=== 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: 069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800 ################# 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 2.6.6, ELPA 2023.05.001, LIBINT 2.6.0, LIBPEXSI 1.2.0, # LIBXC 6.2.2, LIBVORI 220621, LIBXSMM 1.17, PLUMED 2.8.2, # 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. # # Last update: 21.06.2023 # # \ 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 ${PWD}/tools/toolchain/install/setup; \ printf "done\n"; \ echo "Check the output above for error messages and consistency!"; \ echo; \ echo "If everything is OK, you can build a CP2K production binary with"; \ echo " make -j ARCH=${this_file%.*} VERSION=${this_file##*.}"; \ echo; \ echo "Alternatively, you can add further checks, e.g. for regression testing, with"; \ echo " make -j ARCH=${this_file%.*} VERSION=${this_file##*.} DO_CHECKS=yes"; \ echo "or build CP2K as a library with"; \ echo " make -j ARCH=${this_file%.*} VERSION=${this_file##*.} libcp2k"; \ echo; \ return # Set options DO_CHECKS := no USE_ACC := yes USE_COSMA := 2.6.6 USE_ELPA := 2023.05.001 USE_LIBINT := 2.6.0 USE_LIBPEXSI := 1.2.0 USE_LIBVORI := 220621 USE_LIBXC := 6.2.2 USE_LIBXSMM := 1.17 USE_PLUMED := 2.8.2 #USE_QUIP := 0.9.10 USE_SIRIUS := 7.4.3 USE_SPGLIB := 1.16.2 # Only needed for SIRIUS LIBVDWXC_VER := 0.4.0 SPFFT_VER := 1.0.6 SPLA_VER := 1.5.5 HDF5_VER := 1.12.0 # Only needed for LIBPEXSI SCOTCH_VER := 6.0.0 SUPERLU_VER := 6.1.0 LMAX := 5 MAX_CONTR := 4 GPUVER := P100 OFFLOAD_TARGET := cuda CC := cc CXX := CC OFFLOAD_CC := nvcc FC := ftn LD := ftn AR := ar -r # cc, CC, and ftn include already the proper -march flag CFLAGS := -O2 -fopenmp -fopenmp-simd -ftree-vectorize -funroll-loops -g DFLAGS := -D__parallel DFLAGS += -D__SCALAPACK DFLAGS += -D__FFTW3 DFLAGS += -D__MAX_CONTR=$(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.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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/01 job id: 48511549 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/02 job id: 48511551 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/03 job id: 48511553 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/04 job id: 48511555 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/05 job id: 48511557 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/06 job id: 48511558 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/07 job id: 48511559 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/08 job id: 48511561 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/09 job id: 48511562 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/10 job id: 48511564 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/11 job id: 48511565 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/12 job id: 48511566 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/13 job id: 48511567 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/14 job id: 48511569 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/15 job id: 48511571 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/16 job id: 48511574 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/17 job id: 48511575 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/18 job id: 48511576 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/19 job id: 48511577 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/20 job id: 48511578 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/21 job id: 48511579 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/22 job id: 48511580 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/23 job id: 48511581 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/24 job id: 48511582 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/25 job id: 48511583 --- 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/26 job id: 48511587 --- 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 ~~~~~~~ ~~~~~~~~~ TEST ~~~~~~~~~ description: 512 H2O (4 NVE MD steps on 64 nodes) input file: benchmarks/QS/00512_H2O/H2O-512_md.inp required files: [] output file: result.log # nodes = 64 # ranks/node = 12 # threads/rank = 1 nrepeat = 1 time[min] = 10 run dir: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/27 job id: 48511588 --- Point --- name: 601 plot: h2o_512_md regex: CP2K label: (64n/12r/1t) --- Point --- name: 602 plot: h2o_512_md_mem regex: Estimated peak process memory label: (64n/12r/1t) ~~~~~~~ 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]" ~~~~~~~~~ PLOT ~~~~~~~~~ Plot: name="h2o_512_md", title="512 H2O (4 NVE MD steps on 64 nodes)", xlabel="Revision", ylabel="Time [s]" ~~~~~~~~~ PLOT ~~~~~~~~~ Plot: name="h2o_512_md_mem", title="512 H2O (4 NVE MD steps on 64 nodes)", xlabel="Revision", ylabel="Est. peak process memory [MiB]" === END PLOTS (description) === ============ RESULTS ============ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 424 8. 