 4.289 sum_up_and_integrate 119 10.3 0.035 0.038 3.897 3.906 integrate_v_rspace 119 11.3 0.002 0.004 3.862 3.871 qs_rho_update_rho_low 119 7.7 0.001 0.001 3.694 3.728 calculate_rho_elec 119 8.7 0.040 0.046 3.693 3.727 init_scf_loop 11 6.9 0.001 0.008 3.640 3.640 qs_ot_get_p 119 10.4 0.001 0.001 3.270 3.347 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 2.764 3.165 calculate_first_density_matrix 1 7.0 0.000 0.003 2.874 2.878 prepare_preconditioner 11 7.9 0.000 0.001 2.589 2.598 make_preconditioner 11 8.9 0.000 0.001 2.589 2.598 make_full_inverse_cholesky 11 9.9 0.000 0.000 2.188 2.516 rs_pw_transfer 974 11.9 0.009 0.010 2.172 2.377 mp_waitall_1 137007 16.6 1.854 2.362 1.854 2.362 density_rs2pw 119 9.7 0.004 0.004 2.101 2.299 make_m2s 4572 13.5 0.054 0.056 2.192 2.297 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.003 2.252 2.253 pw_transfer 1439 11.6 0.066 0.072 2.166 2.206 make_images 4572 14.5 0.269 0.330 2.084 2.187 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 2.029 2.160 apply_single 119 13.6 0.000 0.000 2.029 2.160 calculate_dm_sparse 119 9.5 0.000 0.000 2.104 2.153 fft_wrap_pw1pw2 1201 12.6 0.008 0.008 2.072 2.118 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 2.036 2.049 qs_ot_p2m_diag 50 11.0 0.015 0.022 2.000 2.009 jit_kernel_multiply 10 16.0 0.817 1.992 0.817 1.992 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 1.919 1.946 grid_integrate_task_list 119 12.3 1.813 1.916 1.813 1.916 fft3d_ps 1201 14.6 0.560 0.611 1.746 1.787 acc_transpose_blocks 27432 15.5 0.112 0.115 1.752 1.785 ot_diis_step 108 11.5 0.012 0.012 1.762 1.762 fft_wrap_pw1pw2_140 487 13.2 0.088 0.097 1.702 1.752 cp_dbcsr_syevd 50 12.0 0.003 0.003 1.684 1.684 potential_pw2rs 119 12.3 0.009 0.009 1.646 1.651 cp_fm_diag_elpa 50 13.0 0.000 0.000 1.445 1.446 cp_fm_redistribute_end 50 14.0 0.957 1.422 0.958 1.423 cp_fm_diag_elpa_base 50 14.0 0.442 1.355 0.462 1.394 grid_collocate_task_list 119 9.7 1.224 1.357 1.224 1.357 wfi_extrapolate 11 7.9 0.001 0.001 1.338 1.338 multiply_cannon_metrocomm3 27432 15.5 0.038 0.039 0.773 1.310 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 1.289 1.298 mp_sum_l 7207 12.9 0.970 1.262 0.970 1.262 qs_energies_init_hamiltonians 11 5.9 0.001 0.002 1.178 1.179 acc_transpose_blocks_kernels 27432 16.5 0.266 0.275 1.130 1.146 qs_ot_get_orbitals 108 10.5 0.000 0.000 1.108 1.130 cp_fm_upper_to_full 72 13.5 0.803 1.116 0.803 1.116 dbcsr_complete_redistribute 329 12.2 0.145 0.194 0.821 1.095 multiply_cannon_sync_h2d 27432 15.5 0.996 1.061 0.996 1.061 make_images_data 4572 15.5 0.045 0.049 0.849 0.973 mp_alltoall_d11v 2130 13.8 0.752 0.927 0.752 0.927 hybrid_alltoall_any 4725 16.4 0.061 0.150 0.721 0.918 copy_fm_to_dbcsr 176 11.2 0.001 0.001 0.618 0.884 jit_kernel_transpose 5 15.6 0.864 0.877 0.864 0.877 build_core_hamiltonian_matrix_ 11 4.9 0.000 0.000 0.793 0.871 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 0.837 0.842 mp_sum_d 4127 12.0 0.577 0.828 0.577 0.828 mp_alltoall_z22v 1201 16.6 0.769 0.809 0.769 0.809 cp_fm_cholesky_invert 11 10.9 0.773 0.777 0.773 0.777 mp_waitany 7280 13.7 0.551 0.758 0.551 0.758 rs_pw_transfer_RS2PW_140 130 11.5 0.106 0.112 0.551 0.741 mp_alltoall_i22 627 13.8 0.431 0.727 0.431 0.727 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="106", plot="h2o_64_md", label="(8n/3r/4t)", y=36.214000, yerr=0.000000 PlotPoint: name="107", plot="h2o_64_md_mem", label="(8n/3r/4t)", y=524.000000, yerr=4.134115 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 605.483008E+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_SendRecv 5355 94533. MP_ISendRecv 5355 94533. MP_Wait 11335 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 0.011 0.028 30.078 30.080 qs_mol_dyn_low 1 2.0 0.003 0.004 29.793 29.801 qs_forces 11 3.9 0.014 0.017 29.735 29.737 qs_energies 11 4.9 0.001 0.001 28.006 28.009 scf_env_do_scf 11 5.9 0.000 0.001 22.594 22.594 scf_env_do_scf_inner_loop 108 6.5 0.002 0.006 20.045 20.046 velocity_verlet 10 3.0 0.002 0.002 14.674 14.676 dbcsr_multiply_generic 2286 12.5 0.094 0.097 13.338 13.410 qs_scf_new_mos 108 7.5 0.001 0.001 11.911 11.943 qs_scf_loop_do_ot 108 8.5 0.001 0.001 11.911 11.942 ot_scf_mini 108 9.5 0.002 0.002 11.226 11.256 multiply_cannon 2286 13.5 0.229 0.234 10.582 10.969 multiply_cannon_loop 2286 14.5 0.330 0.342 9.620 9.986 multiply_cannon_multrec 9144 15.5 1.595 1.866 6.449 6.879 ot_mini 108 10.5 0.001 0.001 6.618 6.654 rebuild_ks_matrix 119 8.3 0.000 0.000 6.067 6.093 qs_ks_build_kohn_sham_matrix 119 9.3 0.012 0.013 6.066 6.093 qs_ks_update_qs_env 119 7.6 0.001 0.001 5.411 5.435 qs_ot_get_derivative 108 11.5 0.001 0.001 5.303 5.333 dbcsr_mm_accdrv_process 12550 15.8 3.688 4.774 4.755 5.086 init_scf_run 11 5.9 0.000 0.001 3.968 3.968 scf_env_initial_rho_setup 11 6.9 0.001 0.001 3.968 3.968 sum_up_and_integrate 119 10.3 0.038 0.041 3.726 3.731 integrate_v_rspace 119 11.3 0.003 0.003 3.688 3.693 qs_rho_update_rho_low 119 7.7 0.001 0.001 3.608 3.617 calculate_rho_elec 119 8.7 0.060 0.061 3.607 3.616 qs_ot_get_p 119 10.4 0.001 0.001 2.807 2.852 calculate_first_density_matrix 1 7.0 0.000 0.000 2.802 2.804 init_scf_loop 11 6.9 0.000 0.000 2.529 2.530 mp_waitall_1 115863 16.7 1.600 2.412 1.600 2.412 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 2.215 2.389 pw_transfer 1439 11.6 0.066 0.069 2.178 2.188 fft_wrap_pw1pw2 1201 12.6 0.008 0.008 2.084 2.097 density_rs2pw 119 9.7 0.004 0.004 1.932 2.068 calculate_dm_sparse 119 9.5 0.000 0.000 1.988 2.004 make_m2s 4572 13.5 0.034 0.036 1.764 1.978 grid_integrate_task_list 119 12.3 1.847 1.921 1.847 1.921 make_images 4572 14.5 0.267 0.297 1.674 1.886 rs_pw_transfer 974 11.9 0.008 0.009 1.740 1.883 qs_ot_p2m_diag 50 11.0 0.022 0.023 1.870 1.872 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 1.339 1.836 apply_single 119 13.6 0.000 0.000 1.339 1.835 prepare_preconditioner 11 7.9 0.000 0.000 1.773 1.778 make_preconditioner 11 8.9 0.000 0.000 1.773 1.778 fft3d_ps 1201 14.6 0.561 0.571 1.743 1.754 fft_wrap_pw1pw2_140 487 13.2 0.086 0.090 1.715 1.727 jit_kernel_multiply 6 15.9 1.030 1.704 1.030 1.704 make_full_inverse_cholesky 11 9.9 0.000 0.000 1.661 1.688 cp_dbcsr_syevd 50 12.0 0.003 0.003 1.673 1.674 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 1.569 1.570 potential_pw2rs 119 12.3 0.010 0.011 1.468 1.471 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 1.406 1.412 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 1.398 1.403 grid_collocate_task_list 119 9.7 1.272 1.393 1.272 1.393 cp_fm_diag_elpa 50 13.0 0.000 0.000 1.374 1.375 cp_fm_redistribute_end 50 14.0 0.687 1.354 0.688 1.355 cp_fm_diag_elpa_base 50 14.0 0.622 1.281 0.665 1.337 ot_diis_step 108 11.5 0.012 0.013 1.302 1.302 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 1.231 1.233 multiply_cannon_metrocomm3 9144 15.5 0.019 0.019 0.453 1.201 mp_sum_l 7207 12.9 0.559 1.182 0.559 1.182 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 1.143 1.158 wfi_extrapolate 11 7.9 0.001 0.001 1.116 1.116 acc_transpose_blocks 9144 15.5 0.038 0.039 1.077 1.100 hybrid_alltoall_any 4725 16.4 0.062 0.175 0.789 1.094 make_images_data 4572 15.5 0.039 0.042 0.814 1.058 build_core_hamiltonian_matrix_ 11 4.9 0.000 0.000 0.869 0.919 mp_alltoall_d11v 2130 13.8 0.787 0.900 0.787 0.900 cp_fm_cholesky_invert 11 10.9 0.886 0.889 0.886 0.889 acc_transpose_blocks_kernels 9144 16.5 0.117 0.121 0.835 0.859 multiply_cannon_sync_h2d 9144 15.5 0.710 0.797 0.710 0.797 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 0.779 0.781 qs_ot_get_orbitals 108 10.5 0.000 0.000 0.751 0.758 jit_kernel_transpose 5 15.6 0.718 0.744 0.718 0.744 qs_env_update_s_mstruct 11 6.9 0.000 0.000 0.685 0.735 parallel_gemm_fm 81 9.0 0.000 0.000 0.733 0.734 parallel_gemm_fm_cosma 81 10.0 0.733 0.734 0.733 0.734 mp_allgather_i34 2286 14.5 0.294 0.707 0.294 0.707 make_basis_sm 11 9.8 0.000 0.000 0.703 0.704 mp_alltoall_z22v 1201 16.6 0.630 0.682 0.630 0.682 yz_to_x 606 15.1 0.265 0.275 0.590 0.606 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="108", plot="h2o_64_md", label="(8n/2r/6t)", y=30.080000, yerr=0.000000 PlotPoint: name="109", plot="h2o_64_md_mem", label="(8n/2r/6t)", y=572.000000, yerr=6.266796 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 741.785600E+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_SendRecv 2499 189067. MP_ISendRecv 2499 189067. MP_Wait 6399 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 0.014 0.036 41.972 41.974 qs_mol_dyn_low 1 2.0 0.003 0.003 41.750 41.758 qs_forces 11 3.9 0.002 0.002 41.675 41.676 qs_energies 11 4.9 0.001 0.001 39.675 39.679 scf_env_do_scf 11 5.9 0.001 0.001 33.795 33.795 scf_env_do_scf_inner_loop 108 6.5 0.003 0.007 25.759 25.761 velocity_verlet 10 3.0 0.012 0.026 23.645 23.650 dbcsr_multiply_generic 2286 12.5 0.101 0.102 17.662 17.913 qs_scf_new_mos 108 7.5 0.001 0.001 15.843 15.940 qs_scf_loop_do_ot 108 8.5 0.001 0.001 15.842 15.940 ot_scf_mini 108 9.5 0.002 0.002 14.761 14.866 multiply_cannon 2286 13.5 0.299 0.306 13.804 14.847 multiply_cannon_loop 2286 14.5 0.343 0.348 12.523 13.518 ot_mini 108 10.5 0.001 0.001 8.770 8.893 multiply_cannon_multrec 9144 15.5 3.346 4.677 8.530 8.681 init_scf_loop 11 6.9 0.000 0.000 8.005 8.009 rebuild_ks_matrix 119 8.3 0.000 0.000 7.419 7.565 qs_ks_build_kohn_sham_matrix 119 9.3 0.013 0.013 7.418 7.565 prepare_preconditioner 11 7.9 0.000 0.000 6.919 6.934 make_preconditioner 11 8.9 0.000 0.000 6.919 6.934 qs_ot_get_derivative 108 11.5 0.001 0.001 6.748 6.853 qs_ks_update_qs_env 119 7.6 0.001 0.001 6.707 6.839 make_full_inverse_cholesky 11 9.9 0.000 0.000 5.454 6.813 dbcsr_mm_accdrv_process 12550 15.8 4.233 6.294 5.064 6.399 cp_fm_upper_to_full 72 14.2 3.201 4.610 3.201 4.610 qs_rho_update_rho_low 119 7.7 0.001 0.001 4.293 4.327 calculate_rho_elec 119 8.7 0.118 0.121 4.292 4.326 sum_up_and_integrate 119 10.3 0.065 0.066 4.063 4.069 integrate_v_rspace 119 11.3 0.003 0.003 3.998 4.005 init_scf_run 11 5.9 0.000 0.001 3.849 3.849 scf_env_initial_rho_setup 11 6.9 0.001 0.001 3.848 3.849 mp_waitall_1 94719 16.7 2.439 3.512 2.439 3.512 qs_ot_get_p 119 10.4 0.001 0.001 3.313 3.458 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 2.484 2.869 dbcsr_complete_redistribute 329 12.2 0.287 0.293 2.013 2.842 pw_transfer 1439 11.6 0.069 0.070 2.809 2.814 fft_wrap_pw1pw2 1201 12.6 0.009 0.009 2.710 2.715 copy_fm_to_dbcsr 176 11.2 0.001 0.001 1.694 2.515 make_m2s 4572 13.5 0.038 0.038 2.316 2.510 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 2.146 2.443 apply_single 119 13.6 0.000 0.000 2.146 2.443 make_images 4572 14.5 0.350 0.379 2.194 2.386 calculate_first_density_matrix 1 7.0 0.000 0.000 2.356 2.363 fft3d_ps 1201 14.6 0.596 0.606 2.333 2.339 mp_alltoall_i22 627 13.8 1.440 2.303 1.440 2.303 calculate_dm_sparse 119 9.5 0.000 0.000 2.277 2.299 multiply_cannon_metrocomm3 9144 15.5 0.020 0.020 1.341 2.296 fft_wrap_pw1pw2_140 487 13.2 0.095 0.097 2.274 2.283 transfer_fm_to_dbcsr 11 9.9 0.000 0.000 1.460 2.274 density_rs2pw 119 9.7 0.004 0.004 2.219 2.240 grid_integrate_task_list 119 12.3 2.058 2.077 2.058 2.077 ot_diis_step 108 11.5 0.014 0.014 1.997 1.997 qs_ot_p2m_diag 50 11.0 0.042 0.043 1.956 1.957 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 