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 0.023 0.036 135.519 135.519 farming_run 1 2.0 135.014 135.031 135.488 135.491 ------------------------------------------------------------------------------- @@@@@@@@@@ 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 1.483346E+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 230 1134128. MP_Allreduce 571 1938539. MP_Sync 25 MP_Alltoall 38 9316958. MP_SendRecv 120 384007. MP_ISendRecv 45 235435. MP_Wait 191 MP_comm_split 10 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 0.007 0.027 114.406 114.406 qs_energies 1 2.0 0.000 0.000 114.156 114.161 mp2_main 1 3.0 0.000 0.000 112.105 112.109 mp2_gpw_main 1 4.0 0.028 0.034 111.120 111.125 mp2_ri_gpw_compute_in 1 5.0 0.173 0.183 92.180 92.602 mp2_ri_gpw_compute_in_loop 1 6.0 0.004 0.004 54.638 55.062 mp2_eri_3c_integrate_gpw 272 7.0 0.154 0.165 41.022 46.150 get_2c_integrals 1 6.0 0.008 0.009 36.628 37.368 integrate_v_rspace 273 8.0 0.441 0.457 24.592 29.447 pw_transfer 6555 10.6 0.374 0.388 26.773 27.351 fft_wrap_pw1pw2 5465 11.4 0.045 0.048 25.406 26.068 grid_integrate_task_list 273 9.0 20.504 25.816 20.504 25.816 fft_wrap_pw1pw2_100 2178 12.4 1.102 1.294 22.949 23.604 rpa_ri_compute_en 1 5.0 0.021 0.064 18.824 18.980 compute_2c_integrals 1 7.0 0.002 0.003 18.973 18.973 compute_2c_integrals_loop_lm 1 8.0 0.003 0.004 18.581 18.713 mp2_eri_2c_integrate_gpw 1 9.0 2.388 2.422 18.579 18.709 cp_fm_cholesky_decompose 12 8.2 17.580 18.368 17.580 18.368 cholesky_decomp 1 7.0 0.000 0.000 16.506 17.291 fft3d_s 5443 13.4 16.120 16.516 16.142 16.538 ao_to_mo_and_store_B_mult_1 272 7.0 10.778 15.312 10.778 15.312 calculate_wavefunction 272 8.0 5.430 5.577 12.338 13.028 rpa_num_int 1 6.0 0.001 0.007 10.606 10.620 rpa_num_int_RPA_matrix_operati 8 7.0 0.000 0.000 10.578 10.582 calc_mat_Q 8 8.0 0.000 0.000 9.458 9.538 contract_S_to_Q 8 9.0 0.000 0.000 8.878 8.963 calc_potential_gpw 544 9.5 0.005 0.006 8.202 8.665 parallel_gemm_fm 14 9.1 0.000 0.000 8.472 8.564 parallel_gemm_fm_cosma 14 10.1 8.472 8.564 8.472 8.564 mp2_eri_2c_integrate_gpw_pot_l 272 10.0 0.001 0.001 8.056 8.285 potential_pw2rs 545 10.0 0.107 0.109 7.487 8.137 create_integ_mat 1 6.0 0.014 0.026 7.663 7.672 collocate_single_gaussian 272 10.0 0.039 0.041 7.332 7.582 array2fm 1 7.0 0.000 0.000 6.652 7.004 pw_scatter_s 2720 13.7 4.351 4.515 4.351 4.515 pw_gather_s 2722 13.2 3.428 3.772 3.428 3.772 array2fm_buffer_send 1 8.0 2.926 3.119 2.926 3.119 pw_poisson_solve 545 10.5 1.121 1.184 2.195 2.344 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="10", plot="h2o_32_ri_rpa_mp2", label="RI-RPA (8n/2r/6t)", y=111.124885, yerr=0.000000 PlotPoint: name="11", plot="h2o_32_ri_rpa_mp2_mem", label="RI-RPA (8n/2r/6t)", y=2796.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 424 9. 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 0.054 0.061 426.353 426.355 farming_run 1 2.0 425.603 425.607 426.292 426.321 ------------------------------------------------------------------------------- @@@@@@@@@@ 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 897827128576 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 249788821 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 1.239892E+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 757 478553. MP_Allreduce 2021 21391. MP_Sync 37 MP_Alltoall 77 28055074. 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 0.034 216.831 216.832 qs_energies 1 2.0 0.000 0.000 216.589 216.607 scf_env_do_scf 1 3.0 0.000 0.000 114.956 114.956 qs_ks_update_qs_env 5 5.0 0.000 0.000 114.035 114.043 rebuild_ks_matrix 4 6.0 0.000 0.000 114.033 114.041 qs_ks_build_kohn_sham_matrix 4 7.0 0.054 0.062 114.033 114.041 hfx_ks_matrix 4 8.0 0.001 0.001 113.687 113.692 integrate_four_center 4 9.0 0.154 0.471 113.686 113.691 integrate_four_center_main 4 10.0 0.158 0.732 102.185 105.439 integrate_four_center_bin 263 11.0 102.027 105.386 102.027 105.386 mp2_main 1 3.0 0.000 0.000 101.339 101.357 mp2_gpw_main 1 4.0 0.052 0.081 100.444 100.465 init_scf_loop 1 4.0 0.000 0.000 97.098 97.098 mp2_ri_gpw_compute_in 1 5.0 0.066 0.072 73.889 74.882 mp2_ri_gpw_compute_in_loop 1 6.0 0.002 0.002 53.739 54.730 mp2_eri_3c_integrate_gpw 91 7.0 0.146 0.162 41.457 46.441 integrate_v_rspace 95 8.0 0.397 0.563 27.928 32.778 pw_transfer 2240 10.6 0.145 0.178 29.522 29.921 fft_wrap_pw1pw2 1868 11.4 0.018 0.023 28.529 28.938 ao_to_mo_and_store_B_mult_1 91 7.0 10.608 28.732 10.608 28.732 mp2_ri_gpw_compute_en 1 5.0 0.062 0.081 26.389 28.439 grid_integrate_task_list 95 9.0 23.346 28.411 23.346 28.411 fft_wrap_pw1pw2_100 730 12.4 1.291 1.427 26.290 26.692 mp2_ri_gpw_compute_en_RI_loop 1 6.0 1.845 1.898 24.743 24.754 get_2c_integrals 1 6.0 0.000 0.000 20.054 20.074 compute_2c_integrals 1 7.0 0.003 0.004 19.036 19.044 compute_2c_integrals_loop_lm 1 8.0 0.003 0.008 18.662 18.917 mp2_eri_2c_integrate_gpw 1 9.0 1.729 1.850 18.659 18.915 fft3d_s 1823 13.4 18.503 18.818 18.517 18.832 scf_env_do_scf_inner_loop 4 4.0 0.000 0.001 17.856 17.856 calculate_wavefunction 91 8.0 2.008 2.038 9.617 9.845 mp2_ri_gpw_compute_en_expansio 172 7.0 0.556 0.581 8.732 9.166 potential_pw2rs 186 10.0 0.033 0.035 8.392 8.925 local_gemm 172 8.0 8.176 8.608 8.176 8.608 mp2_eri_2c_integrate_gpw_pot_l 91 10.0 0.001 0.001 8.163 8.421 mp2_ri_gpw_compute_en_comm 22 7.0 0.499 0.519 7.775 8.229 calc_potential_gpw 182 9.5 0.002 0.002 7.945 8.172 collocate_single_gaussian 91 10.0 0.016 0.023 7.773 7.957 integrate_four_center_load 4 10.0 0.000 0.000 6.769 6.773 hfx_load_balance 1 11.0 0.000 0.000 6.769 6.773 mp2_ri_gpw_compute_en_ener 172 7.0 6.346 6.410 6.346 6.410 mp_sendrecv_dm3 2068 8.0 5.808 6.252 5.808 6.252 mp_sync 37 10.5 3.467 5.717 3.467 5.717 pw_gather_s 912 13.2 4.452 4.894 4.452 4.894 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="20", plot="h2o_32_ri_rpa_mp2", label="RI-MP2 (8n/6r/2t)", y=100.451350, yerr=0.000000 PlotPoint: name="21", plot="h2o_32_ri_rpa_mp2_mem", label="RI-MP2 (8n/6r/2t)", y=1500.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 452.481024E+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 62379. MP_Allreduce 10329 270. MP_Sync 530 MP_Alltoall 2083 592243. 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 0.016 0.048 58.856 58.857 qs_mol_dyn_low 1 2.0 0.003 0.006 58.472 58.479 qs_forces 11 3.9 0.002 0.003 58.410 58.411 qs_energies 11 4.9 0.002 0.003 56.905 56.914 scf_env_do_scf 11 5.9 0.000 0.001 45.231 45.231 scf_env_do_scf_inner_loop 108 6.5 0.003 0.011 43.158 43.158 dbcsr_multiply_generic 2286 12.5 0.094 0.099 38.692 39.154 qs_scf_new_mos 108 7.5 0.000 0.001 33.276 33.545 qs_scf_loop_do_ot 108 8.5 0.000 0.001 33.275 33.545 multiply_cannon 2286 13.5 0.189 0.202 31.737 33.359 multiply_cannon_loop 2286 14.5 1.862 1.971 31.097 32.728 ot_scf_mini 108 9.5 0.002 0.002 31.624 31.809 velocity_verlet 10 3.0 0.001 0.002 26.697 26.697 ot_mini 108 10.5 0.001 0.001 18.891 19.127 mp_waitall_1 245248 16.5 10.571 16.942 10.571 16.942 qs_ot_get_derivative 108 11.5 0.001 0.001 15.860 16.022 multiply_cannon_metrocomm3 54864 15.5 0.074 0.080 7.803 15.401 multiply_cannon_multrec 54864 15.5 3.614 5.653 7.579 11.125 init_scf_run 11 5.9 0.000 0.001 10.375 10.375 scf_env_initial_rho_setup 11 6.9 0.001 0.001 10.374 10.375 qs_ot_get_p 119 10.4 0.001 0.002 8.132 8.437 calculate_first_density_matrix 1 7.0 0.000 0.000 8.109 8.116 rebuild_ks_matrix 119 8.3 0.000 0.000 7.783 7.935 qs_ks_build_kohn_sham_matrix 119 9.3 0.010 0.012 7.782 7.934 acc_transpose_blocks 54864 15.5 0.232 0.262 5.595 7.424 qs_ks_update_qs_env 119 7.6 0.001 0.001 6.836 6.970 mp_sum_l 7287 12.8 5.101 6.761 5.101 6.761 multiply_cannon_sync_h2d 54864 15.5 5.148 6.294 5.148 6.294 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 5.826 6.266 calculate_dm_sparse 119 9.5 0.000 0.000 5.715 5.843 acc_transpose_blocks_kernels 54864 16.5 0.251 0.388 4.088 5.721 dbcsr_mm_accdrv_process 76910 16.1 1.856 2.927 3.877 5.716 jit_kernel_transpose 5 15.5 3.837 5.391 3.837 5.391 qs_ot_p2m_diag 50 11.0 0.004 0.006 5.156 5.197 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 5.103 5.106 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 4.803 4.865 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 4.717 4.828 multiply_cannon_metrocomm1 54864 15.5 0.059 0.064 2.160 4.739 sum_up_and_integrate 119 10.3 0.012 0.014 4.377 4.383 integrate_v_rspace 119 11.3 0.002 0.003 4.365 4.372 cp_dbcsr_syevd 50 12.0 0.003 0.003 4.371 4.372 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.001 4.223 4.250 qs_rho_update_rho_low 119 7.7 0.000 0.001 4.080 4.185 calculate_rho_elec 119 8.7 0.011 0.016 4.079 4.184 cp_fm_diag_elpa 50 13.0 0.000 0.000 4.182 4.182 cp_fm_redistribute_end 50 14.0 2.135 4.159 2.140 4.161 cp_fm_diag_elpa_base 50 14.0 2.016 4.064 2.020 4.072 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 2.925 3.120 apply_single 119 13.6 0.000 0.000 2.924 3.120 ot_diis_step 108 11.5 0.006 0.006 2.722 2.722 jit_kernel_multiply 13 15.8 1.955 2.710 1.955 2.710 qs_ot_get_orbitals 108 10.5 0.000 0.000 2.403 2.465 wfi_extrapolate 11 7.9 0.001 0.001 2.195 2.195 density_rs2pw 119 9.7 0.004 0.004 2.073 2.187 grid_integrate_task_list 119 12.3 2.016 2.095 2.016 2.095 mp_sum_d 4135 12.0 1.404 2.072 1.404 2.072 init_scf_loop 11 6.9 0.000 0.000 2.056 2.056 potential_pw2rs 119 12.3 0.004 0.004 1.759 1.767 pw_transfer 1439 11.6 0.051 0.056 1.595 1.660 make_m2s 4572 13.5 0.054 0.057 1.558 1.594 fft_wrap_pw1pw2 1201 12.6 0.007 0.007 1.520 1.587 mp_alltoall_d11v 2130 13.8 1.316 1.538 1.316 1.538 make_images 4572 14.5 0.133 0.138 1.475 1.512 mp_waitany 12084 13.8 1.256 1.492 1.256 1.492 acc_transpose_blocks_sync 164592 16.5 1.228 1.472 1.228 1.472 transfer_rs2pw 487 10.6 0.005 0.006 1.340 1.463 grid_collocate_task_list 119 9.7 1.344 1.417 1.344 1.417 transfer_pw2rs 487 13.2 0.006 0.007 1.293 1.300 fft3d_ps 1201 14.6 0.368 0.473 1.234 1.300 fft_wrap_pw1pw2_140 487 13.2 0.141 0.154 1.180 1.248 dbcsr_dot_sd 1205 11.9 0.048 0.059 0.784 1.215 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="100", plot="h2o_64_md", label="(8n/12r/1t)", y=58.857000, 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/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 488.042496E+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 62658. MP_Allreduce 10306 303. MP_Sync 54 MP_Alltoall 2060 1236338. 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 0.016 0.031 47.801 47.803 qs_mol_dyn_low 1 2.0 0.003 0.003 47.568 47.576 qs_forces 11 3.9 0.006 0.030 46.722 46.723 qs_energies 11 4.9 0.003 0.011 44.994 44.998 scf_env_do_scf 11 5.9 0.001 0.001 32.799 32.800 scf_env_do_scf_inner_loop 108 6.5 0.003 0.006 30.244 30.245 dbcsr_multiply_generic 2286 12.5 0.101 0.103 29.259 29.666 multiply_cannon 2286 13.5 0.210 0.219 23.932 25.434 multiply_cannon_loop 2286 14.5 1.204 1.273 22.714 24.282 qs_scf_new_mos 108 7.5 0.001 0.001 21.352 21.584 qs_scf_loop_do_ot 108 8.5 0.001 0.001 21.352 21.583 ot_scf_mini 108 9.5 0.003 0.003 20.409 20.579 velocity_verlet 10 3.0 0.001 0.002 19.596 19.613 mp_waitall_1 200699 16.5 9.477 14.996 9.477 14.996 multiply_cannon_metrocomm3 27432 15.5 0.071 0.075 7.473 13.287 ot_mini 108 10.5 0.001 0.001 12.470 12.703 init_scf_run 11 5.9 0.000 0.001 10.873 10.873 scf_env_initial_rho_setup 11 6.9 0.001 0.001 10.873 10.873 qs_ot_get_derivative 108 11.5 0.001 0.001 10.086 10.254 multiply_cannon_multrec 27432 15.5 1.807 4.169 6.385 9.324 calculate_first_density_matrix 1 7.0 0.001 0.002 9.236 9.273 rebuild_ks_matrix 119 8.3 0.000 0.000 6.945 7.113 qs_ks_build_kohn_sham_matrix 119 9.3 0.016 0.048 6.944 7.112 dbcsr_mm_accdrv_process 47894 16.0 3.661 6.196 4.497 6.736 qs_ks_update_qs_env 119 7.6 0.001 0.001 6.098 6.253 calculate_dm_sparse 119 9.5 0.000 0.001 5.584 5.699 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 5.274 5.277 qs_ot_get_p 119 10.4 0.001 0.001 4.975 5.193 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 5.022 5.063 acc_transpose_blocks 27432 15.5 0.114 0.120 4.561 4.915 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 4.759 4.806 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 3.872 4.725 mp_sum_l 7287 12.8 2.126 4.232 2.126 4.232 qs_rho_update_rho_low 119 7.7 0.001 0.001 3.838 4.174 calculate_rho_elec 119 8.7 0.021 0.024 3.837 4.174 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 