1.931 1.932 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 1.760 1.761 mp_sum_l 7207 12.9 0.997 1.739 0.997 1.739 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 1.680 1.732 jit_kernel_multiply 10 15.5 0.804 1.720 0.804 1.720 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 1.673 1.719 cp_dbcsr_syevd 50 12.0 0.003 0.003 1.682 1.682 grid_collocate_task_list 119 9.7 1.490 1.508 1.490 1.508 rs_pw_transfer 974 11.9 0.009 0.010 1.477 1.503 cp_fm_cholesky_invert 11 10.9 1.470 1.473 1.470 1.473 potential_pw2rs 119 12.3 0.014 0.015 1.464 1.467 hybrid_alltoall_any 4725 16.4 0.087 0.148 1.171 1.424 wfi_extrapolate 11 7.9 0.001 0.001 1.422 1.422 cp_fm_diag_elpa 50 13.0 0.000 0.000 1.390 1.391 cp_fm_diag_elpa_base 50 14.0 1.245 1.299 1.389 1.389 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 1.346 1.358 make_images_data 4572 15.5 0.043 0.046 1.113 1.354 qs_ot_get_orbitals 108 10.5 0.000 0.000 1.152 1.172 acc_transpose_blocks 9144 15.5 0.038 0.039 1.105 1.116 qs_env_update_s_mstruct 11 6.9 0.000 0.000 1.095 1.110 mp_alltoall_d11v 2130 13.8 1.067 1.108 1.067 1.108 multiply_cannon_sync_h2d 9144 15.5 1.042 1.046 1.042 1.046 build_core_hamiltonian_matrix_ 11 4.9 0.000 0.000 0.985 1.026 qs_create_task_list 11 7.9 0.000 0.001 0.939 0.951 generate_qs_task_list 11 8.9 0.371 0.390 0.939 0.951 yz_to_x 606 15.1 0.461 0.472 0.929 0.938 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 0.923 0.936 acc_transpose_blocks_kernels 9144 16.5 0.117 0.119 0.855 0.864 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="110", plot="h2o_64_md", label="(8n/1r/12t)", y=41.974000, yerr=0.000000 PlotPoint: name="111", plot="h2o_64_md_mem", label="(8n/1r/12t)", y=700.727273, yerr=11.497754 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 503.750656E+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 1379060. MP_SendRecv 20900 9096. MP_ISendRecv 20900 9096. MP_Wait 37268 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 0.018 0.066 85.805 85.807 qs_mol_dyn_low 1 2.0 0.004 0.012 85.476 85.496 qs_forces 11 3.9 0.003 0.006 84.661 84.664 qs_energies 11 4.9 0.006 0.041 81.772 81.782 scf_env_do_scf 11 5.9 0.000 0.001 72.614 72.617 scf_env_do_scf_inner_loop 99 6.5 0.002 0.006 66.934 66.934 dbcsr_multiply_generic 2055 12.4 0.108 0.112 53.005 53.298 qs_scf_new_mos 99 7.5 0.000 0.001 49.365 49.471 qs_scf_loop_do_ot 99 8.5 0.001 0.001 49.364 49.470 ot_scf_mini 99 9.5 0.002 0.003 46.885 46.959 multiply_cannon 2055 13.4 0.173 0.179 43.239 44.069 velocity_verlet 10 3.0 0.001 0.002 43.651 43.666 multiply_cannon_loop 2055 14.4 1.529 1.566 42.272 43.142 ot_mini 99 10.5 0.001 0.001 28.571 28.646 qs_ot_get_derivative 99 11.5 0.009 0.392 21.716 21.810 multiply_cannon_multrec 49320 15.4 12.084 12.760 17.609 18.491 rebuild_ks_matrix 110 8.3 0.000 0.000 14.663 14.811 qs_ks_build_kohn_sham_matrix 110 9.3 0.011 0.016 14.663 14.811 qs_ks_update_qs_env 110 7.6 0.001 0.001 12.871 12.993 mp_waitall_1 220248 16.4 11.848 12.876 11.848 12.876 multiply_cannon_sync_h2d 49320 15.4 9.903 10.541 9.903 10.541 qs_ot_get_p 110 10.4 0.001 0.001 9.810 9.905 multiply_cannon_metrocomm3 49320 15.4 0.077 0.081 7.218 8.532 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 7.505 8.056 apply_preconditioner_dbcsr 110 12.6 0.000 0.001 7.145 7.602 apply_single 110 13.6 0.000 0.000 7.145 7.602 sum_up_and_integrate 110 10.3 0.037 0.046 7.198 7.211 integrate_v_rspace 110 11.3 0.003 0.004 7.161 7.183 init_scf_run 11 5.9 0.000 0.002 6.980 6.980 scf_env_initial_rho_setup 11 6.9 0.001 0.002 6.980 6.980 qs_rho_update_rho_low 110 7.6 0.001 0.001 6.716 6.872 calculate_rho_elec 110 8.6 0.021 0.026 6.715 6.872 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 6.806 6.850 ot_diis_step 99 11.5 0.012 0.057 6.577 6.577 qs_ot_p2m_diag 48 11.0 0.012 0.018 6.446 6.477 mp_sum_l 6514 12.8 5.442 6.300 5.442 6.300 dbcsr_mm_accdrv_process 87628 16.1 2.071 2.181 5.404 5.826 init_scf_loop 11 6.9 0.000 0.000 5.652 5.653 cp_dbcsr_syevd 48 12.0 0.002 0.003 5.436 5.437 cp_fm_diag_elpa 48 13.0 0.000 0.000 4.936 4.938 cp_fm_redistribute_end 48 14.0 4.306 4.908 4.310 4.910 cp_fm_diag_elpa_base 48 14.0 0.593 4.758 0.596 4.787 rs_pw_transfer 902 11.9 0.012 0.013 3.753 4.346 wfi_extrapolate 11 7.9 0.001 0.003 4.048 4.048 density_rs2pw 110 9.6 0.004 0.005 3.500 4.037 make_m2s 4110 13.4 0.061 0.065 3.894 4.015 calculate_dm_sparse 110 9.5 0.000 0.001 3.867 3.967 multiply_cannon_metrocomm1 49320 15.4 0.059 0.063 2.776 3.934 make_images 4110 14.4 0.177 0.191 3.798 3.922 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.004 3.695 3.699 prepare_preconditioner 11 7.9 0.000 0.000 3.456 3.480 make_preconditioner 11 8.9 0.000 0.000 3.456 3.480 grid_integrate_task_list 110 12.3 3.263 3.403 3.263 3.403 pw_transfer 1331 11.6 0.054 0.066 3.332 3.396 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 3.336 3.373 qs_ot_get_orbitals 99 10.5 0.000 0.001 3.280 3.344 fft_wrap_pw1pw2 1111 12.6 0.008 0.009 3.245 3.308 make_full_inverse_cholesky 11 9.9 0.000 0.000 3.216 3.264 jit_kernel_multiply 13 15.9 3.052 3.163 3.052 3.163 calculate_first_density_matrix 1 7.0 0.000 0.002 2.843 2.848 fft3d_ps 1111 14.6 0.778 0.880 2.772 2.826 fft_wrap_pw1pw2_140 451 13.1 0.171 0.191 2.732 2.798 potential_pw2rs 110 12.3 0.006 0.007 2.722 2.748 mp_alltoall_d11v 2046 13.8 2.106 2.492 2.106 2.492 mp_waitany 14300 13.8 1.809 2.433 1.809 2.433 grid_collocate_task_list 110 9.6 2.082 2.280 2.082 2.280 acc_transpose_blocks 49320 15.4 0.226 0.237 2.162 2.273 mp_sum_d 3883 11.9 1.610 2.191 1.610 2.191 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.936 1.957 cp_fm_cholesky_invert 11 10.9 1.891 1.895 1.891 1.895 make_images_data 4110 15.4 0.042 0.045 1.736 1.872 hybrid_alltoall_any 4261 16.3 0.081 0.479 1.508 1.734 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="200", plot="h2o_128_md", label="(8n/12r/1t)", y=85.807000, yerr=0.000000 PlotPoint: name="201", plot="h2o_128_md_mem", label="(8n/12r/1t)", y=476.909091, yerr=2.937363 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 587.370496E+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 2913700. MP_SendRecv 10340 26400. MP_ISendRecv 10340 26400. MP_Wait 22352 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 0.117 0.405 73.385 73.389 qs_mol_dyn_low 1 2.0 0.004 0.041 72.063 72.072 qs_forces 11 3.9 0.020 0.029 71.075 71.080 qs_energies 11 4.9 0.002 0.005 67.628 67.648 scf_env_do_scf 11 5.9 0.001 0.001 58.036 58.039 scf_env_do_scf_inner_loop 99 6.5 0.002 0.007 50.096 50.123 dbcsr_multiply_generic 2055 12.4 0.116 0.120 38.273 38.432 velocity_verlet 10 3.0 0.002 0.002 37.870 37.891 qs_scf_new_mos 99 7.5 0.001 0.001 33.385 33.543 qs_scf_loop_do_ot 99 8.5 0.001 0.001 33.384 33.543 multiply_cannon 2055 13.4 0.221 0.242 31.448 32.738 ot_scf_mini 99 9.5 0.003 0.004 31.708 31.864 multiply_cannon_loop 2055 14.4 0.931 0.964 30.062 30.873 ot_mini 99 10.5 0.001 0.001 18.510 18.660 multiply_cannon_multrec 24660 15.4 7.630 9.648 14.070 16.237 rebuild_ks_matrix 110 8.3 0.000 0.000 14.108 14.161 qs_ks_build_kohn_sham_matrix 110 9.3 0.015 0.032 14.108 14.161 qs_ot_get_derivative 99 11.5 0.001 0.001 12.624 12.782 qs_ks_update_qs_env 110 7.6 0.001 0.001 12.410 12.458 mp_waitall_1 176588 16.5 7.738 10.745 7.738 10.745 multiply_cannon_sync_h2d 24660 15.4 7.140 8.263 7.140 8.263 multiply_cannon_metrocomm3 24660 15.4 0.070 0.072 5.206 8.058 init_scf_loop 11 6.9 0.004 0.016 7.903 7.926 apply_preconditioner_dbcsr 110 12.6 0.000 0.000 6.604 7.340 apply_single 110 13.6 0.000 0.001 6.604 7.340 init_scf_run 11 5.9 0.000 0.001 7.052 7.053 scf_env_initial_rho_setup 11 6.9 0.001 0.001 7.052 7.053 sum_up_and_integrate 110 10.3 0.051 0.058 6.820 6.836 integrate_v_rspace 110 11.3 0.002 0.003 6.769 6.783 qs_ot_get_p 110 10.4 0.001 0.001 6.543 6.655 dbcsr_mm_accdrv_process 52282 16.1 4.701 5.629 6.281 6.567 qs_rho_update_rho_low 110 7.6 0.001 0.001 6.426 6.436 calculate_rho_elec 110 8.6 0.039 0.047 6.425 6.436 ot_diis_step 99 11.5 0.010 0.010 5.835 5.835 prepare_preconditioner 11 7.9 0.000 0.001 5.736 5.750 make_preconditioner 11 8.9 0.000 0.001 5.736 5.750 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 4.742 5.470 make_full_inverse_cholesky 11 9.9 0.000 0.000 5.288 5.444 make_m2s 4110 13.4 0.057 0.060 4.300 4.814 make_images 4110 14.4 0.395 0.438 4.190 4.702 qs_ot_p2m_diag 48 11.0 0.028 0.043 4.543 4.564 cp_dbcsr_syevd 48 12.0 0.003 0.003 4.109 4.109 pw_transfer 1331 11.6 0.066 0.074 3.928 4.080 density_rs2pw 110 9.6 0.004 0.005 3.533 4.002 fft_wrap_pw1pw2 1111 12.6 0.009 0.009 3.821 3.977 rs_pw_transfer 902 11.9 0.012 0.014 3.197 3.729 wfi_extrapolate 11 7.9 0.001 0.001 3.584 3.585 cp_fm_diag_elpa 48 13.0 0.000 0.000 3.519 3.521 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.002 3.517 3.519 cp_fm_redistribute_end 48 14.0 2.628 3.486 2.630 3.486 cp_fm_diag_elpa_base 48 14.0 0.820 3.334 0.852 3.420 fft_wrap_pw1pw2_140 451 13.1 0.203 0.220 3.199 3.360 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 3.277 3.353 grid_integrate_task_list 110 12.3 3.155 3.344 3.155 3.344 calculate_first_density_matrix 1 7.0 0.000 0.002 3.322 3.325 fft3d_ps 1111 14.6 1.103 1.322 3.170 3.316 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 3.177 3.211 calculate_dm_sparse 110 9.5 0.001 0.001 2.980 3.012 make_images_data 4110 15.4 0.046 0.050 2.389 2.909 mp_sum_l 6514 12.8 2.107 2.871 2.107 2.871 hybrid_alltoall_any 4261 16.3 0.101 0.439 2.056 2.838 cp_fm_cholesky_invert 11 10.9 2.698 2.706 2.698 2.706 potential_pw2rs 110 12.3 0.008 0.009 2.576 2.590 jit_kernel_multiply 12 16.2 1.229 2.544 1.229 2.544 grid_collocate_task_list 110 9.6 2.056 2.508 2.056 2.508 mp_alltoall_d11v 2046 13.8 1.863 2.265 1.863 2.265 qs_energies_init_hamiltonians 11 5.9 0.000 0.001 2.005 2.024 qs_ot_get_orbitals 99 10.5 0.001 0.001 1.979 1.996 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.872 1.883 mp_waitany 10164 13.8 1.311 1.808 1.311 1.808 mp_allgather_i34 2055 14.4 0.716 1.770 0.716 1.770 rs_pw_transfer_RS2PW_140 121 11.5 0.223 0.273 1.069 1.607 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.001 0.001 1.540 1.552 acc_transpose_blocks 24660 15.4 0.112 0.115 1.518 1.546 multiply_cannon_metrocomm4 22605 15.4 0.073 0.077 0.765 1.536 make_basis_sm 11 9.8 0.000 0.000 1.515 1.518 parallel_gemm_fm 81 9.0 0.000 0.000 1.494 1.500 parallel_gemm_fm_cosma 81 10.0 1.493 1.499 1.493 1.499 mp_alltoall_z22v 1111 16.6 1.380 1.497 1.380 1.497 build_core_hamiltonian_matrix_ 11 4.9 0.001 0.001 1.380 1.484 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="202", plot="h2o_128_md", label="(8n/6r/2t)", y=73.389000, yerr=0.000000 PlotPoint: name="203", plot="h2o_128_md_mem", label="(8n/6r/2t)", y=556.181818, yerr=7.246544 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 660.025344E+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_SendRecv 6820 27424. MP_ISendRecv 6820 27424. MP_Wait 25498 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 0.046 0.143 62.586 62.595 qs_mol_dyn_low 1 2.0 0.003 0.004 61.957 62.048 qs_forces 11 3.9 0.005 0.029 61.584 61.590 qs_energies 11 4.9 0.006 0.022 58.344 58.356 scf_env_do_scf 11 5.9 0.001 0.001 49.234 49.235 scf_env_do_scf_inner_loop 