3.027 4.095 apply_single 119 13.6 0.000 0.000 3.027 4.095 sum_up_and_integrate 119 10.3 0.025 0.028 3.835 3.844 integrate_v_rspace 119 11.3 0.002 0.003 3.811 3.820 acc_transpose_blocks_kernels 27432 16.5 0.190 0.279 3.597 3.814 jit_kernel_transpose 5 15.5 3.407 3.537 3.407 3.537 multiply_cannon_metrocomm1 27432 15.5 0.033 0.036 0.671 3.438 qs_ot_p2m_diag 50 11.0 0.009 0.014 3.210 3.230 make_m2s 4572 13.5 0.052 0.054 2.757 3.097 make_images 4572 14.5 0.206 0.244 2.669 3.013 cp_dbcsr_syevd 50 12.0 0.003 0.003 2.794 2.794 init_scf_loop 11 6.9 0.010 0.038 2.533 2.539 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 2.398 2.485 cp_fm_diag_elpa 50 13.0 0.000 0.000 2.428 2.429 cp_fm_redistribute_end 50 14.0 1.228 2.395 1.232 2.397 cp_fm_diag_elpa_base 50 14.0 1.131 2.290 1.160 2.336 ot_diis_step 108 11.5 0.011 0.011 2.334 2.335 jit_kernel_multiply 9 16.2 0.777 2.308 0.777 2.308 multiply_cannon_sync_h2d 27432 15.5 1.675 2.203 1.675 2.203 density_rs2pw 119 9.7 0.004 0.004 2.087 2.157 pw_transfer 1439 11.6 0.064 0.068 1.951 1.996 grid_integrate_task_list 119 12.3 1.837 1.933 1.837 1.933 fft_wrap_pw1pw2 1201 12.6 0.008 0.008 1.860 1.906 mp_alltoall_d11v 2130 13.8 1.329 1.682 1.329 1.682 prepare_preconditioner 11 7.9 0.000 0.001 1.606 1.645 make_preconditioner 11 8.9 0.000 0.001 1.605 1.645 make_images_data 4572 15.5 0.047 0.054 1.284 1.628 potential_pw2rs 119 12.3 0.006 0.006 1.559 1.574 make_full_inverse_cholesky 11 9.9 0.000 0.001 1.503 1.571 fft3d_ps 1201 14.6 0.519 0.577 1.496 1.534 wfi_extrapolate 11 7.9 0.001 0.001 1.527 1.527 hybrid_alltoall_any 4725 16.4 0.054 0.115 1.142 1.517 fft_wrap_pw1pw2_140 487 13.2 0.159 0.169 1.442 1.487 grid_collocate_task_list 119 9.7 1.291 1.346 1.291 1.346 transfer_rs2pw 487 10.6 0.005 0.005 1.233 1.310 qs_ot_get_orbitals 108 10.5 0.000 0.000 1.241 1.291 mp_sum_d 4135 12.0 0.600 1.089 0.600 1.089 mp_allgather_i34 2286 14.5 0.639 1.067 0.639 1.067 transfer_pw2rs 487 13.2 0.005 0.005 0.985 0.993 qs_energies_init_hamiltonians 11 5.9 0.001 0.003 0.988 0.990 acc_transpose_blocks_sync 82296 16.5 0.825 0.956 0.825 0.956 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="102", plot="h2o_64_md", label="(8n/6r/2t)", y=47.803000, yerr=0.000000 PlotPoint: name="103", plot="h2o_64_md_mem", label="(8n/6r/2t)", y=464.727273, yerr=1.710444 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 520.757248E+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 3622 63490. MP_Allreduce 10155 305. MP_Sync 54 MP_Alltoall 1821 1607811. MP_SendRecv 11067 57667. MP_ISendRecv 11067 57667. MP_Wait 21987 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 0.019 0.029 38.401 38.402 qs_mol_dyn_low 1 2.0 0.003 0.003 36.788 38.136 qs_forces 11 3.9 0.002 0.004 36.689 36.689 qs_energies 11 4.9 0.004 0.011 35.131 35.134 scf_env_do_scf 11 5.9 0.001 0.001 25.660 25.660 scf_env_do_scf_inner_loop 108 6.5 0.004 0.008 23.242 23.243 dbcsr_multiply_generic 2286 12.5 0.095 0.097 21.565 21.645 multiply_cannon 2286 13.5 0.194 0.199 18.242 19.025 multiply_cannon_loop 2286 14.5 0.858 0.904 17.463 18.289 qs_scf_new_mos 108 7.5 0.001 0.001 15.574 15.590 qs_scf_loop_do_ot 108 8.5 0.001 0.001 15.574 15.589 velocity_verlet 10 3.0 0.002 0.002 15.050 15.052 ot_scf_mini 108 9.5 0.003 0.003 14.817 14.831 ot_mini 108 10.5 0.001 0.001 9.025 9.042 init_scf_run 11 5.9 0.000 0.001 8.208 8.208 scf_env_initial_rho_setup 11 6.9 0.001 0.001 8.208 8.208 qs_ot_get_derivative 108 11.5 0.001 0.001 7.545 7.560 multiply_cannon_multrec 18288 15.5 1.883 2.904 7.186 7.537 calculate_first_density_matrix 1 7.0 0.001 0.002 6.981 6.982 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 6.404 6.406 acc_transpose_blocks 18288 15.5 0.079 0.081 4.819 6.265 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 6.206 6.217 dbcsr_mm_accdrv_process 38222 16.0 5.117 5.950 5.208 6.015 rebuild_ks_matrix 119 8.3 0.000 0.000 5.994 6.011 qs_ks_build_kohn_sham_matrix 119 9.3 0.013 0.016 5.993 6.011 acc_transpose_blocks_kernels 18288 16.5 0.220 0.227 3.985 5.403 qs_ks_update_qs_env 119 7.6 0.001 0.001 5.280 5.298 jit_kernel_transpose 5 15.6 3.765 5.185 3.765 5.185 mp_waitall_1 158411 16.6 3.769 4.791 3.769 4.791 qs_ot_get_p 119 10.4 0.001 0.002 3.562 3.579 sum_up_and_integrate 119 10.3 0.030 0.031 3.571 3.578 integrate_v_rspace 119 11.3 0.003 0.003 3.540 3.551 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 2.820 3.422 multiply_cannon_metrocomm3 18288 15.5 0.047 0.049 2.507 3.362 qs_rho_update_rho_low 119 7.7 0.001 0.001 3.241 3.248 calculate_rho_elec 119 8.7 0.031 0.031 3.241 3.248 qs_ot_p2m_diag 50 11.0 0.050 0.165 2.406 2.413 init_scf_loop 11 6.9 0.001 0.004 2.401 2.402 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 2.005 2.335 apply_single 119 13.6 0.000 0.000 2.005 2.334 cp_dbcsr_syevd 50 12.0 0.003 0.003 2.069 2.108 make_m2s 4572 13.5 0.046 0.061 1.819 1.968 density_rs2pw 119 9.7 0.004 0.004 1.808 1.926 grid_integrate_task_list 119 12.3 1.797 1.897 1.797 1.897 calculate_dm_sparse 119 9.5 0.000 0.000 1.874 1.881 make_images 4572 14.5 0.194 0.219 1.732 1.879 multiply_cannon_metrocomm1 18288 15.5 0.030 0.031 0.527 1.865 pw_transfer 1439 11.6 0.064 0.068 1.828 1.836 fft_wrap_pw1pw2 1201 12.6 0.008 0.008 1.736 1.746 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 1.720 1.728 cp_fm_diag_elpa 50 13.0 0.000 0.000 1.705 1.715 cp_fm_diag_elpa_base 50 14.0 1.682 1.693 1.703 1.713 prepare_preconditioner 11 7.9 0.000 0.000 1.629 1.633 make_preconditioner 11 8.9 0.000 0.001 1.629 1.632 make_full_inverse_cholesky 11 9.9 0.000 0.000 1.491 1.572 mp_sum_l 7287 12.8 1.119 1.499 1.119 1.499 ot_diis_step 108 11.5 0.011 0.012 1.465 1.465 md_output 10 3.0 0.000 0.000 0.047 1.427 fft_wrap_pw1pw2_140 487 13.2 0.212 0.219 1.402 1.412 mp_max_l 33 2.8 1.349 1.393 1.349 1.393 potential_pw2rs 119 12.3 0.007 0.008 1.353 1.360 fft3d_ps 1201 14.6 0.526 0.544 1.303 1.312 grid_collocate_task_list 119 9.7 1.234 1.303 1.234 1.303 multiply_cannon_sync_h2d 18288 15.5 1.042 1.220 1.042 1.220 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 1.214 1.216 wfi_extrapolate 11 7.9 0.001 0.001 1.171 1.171 transfer_rs2pw 487 10.6 0.005 0.005 1.029 1.134 md_write_output 11 3.9 0.002 0.033 0.037 1.131 write_trajectory 44 4.9 0.035 1.092 0.035 1.098 qs_energies_init_hamiltonians 11 5.9 0.001 0.002 1.028 1.029 jit_kernel_multiply 4 15.0 0.032 1.017 0.032 1.017 qs_ot_get_orbitals 108 10.5 0.000 0.000 0.954 0.972 make_images_data 4572 15.5 0.047 0.051 0.762 0.930 hybrid_alltoall_any 4725 16.4 0.058 0.114 0.666 0.856 transfer_pw2rs 487 13.2 0.004 0.004 0.823 0.829 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 0.785 0.787 mp_alltoall_d11v 2130 13.8 0.670 0.779 0.670 0.779 acc_transpose_blocks_sync 54864 16.5 0.735 0.776 0.735 0.776 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="104", plot="h2o_64_md", label="(8n/4r/3t)", y=38.402000, yerr=0.000000 PlotPoint: name="105", plot="h2o_64_md_mem", label="(8n/4r/3t)", y=494.909091, yerr=1.831767 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 555.888640E+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 3622 63489. MP_Allreduce 10154 346. MP_Sync 54 MP_Alltoall 1582 2412273. MP_SendRecv 8211 74133. MP_ISendRecv 8211 74133. MP_Wait 16271 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 0.043 43.018 43.018 qs_mol_dyn_low 1 2.0 0.003 0.003 42.793 42.801 qs_forces 11 3.9 0.002 0.003 42.734 42.735 qs_energies 11 4.9 0.001 0.002 40.933 40.942 scf_env_do_scf 11 5.9 0.000 0.001 32.735 32.737 scf_env_do_scf_inner_loop 108 6.5 0.003 0.007 28.888 28.889 dbcsr_multiply_generic 2286 12.5 0.099 0.103 23.922 24.120 velocity_verlet 10 3.0 0.002 0.002 20.487 20.489 multiply_cannon 2286 13.5 0.218 0.226 19.374 19.873 qs_scf_new_mos 108 7.5 0.001 0.001 19.755 19.816 qs_scf_loop_do_ot 108 8.5 0.001 0.001 19.755 19.815 ot_scf_mini 108 9.5 0.002 0.003 18.678 18.730 multiply_cannon_loop 2286 14.5 1.560 1.641 18.293 18.654 ot_mini 108 10.5 0.001 0.001 11.334 11.395 qs_ot_get_derivative 108 11.5 0.001 0.001 9.418 9.472 multiply_cannon_multrec 27432 15.5 2.464 3.144 9.016 9.305 dbcsr_mm_accdrv_process 47916 15.9 6.077 7.526 6.448 7.809 rebuild_ks_matrix 119 8.3 0.000 0.000 6.919 6.965 qs_ks_build_kohn_sham_matrix 119 9.3 0.012 0.013 6.918 6.965 init_scf_run 11 5.9 0.000 0.001 6.788 6.789 scf_env_initial_rho_setup 11 6.9 0.001 0.001 6.788 6.789 qs_ks_update_qs_env 119 7.6 0.001 0.001 6.130 6.172 acc_transpose_blocks 27432 15.5 0.120 0.125 5.323 5.418 calculate_first_density_matrix 1 7.0 0.000 0.000 5.188 5.189 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 4.820 4.823 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 4.549 4.562 qs_ot_get_p 119 10.4 0.006 0.007 4.419 4.489 sum_up_and_integrate 119 10.3 0.035 0.037 4.018 4.027 qs_rho_update_rho_low 119 7.7 0.001 0.001 3.983 4.025 calculate_rho_elec 119 8.7 0.040 0.046 3.982 4.024 integrate_v_rspace 119 11.3 0.003 0.003 3.983 3.992 init_scf_loop 11 6.9 0.000 0.000 3.815 3.816 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 3.328 3.808 acc_transpose_blocks_kernels 27432 16.5 0.272 0.277 3.713 3.728 jit_kernel_transpose 5 15.6 3.440 3.455 3.440 3.455 mp_waitall_1 137007 16.6 2.576 3.238 2.576 3.238 make_m2s 4572 13.5 0.055 0.057 2.882 3.011 make_images 4572 14.5 0.273 0.334 2.773 2.901 prepare_preconditioner 11 7.9 0.000 0.000 2.884 2.893 make_preconditioner 11 8.9 0.000 0.000 2.884 2.893 make_full_inverse_cholesky 11 9.9 0.000 0.000 2.466 2.807 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 2.227 2.716 apply_single 119 13.6 0.000 0.000 2.226 2.716 qs_ot_p2m_diag 50 11.0 0.015 0.023 2.694 2.706 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 2.610 2.640 density_rs2pw 119 9.7 0.004 0.004 2.398 2.514 calculate_dm_sparse 119 9.5 0.000 0.000 2.204 2.267 cp_dbcsr_syevd 50 12.0 0.003 0.003 2.261 2.262 pw_transfer 1439 11.6 0.064 0.068 2.182 2.225 fft_wrap_pw1pw2 1201 12.6 0.008 0.008 2.091 2.138 cp_fm_diag_elpa 50 13.0 0.000 0.000 1.945 1.956 cp_fm_diag_elpa_base 50 14.0 1.906 1.925 1.943 1.953 grid_integrate_task_list 119 12.3 1.827 1.932 1.827 1.932 mp_sum_l 7287 12.8 1.244 1.896 1.244 1.896 ot_diis_step 108 11.5 0.012 0.012 1.872 1.872 potential_pw2rs 119 12.3 0.009 0.009 1.758 1.773 fft_wrap_pw1pw2_140 487 13.2 0.245 0.258 1.649 1.698 fft3d_ps 1201 14.6 0.554 0.605 1.611 1.643 acc_transpose_blocks_sync 82296 16.5 1.461 1.562 1.461 1.562 wfi_extrapolate 11 7.9 0.001 0.001 1.537 1.538 transfer_rs2pw 487 10.6 0.004 0.005 1.380 1.504 multiply_cannon_metrocomm3 27432 15.5 0.040 0.041 0.858 1.484 qs_ot_get_orbitals 108 10.5 0.000 0.000 1.350 1.367 grid_collocate_task_list 119 9.7 1.248 1.352 1.248 1.352 make_images_data 4572 15.5 0.047 0.051 1.098 1.293 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 1.269 1.284 dbcsr_complete_redistribute 329 12.2 0.146 0.206 0.966 1.244 cp_fm_upper_to_full 72 14.2 0.832 1.194 0.832 1.194 hybrid_alltoall_any 4725 16.4 0.065 0.153 0.914 1.193 mp_alltoall_d11v 2130 13.8 0.979 1.175 0.979 1.175 transfer_pw2rs 487 13.2 0.004 0.004 1.107 1.121 jit_kernel_multiply 6 16.5 0.304 1.102 0.304 1.102 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 1.093 1.095 copy_fm_to_dbcsr 176 11.2 0.001 0.001 0.701 0.977 mp_alltoall_z22v 1201 16.6 0.938 0.973 0.938 0.973 cp_fm_cholesky_invert 11 10.9 0.934 0.938 0.934 0.938 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 0.880 0.885 build_core_hamiltonian_matrix_ 11 4.9 0.000 0.000 0.801 0.875 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="106", plot="h2o_64_md", label="(8n/3r/4t)", y=43.018000, yerr=0.000000 PlotPoint: name="107", plot="h2o_64_md_mem", label="(8n/3r/4t)", y=527.454545, yerr=3.201239 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 623.079424E+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 3622 63488. MP_Allreduce 10154 346. MP_Sync 54 MP_Alltoall 1582 3682667. MP_SendRecv 5355 94533. MP_ISendRecv 5355 94533. MP_Wait 11335 MP_ISend 5200 225425. MP_IRecv 5200 225425. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.015 0.033 33.838 33.838 qs_mol_dyn_low 1 2.0 0.003 0.003 33.657 33.669 qs_forces 11 3.9 0.002 0.003 33.600 33.601 qs_energies 11 4.9 0.001 0.001 31.889 31.892 scf_env_do_scf 11 5.9 0.000 0.001 22.365 22.365 scf_env_do_scf_inner_loop 108 6.5 0.003 0.007 19.831 19.832 dbcsr_multiply_generic 2286 12.5 0.092 0.097 17.872 18.007 multiply_cannon 2286 13.5 0.224 0.235 15.197 15.737 velocity_verlet 10 3.0 0.001 0.002 14.939 14.941 multiply_cannon_loop 2286 14.5 0.648 0.669 14.270 14.548 qs_scf_new_mos 108 7.5 0.001 0.001 12.143 12.173 qs_scf_loop_do_ot 108 8.5 0.001 0.001 12.143 12.172 ot_scf_mini 108 9.5 0.002 0.002 11.428 11.452 init_scf_run 11 5.9 0.000 0.001 8.039 8.039 scf_env_initial_rho_setup 11 6.9 0.001 0.001 8.038 8.038 multiply_cannon_multrec 9144 15.5 1.727 1.956 6.464 7.918 calculate_first_density_matrix 1 7.0 0.000 0.000 6.842 6.844 ot_mini 108 10.5 0.001 0.001 6.623 6.652 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 6.157 6.158 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 5.974 5.980 dbcsr_mm_accdrv_process 12550 15.8 3.785 5.933 4.628 5.975 rebuild_ks_matrix 119 8.3 0.000 0.000 5.824 5.847 qs_ks_build_kohn_sham_matrix 119 9.3 0.012 0.013 5.824 5.847 qs_ot_get_derivative 108 11.5 0.001 0.001 5.329 5.352 qs_ks_update_qs_env 119 7.6 0.001 0.001 5.181 5.203 acc_transpose_blocks 9144 15.5 0.042 0.043 4.259 4.292 mp_waitall_1 115863 16.7 2.827 3.534 2.827 3.534 sum_up_and_integrate 119 10.3 0.037 0.041 3.497 3.502 acc_transpose_blocks_kernels 9144 16.5 0.118 0.122 3.474 3.476 integrate_v_rspace 119 11.3 0.003 0.003 3.459 3.465 qs_rho_update_rho_low 119 7.7 0.001 0.001 3.353 3.363 calculate_rho_elec 119 8.7 0.060 0.061 3.353 3.362 jit_kernel_transpose 5 15.6 3.356 3.359 3.356 3.359 qs_ot_get_p 119 10.4 0.001 0.001 2.908 2.943 init_scf_loop 11 6.9 0.000 0.000 2.514 2.515 jit_kernel_multiply 6 15.3 0.803 2.490 0.803 2.490 pw_transfer 1439 11.6 0.065 0.068 2.002 2.013 multiply_cannon_metrocomm1 9144 15.5 0.023 0.024 1.070 1.999 make_m2s 4572 13.5 0.035 0.036 1.822 1.979 fft_wrap_pw1pw2 1201 12.6 0.008 0.008 1.910 1.921 grid_integrate_task_list 119 12.3 1.867 1.916 1.867 1.916 qs_ot_p2m_diag 50 11.0 0.022 0.023 1.887 1.890 make_images 4572 14.5 0.269 0.301 1.733 1.887 density_rs2pw 119 9.7 0.003 0.004 1.740 1.862 multiply_cannon_metrocomm3 9144 15.5 0.020 0.021 0.969 1.846 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 1.754 1.784 prepare_preconditioner 11 7.9 0.000 0.000 1.772 1.777 make_preconditioner 11 8.9 0.000 0.000 1.772 1.777 calculate_dm_sparse 119 9.5 0.000 0.000 1.748 1.767 make_full_inverse_cholesky 11 9.9 0.000 0.000 1.658 1.685 cp_dbcsr_syevd 50 12.0 0.003 0.003 1.640 1.641 fft_wrap_pw1pw2_140 487 13.2 0.323 0.333 1.555 1.567 grid_collocate_task_list 119 9.7 1.295 1.406 1.295 1.406 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 1.360 1.373 cp_fm_diag_elpa 50 13.0 0.000 0.000 1.352 1.360 cp_fm_diag_elpa_base 50 14.0 1.325 1.343 1.350 1.359 fft3d_ps 1201 14.6 0.558 0.568 1.324 1.334 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 1.235 1.283 apply_single 119 13.6 0.000 0.000 1.235 1.283 ot_diis_step 108 11.5 0.013 0.013 1.279 1.280 potential_pw2rs 119 12.3 0.010 0.011 1.272 1.275 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 1.264 1.265 wfi_extrapolate 11 7.9 0.001 0.001 1.144 1.144 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 1.133 1.140 hybrid_alltoall_any 4725 16.4 0.064 0.175 0.774 0.988 make_images_data 4572 15.5 0.042 0.045 0.781 0.949 build_core_hamiltonian_matrix_ 11 4.9 0.000 0.000 0.876 0.924 transfer_rs2pw 487 10.6 0.004 0.005 0.821 0.909 mp_alltoall_d11v 2130 13.8 0.814 0.896 0.814 0.896 cp_fm_cholesky_invert 11 10.9 0.881 0.884 0.881 0.884 qs_ot_get_orbitals 108 10.5 0.000 0.000 0.834 0.844 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 0.766 0.768 acc_transpose_blocks_sync 27432 16.5 0.728 0.760 0.728 0.760 qs_env_update_s_mstruct 11 6.9 0.000 0.001 0.708 0.757 transfer_pw2rs 487 13.2 0.003 0.004 0.689 0.691 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="108", plot="h2o_64_md", label="(8n/2r/6t)", y=33.838000, yerr=0.000000 PlotPoint: name="109", plot="h2o_64_md_mem", label="(8n/2r/6t)", y=591.000000, yerr=6.660603 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 777.371648E+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 63723. MP_Allreduce 10154 429. 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 0.033 46.476 46.476 qs_mol_dyn_low 1 2.0 0.003 0.003 46.264 46.272 qs_forces 11 3.9 0.002 0.002 46.204 46.205 qs_energies 11 4.9 0.001 0.001 44.168 44.172 scf_env_do_scf 11 5.9 0.001 0.001 35.630 35.630 scf_env_do_scf_inner_loop 108 6.5 0.005 0.010 27.617 27.619 velocity_verlet 10 3.0 0.002 0.002 24.527 24.532 dbcsr_multiply_generic 2286 12.5 0.101 0.104 22.239 22.404 multiply_cannon 2286 13.5 0.297 0.309 18.113 19.055 qs_scf_new_mos 108 7.5 0.001 0.001 18.022 18.128 qs_scf_loop_do_ot 108 8.5 0.001 0.001 18.021 18.127 multiply_cannon_loop 2286 14.5 0.871 0.892 16.815 17.696 ot_scf_mini 108 9.5 0.002 0.002 16.836 16.942 ot_mini 108 10.5 0.001 0.001 10.220 10.342 multiply_cannon_multrec 9144 15.5 3.436 4.807 8.840 8.961 qs_ot_get_derivative 108 11.5 0.001 0.001 8.123 8.230 init_scf_loop 11 6.9 0.000 0.000 