99 6.5 0.002 0.007 40.471 40.471 velocity_verlet 10 3.0 0.013 0.025 33.226 33.237 dbcsr_multiply_generic 2055 12.4 0.110 0.114 28.712 28.968 qs_scf_new_mos 99 7.5 0.001 0.001 25.219 25.340 qs_scf_loop_do_ot 99 8.5 0.001 0.001 25.218 25.340 ot_scf_mini 99 9.5 0.003 0.003 23.992 24.127 multiply_cannon 2055 13.4 0.211 0.224 22.476 23.686 multiply_cannon_loop 2055 14.4 0.615 0.634 21.266 22.367 ot_mini 99 10.5 0.001 0.001 13.701 13.839 rebuild_ks_matrix 110 8.3 0.000 0.000 12.568 12.740 qs_ks_build_kohn_sham_matrix 110 9.3 0.012 0.015 12.568 12.740 qs_ks_update_qs_env 110 7.6 0.001 0.001 11.083 11.237 multiply_cannon_multrec 16440 15.4 3.825 4.703 10.019 10.859 mp_waitall_1 139946 16.5 6.951 10.151 6.951 10.151 qs_ot_get_derivative 99 11.5 0.001 0.001 9.231 9.368 init_scf_loop 11 6.9 0.001 0.002 8.727 8.729 multiply_cannon_metrocomm3 16440 15.4 0.044 0.045 4.334 7.314 prepare_preconditioner 11 7.9 0.000 0.000 6.877 6.897 make_preconditioner 11 8.9 0.000 0.001 6.877 6.897 sum_up_and_integrate 110 10.3 0.060 0.061 6.701 6.717 integrate_v_rspace 110 11.3 0.003 0.003 6.641 6.657 make_full_inverse_cholesky 11 9.9 0.000 0.000 6.228 6.587 init_scf_run 11 5.9 0.000 0.001 6.439 6.443 scf_env_initial_rho_setup 11 6.9 0.001 0.001 6.438 6.443 dbcsr_mm_accdrv_process 34862 16.1 4.838 5.752 6.048 6.244 qs_rho_update_rho_low 110 7.6 0.001 0.001 6.122 6.127 calculate_rho_elec 110 8.6 0.058 0.059 6.121 6.126 qs_ot_get_p 110 10.4 0.001 0.001 5.526 5.693 apply_preconditioner_dbcsr 110 12.6 0.000 0.000 4.987 5.415 apply_single 110 13.6 0.000 0.000 4.987 5.415 make_m2s 4110 13.4 0.050 0.052 4.110 4.504 ot_diis_step 99 11.5 0.010 0.011 4.439 4.439 density_rs2pw 110 9.6 0.004 0.004 3.223 4.420 make_images 4110 14.4 0.388 0.508 3.994 4.388 multiply_cannon_sync_h2d 16440 15.4 3.697 4.246 3.697 4.246 rs_pw_transfer 902 11.9 0.011 0.011 2.943 4.174 qs_ot_p2m_diag 48 11.0 0.042 0.044 3.864 3.868 pw_transfer 1331 11.6 0.066 0.073 3.759 3.769 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 3.082 3.715 fft_wrap_pw1pw2 1111 12.6 0.008 0.009 3.652 3.664 cp_dbcsr_syevd 48 12.0 0.003 0.003 3.522 3.523 grid_integrate_task_list 110 12.3 3.188 3.422 3.188 3.422 calculate_first_density_matrix 1 7.0 0.000 0.002 3.310 3.317 fft_wrap_pw1pw2_140 451 13.1 0.212 0.216 3.136 3.145 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 3.038 3.040 cp_fm_diag_elpa 48 13.0 0.000 0.000 2.979 2.981 wfi_extrapolate 11 7.9 0.001 0.001 2.972 2.972 fft3d_ps 1111 14.6 1.091 1.103 2.951 2.966 cp_fm_redistribute_end 48 14.0 1.858 2.952 1.860 2.953 cp_fm_diag_elpa_base 48 14.0 1.030 2.802 1.088 2.912 make_images_data 4110 15.4 0.043 0.048 2.361 2.835 cp_fm_cholesky_invert 11 10.9 2.752 2.758 2.752 2.758 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 2.692 2.744 hybrid_alltoall_any 4261 16.3 0.105 0.375 2.104 2.686 mp_waitany 17072 13.8 1.386 2.671 1.386 2.671 calculate_dm_sparse 110 9.5 0.001 0.001 2.496 2.529 grid_collocate_task_list 110 9.6 2.100 2.481 2.100 2.481 multiply_cannon_metrocomm4 14385 15.4 0.044 0.048 0.908 2.404 potential_pw2rs 110 12.3 0.011 0.011 2.395 2.403 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 2.333 2.398 rs_pw_transfer_RS2PW_140 121 11.5 0.176 0.182 1.067 2.290 mp_irecv_dv 48980 15.7 0.839 2.280 0.839 2.280 mp_sum_l 6514 12.8 1.585 2.201 1.585 2.201 qs_energies_init_hamiltonians 11 5.9 0.022 0.032 2.177 2.186 mp_alltoall_d11v 2046 13.8 1.769 2.162 1.769 2.162 dbcsr_complete_redistribute 325 12.2 0.326 0.383 1.458 1.917 cp_fm_upper_to_full 70 13.6 1.385 1.862 1.385 1.862 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.685 1.703 jit_kernel_multiply 8 16.3 0.821 1.683 0.821 1.683 cp_fm_cholesky_decompose 22 10.9 1.610 1.629 1.610 1.629 mp_allgather_i34 2055 14.4 0.537 1.565 0.537 1.565 build_core_hamiltonian_matrix_ 11 4.9 0.001 0.001 1.370 1.505 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.001 0.001 1.453 1.464 copy_fm_to_dbcsr 174 11.2 0.001 0.001 0.963 1.403 make_basis_sm 11 9.8 0.000 0.000 1.378 1.381 parallel_gemm_fm 81 9.0 0.000 0.000 1.373 1.375 parallel_gemm_fm_cosma 81 10.0 1.373 1.374 1.373 1.374 rs_gather_matrices 110 12.3 0.233 0.260 0.982 1.336 qs_env_update_s_mstruct 11 6.9 0.000 0.000 1.239 1.327 qs_ot_get_orbitals 99 10.5 0.000 0.001 1.270 1.278 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="204", plot="h2o_128_md", label="(8n/4r/3t)", y=62.595000, yerr=0.000000 PlotPoint: name="205", plot="h2o_128_md_mem", label="(8n/4r/3t)", y=623.727273, yerr=9.126295 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 726.761472E+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_SendRecv 5060 43184. MP_ISendRecv 5060 43184. MP_Wait 20042 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 0.249 0.254 67.099 67.101 qs_mol_dyn_low 1 2.0 0.003 0.003 66.304 66.313 qs_forces 11 3.9 0.061 0.062 65.915 65.916 qs_energies 11 4.9 0.001 0.001 62.433 62.436 scf_env_do_scf 11 5.9 0.000 0.001 53.942 53.946 scf_env_do_scf_inner_loop 99 6.5 0.002 0.007 42.051 42.052 velocity_verlet 10 3.0 0.001 0.002 37.660 37.676 dbcsr_multiply_generic 2055 12.4 0.116 0.119 29.953 30.184 qs_scf_new_mos 99 7.5 0.001 0.001 26.799 26.900 qs_scf_loop_do_ot 99 8.5 0.001 0.001 26.798 26.899 ot_scf_mini 99 9.5 0.003 0.003 25.170 25.266 multiply_cannon 2055 13.4 0.243 0.263 23.096 23.915 multiply_cannon_loop 2055 14.4 0.879 0.899 21.559 22.154 ot_mini 99 10.5 0.001 0.001 14.532 14.649 multiply_cannon_multrec 24660 15.4 4.247 6.889 13.013 14.206 rebuild_ks_matrix 110 8.3 0.000 0.000 12.340 12.442 qs_ks_build_kohn_sham_matrix 110 9.3 0.012 0.013 12.339 12.441 init_scf_loop 11 6.9 0.000 0.000 11.849 11.850 qs_ks_update_qs_env 110 7.6 0.001 0.001 10.900 10.991 qs_ot_get_derivative 99 11.5 0.001 0.001 10.394 10.498 prepare_preconditioner 11 7.9 0.000 0.000 10.065 10.079 make_preconditioner 11 8.9 0.000 0.000 10.065 10.079 make_full_inverse_cholesky 11 9.9 0.000 0.000 8.265 9.749 dbcsr_mm_accdrv_process 52304 16.0 7.159 8.577 8.621 9.640 sum_up_and_integrate 110 10.3 0.066 0.070 6.612 6.623 integrate_v_rspace 110 11.3 0.003 0.003 6.545 6.560 mp_waitall_1 121746 16.5 4.414 6.473 4.414 6.473 qs_rho_update_rho_low 110 7.6 0.001 0.001 6.219 6.227 calculate_rho_elec 110 8.6 0.077 0.081 6.219 6.226 qs_ot_get_p 110 10.4 0.001 0.001 5.657 5.783 init_scf_run 11 5.9 0.000 0.003 5.754 5.755 scf_env_initial_rho_setup 11 6.9 0.001 0.003 5.754 5.754 make_m2s 4110 13.4 0.060 0.062 5.256 5.729 make_images 4110 14.4 0.572 0.695 5.115 5.582 cp_fm_upper_to_full 70 13.8 3.281 4.686 3.281 4.686 ot_diis_step 99 11.5 0.011 0.011 4.101 4.101 apply_preconditioner_dbcsr 110 12.6 0.000 0.000 4.020 4.093 apply_single 110 13.6 0.000 0.000 4.020 4.093 pw_transfer 1331 11.6 0.065 0.075 3.929 3.955 qs_ot_p2m_diag 48 11.0 0.054 0.063 3.849 3.863 dbcsr_complete_redistribute 325 12.2 0.418 0.461 2.705 3.855 fft_wrap_pw1pw2 1111 12.6 0.008 0.008 3.823 3.852 density_rs2pw 110 9.6 0.004 0.004 3.161 3.676 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 3.611 3.658 grid_integrate_task_list 110 12.3 3.270 3.453 3.270 3.453 multiply_cannon_metrocomm3 24660 15.4 0.036 0.037 1.613 3.378 cp_dbcsr_syevd 48 12.0 0.003 0.003 3.371 3.371 multiply_cannon_sync_h2d 24660 15.4 3.192 3.311 3.192 3.311 fft_wrap_pw1pw2_140 451 13.1 0.203 0.214 3.257 3.291 copy_fm_to_dbcsr 174 11.2 0.001 0.001 2.153 3.288 fft3d_ps 1111 14.6 1.092 1.131 3.114 3.137 calculate_dm_sparse 110 9.5 0.001 0.001 3.071 3.098 make_images_data 4110 15.4 0.047 0.050 2.593 3.073 hybrid_alltoall_any 4261 16.3 0.119 0.459 2.272 3.070 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 3.002 3.059 wfi_extrapolate 11 7.9 0.001 0.001 3.006 3.006 rs_pw_transfer 902 11.9 0.010 0.011 2.471 2.999 transfer_fm_to_dbcsr 11 9.9 0.000 0.000 1.791 2.912 mp_alltoall_i22 605 13.7 1.672 2.832 1.672 2.832 cp_fm_diag_elpa 48 13.0 0.000 0.000 2.806 2.808 cp_fm_cholesky_invert 11 10.9 2.787 2.797 2.787 2.797 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 2.793 2.795 cp_fm_redistribute_end 48 14.0 1.403 2.775 1.404 2.776 cp_fm_diag_elpa_base 48 14.0 1.288 2.634 1.368 2.748 calculate_first_density_matrix 1 7.0 0.000 0.000 2.651 2.653 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 2.435 2.468 grid_collocate_task_list 110 9.6 2.182 2.440 2.182 2.440 potential_pw2rs 110 12.3 0.012 0.013 2.287 2.295 qs_energies_init_hamiltonians 11 5.9 0.000 0.001 2.244 2.244 mp_alltoall_d11v 2046 13.8 1.712 2.057 1.712 2.057 jit_kernel_multiply 10 15.7 1.132 2.004 1.132 2.004 cp_fm_cholesky_decompose 22 10.9 1.715 1.769 1.715 1.769 qs_ot_get_orbitals 99 10.5 0.001 0.001 1.701 1.733 build_core_hamiltonian_matrix_ 11 4.9 0.001 0.001 1.600 1.699 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.001 0.001 1.680 1.696 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.684 1.696 mp_waitany 13376 13.8 1.054 1.573 1.054 1.573 multiply_cannon_metrocomm4 20550 15.4 0.057 0.059 0.834 1.524 acc_transpose_blocks 24660 15.4 0.106 0.108 1.492 1.510 mp_sum_l 6514 12.8 0.915 1.480 0.915 1.480 mp_irecv_dv 62702 16.1 0.736 1.447 0.736 1.447 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="206", plot="h2o_128_md", label="(8n/3r/4t)", y=67.101000, yerr=0.000000 PlotPoint: name="207", plot="h2o_128_md_mem", label="(8n/3r/4t)", y=686.727273, yerr=6.876718 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 842.035200E+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_SendRecv 3300 54848. MP_ISendRecv 3300 54848. MP_Wait 13926 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 0.076 0.098 58.946 58.953 qs_mol_dyn_low 1 2.0 0.003 0.003 57.955 57.964 qs_forces 11 3.9 0.057 0.066 57.847 57.848 qs_energies 11 4.9 0.001 0.001 54.076 54.135 scf_env_do_scf 11 5.9 0.001 0.001 44.276 44.277 scf_env_do_scf_inner_loop 99 6.5 0.002 0.007 36.228 36.229 velocity_verlet 10 3.0 0.002 0.002 31.799 31.805 dbcsr_multiply_generic 2055 12.4 0.116 0.127 23.651 23.835 qs_scf_new_mos 99 7.5 0.001 0.001 20.930 20.979 qs_scf_loop_do_ot 99 8.5 0.001 0.001 20.930 20.979 ot_scf_mini 99 9.5 0.002 0.002 19.693 19.715 multiply_cannon 2055 13.4 0.246 0.269 18.001 19.528 multiply_cannon_loop 2055 14.4 0.321 0.332 16.598 17.033 rebuild_ks_matrix 110 8.3 0.000 0.000 12.133 12.177 qs_ks_build_kohn_sham_matrix 110 9.3 0.012 0.013 12.133 12.176 qs_ks_update_qs_env 110 7.6 0.001 0.001 10.790 10.830 ot_mini 99 10.5 0.001 0.001 10.581 10.598 multiply_cannon_multrec 8220 15.4 3.213 4.479 8.072 8.958 init_scf_loop 11 6.9 0.000 0.000 7.997 7.999 mp_waitall_1 103326 16.6 6.078 7.663 6.078 7.663 qs_ot_get_derivative 99 11.5 0.001 0.001 6.777 6.798 sum_up_and_integrate 110 10.3 0.079 0.081 6.628 6.639 init_scf_run 11 5.9 0.001 0.013 6.601 6.601 scf_env_initial_rho_setup 11 6.9 0.009 0.012 6.600 6.601 integrate_v_rspace 110 11.3 0.003 0.003 6.549 6.561 qs_rho_update_rho_low 110 7.6 0.001 0.001 6.364 6.378 calculate_rho_elec 110 8.6 0.115 0.116 6.364 6.377 prepare_preconditioner 11 7.9 0.000 0.000 6.276 6.281 make_preconditioner 11 8.9 0.000 0.000 6.276 6.281 make_full_inverse_cholesky 11 9.9 0.000 0.000 5.851 5.927 dbcsr_mm_accdrv_process 17442 15.9 3.066 4.162 4.729 5.637 qs_ot_get_p 110 10.4 0.001 0.001 5.189 5.209 make_m2s 4110 13.4 0.039 0.040 4.286 4.569 make_images 4110 14.4 