7.983 7.987 rebuild_ks_matrix 119 8.3 0.000 0.000 7.277 7.428 qs_ks_build_kohn_sham_matrix 119 9.3 0.013 0.013 7.276 7.428 prepare_preconditioner 11 7.9 0.000 0.000 6.936 6.951 make_preconditioner 11 8.9 0.000 0.000 6.936 6.951 make_full_inverse_cholesky 11 9.9 0.000 0.000 5.462 6.829 qs_ks_update_qs_env 119 7.6 0.001 0.001 6.547 6.682 dbcsr_mm_accdrv_process 12550 15.8 4.775 6.563 5.267 6.608 init_scf_run 11 5.9 0.000 0.001 6.462 6.463 scf_env_initial_rho_setup 11 6.9 0.001 0.001 6.462 6.462 calculate_first_density_matrix 1 7.0 0.000 0.000 4.928 4.930 acc_transpose_blocks 9144 15.5 0.044 0.045 4.657 4.705 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 4.703 4.704 cp_fm_upper_to_full 72 14.2 3.213 4.630 3.213 4.630 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 4.422 4.473 qs_rho_update_rho_low 119 7.7 0.001 0.001 4.097 4.101 calculate_rho_elec 119 8.7 0.118 0.121 4.097 4.100 sum_up_and_integrate 119 10.3 0.065 0.066 3.930 3.936 integrate_v_rspace 119 11.3 0.004 0.004 3.866 3.872 qs_ot_get_p 119 10.4 0.009 0.016 3.668 3.810 mp_waitall_1 94719 16.7 2.515 3.561 2.515 3.561 acc_transpose_blocks_kernels 9144 16.5 0.119 0.121 3.491 3.495 jit_kernel_transpose 5 15.6 3.372 3.375 3.372 3.375 qs_ot_get_derivative_taylor 59 13.0 0.001 0.001 2.886 3.297 dbcsr_complete_redistribute 329 12.2 0.284 0.291 2.059 2.917 make_m2s 4572 13.5 0.038 0.039 2.473 2.640 pw_transfer 1439 11.6 0.068 0.069 2.633 2.639 copy_fm_to_dbcsr 176 11.2 0.001 0.001 1.730 2.591 fft_wrap_pw1pw2 1201 12.6 0.009 0.009 2.535 2.541 make_images 4572 14.5 0.354 0.386 2.352 2.520 apply_preconditioner_dbcsr 119 12.6 0.000 0.000 2.234 2.503 apply_single 119 13.6 0.000 0.000 2.234 2.502 multiply_cannon_metrocomm3 9144 15.5 0.021 0.021 1.475 2.431 mp_alltoall_i22 627 13.8 1.460 2.368 1.460 2.368 transfer_fm_to_dbcsr 11 9.9 0.000 0.000 1.469 2.323 calculate_dm_sparse 119 9.5 0.000 0.000 2.263 2.280 qs_ot_get_derivative_diag 49 12.0 0.001 0.001 2.117 2.171 grid_integrate_task_list 119 12.3 2.093 2.157 2.093 2.157 qs_ot_p2m_diag 50 11.0 0.043 0.044 2.146 2.147 fft_wrap_pw1pw2_140 487 13.2 0.575 0.579 2.132 2.140 density_rs2pw 119 9.7 0.003 0.004 2.094 2.127 ot_diis_step 108 11.5 0.014 0.015 2.066 2.066 mp_sum_l 7287 12.8 1.111 1.837 1.111 1.837 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 1.777 1.778 cp_dbcsr_syevd 50 12.0 0.003 0.003 1.772 1.773 fft3d_ps 1201 14.6 0.594 0.605 1.631 1.636 grid_collocate_task_list 119 9.7 1.532 1.561 1.532 1.561 wfi_extrapolate 11 7.9 0.001 0.001 1.469 1.469 cp_fm_diag_elpa 50 13.0 0.000 0.000 1.465 1.465 cp_fm_cholesky_invert 11 10.9 1.461 1.464 1.461 1.464 cp_fm_diag_elpa_base 50 14.0 1.317 1.369 1.463 1.463 qs_ot_get_orbitals 108 10.5 0.000 0.000 1.355 1.379 potential_pw2rs 119 12.3 0.014 0.014 1.340 1.343 hybrid_alltoall_any 4725 16.4 0.090 0.151 1.106 1.316 mp_alltoall_d11v 2130 13.8 1.241 1.278 1.241 1.278 make_images_data 4572 15.5 0.046 0.049 1.058 1.273 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.000 0.000 1.249 1.255 acc_transpose_blocks_sync 27432 16.5 1.106 1.151 1.106 1.151 qs_env_update_s_mstruct 11 6.9 0.001 0.001 1.106 1.123 build_core_hamiltonian_matrix_ 11 4.9 0.000 0.001 0.999 1.040 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 0.943 0.959 qs_create_task_list 11 7.9 0.000 0.001 0.934 0.947 generate_qs_task_list 11 8.9 0.368 0.387 0.933 0.947 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="110", plot="h2o_64_md", label="(8n/1r/12t)", y=46.476000, yerr=0.000000 PlotPoint: name="111", plot="h2o_64_md_mem", label="(8n/1r/12t)", y=731.818182, yerr=12.967219 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/09/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 32 x 32 x 32 184415158272 0.0% 0.0% 100.0% flops 9 x 9 x 32 269180485632 0.0% 0.0% 100.0% flops 9 x 22 x 32 349395425280 0.0% 0.0% 100.0% flops 22 x 9 x 32 350042406912 0.0% 0.0% 100.0% flops 22 x 22 x 32 453581815808 0.0% 0.0% 100.0% flops 32 x 32 x 9 465064427520 0.0% 0.0% 100.0% flops 32 x 32 x 22 568412078080 0.0% 0.0% 100.0% flops 9 x 32 x 32 572195340288 0.0% 0.0% 100.0% flops 22 x 32 x 32 699349860352 0.0% 0.0% 100.0% flops 9 x 32 x 9 1735942275072 0.0% 0.0% 100.0% flops 22 x 32 x 9 2216407818240 0.0% 0.0% 100.0% flops 9 x 32 x 22 2216407818240 0.0% 0.0% 100.0% flops 22 x 32 x 22 2803661053952 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 12.884056E+12 0.0% 0.0% 100.0% flops max/rank 198.287135E+09 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 984178160 0.0% 0.0% 100.0% number of processed stacks 8410880 0.0% 0.0% 100.0% average stack size 0.0 0.0 117.0 marketing flops 15.646302E+12 ------------------------------------------------------------------------------- # multiplications 2055 max memory usage/rank 501.346304E+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 3473 66212. MP_Allreduce 9776 488. MP_Sync 52 MP_Alltoall 1938 1383689. MP_SendRecv 20900 9096. MP_ISendRecv 20900 9096. MP_Wait 37268 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 0.015 0.029 84.897 84.898 qs_mol_dyn_low 1 2.0 0.003 0.003 84.628 84.637 qs_forces 11 3.9 0.003 0.004 84.562 84.562 qs_energies 11 4.9 0.001 0.002 81.672 81.685 scf_env_do_scf 11 5.9 0.000 0.001 69.953 69.955 scf_env_do_scf_inner_loop 99 6.5 0.002 0.007 64.456 64.457 dbcsr_multiply_generic 2055 12.4 0.109 0.112 53.640 53.891 qs_scf_new_mos 99 7.5 0.000 0.001 47.261 47.407 qs_scf_loop_do_ot 99 8.5 0.000 0.001 47.260 47.406 multiply_cannon 2055 13.4 0.182 0.188 45.375 46.127 multiply_cannon_loop 2055 14.4 1.819 1.859 44.353 45.091 ot_scf_mini 99 9.5 0.002 0.002 44.868 44.963 velocity_verlet 10 3.0 0.002 0.004 43.381 43.381 ot_mini 99 10.5 0.001 0.001 26.421 26.517 qs_ot_get_derivative 99 11.5 0.001 0.001 19.602 19.710 