0.636 0.695 4.155 4.438 multiply_cannon_metrocomm3 8220 15.4 0.018 0.018 2.983 4.367 pw_transfer 1331 11.6 0.066 0.071 4.106 4.113 fft_wrap_pw1pw2 1111 12.6 0.008 0.008 3.998 4.009 ot_diis_step 99 11.5 0.012 0.012 3.781 3.781 qs_ot_p2m_diag 48 11.0 0.081 0.084 3.731 3.735 apply_preconditioner_dbcsr 110 12.6 0.000 0.000 3.683 3.733 apply_single 110 13.6 0.000 0.000 3.683 3.732 calculate_first_density_matrix 1 7.0 0.000 0.000 3.711 3.721 grid_integrate_task_list 110 12.3 3.349 3.527 3.349 3.527 density_rs2pw 110 9.6 0.004 0.004 3.138 3.509 fft_wrap_pw1pw2_140 451 13.1 0.219 0.224 3.441 3.456 cp_dbcsr_syevd 48 12.0 0.003 0.003 3.421 3.421 fft3d_ps 1111 14.6 1.151 1.177 3.236 3.248 cp_fm_cholesky_invert 11 10.9 3.077 3.081 3.077 3.081 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 3.074 3.075 multiply_cannon_sync_h2d 8220 15.4 2.906 3.025 2.906 3.025 cp_fm_diag_elpa 48 13.0 0.000 0.000 2.874 2.875 make_images_data 4110 15.4 0.038 0.044 2.404 2.849 cp_fm_redistribute_end 48 14.0 0.733 2.841 0.737 2.842 cp_fm_diag_elpa_base 48 14.0 1.919 2.642 2.097 2.805 wfi_extrapolate 11 7.9 0.001 0.001 2.785 2.785 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 2.697 2.753 hybrid_alltoall_any 4261 16.3 0.200 0.866 2.311 2.743 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 2.678 2.697 rs_pw_transfer 902 11.9 0.010 0.011 2.292 2.683 calculate_dm_sparse 110 9.5 0.001 0.001 2.603 2.634 grid_collocate_task_list 110 9.6 2.277 2.544 2.277 2.544 potential_pw2rs 110 12.3 0.015 0.016 2.191 2.195 jit_kernel_multiply 9 15.6 1.356 2.062 1.356 2.062 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 2.029 2.048 build_core_hamiltonian_matrix_ 11 4.9 0.001 0.001 1.775 2.001 mp_alltoall_d11v 2046 13.8 1.604 1.889 1.604 1.889 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 1.815 1.828 cp_fm_cholesky_decompose 22 10.9 1.759 1.783 1.759 1.783 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.677 1.681 mp_allgather_i34 2055 14.4 0.555 1.649 0.555 1.649 qs_env_update_s_mstruct 11 6.9 0.000 0.000 1.532 1.643 dbcsr_complete_redistribute 325 12.2 0.569 0.596 1.510 1.610 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.001 0.001 1.568 1.578 mp_waitany 9240 13.8 1.106 1.511 1.106 1.511 make_basis_sm 11 9.8 0.000 0.000 1.446 1.448 parallel_gemm_fm 81 9.0 0.000 0.000 1.409 1.410 parallel_gemm_fm_cosma 81 10.0 1.409 1.410 1.409 1.410 qs_create_task_list 11 7.9 0.000 0.001 1.231 1.335 generate_qs_task_list 11 8.9 0.380 0.447 1.231 1.335 multiply_cannon_metrocomm1 8220 15.4 0.021 0.022 0.824 1.265 rs_pw_transfer_RS2PW_140 121 11.5 0.162 0.166 0.808 1.200 rs_gather_matrices 110 12.3 0.323 0.363 0.923 1.197 copy_dbcsr_to_fm 151 11.3 0.003 0.003 1.155 1.183 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="208", plot="h2o_128_md", label="(8n/2r/6t)", y=58.953000, yerr=0.000000 PlotPoint: name="209", plot="h2o_128_md_mem", label="(8n/2r/6t)", y=788.545455, yerr=12.521717 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 1.417007E+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_SendRecv 2310 360267. MP_ISendRecv 2310 360267. MP_Wait 5214 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 0.026 0.059 89.690 89.700 qs_mol_dyn_low 1 2.0 0.003 0.006 89.340 89.348 qs_forces 11 3.9 0.002 0.002 88.837 88.839 qs_energies 11 4.9 0.001 0.001 84.667 84.670 scf_env_do_scf 11 5.9 0.001 0.001 73.217 73.217 velocity_verlet 10 3.0 0.002 0.002 56.397 56.457 scf_env_do_scf_inner_loop 99 6.5 0.002 0.007 44.628 44.630 dbcsr_multiply_generic 2055 12.4 0.120 0.121 28.903 28.947 init_scf_loop 11 6.9 0.000 0.000 28.515 28.518 prepare_preconditioner 11 7.9 0.000 0.000 26.475 26.479 make_preconditioner 11 8.9 0.000 0.000 26.475 26.478 qs_scf_new_mos 99 7.5 0.001 0.001 26.222 26.239 qs_scf_loop_do_ot 99 8.5 0.001 0.001 26.222 26.239 make_full_inverse_cholesky 11 9.9 0.000 0.000 20.645 25.956 ot_scf_mini 99 9.5 0.002 0.002 24.465 24.481 multiply_cannon 2055 13.4 0.353 0.387 21.869 22.943 multiply_cannon_loop 2055 14.4 0.341 0.344 20.088 20.658 cp_fm_upper_to_full 70 14.2 12.682 18.239 12.682 18.239 rebuild_ks_matrix 110 8.3 0.000 0.001 14.184 14.216 qs_ks_build_kohn_sham_matrix 110 9.3 0.014 0.014 14.184 14.216 ot_mini 99 10.5 0.001 0.001 13.588 13.606 qs_ks_update_qs_env 110 7.6 0.001 0.001 12.851 12.880 dbcsr_complete_redistribute 325 12.2 1.051 1.092 7.499 10.782 multiply_cannon_multrec 8220 15.4 4.372 4.571 10.027 10.252 copy_fm_to_dbcsr 174 11.2 0.001 0.001 6.445 9.719 qs_ot_get_derivative 99 11.5 0.001 0.001 9.114 9.132 transfer_fm_to_dbcsr 11 9.9 0.000 0.000 5.814 9.051 mp_alltoall_i22 605 13.7 5.455 8.720 5.455 8.720 mp_waitall_1 84994 16.7 7.384 8.304 7.384 8.304 qs_rho_update_rho_low 110 7.6 0.001 0.001 7.811 7.843 calculate_rho_elec 110 8.6 0.227 0.227 7.811 7.843 sum_up_and_integrate 110 10.3 0.150 0.151 7.392 7.404 integrate_v_rspace 110 11.3 0.004 0.004 7.242 7.254 init_scf_run 11 5.9 0.000 0.001 6.980 6.980 scf_env_initial_rho_setup 11 6.9 0.005 0.010 6.979 6.980 dbcsr_mm_accdrv_process 11614 15.7 3.350 3.663 5.512 5.845 qs_ot_get_p 110 10.4 0.001 0.001 5.736 5.765 make_m2s 4110 13.4 0.043 0.044 5.171 5.652 make_images 4110 14.4 0.880 0.938 4.981 5.461 cp_fm_cholesky_invert 11 10.9 5.378 5.382 5.378 5.382 pw_transfer 1331 11.6 0.075 0.075 5.318 5.326 multiply_cannon_metrocomm3 8220 15.4 0.018 0.018 4.702 5.314 fft_wrap_pw1pw2 1111 12.6 0.009 0.009 5.201 5.208 apply_preconditioner_dbcsr 110 12.6 0.000 0.000 4.679 5.112 apply_single 110 13.6 0.000 0.000 4.679 5.112 fft_wrap_pw1pw2_140 451 13.1 0.223 0.228 4.513 4.531 ot_diis_step 99 11.5 0.015 0.015 4.452 4.452 fft3d_ps 1111 14.6 1.298 1.314 4.392 4.400 qs_ot_p2m_diag 48 11.0 0.151 0.156 4.197 4.204 multiply_cannon_sync_h2d 8220 15.4 3.948 3.954 3.948 3.954 density_rs2pw 110 9.6 0.004 0.004 3.850 3.875 qs_energies_init_hamiltonians 11 5.9 0.006 0.011 3.781 3.783 cp_dbcsr_syevd 48 12.0 0.003 0.003 3.752 3.753 grid_integrate_task_list 110 12.3 3.670 3.723 3.670 3.723 calculate_first_density_matrix 1 7.0 0.000 0.000 3.543 3.548 hybrid_alltoall_any 4261 16.3 0.260 0.571 2.825 3.535 make_images_data 4110 15.4 0.042 0.045 2.797 3.414 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 2.850 3.385 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 3.366 3.369 wfi_extrapolate 11 7.9 0.001 0.001 3.306 3.306 calculate_dm_sparse 110 9.5 0.001 0.001 3.222 3.246 cp_fm_diag_elpa 48 13.0 0.000 0.000 3.128 3.128 cp_fm_diag_elpa_base 48 14.0 2.593 2.796 3.126 3.126 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 2.798 2.814 grid_collocate_task_list 110 9.6 2.636 2.661 2.636 2.661 potential_pw2rs 110 12.3 0.021 0.021 2.508 2.511 rs_pw_transfer 902 11.9 0.011 0.011 2.335 2.417 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 2.379 2.388 qs_env_update_s_mstruct 11 6.9 0.000 0.000 2.301 2.351 jit_kernel_multiply 10 15.2 1.962 2.305 1.962 2.305 build_core_hamiltonian_matrix_ 11 4.9 0.001 0.001 2.092 2.177 cp_fm_cholesky_decompose 22 10.9 2.046 2.063 2.046 2.063 mp_alltoall_d11v 2046 13.8 1.937 2.040 1.937 2.040 qs_create_task_list 11 7.9 0.001 0.001 1.902 1.949 generate_qs_task_list 11 8.9 0.736 0.789 1.901 1.949 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.940 1.943 copy_dbcsr_to_fm 151 11.3 0.003 0.003 1.802 1.823 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="210", plot="h2o_128_md", label="(8n/1r/12t)", y=89.700000, yerr=0.000000 PlotPoint: name="211", plot="h2o_128_md_mem", label="(8n/1r/12t)", y=1266.000000, yerr=76.537572 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/15/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1420239992832 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 11528891191296 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 93.514751E+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 6755938624 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 626.864128E+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 3175954870160 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 2523700. MP_SendRecv 24320 18752. MP_ISendRecv 24320 18752. MP_Wait 42476 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 0.067 0.451 206.344 206.345 qs_mol_dyn_low 1 2.0 0.003 0.006 205.461 205.474 qs_forces 11 3.9 0.003 0.004 205.266 205.267 qs_energies 11 4.9 0.001 0.001 199.378 199.394 scf_env_do_scf 11 5.9 0.001 0.001 181.836 181.840 scf_env_do_scf_inner_loop 117 6.6 0.003 0.009 161.108 161.109 dbcsr_multiply_generic 2507 12.6 0.179 0.183 123.918 124.987 velocity_verlet 10 3.0 0.001 0.002 123.232 123.237 qs_scf_new_mos 117 7.6 0.001 0.001 120.454 120.648 qs_scf_loop_do_ot 117 8.6 0.001 0.001 120.453 120.647 ot_scf_mini 117 9.6 0.003 0.003 113.875 114.042 multiply_cannon 2507 13.6 0.236 0.243 100.710 102.452 multiply_cannon_loop 2507 14.6 2.102 2.172 98.561 100.160 ot_mini 117 10.6 0.001 0.001 65.305 65.494 multiply_cannon_multrec 60168 15.6 33.144 35.088 41.730 43.987 qs_ot_get_derivative 117 11.6 0.001 0.001 40.421 40.584 rebuild_ks_matrix 128 8.3 0.001 0.001 34.081 34.289 qs_ks_build_kohn_sham_matrix 128 9.3 0.015 0.017 34.080 34.289 mp_waitall_1 267128 16.5 28.265 31.215 28.265 31.215 qs_ks_update_qs_env 128 7.6 0.001 0.001 30.625 30.804 multiply_cannon_sync_h2d 60168 15.6 27.282 29.075 27.282 29.075 qs_ot_get_p 128 10.4 0.001 0.001 26.710 26.907 apply_preconditioner_dbcsr 128 12.6 0.000 0.001 24.315 25.611 apply_single 128 13.6 0.001 0.001 24.315 25.610 ot_diis_step 117 11.6 0.007 0.008 24.529 24.530 init_scf_loop 11 6.9 0.000 0.000 20.651 20.652 qs_ot_p2m_diag 83 11.4 0.077 0.091 20.057 20.132 qs_ot_get_derivative_diag 77 12.4 0.002 0.002 18.772 18.951 multiply_cannon_metrocomm3 60168 15.6 0.113 0.117 15.488 18.163 cp_dbcsr_syevd 83 12.4 0.004 0.005 17.385 17.387 prepare_preconditioner 11 7.9 0.000 0.000 16.007 16.049 make_preconditioner 11 8.9 0.000 0.000 16.007 16.049 make_full_inverse_cholesky 11 9.9 0.000 0.000 15.183 15.365 sum_up_and_integrate 128 10.3 0.090 0.108 14.526 14.541 cp_fm_diag_elpa 83 13.4 0.000 0.000 14.513 14.519 make_m2s 5014 13.6 0.105 0.113 13.984 14.501 integrate_v_rspace 128 11.3 0.003 0.004 14.436 14.454 cp_fm_redistribute_end 83 14.4 11.477 14.444 11.488 14.446 make_images 5014 14.6 0.400 0.418 13.801 14.329 cp_fm_diag_elpa_base 83 14.4 2.915 14.062 2.946 14.180 qs_rho_update_rho_low 128 7.7 0.001 0.001 14.062 14.155 calculate_rho_elec 128 8.7 0.045 0.064 14.061 14.155 init_scf_run 11 5.9 0.000 0.002 13.137 13.138 scf_env_initial_rho_setup 11 6.9 0.001 0.001 13.137 13.138 density_rs2pw 128 9.7 0.006 0.007 7.395 10.569 mp_sum_l 7870 13.0 8.086 9.202 8.086 9.202 wfi_extrapolate 11 7.9 0.001 0.001 9.193 9.193 rs_pw_transfer 1046 11.9 0.017 0.019 5.953 9.148 cp_fm_cholesky_invert 11 10.9 9.029 9.037 9.029 9.037 dbcsr_mm_accdrv_process 124484 16.2 3.321 3.444 8.148 8.649 calculate_dm_sparse 128 9.5 0.001 0.001 8.396 8.531 pw_transfer 1547 11.6 0.076 0.099 8.207 8.461 fft_wrap_pw1pw2 1291 12.7 0.011 0.013 8.001 8.264 qs_ot_get_derivative_taylor 40 13.0 0.001 0.001 7.853 7.957 qs_ot_get_orbitals 117 10.6 0.001 0.001 7.866 7.934 make_images_data 5014 15.6 0.064 0.071 6.899 7.878 multiply_cannon_metrocomm1 60168 15.6 0.089 0.094 6.042 7.562 grid_integrate_task_list 128 12.3 7.068 7.508 7.068 7.508 hybrid_alltoall_any 5200 16.5 0.288 2.254 6.027 7.285 fft_wrap_pw1pw2_140 523 13.2 0.444 0.511 6.922 7.225 fft3d_ps 1291 14.7 2.103 2.871 6.789 6.971 cp_dbcsr_sm_fm_multiply 37 9.5 0.005 0.029 6.901 6.911 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 6.004 6.073 mp_waitany 16020 13.9 2.674 5.843 2.674 5.843 grid_collocate_task_list 128 9.7 4.563 5.813 4.563 5.813 mp_alltoall_d11v 2415 14.1 4.372 5.615 4.372 5.615 rs_pw_transfer_RS2PW_140 139 11.5 0.280 0.295 2.142 5.364 potential_pw2rs 128 12.3 0.009 0.011 4.850 4.882 mp_sum_d 4459 12.1 3.947 4.760 3.947 4.760 cp_fm_cholesky_decompose 22 10.9 4.715 4.730 4.715 4.730 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="400", plot="h2o_256_md", label="(8n/12r/1t)", y=206.345000, yerr=0.000000 PlotPoint: name="401", plot="h2o_256_md_mem", label="(8n/12r/1t)", y=593.727273, yerr=7.085079 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/16/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1420239992832 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 11528891191296 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 93.514751E+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 6755938624 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 831.188992E+06 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 2406720 MPI messages size (bytes): total size 4.100942E+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 1463834332048 16777216 < size 67584 1181116006400 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 4075 57341. MP_Allreduce 11227 947. MP_Sync 170 MP_Alltoall 1969 5619023. MP_SendRecv 12032 47072. MP_ISendRecv 12032 47072. MP_Wait 25916 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 0.142 0.222 194.329 194.338 qs_mol_dyn_low 1 2.0 0.003 0.004 193.730 193.743 qs_forces 11 3.9 0.004 0.006 193.649 193.649 qs_energies 11 4.9 0.005 0.022 186.791 186.801 scf_env_do_scf 11 5.9 0.001 0.002 169.451 169.461 scf_env_do_scf_inner_loop 117 6.6 0.003 0.009 136.161 136.164 velocity_verlet 10 3.0 0.002 0.003 121.141 121.142 dbcsr_multiply_generic 2507 12.6 0.190 0.207 97.845 99.044 qs_scf_new_mos 117 7.6 0.001 0.001 95.586 96.076 qs_scf_loop_do_ot 117 8.6 0.001 0.002 95.585 96.075 ot_scf_mini 117 9.6 0.004 0.005 90.796 91.347 multiply_cannon 2507 13.6 0.479 0.535 77.911 81.972 multiply_cannon_loop 2507 14.6 1.261 1.307 74.608 77.157 ot_mini 117 10.6 0.001 0.001 50.251 50.750 mp_waitall_1 214728 16.6 24.694 38.957 24.694 38.957 multiply_cannon_multrec 30084 15.6 22.045 26.283 31.898 36.349 rebuild_ks_matrix 128 8.3 0.001 0.001 33.630 34.110 qs_ks_build_kohn_sham_matrix 128 9.3 0.017 0.019 33.630 34.109 init_scf_loop 11 6.9 0.001 0.003 33.200 33.201 qs_ks_update_qs_env 128 7.6 0.001 0.001 30.260 30.686 qs_ot_get_derivative 117 11.6 0.001 0.002 28.383 28.921 prepare_preconditioner 11 7.9 0.000 0.002 28.766 28.823 make_preconditioner 11 8.9 0.000 0.001 28.765 28.823 multiply_cannon_metrocomm3 30084 15.6 0.095 0.100 15.917 28.573 make_full_inverse_cholesky 11 9.9 0.000 0.000 27.462 28.010 apply_preconditioner_dbcsr 128 12.6 0.000 0.000 21.946 23.077 apply_single 128 13.6 0.001 0.001 21.945 23.076 qs_ot_get_p 128 10.4 0.001 0.001 22.274 22.889 ot_diis_step 117 11.6 0.014 0.015 21.692 21.694 multiply_cannon_sync_h2d 30084 15.6 19.310 21.296 19.310 21.296 qs_ot_p2m_diag 83 11.4 0.187 0.216 17.458 17.500 cp_fm_cholesky_invert 11 10.9 16.772 16.785 16.772 16.785 cp_dbcsr_syevd 83 12.4 0.005 0.006 16.298 16.300 make_m2s 5014 13.6 0.090 0.096 14.189 15.763 make_images 5014 14.6 1.145 1.332 13.977 15.549 sum_up_and_integrate 128 10.3 0.114 0.133 14.965 14.996 integrate_v_rspace 128 11.3 0.003 0.004 14.851 14.888 qs_rho_update_rho_low 128 7.7 0.001 0.001 14.345 14.380 calculate_rho_elec 128 8.7 0.088 0.106 14.344 14.379 cp_fm_diag_elpa 83 13.4 0.000 0.001 13.132 13.140 cp_fm_redistribute_end 83 14.4 7.666 13.059 7.681 13.064 cp_fm_diag_elpa_base 83 14.4 5.143 12.546 5.366 12.919 init_scf_run 11 5.9 0.000 0.009 12.138 12.140 scf_env_initial_rho_setup 11 6.9 0.001 0.001 12.137 12.140 qs_ot_get_derivative_diag 77 12.4 0.002 0.002 10.964 11.331 multiply_cannon_metrocomm4 27577 15.6 0.094 0.107 3.718 10.796 make_images_data 5014 15.6 0.063 0.072 8.556 10.439 mp_irecv_dv 69486 16.3 3.527 10.412 3.527 10.412 density_rs2pw 128 9.7 0.006 0.006 7.708 10.101 dbcsr_mm_accdrv_process 62242 16.2 4.705 5.388 9.313 9.924 hybrid_alltoall_any 5200 16.5 0.341 1.515 7.249 9.892 pw_transfer 1547 11.6 0.087 0.107 9.130 9.195 fft_wrap_pw1pw2 1291 12.7 0.011 0.012 8.904 8.973 rs_pw_transfer 1046 11.9 0.014 0.016 6.124 8.602 wfi_extrapolate 11 7.9 0.001 0.001 8.451 8.451 fft_wrap_pw1pw2_140 523 13.2 0.472 0.527 7.810 7.901 grid_integrate_task_list 128 12.3 7.175 7.586 7.175 7.586 qs_ot_get_derivative_taylor 40 13.0 0.001 0.001 6.490 7.317 fft3d_ps 1291 14.7 2.755 2.926 7.265 7.316 cp_fm_cholesky_decompose 22 10.9 7.024 7.106 7.024 7.106 calculate_dm_sparse 128 9.5 0.001 0.001 6.439 6.592 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 6.238 6.250 mp_sum_l 7870 13.0 4.014 5.921 4.014 5.921 grid_collocate_task_list 128 9.7 4.704 5.840 4.704 5.840 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 5.354 5.495 qs_ot_get_orbitals 117 10.6 0.001 0.001 5.336 5.400 potential_pw2rs 128 12.3 0.015 0.018 5.278 5.299 mp_waitany 11748 13.9 2.656 5.083 2.656 5.083 mp_alltoall_d11v 2415 14.1 4.071 5.004 4.071 5.004 mp_allgather_i34 2507 14.6 1.869 4.929 1.869 4.929 rs_pw_transfer_RS2PW_140 139 11.5 0.354 0.380 2.242 4.707 dbcsr_complete_redistribute 395 12.7 0.773 0.867 3.233 4.120 mp_sum_d 4459 12.1 2.735 4.003 2.735 4.003 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="402", plot="h2o_256_md", label="(8n/6r/2t)", y=194.338000, yerr=0.000000 PlotPoint: name="403", plot="h2o_256_md_mem", label="(8n/6r/2t)", y=791.545455, yerr=3.201239 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/17/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1410022121472 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 11444702699520 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 92.796573E+12 0.0% 0.0% 100.0% flops max/rank 2.906045E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 6705499744 0.0% 0.0% 100.0% number of processed stacks 3951168 0.0% 0.0% 100.0% average stack size 0.0 0.0 1697.1 marketing flops 143.507742E+12 ------------------------------------------------------------------------------- # multiplications 2485 max memory usage/rank 940.163072E+06 # max total images/rank 2 # max 3D layers 1 # MPI messages exchanged 1033760 MPI messages size (bytes): total size 2.695213E+12 min size 0.000000E+00 max size 26.214400E+06 average size 2.607194E+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 279168 36591108096 131072 < size <= 4194304 654272 987691483136 4194304 < size <= 16777216 65184 925172769472 16777216 < size 28448 745747251200 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 4089 57139. MP_Allreduce 11261 985. MP_Sync 168 MP_Alltoall 1700 9383497. MP_SendRecv 7874 75008. MP_ISendRecv 7874 75008. MP_Wait 21654 MP_comm_split 82 MP_ISend 11660 275234. MP_IRecv 11660 275234. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.041 0.085 178.269 178.272 qs_mol_dyn_low 1 2.0 0.004 0.010 177.758 177.771 qs_forces 11 3.9 0.004 0.009 177.645 177.651 qs_energies 11 4.9 0.002 0.010 170.970 170.980 scf_env_do_scf 11 5.9 0.001 0.002 154.738 154.739 scf_env_do_scf_inner_loop 116 6.6 0.003 0.008 119.346 119.347 velocity_verlet 10 3.0 0.002 0.002 113.190 113.193 dbcsr_multiply_generic 2485 12.5 0.181 0.185 81.721 82.730 qs_scf_new_mos 116 7.6 0.001 0.001 80.944 81.232 qs_scf_loop_do_ot 116 8.6 0.001 0.002 80.943 81.231 ot_scf_mini 116 9.6 0.004 0.005 76.831 77.163 multiply_cannon 2485 13.5 0.494 0.509 62.027 66.877 multiply_cannon_loop 2485 14.5 0.847 0.877 58.991 61.524 ot_mini 116 10.6 0.001 0.002 42.337 42.674 init_scf_loop 11 6.9 0.001 0.003 35.289 35.290 mp_waitall_1 169034 16.6 25.286 35.221 25.286 35.221 rebuild_ks_matrix 127 8.3 0.001 0.001 30.884 31.350 qs_ks_build_kohn_sham_matrix 127 9.3 0.016 0.021 30.884 31.350 prepare_preconditioner 11 7.9 0.000 0.000 31.194 31.255 make_preconditioner 11 8.9 0.000 0.001 31.194 31.255 make_full_inverse_cholesky 11 9.9 0.000 0.000 28.793 30.195 qs_ks_update_qs_env 127 7.6 0.001 0.001 27.804 28.233 multiply_cannon_multrec 19880 15.5 13.304 16.480 22.449 25.665 multiply_cannon_metrocomm3 19880 15.5 0.059 0.062 15.431 25.192 qs_ot_get_derivative 116 11.6 0.002 0.002 22.811 23.134 apply_preconditioner_dbcsr 127 12.6 0.000 0.001 19.683 20.681 apply_single 127 13.6 0.001 0.001 19.683 20.681 qs_ot_get_p 127 10.4 0.001 0.001 19.912 20.351 ot_diis_step 116 11.6 0.018 0.021 19.422 19.422 multiply_cannon_sync_h2d 19880 15.5 14.136 15.933 14.136 15.933 qs_ot_p2m_diag 82 11.4 0.262 0.269 15.698 15.710 make_m2s 4970 13.5 0.081 0.086 14.740 15.516 make_images 4970 14.5 1.153 1.245 14.507 15.274 sum_up_and_integrate 127 10.3 0.131 0.144 14.757 14.785 cp_dbcsr_syevd 82 12.4 0.005 0.005 14.693 14.694 integrate_v_rspace 127 11.3 0.004 0.004 14.626 14.651 cp_fm_cholesky_invert 11 10.9 14.541 14.552 14.541 14.552 qs_rho_update_rho_low 127 7.7 0.001 0.001 14.453 14.486 calculate_rho_elec 127 8.7 0.131 0.146 14.453 14.485 cp_fm_diag_elpa 82 13.4 0.000 0.001 11.507 11.509 cp_fm_redistribute_end 82 14.4 4.349 11.440 4.364 11.442 cp_fm_diag_elpa_base 82 14.4 6.653 10.845 7.056 11.316 init_scf_run 11 5.9 0.000 0.002 11.069 11.070 scf_env_initial_rho_setup 11 6.9 0.001 0.002 11.069 11.070 make_images_data 4970 15.5 0.059 0.068 9.132 10.289 hybrid_alltoall_any 5155 16.4 0.430 1.968 7.982 9.552 pw_transfer 1535 11.6 0.086 0.110 9.170 9.282 density_rs2pw 127 9.7 0.006 0.006 7.388 9.278 multiply_cannon_metrocomm4 17395 15.5 0.060 0.070 3.465 9.140 fft_wrap_pw1pw2 1281 12.7 0.010 0.012 8.947 9.064 qs_ot_get_derivative_diag 76 12.4 0.002 0.002 8.820 9.057 mp_irecv_dv 49801 16.2 3.346 8.902 3.346 8.902 dbcsr_mm_accdrv_process 41158 16.2 4.536 5.696 8.603 8.761 fft_wrap_pw1pw2_140 519 13.2 0.473 0.523 7.871 7.998 grid_integrate_task_list 127 12.3 7.234 7.758 7.234 7.758 cp_fm_cholesky_decompose 22 10.9 7.506 7.529 7.506 7.529 wfi_extrapolate 11 7.9 0.002 0.013 7.461 7.461 cp_fm_upper_to_full 104 14.5 5.702 7.404 5.702 7.404 fft3d_ps 1281 14.7 2.646 2.878 7.222 7.305 rs_pw_transfer 1038 11.9 0.014 0.014 5.371 7.299 dbcsr_complete_redistribute 393 12.7 1.166 1.194 4.637 6.384 calculate_dm_sparse 127 9.5 0.001 0.001 5.809 5.900 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 5.866 5.871 grid_collocate_task_list 127 9.7 4.857 5.691 4.857 5.691 qs_ot_get_derivative_taylor 40 13.0 0.001 0.001 4.775 5.466 mp_alltoall_d11v 2401 14.1 4.378 5.203 4.378 5.203 copy_fm_to_dbcsr 208 11.6 0.002 0.002 3.437 5.180 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 5.001 5.138 potential_pw2rs 127 12.3 0.020 0.022 4.903 4.921 mp_allgather_i34 2485 14.5 1.575 4.575 1.575 4.575 mp_waitany 11660 13.9 2.423 4.427 2.423 4.427 mp_sum_l 7804 13.0 3.159 4.403 3.159 4.403 transfer_fm_to_dbcsr 11 9.9 0.000 0.000 2.381 4.095 qs_ot_get_orbitals 116 10.6 0.001 0.001 4.005 4.032 rs_pw_transfer_RS2PW_140 138 11.5 0.329 0.353 2.018 3.953 qs_energies_init_hamiltonians 11 5.9 0.001 0.004 3.899 3.900 mp_alltoall_i22 712 14.1 1.935 3.808 1.935 3.808 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 3.592 3.632 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="404", plot="h2o_256_md", label="(8n/4r/3t)", y=178.272000, yerr=0.000000 PlotPoint: name="405", plot="h2o_256_md_mem", label="(8n/4r/3t)", y=893.909091, yerr=9.774313 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/18/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1410022121472 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 11444702699520 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 92.796573E+12 0.0% 0.0% 100.0% flops max/rank 4.320339E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 6705499744 0.0% 0.0% 100.0% number of processed stacks 5927808 0.0% 0.0% 100.0% average stack size 0.0 0.0 1131.2 marketing flops 143.508480E+12 ------------------------------------------------------------------------------- # multiplications 2485 max memory usage/rank 1.131250E+09 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 1133160 MPI messages size (bytes): total size 2.008142E+12 min size 0.000000E+00 max size 17.653760E+06 average size 1.772161E+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 315952 35695099904 131072 < size <= 4194304 709496 778939400192 4194304 < size <= 16777216 69840 660837692480 16777216 < size 30480 532676608000 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 4089 57136. MP_Allreduce 11261 1067. MP_Sync 168 MP_Alltoall 1700 12496371. MP_SendRecv 5842 75008. MP_ISendRecv 5842 75008. MP_Wait 22272 MP_comm_split 82 MP_ISend 14840 244848. MP_IRecv 14840 244848. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.049 0.065 190.434 190.436 qs_mol_dyn_low 1 2.0 0.019 0.026 189.938 189.951 qs_forces 11 3.9 0.004 0.006 189.199 189.221 qs_energies 11 4.9 0.010 0.029 181.941 181.963 scf_env_do_scf 11 5.9 0.013 0.017 164.399 164.406 velocity_verlet 10 3.0 0.003 0.015 125.756 125.785 scf_env_do_scf_inner_loop 116 6.6 0.004 0.009 117.404 117.406 qs_scf_new_mos 116 7.6 0.001 0.001 79.755 80.033 qs_scf_loop_do_ot 116 8.6 0.001 0.002 79.754 80.032 dbcsr_multiply_generic 2485 12.5 0.189 0.193 79.240 79.759 ot_scf_mini 116 9.6 0.004 0.004 75.310 75.586 multiply_cannon 2485 13.5 0.548 0.590 54.608 57.156 multiply_cannon_loop 2485 14.5 1.167 1.194 50.918 52.203 init_scf_loop 11 6.9 0.006 0.024 46.855 46.856 prepare_preconditioner 11 7.9 0.000 0.001 42.679 42.699 make_preconditioner 11 8.9 0.000 0.002 42.679 42.699 ot_mini 116 10.6 0.001 0.001 42.129 42.393 make_full_inverse_cholesky 11 9.9 0.000 0.000 36.274 41.328 multiply_cannon_multrec 29820 15.5 14.241 19.122 26.418 30.761 rebuild_ks_matrix 127 8.3 0.001 0.001 30.065 30.263 qs_ks_build_kohn_sham_matrix 127 9.3 0.017 0.020 30.065 30.262 qs_ks_update_qs_env 127 7.6 0.001 0.001 27.062 27.246 mp_waitall_1 146592 16.7 17.275 26.728 17.275 26.728 qs_ot_get_derivative 116 11.6 0.001 0.002 22.602 22.878 make_m2s 4970 13.5 0.094 0.099 20.306 21.333 make_images 4970 14.5 1.932 2.195 20.001 21.028 qs_ot_get_p 127 10.4 0.001 0.001 19.511 19.819 apply_preconditioner_dbcsr 127 12.6 0.000 0.001 19.021 19.521 apply_single 127 13.6 0.001 0.001 19.021 19.520 ot_diis_step 116 11.6 0.017 0.018 19.396 19.398 cp_fm_upper_to_full 104 14.7 11.086 16.427 11.086 16.427 cp_fm_cholesky_invert 11 10.9 16.402 16.411 16.402 16.411 qs_ot_p2m_diag 82 11.4 0.339 0.386 15.440 15.492 multiply_cannon_metrocomm3 29820 15.5 0.047 0.050 6.244 15.116 sum_up_and_integrate 127 10.3 0.139 0.150 14.870 14.899 integrate_v_rspace 127 11.3 0.004 0.004 14.731 14.764 qs_rho_update_rho_low 127 7.7 0.001 0.001 14.700 14.736 calculate_rho_elec 127 8.7 0.174 0.189 14.699 14.735 cp_dbcsr_syevd 82 12.4 0.005 0.005 14.152 14.154 make_images_data 4970 15.5 0.063 0.068 10.994 12.790 dbcsr_complete_redistribute 393 12.7 1.508 1.628 8.982 12.765 multiply_cannon_sync_h2d 29820 15.5 11.585 12.597 11.585 12.597 dbcsr_mm_accdrv_process 61748 16.2 7.404 8.130 11.755 12.252 hybrid_alltoall_any 5155 16.4 0.520 2.188 9.832 11.796 copy_fm_to_dbcsr 208 11.6 0.002 0.002 7.617 11.385 init_scf_run 11 5.9 0.000 0.001 11.305 11.307 scf_env_initial_rho_setup 11 6.9 0.001 0.001 11.305 11.307 cp_fm_diag_elpa 82 13.4 0.000 0.001 11.013 11.016 cp_fm_redistribute_end 82 14.4 1.909 10.928 1.928 10.932 cp_fm_diag_elpa_base 82 14.4 8.387 10.295 8.967 10.778 transfer_fm_to_dbcsr 11 9.9 0.000 0.000 6.380 10.088 pw_transfer 1535 11.6 0.088 0.106 9.522 9.612 qs_ot_get_derivative_diag 76 12.4 0.002 0.002 9.252 9.448 mp_alltoall_i22 712 14.1 5.627 9.403 5.627 9.403 fft_wrap_pw1pw2 1281 12.7 0.010 0.012 9.295 9.392 density_rs2pw 127 9.7 0.006 0.006 7.204 8.581 fft_wrap_pw1pw2_140 519 13.2 0.480 0.491 8.230 8.336 grid_integrate_task_list 127 12.3 7.420 7.863 7.420 7.863 cp_fm_cholesky_decompose 22 10.9 7.694 7.805 7.694 7.805 wfi_extrapolate 11 7.9 0.001 0.001 7.739 7.739 fft3d_ps 1281 14.7 2.763 2.873 7.527 7.607 multiply_cannon_metrocomm4 24850 15.5 0.075 0.083 2.681 7.322 mp_irecv_dv 75445 16.2 2.540 7.059 2.540 7.059 rs_pw_transfer 1038 11.9 0.014 0.014 4.879 6.478 calculate_dm_sparse 127 9.5 0.001 0.001 6.196 6.297 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 5.776 5.840 grid_collocate_task_list 127 9.7 5.045 5.819 5.045 5.819 mp_alltoall_d11v 2401 14.1 4.842 5.765 4.842 5.765 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 4.711 4.803 potential_pw2rs 127 12.3 0.022 0.023 4.751 4.765 qs_energies_init_hamiltonians 11 5.9 0.001 0.002 4.708 4.725 qs_ot_get_derivative_taylor 40 13.0 0.001 0.001 4.357 4.440 qs_ot_get_orbitals 116 10.6 0.001 0.001 4.164 4.216 mp_waitany 14840 13.9 2.333 3.961 2.333 3.961 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="406", plot="h2o_256_md", label="(8n/3r/4t)", y=190.436000, yerr=0.000000 PlotPoint: name="407", plot="h2o_256_md_mem", label="(8n/3r/4t)", y=1073.636364, yerr=12.234621 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/19/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1410022950912 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 1963542011904 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 11444706349056 0.0% 0.0% 100.0% flops 22 x 32 x 9 15019182452736 0.0% 0.0% 100.0% flops 9 x 32 x 22 15019182452736 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 92.796564E+12 0.0% 0.0% 100.0% flops max/rank 5.820057E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 6705499488 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 1.510289E+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 716108490000 ------------------------------------------------------------------------------- - - - 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 3989 58394. MP_Allreduce 10970 1176. MP_Sync 86 MP_Alltoall 1700 18828174. MP_SendRecv 3810 122880. MP_ISendRecv 3810 122880. MP_Wait 16000 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 0.067 0.129 175.075 175.081 qs_mol_dyn_low 1 2.0 0.004 0.018 174.469 174.485 qs_forces 11 3.9 0.010 0.016 174.003 174.013 qs_energies 11 4.9 0.001 0.001 166.408 166.426 scf_env_do_scf 11 5.9 0.001 0.001 148.672 148.682 velocity_verlet 10 3.0 0.307 0.365 113.243 113.260 scf_env_do_scf_inner_loop 116 6.6 0.003 0.008 112.309 112.311 qs_scf_new_mos 116 7.6 0.001 0.001 73.900 73.948 qs_scf_loop_do_ot 116 8.6 0.001 0.001 73.899 73.948 dbcsr_multiply_generic 2485 12.5 0.186 0.193 71.805 72.224 ot_scf_mini 116 9.6 0.003 0.004 69.585 69.634 multiply_cannon 2485 13.5 0.582 0.607 53.005 57.465 multiply_cannon_loop 2485 14.5 0.440 0.450 49.102 49.893 ot_mini 116 10.6 0.001 0.001 38.276 38.329 init_scf_loop 11 6.9 0.000 0.000 36.215 36.217 prepare_preconditioner 11 7.9 0.000 0.000 32.254 32.280 make_preconditioner 11 8.9 0.000 0.000 32.254 32.280 mp_waitall_1 124680 16.7 24.629 31.900 24.629 31.900 make_full_inverse_cholesky 11 9.9 0.000 0.000 30.022 30.300 rebuild_ks_matrix 127 8.3 0.001 0.001 29.751 29.845 qs_ks_build_kohn_sham_matrix 127 9.3 0.017 0.018 29.751 29.844 qs_ks_update_qs_env 127 7.6 0.001 0.001 26.981 27.063 multiply_cannon_multrec 9940 15.5 10.332 14.230 17.561 20.285 qs_ot_get_derivative 116 11.6 0.001 0.002 19.422 19.473 qs_ot_get_p 127 10.4 0.001 0.001 18.826 18.901 multiply_cannon_metrocomm3 9940 15.5 0.023 0.025 12.117 18.869 ot_diis_step 116 11.6 0.019 0.020 18.784 18.784 apply_preconditioner_dbcsr 127 12.6 0.000 0.000 18.387 18.580 apply_single 127 13.6 0.001 0.001 18.386 18.579 cp_fm_cholesky_invert 11 10.9 18.410 18.415 18.410 18.415 make_m2s 4970 13.5 0.065 0.071 15.185 16.762 make_images 4970 14.5 2.251 2.542 14.879 16.457 qs_rho_update_rho_low 127 7.7 0.001 0.001 15.354 15.379 calculate_rho_elec 127 8.7 0.257 0.266 15.354 15.378 sum_up_and_integrate 127 10.3 0.178 0.189 15.098 15.147 qs_ot_p2m_diag 82 11.4 0.489 0.495 15.117 15.132 integrate_v_rspace 127 11.3 0.004 0.004 14.920 14.977 cp_dbcsr_syevd 82 12.4 0.005 0.005 13.985 13.987 multiply_cannon_sync_h2d 9940 15.5 11.592 11.799 11.592 11.799 make_images_data 4970 15.5 0.051 0.060 9.014 11.172 cp_fm_diag_elpa 82 13.4 0.000 0.000 10.891 10.899 cp_fm_diag_elpa_base 82 14.4 10.635 10.731 10.883 10.893 init_scf_run 11 5.9 0.000 0.002 10.742 10.743 scf_env_initial_rho_setup 11 6.9 0.001 0.001 10.742 10.742 hybrid_alltoall_any 5155 16.4 0.811 3.636 8.919 10.732 pw_transfer 1535 11.6 0.085 0.096 9.754 9.793 fft_wrap_pw1pw2 1281 12.7 0.010 0.011 9.531 9.577 fft_wrap_pw1pw2_140 519 13.2 0.494 0.517 8.369 8.421 cp_fm_cholesky_decompose 22 10.9 8.065 8.172 8.065 8.172 grid_integrate_task_list 127 12.3 7.667 8.151 7.667 8.151 density_rs2pw 127 9.7 0.005 0.006 7.014 7.902 qs_ot_get_derivative_diag 76 12.4 0.002 0.002 7.729 7.778 fft3d_ps 1281 14.7 2.704 2.802 7.697 7.724 dbcsr_mm_accdrv_process 20590 16.1 2.687 3.567 6.861 7.471 wfi_extrapolate 11 7.9 0.001 0.001 7.331 7.331 multiply_cannon_metrocomm1 9940 15.5 0.028 0.029 4.176 6.892 mp_allgather_i34 2485 14.5 2.124 6.254 2.124 6.254 mp_alltoall_d11v 2401 14.1 5.107 6.189 5.107 6.189 calculate_dm_sparse 127 9.5 0.001 0.001 5.959 6.048 grid_collocate_task_list 127 9.7 5.352 6.013 5.352 6.013 dbcsr_complete_redistribute 393 12.7 2.127 2.251 5.323 5.756 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 5.483 5.495 rs_pw_transfer 1038 11.9 0.013 0.014 4.319 5.263 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 5.082 5.091 potential_pw2rs 127 12.3 0.026 0.026 4.635 4.650 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 3.996 4.026 copy_fm_to_dbcsr 208 11.6 0.002 0.002 3.517 3.864 build_core_hamiltonian_matrix_ 11 4.9 0.001 0.001 3.561 3.863 multiply_cannon_metrocomm4 7455 15.5 0.023 0.025 1.817 3.793 mp_irecv_dv 28618 15.9 1.781 3.717 1.781 3.717 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 3.674 3.690 copy_dbcsr_to_fm 185 11.7 0.004 0.004 3.566 3.634 qs_ot_get_orbitals 116 10.6 0.001 0.001 3.578 3.608 qs_ot_get_derivative_taylor 40 13.0 0.001 0.001 3.515 3.545 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="408", plot="h2o_256_md", label="(8n/2r/6t)", y=175.081000, yerr=0.000000 PlotPoint: name="409", plot="h2o_256_md_mem", label="(8n/2r/6t)", y=1400.000000, yerr=59.865000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_performance_tests/20/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1430456868864 0.0% 0.0% 100.0% flops 32 x 32 x 32 1971389988864 0.0% 0.0% 100.0% flops 22 x 9 x 32 1986252263424 0.0% 0.0% 100.0% flops 9 x 22 x 32 1992007581696 0.0% 0.0% 100.0% flops 22 x 22 x 32 2753954734080 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 11613074706432 0.0% 0.0% 100.0% flops 22 x 32 x 9 15239176888320 0.0% 0.0% 100.0% flops 9 x 32 x 22 15239176888320 0.0% 0.0% 100.0% flops 22 x 32 x 22 19911117062144 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 94.241596E+12 0.0% 0.0% 100.0% flops max/rank 11.787141E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 6806515136 0.0% 0.0% 100.0% number of processed stacks 1981760 0.0% 0.0% 100.0% average stack size 0.0 0.0 3434.6 marketing flops 145.659521E+12 ------------------------------------------------------------------------------- # multiplications 2533 max memory usage/rank 3.017396E+09 # max total images/rank 2 # max 3D layers 1 # MPI messages exchanged 101320 MPI messages size (bytes): total size 1.145104E+12 min size 0.000000E+00 max size 104.857600E+06 average size 11.301851E+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 45808 35567697920 4194304 < size <= 16777216 44720 382939955200 16777216 < size 10176 726592330608 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 4069 58204. MP_Allreduce 11170 1510. MP_Sync 86 MP_Alltoall 1724 36993608. MP_SendRecv 1806 218624. MP_ISendRecv 1806 218624. MP_Wait 9876 MP_ISend 6456 1080169. MP_IRecv 6456 1080169. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.038 0.090 300.241 300.251 qs_mol_dyn_low 1 2.0 0.003 0.003 299.688 299.700 qs_forces 11 3.9 0.003 0.003 299.598 299.602 qs_energies 11 4.9 0.001 0.001 290.260 290.272 scf_env_do_scf 11 5.9 0.001 0.001 267.408 267.423 velocity_verlet 10 3.0 0.002 0.003 215.402 215.409 scf_env_do_scf_inner_loop 118 6.6 0.003 0.008 143.072 143.075 init_scf_loop 11 6.9 0.000 0.000 124.083 124.087 prepare_preconditioner 11 7.9 0.000 0.000 119.000 119.030 make_preconditioner 11 8.9 0.000 0.000 119.000 119.030 make_full_inverse_cholesky 11 9.9 0.000 0.000 95.286 116.147 qs_scf_new_mos 118 7.6 0.001 0.001 91.335 91.445 qs_scf_loop_do_ot 118 8.6 0.001 0.001 91.334 91.444 ot_scf_mini 118 9.6 0.004 0.004 86.512 86.565 dbcsr_multiply_generic 2533 12.6 0.216 0.224 84.591 85.143 cp_fm_upper_to_full 104 14.8 52.603 75.266 52.603 75.266 multiply_cannon 2533 13.6 0.714 0.746 60.121 60.548 multiply_cannon_loop 2533 14.6 0.476 0.489 56.447 57.763 ot_mini 118 10.6 0.001 0.001 45.075 45.125 dbcsr_complete_redistribute 393 12.7 3.972 4.006 29.362 41.936 rebuild_ks_matrix 129 8.3 0.001 0.001 38.883 38.909 qs_ks_build_kohn_sham_matrix 129 9.3 0.018 0.018 38.882 38.908 copy_fm_to_dbcsr 208 11.6 0.002 0.002 25.967 38.576 transfer_fm_to_dbcsr 11 9.9 0.000 0.000 23.670 36.122 qs_ks_update_qs_env 129 7.6 0.001 0.001 35.910 35.945 mp_alltoall_i22 716 14.1 21.520 34.301 21.520 34.301 cp_fm_cholesky_invert 11 10.9 33.501 33.507 33.501 33.507 mp_waitall_1 104740 16.8 28.777 32.733 28.777 32.733 qs_ot_get_p 129 10.4 0.001 0.001 25.744 25.789 qs_ot_get_derivative 118 11.6 0.002 0.002 24.600 24.649 qs_ot_p2m_diag 82 11.4 0.868 0.874 21.701 21.733 qs_rho_update_rho_low 129 7.7 0.001 0.001 20.994 21.019 calculate_rho_elec 129 8.7 0.487 0.487 20.994 21.019 make_m2s 5066 13.6 0.077 0.079 19.780 20.735 multiply_cannon_metrocomm3 10132 15.6 0.024 0.024 19.496 20.732 ot_diis_step 118 11.6 0.022 0.023 20.444 20.444 make_images 5066 14.6 3.756 3.921 19.295 20.252 sum_up_and_integrate 129 10.3 0.323 0.325 20.017 20.088 cp_dbcsr_syevd 82 12.4 0.006 0.006 19.976 19.978 apply_preconditioner_dbcsr 129 12.6 0.000 0.000 19.645 19.882 apply_single 129 13.6 0.001 0.001 19.644 19.882 integrate_v_rspace 129 11.3 0.004 0.004 19.693 19.763 multiply_cannon_multrec 10132 15.6 10.560 12.389 18.599 18.783 cp_fm_diag_elpa 82 13.4 0.000 0.000 16.768 16.770 cp_fm_diag_elpa_base 82 14.4 12.425 14.006 16.764 16.765 multiply_cannon_sync_h2d 10132 15.6 15.782 15.808 15.782 15.808 init_scf_run 11 5.9 0.000 0.001 12.807 12.807 scf_env_initial_rho_setup 11 6.9 0.001 0.001 12.806 12.807 pw_transfer 1559 11.6 0.094 0.095 12.529 12.545 hybrid_alltoall_any 5251 16.5 1.323 3.107 10.651 12.368 make_images_data 5066 15.6 0.060 0.065 10.533 12.309 fft_wrap_pw1pw2 1301 12.7 0.011 0.012 12.290 12.305 fft_wrap_pw1pw2_140 527 13.2 0.550 0.556 10.894 10.929 fft3d_ps 1301 14.7 2.789 2.821 10.304 10.308 dbcsr_mm_accdrv_process 20950 16.1 3.890 6.006 7.797 9.775 qs_ot_get_derivative_diag 76 12.4 0.002 0.002 9.335 9.375 mp_alltoall_d11v 2417 14.1 8.364 9.360 8.364 9.360 wfi_extrapolate 11 7.9 0.001 0.001 9.071 9.071 cp_fm_cholesky_decompose 22 10.9 8.949 8.982 8.949 8.982 grid_integrate_task_list 129 12.3 8.647 8.821 8.647 8.821 density_rs2pw 129 9.7 0.005 0.005 8.467 8.619 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 8.064 8.064 calculate_dm_sparse 129 9.5 0.001 0.001 6.794 6.895 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 6.725 6.787 grid_collocate_task_list 129 9.7 6.420 6.448 6.420 6.448 rs_scatter_matrices 140 9.7 3.604 4.614 6.073 6.316 copy_dbcsr_to_fm 185 11.7 0.004 0.004 6.043 6.109 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="410", plot="h2o_256_md", label="(8n/1r/12t)", y=300.251000, yerr=0.000000 PlotPoint: name="411", plot="h2o_256_md_mem", label="(8n/1r/12t)", y=2687.090909, yerr=165.347818 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 1.261900E+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 6805335. MP_SendRecv 285 19200. MP_ISendRecv 285 19200. MP_Wait 1017 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 0.023 0.046 84.895 84.897 qs_energies 1 2.0 0.000 0.000 84.357 84.364 ls_scf 1 3.0 0.000 0.000 83.448 83.455 dbcsr_multiply_generic 111 6.7 0.014 0.015 72.410 72.590 multiply_cannon 111 7.7 0.017 0.020 55.768 57.115 multiply_cannon_loop 111 8.7 0.210 0.224 52.396 53.904 ls_scf_main 1 4.0 0.000 0.000 52.289 52.289 density_matrix_trs4 2 5.0 0.002 0.003 46.764 46.841 ls_scf_init_scf 1 4.0 0.000 0.000 28.127 28.128 ls_scf_init_matrix_S 1 5.0 0.000 0.000 26.990 27.042 mp_waitall_1 11031 10.9 22.471 25.605 22.471 25.605 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.001 24.848 24.868 multiply_cannon_multrec 2664 9.7 8.180 9.011 15.513 17.262 multiply_cannon_sync_h2d 2664 9.7 13.523 14.768 13.523 14.768 make_m2s 222 7.7 0.008 0.011 13.048 13.480 make_images 222 8.7 0.099 0.110 13.026 13.460 multiply_cannon_metrocomm1 2664 9.7 0.009 0.010 9.764 12.857 make_images_data 222 9.7 0.004 0.005 7.609 8.187 hybrid_alltoall_any 227 10.6 0.215 1.836 6.552 7.903 multiply_cannon_metrocomm3 2664 9.7 0.009 0.010 5.430 7.901 dbcsr_mm_accdrv_process 4760 10.4 0.508 0.622 6.951 7.894 dbcsr_mm_accdrv_process_sort 4760 11.4 6.244 7.108 6.244 7.108 calculate_norms 4752 9.8 5.522 6.153 5.522 6.153 apply_matrix_preconditioner 6 5.3 0.000 0.000 5.013 5.164 mp_sum_l 807 5.4 3.076 4.690 3.076 4.690 make_images_sizes 222 9.7 0.000 0.000 0.785 3.717 mp_alltoall_i44 222 10.7 0.784 3.717 0.784 3.717 multiply_cannon_metrocomm4 2442 9.7 0.012 0.014 2.056 3.642 mp_irecv_dv 6231 10.9 2.038 3.613 2.038 3.613 dbcsr_multiply_generic_mpsum_f 86 7.8 0.000 0.000 2.290 3.398 arnoldi_extremal 4 6.8 0.000 0.000 3.257 3.284 arnoldi_normal_ev 4 7.8 0.001 0.002 3.257 3.283 build_subspace 16 8.4 0.009 0.012 3.162 3.165 ls_scf_post 1 4.0 0.000 0.000 3.033 3.040 ls_scf_store_result 1 5.0 0.000 0.000 2.854 2.892 dbcsr_special_finalize 555 9.7 0.005 0.006 2.301 2.872 dbcsr_merge_single_wm 555 10.7 0.453 0.571 2.293 2.864 make_images_pack 222 9.7 2.208 2.634 2.209 2.636 dbcsr_matrix_vector_mult 304 9.0 0.005 0.013 2.332 2.596 dbcsr_sort_data 658 11.4 2.100 2.590 2.100 2.590 dbcsr_matrix_vector_mult_local 304 10.0 2.074 2.490 2.076 2.492 ls_scf_dm_to_ks 2 5.0 0.000 0.000 2.296 2.360 buffer_matrices_ensure_size 222 8.7 1.744 2.077 1.744 2.077 qs_ks_update_qs_env 3 6.3 0.000 0.000 1.787 1.788 rebuild_ks_matrix 3 7.3 0.000 0.000 1.777 1.779 qs_ks_build_kohn_sham_matrix 3 8.3 0.000 0.001 1.777 1.779 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="500", plot="h2o_32_nrep3_ls", label="(8n/12r/1t)", y=84.897000, yerr=0.000000 PlotPoint: name="501", plot="h2o_32_nrep3_ls_mem", label="(8n/12r/1t)", y=1142.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 2.111095E+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_SendRecv 141 57600. MP_ISendRecv 141 57600. MP_Wait 687 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 0.048 0.111 91.618 91.621 qs_energies 1 2.0 0.002 0.020 91.000 91.014 ls_scf 1 3.0 0.009 0.062 89.677 89.689 dbcsr_multiply_generic 111 6.7 0.015 0.016 75.345 75.670 multiply_cannon 111 7.7 0.028 0.045 53.349 56.758 ls_scf_main 1 4.0 0.004 0.046 55.056 55.062 multiply_cannon_loop 111 8.7 0.116 0.122 50.052 52.698 density_matrix_trs4 2 5.0 0.003 0.016 49.307 49.511 ls_scf_init_scf 1 4.0 0.001 0.002 30.731 30.735 mp_waitall_1 9105 10.9 21.278 29.598 21.278 29.598 ls_scf_init_matrix_S 1 5.0 0.000 0.001 29.496 29.596 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.002 27.091 27.107 multiply_cannon_multrec 1332 9.7 13.150 16.654 22.368 26.992 multiply_cannon_metrocomm3 1332 9.7 0.006 0.007 11.794 20.376 make_m2s 222 7.7 0.006 0.007 15.418 16.069 make_images 222 8.7 1.569 1.938 15.388 16.041 dbcsr_mm_accdrv_process 4041 10.4 0.293 0.453 8.820 10.358 dbcsr_mm_accdrv_process_sort 4041 11.4 8.412 9.911 8.412 9.911 make_images_data 222 9.7 0.004 0.004 8.909 9.811 hybrid_alltoall_any 227 10.6 0.521 2.437 8.338 9.362 mp_sum_l 807 5.4 5.287 8.056 5.287 8.056 multiply_cannon_metrocomm4 1221 9.7 0.006 0.008 3.214 7.633 mp_irecv_dv 3311 11.0 3.194 7.569 3.194 7.569 calculate_norms 2376 9.8 5.987 6.738 5.987 6.738 dbcsr_multiply_generic_mpsum_f 86 7.8 0.000 0.000 4.023 6.545 multiply_cannon_sync_h2d 1332 9.7 4.824 6.001 4.824 6.001 apply_matrix_preconditioner 6 5.3 0.000 0.000 5.075 5.259 arnoldi_extremal 4 6.8 0.000 0.000 4.625 4.649 arnoldi_normal_ev 4 7.8 0.001 0.005 4.625 4.649 build_subspace 16 8.4 0.014 0.021 4.369 4.371 ls_scf_post 1 4.0 0.004 0.025 3.880 3.890 ls_scf_store_result 1 5.0 0.000 0.000 3.254 3.377 dbcsr_matrix_vector_mult 304 9.0 0.009 0.021 3.133 3.362 dbcsr_matrix_vector_mult_local 304 10.0 2.741 3.220 2.743 3.222 multiply_cannon_metrocomm1 1332 9.7 0.003 0.004 1.206 3.057 ls_scf_dm_to_ks 2 5.0 0.002 0.008 2.621 2.706 mp_allgather_i34 111 8.7 0.998 2.469 0.998 2.469 make_images_pack 222 9.7 2.021 2.428 2.023 2.430 dbcsr_sort_data 436 11.2 1.826 2.023 1.826 2.023 qs_ks_update_qs_env 3 6.3 0.000 0.000 1.895 1.900 rebuild_ks_matrix 3 7.3 0.000 0.000 1.882 1.888 qs_ks_build_kohn_sham_matrix 3 8.3 0.001 0.008 1.882 1.888 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="502", plot="h2o_32_nrep3_ls", label="(8n/6r/2t)", y=91.621000, yerr=0.000000 PlotPoint: name="503", plot="h2o_32_nrep3_ls_mem", label="(8n/6r/2t)", y=1762.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 2.762412E+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_SendRecv 93 57600. MP_ISendRecv 93 57600. MP_Wait 639 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 0.034 0.054 94.815 94.817 qs_energies 1 2.0 0.000 0.000 94.177 94.183 ls_scf 1 3.0 0.000 0.001 92.550 92.555 dbcsr_multiply_generic 111 6.7 0.016 0.022 76.564 76.802 ls_scf_main 1 4.0 0.000 0.001 57.547 57.552 multiply_cannon 111 7.7 0.039 0.099 52.821 56.709 multiply_cannon_loop 111 8.7 0.099 0.107 49.214 51.875 density_matrix_trs4 2 5.0 0.002 0.003 51.551 51.697 mp_waitall_1 7281 11.0 24.031 33.034 24.031 33.034 ls_scf_init_scf 1 4.0 0.000 0.001 31.447 31.449 ls_scf_init_matrix_S 1 5.0 0.000 0.002 29.649 29.702 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.001 27.228 27.243 multiply_cannon_multrec 888 9.7 12.543 15.440 21.090 24.557 multiply_cannon_metrocomm3 888 9.7 0.004 0.004 11.115 21.948 make_m2s 222 7.7 0.006 0.007 17.102 18.396 make_images 222 8.7 1.965 2.263 17.064 18.356 make_images_data 222 9.7 0.003 0.004 9.897 10.945 hybrid_alltoall_any 227 10.6 0.620 2.861 9.562 10.800 dbcsr_mm_accdrv_process 3754 10.4 0.258 0.425 8.051 9.314 dbcsr_mm_accdrv_process_sort 3754 11.4 7.680 8.889 7.680 8.889 mp_sum_l 807 5.4 5.094 7.679 5.094 7.679 multiply_cannon_sync_h2d 888 9.7 6.001 7.125 6.001 7.125 multiply_cannon_metrocomm1 888 9.7 0.003 0.003 3.642 7.085 multiply_cannon_metrocomm4 777 9.7 0.004 0.005 2.445 7.049 mp_irecv_dv 2335 11.1 2.431 6.995 2.431 6.995 dbcsr_multiply_generic_mpsum_f 86 7.8 0.000 0.000 3.880 5.989 arnoldi_extremal 4 6.8 0.000 0.000 5.173 5.193 arnoldi_normal_ev 4 7.8 0.001 0.005 5.173 5.193 apply_matrix_preconditioner 6 5.3 0.000 0.000 4.965 5.168 build_subspace 16 8.4 0.014 0.020 4.866 4.872 calculate_norms 1584 9.8 4.294 4.637 4.294 4.637 mp_allgather_i34 111 8.7 1.450 3.907 1.450 3.907 dbcsr_matrix_vector_mult 304 9.0 0.009 0.021 3.443 3.814 dbcsr_matrix_vector_mult_local 304 10.0 3.002 3.588 3.004 3.590 ls_scf_post 1 4.0 0.000 0.001 3.557 3.562 ls_scf_store_result 1 5.0 0.000 0.000 3.286 3.383 ls_scf_dm_to_ks 2 5.0 0.000 0.001 2.864 2.937 qs_ks_update_qs_env 3 6.3 0.000 0.000 2.398 2.400 rebuild_ks_matrix 3 7.3 0.000 0.000 2.380 2.382 qs_ks_build_kohn_sham_matrix 3 8.3 0.000 0.001 2.380 2.382 make_images_pack 222 9.7 1.808 2.119 1.810 2.122 dbcsr_sort_data 325 11.1 1.877 2.104 1.877 2.104 make_images_sizes 222 9.7 0.000 0.000 0.892 2.041 mp_alltoall_i44 222 10.7 0.892 2.041 0.892 2.041 dbcsr_data_release 9322 10.9 1.331 1.953 1.331 1.953 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="504", plot="h2o_32_nrep3_ls", label="(8n/4r/3t)", y=94.817000, yerr=0.000000 PlotPoint: name="505", plot="h2o_32_nrep3_ls_mem", label="(8n/4r/3t)", y=2212.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 3.321065E+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_SendRecv 69 86400. MP_ISendRecv 69 86400. MP_Wait 531 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 0.053 0.079 97.559 97.560 qs_energies 1 2.0 0.000 0.000 96.847 96.853 ls_scf 1 3.0 0.000 0.001 95.195 95.199 dbcsr_multiply_generic 111 6.7 0.017 0.019 78.937 79.193 ls_scf_main 1 4.0 0.000 0.002 59.249 59.250 multiply_cannon 111 7.7 0.046 0.139 51.588 56.046 density_matrix_trs4 2 5.0 0.002 0.003 53.039 53.185 multiply_cannon_loop 111 8.7 0.114 0.125 46.492 50.363 ls_scf_init_scf 1 4.0 0.000 0.001 32.716 32.717 ls_scf_init_matrix_S 1 5.0 0.000 0.001 31.481 31.544 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.001 28.986 29.001 mp_waitall_1 6369 11.0 22.667 28.869 22.667 28.869 multiply_cannon_multrec 1332 9.7 14.246 16.929 22.081 24.860 make_m2s 222 7.7 0.007 0.008 21.081 22.538 make_images 222 8.7 3.132 3.597 21.030 22.490 multiply_cannon_metrocomm3 1332 9.7 0.003 0.003 8.978 16.922 make_images_data 222 9.7 0.004 0.004 11.780 13.367 hybrid_alltoall_any 227 10.6 0.797 3.814 11.016 12.814 mp_sum_l 807 5.4 4.691 9.211 4.691 9.211 dbcsr_mm_accdrv_process 3641 10.4 0.203 0.397 7.484 9.011 dbcsr_mm_accdrv_process_sort 3641 11.4 7.112 8.583 7.112 8.583 dbcsr_multiply_generic_mpsum_f 86 7.8 0.000 0.000 3.638 7.361 multiply_cannon_metrocomm4 1110 9.7 0.004 0.006 2.081 6.016 multiply_cannon_sync_h2d 1332 9.7 5.480 5.980 5.480 5.980 mp_irecv_dv 3229 10.9 2.058 5.938 2.058 5.938 multiply_cannon_metrocomm1 1332 9.7 0.003 0.003 2.739 5.353 arnoldi_extremal 4 6.8 0.000 0.000 5.198 5.210 arnoldi_normal_ev 4 7.8 0.001 0.005 5.198 5.209 build_subspace 16 8.4 0.014 0.021 4.866 4.873 apply_matrix_preconditioner 6 5.3 0.000 0.000 4.618 4.795 mp_allgather_i34 111 8.7 2.221 4.687 2.221 4.687 calculate_norms 2376 9.8 4.197 4.562 4.197 4.562 dbcsr_matrix_vector_mult 304 9.0 0.010 0.020 3.573 3.857 dbcsr_matrix_vector_mult_local 304 10.0 3.162 3.659 3.164 3.661 dbcsr_sort_data 658 11.4 3.040 3.468 3.040 3.468 dbcsr_special_finalize 555 9.7 0.006 0.007 2.787 3.248 dbcsr_merge_single_wm 555 10.7 0.537 0.670 2.779 3.240 ls_scf_post 1 4.0 0.000 0.001 3.230 3.236 ls_scf_dm_to_ks 2 5.0 0.000 0.001 3.044 3.103 ls_scf_store_result 1 5.0 0.000 0.000 2.964 3.033 dbcsr_data_release 10477 10.7 1.574 2.420 1.574 2.420 qs_ks_update_qs_env 3 6.3 0.000 0.000 1.990 1.991 dbcsr_finalize 304 7.8 0.049 0.061 1.795 1.985 rebuild_ks_matrix 3 7.3 0.000 0.000 1.967 1.968 qs_ks_build_kohn_sham_matrix 3 8.3 0.001 0.002 1.967 1.968 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="506", plot="h2o_32_nrep3_ls", label="(8n/3r/4t)", y=97.560000, yerr=0.000000 PlotPoint: name="507", plot="h2o_32_nrep3_ls_mem", label="(8n/3r/4t)", y=2723.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 4.629844E+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_SendRecv 45 115200. MP_ISendRecv 45 115200. MP_Wait 528 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 0.041 0.058 99.713 99.715 qs_energies 1 2.0 0.000 0.000 98.904 98.909 ls_scf 1 3.0 0.000 0.000 96.956 96.966 dbcsr_multiply_generic 111 6.7 0.017 0.018 78.028 78.305 ls_scf_main 1 4.0 0.000 0.000 62.338 62.339 multiply_cannon 111 7.7 0.064 0.109 55.121 60.983 density_matrix_trs4 2 5.0 0.002 0.003 55.072 55.188 multiply_cannon_loop 111 8.7 0.069 0.078 50.540 52.724 mp_waitall_1 5436 11.0 26.266 33.475 26.266 33.475 ls_scf_init_scf 1 4.0 0.000 0.000 30.978 30.983 ls_scf_init_matrix_S 1 5.0 0.000 0.000 29.579 29.648 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.001 27.413 27.422 multiply_cannon_multrec 444 9.7 13.985 16.281 21.007 23.585 make_m2s 222 7.7 0.004 0.005 17.608 20.139 make_images 222 8.7 3.714 4.408 17.546 20.079 multiply_cannon_metrocomm1 444 9.7 0.002 0.002 11.297 16.662 multiply_cannon_metrocomm3 444 9.7 0.001 0.001 5.911 15.234 make_images_data 222 9.7 0.003 0.004 9.862 12.341 hybrid_alltoall_any 227 10.6 0.790 3.760 9.690 12.197 multiply_cannon_sync_h2d 444 9.7 6.474 8.466 6.474 8.466 dbcsr_mm_accdrv_process 3003 10.4 0.176 0.339 6.732 7.878 dbcsr_mm_accdrv_process_sort 3003 11.4 6.418 7.532 6.418 7.532 mp_allgather_i34 111 8.7 2.819 6.965 2.819 6.965 mp_sum_l 807 5.4 3.354 6.508 3.354 6.508 arnoldi_extremal 4 6.8 0.000 0.000 5.892 5.913 arnoldi_normal_ev 4 7.8 0.001 0.004 5.892 5.912 build_subspace 16 8.4 0.015 0.020 5.472 5.480 dbcsr_multiply_generic_mpsum_f 86 7.8 0.000 0.000 2.434 5.168 apply_matrix_preconditioner 6 5.3 0.000 0.000 4.584 4.740 multiply_cannon_metrocomm4 333 9.7 0.001 0.002 1.661 4.713 mp_irecv_dv 1241 11.2 1.641 4.680 1.641 4.680 dbcsr_matrix_vector_mult 304 9.0 0.011 0.021 4.166 4.374 dbcsr_matrix_vector_mult_local 304 10.0 3.647 4.117 3.649 4.119 ls_scf_dm_to_ks 2 5.0 0.000 0.000 3.653 3.734 calculate_norms 792 9.8 3.553 3.713 3.553 3.713 ls_scf_post 1 4.0 0.000 0.000 3.640 3.645 make_images_sizes 222 9.7 0.000 0.000 1.117 3.479 mp_alltoall_i44 222 10.7 1.117 3.479 1.117 3.479 ls_scf_store_result 1 5.0 0.000 0.000 3.369 3.463 dbcsr_finalize 304 7.8 0.062 0.078 2.197 2.254 qs_ks_update_qs_env 3 6.3 0.000 0.000 2.196 2.198 rebuild_ks_matrix 3 7.3 0.000 0.000 2.164 2.165 qs_ks_build_kohn_sham_matrix 3 8.3 0.000 0.001 2.164 2.165 dbcsr_merge_all 275 8.9 0.473 0.523 2.046 2.096 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="508", plot="h2o_32_nrep3_ls", label="(8n/2r/6t)", y=99.715000, yerr=0.000000 PlotPoint: name="509", plot="h2o_32_nrep3_ls_mem", label="(8n/2r/6t)", y=3632.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/843d9feafdfe50065c8ad5624e7456a67c33f1ab_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 8.776348E+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_SendRecv 42 732600. MP_ISendRecv 42 732600. MP_Wait 267 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 0.130 0.171 109.806 109.816 qs_energies 1 2.0 0.000 0.000 108.056 108.069 ls_scf 1 3.0 0.000 0.000 105.080 105.092 dbcsr_multiply_generic 111 6.7 0.023 0.027 77.887 77.981 ls_scf_main 1 4.0 0.000 0.000 66.322 66.323 density_matrix_trs4 2 5.0 0.002 0.003 56.987 57.043 multiply_cannon 111 7.7 0.207 0.330 50.066 51.659 multiply_cannon_loop 111 8.7 0.067 0.069 46.733 47.763 ls_scf_init_scf 1 4.0 0.000 0.000 34.658 34.658 ls_scf_init_matrix_S 1 5.0 0.000 0.000 32.826 32.845 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.001 30.012 30.021 mp_waitall_1 4527 11.1 22.786 26.315 22.786 26.315 make_m2s 222 7.7 0.005 0.005 24.149 25.018 make_images 222 8.7 4.574 4.951 24.043 24.910 multiply_cannon_multrec 444 9.7 17.861 18.661 22.462 23.187 hybrid_alltoall_any 227 10.6 1.660 3.639 13.168 16.059 make_images_data 222 9.7 0.003 0.003 13.395 15.714 multiply_cannon_metrocomm3 444 9.7 0.001 0.001 10.796 11.581 multiply_cannon_sync_h2d 444 9.7 8.870 8.989 8.870 8.989 arnoldi_extremal 4 6.8 0.000 0.000 7.355 7.366 arnoldi_normal_ev 4 7.8 0.002 0.008 7.355 7.366 build_subspace 16 8.4 0.026 0.036 6.802 6.813 dbcsr_matrix_vector_mult 304 9.0 0.016 0.032 5.431 5.581 ls_scf_dm_to_ks 2 5.0 0.000 0.000 5.288 5.381 apply_matrix_preconditioner 6 5.3 0.000 0.000 5.034 5.300 dbcsr_matrix_vector_mult_local 304 10.0 4.935 5.232 4.938 5.234 dbcsr_mm_accdrv_process 1814 10.4 0.231 0.319 4.423 4.549 dbcsr_mm_accdrv_process_sort 1814 11.4 4.124 4.255 4.124 4.255 ls_scf_post 1 4.0 0.000 0.000 4.100 4.113 make_images_sizes 222 9.7 0.000 0.000 1.445 3.509 mp_alltoall_i44 222 10.7 1.444 3.509 1.444 3.509 ls_scf_store_result 1 5.0 0.000 0.000 3.436 3.446 calculate_norms 792 9.8 3.219 3.273 3.219 3.273 dbcsr_finalize 304 7.8 0.082 0.089 3.086 3.166 mp_allgather_i34 111 8.7 0.983 2.972 0.983 2.972 qs_energies_init_hamiltonians 1 3.0 0.001 0.001 2.945 2.946 dbcsr_merge_all 275 8.9 0.893 0.920 2.872 2.945 dbcsr_complete_redistribute 5 7.6 1.433 1.464 2.771 2.883 dbcsr_data_release 12724 10.6 2.340 2.876 2.340 2.876 qs_ks_update_qs_env 3 6.3 0.000 0.000 2.605 2.606 matrix_ls_to_qs 2 6.0 0.000 0.000 2.428 2.557 rebuild_ks_matrix 3 7.3 0.000 0.000 2.537 2.539 qs_ks_build_kohn_sham_matrix 3 8.3 0.001 0.001 2.537 2.539 dbcsr_sort_data 325 11.1 2.439 2.497 2.439 2.497 dbcsr_new_transposed 4 7.5 0.243 0.252 2.287 2.297 dbcsr_frobenius_norm 74 6.6 2.056 2.141 2.206 2.245 dbcsr_add_d 103 6.2 0.000 0.000 2.133 2.207 dbcsr_add_anytype 103 7.2 0.859 0.892 2.133 2.207 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="510", plot="h2o_32_nrep3_ls", label="(8n/1r/12t)", y=109.816000, yerr=0.000000 PlotPoint: name="511", plot="h2o_32_nrep3_ls_mem", label="(8n/1r/12t)", y=6834.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ========= END RESULTS =========== CommitSHA: 843d9feafdfe50065c8ad5624e7456a67c33f1ab Summary: empty Status: OK