multiply_cannon_multrec 49320 15.4 11.379 11.961 17.392 18.112 rebuild_ks_matrix 110 8.3 0.000 0.000 14.589 14.771 qs_ks_build_kohn_sham_matrix 110 9.3 0.011 0.012 14.588 14.770 qs_ks_update_qs_env 110 7.6 0.001 0.001 12.783 12.941 mp_waitall_1 220248 16.4 10.651 11.740 10.651 11.740 multiply_cannon_sync_h2d 49320 15.4 9.634 10.194 9.634 10.194 qs_ot_get_p 110 10.4 0.001 0.001 9.941 10.041 init_scf_run 11 5.9 0.000 0.001 9.670 9.670 scf_env_initial_rho_setup 11 6.9 0.001 0.001 9.670 9.670 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 7.495 8.044 multiply_cannon_metrocomm3 49320 15.4 0.084 0.088 6.373 7.750 apply_preconditioner_dbcsr 110 12.6 0.000 0.000 7.228 7.729 apply_single 110 13.6 0.000 0.001 7.228 7.729 sum_up_and_integrate 110 10.3 0.036 0.042 7.079 7.092 integrate_v_rspace 110 11.3 0.003 0.004 7.043 7.064 qs_ot_p2m_diag 48 11.0 0.012 0.018 6.662 6.676 ot_diis_step 99 11.5 0.006 0.006 6.607 6.607 qs_rho_update_rho_low 110 7.6 0.000 0.001 6.441 6.578 calculate_rho_elec 110 8.6 0.020 0.024 6.441 6.578 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 6.394 6.399 dbcsr_mm_accdrv_process 87628 16.1 3.057 3.133 5.882 6.124 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 6.016 6.083 cp_dbcsr_syevd 48 12.0 0.002 0.003 5.846 5.846 acc_transpose_blocks 49320 15.4 0.217 0.224 5.440 5.513 calculate_first_density_matrix 1 7.0 0.000 0.000 5.482 5.486 init_scf_loop 11 6.9 0.000 0.000 5.468 5.469 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 5.297 5.350 cp_fm_diag_elpa 48 13.0 0.000 0.000 5.253 5.279 cp_fm_diag_elpa_base 48 14.0 5.238 5.263 5.251 5.277 mp_sum_l 6594 12.7 3.872 4.527 3.872 4.527 wfi_extrapolate 11 7.9 0.001 0.001 4.089 4.089 make_m2s 4110 13.4 0.061 0.065 3.931 4.066 make_images 4110 14.4 0.176 0.189 3.835 3.972 acc_transpose_blocks_kernels 49320 16.4 0.427 0.440 3.878 3.909 calculate_dm_sparse 110 9.5 0.001 0.001 3.781 3.900 density_rs2pw 110 9.6 0.004 0.005 3.458 3.627 jit_kernel_transpose 5 15.6 3.451 3.477 3.451 3.477 grid_integrate_task_list 110 12.3 3.247 3.446 3.247 3.446 prepare_preconditioner 11 7.9 0.000 0.000 3.296 3.329 make_preconditioner 11 8.9 0.000 0.000 3.296 3.329 qs_ot_get_orbitals 99 10.5 0.000 0.001 3.177 3.220 multiply_cannon_metrocomm1 49320 15.4 0.068 0.072 2.362 3.207 pw_transfer 1331 11.6 0.055 0.064 3.119 3.198 make_full_inverse_cholesky 11 9.9 0.000 0.000 3.091 3.148 fft_wrap_pw1pw2 1111 12.6 0.008 0.008 3.030 3.111 potential_pw2rs 110 12.3 0.006 0.007 2.659 2.693 fft_wrap_pw1pw2_140 451 13.1 0.389 0.430 2.576 2.666 jit_kernel_multiply 13 15.9 2.543 2.572 2.543 2.572 mp_alltoall_d11v 2046 13.8 2.105 2.540 2.105 2.540 fft3d_ps 1111 14.6 0.790 0.880 2.327 2.386 grid_collocate_task_list 110 9.6 2.159 2.260 2.159 2.260 transfer_rs2pw 451 10.6 0.005 0.006 1.979 2.131 mp_waitany 14300 13.8 1.815 2.065 1.815 2.065 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.936 1.964 make_images_data 4110 15.4 0.043 0.047 1.772 1.902 transfer_pw2rs 451 13.1 0.006 0.007 1.817 1.839 mp_sum_d 3889 11.9 1.366 1.825 1.366 1.825 cp_fm_cholesky_invert 11 10.9 1.803 1.807 1.803 1.807 hybrid_alltoall_any 4261 16.3 0.085 0.486 1.543 1.775 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="200", plot="h2o_128_md", label="(8n/12r/1t)", y=84.898000, yerr=0.000000 PlotPoint: name="201", plot="h2o_128_md_mem", label="(8n/12r/1t)", y=475.909091, yerr=2.712079 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 591.994880E+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 3473 66430. MP_Allreduce 9775 566. MP_Sync 52 MP_Alltoall 1717 2427495. MP_SendRecv 10340 26400. MP_ISendRecv 10340 26400. MP_Wait 22352 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 0.014 0.030 73.213 73.214 qs_mol_dyn_low 1 2.0 0.003 0.003 72.973 72.982 qs_forces 11 3.9 0.003 0.003 71.494 71.495 qs_energies 11 4.9 0.002 0.002 68.157 68.161 scf_env_do_scf 11 5.9 0.000 0.001 56.810 56.813 scf_env_do_scf_inner_loop 99 6.5 0.002 0.007 49.244 49.245 dbcsr_multiply_generic 2055 12.4 0.114 0.118 40.838 41.004 velocity_verlet 10 3.0 0.001 0.002 37.474 37.505 multiply_cannon 2055 13.4 0.225 0.244 34.201 35.364 multiply_cannon_loop 2055 14.4 1.172 1.200 32.931 33.907 qs_scf_new_mos 99 7.5 0.001 0.001 33.332 33.471 qs_scf_loop_do_ot 99 8.5 0.001 0.001 33.332 33.470 ot_scf_mini 99 9.5 0.003 0.003 31.665 31.788 ot_mini 99 10.5 0.001 0.001 18.436 18.573 multiply_cannon_multrec 24660 15.4 7.001 8.319 14.113 15.349 rebuild_ks_matrix 110 8.3 0.000 0.000 13.578 13.704 qs_ks_build_kohn_sham_matrix 110 9.3 0.012 0.014 13.578 13.704 qs_ot_get_derivative 99 11.5 0.001 0.001 12.670 12.795 qs_ks_update_qs_env 110 7.6 0.001 0.001 11.947 12.056 mp_waitall_1 176588 16.5 7.512 10.032 7.512 10.032 init_scf_run 11 5.9 0.000 0.001 8.992 8.992 scf_env_initial_rho_setup 11 6.9 0.001 0.001 8.992 8.992 multiply_cannon_metrocomm3 24660 15.4 0.071 0.073 5.111 7.774 init_scf_loop 11 6.9 0.000 0.000 7.529 7.530 multiply_cannon_sync_h2d 24660 15.4 6.332 7.438 6.332 7.438 dbcsr_mm_accdrv_process 52282 16.1 5.557 6.185 6.941 7.243 apply_preconditioner_dbcsr 110 12.6 0.000 0.000 6.471 7.143 apply_single 110 13.6 0.000 0.001 6.471 7.143 qs_ot_get_p 110 10.4 0.001 0.002 6.587 6.755 sum_up_and_integrate 110 10.3 0.052 0.058 6.431 6.442 integrate_v_rspace 110 11.3 0.003 0.003 6.379 6.391 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 6.071 6.073 qs_rho_update_rho_low 110 7.6 0.001 0.001 5.910 5.918 calculate_rho_elec 110 8.6 0.039 0.047 5.910 5.917 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 5.720 5.764 ot_diis_step 99 11.5 0.010 0.010 5.720 5.720 prepare_preconditioner 11 7.9 0.000 0.000 5.521 5.537 make_preconditioner 11 8.9 0.000 0.000 5.521 5.537 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 4.770 5.470 calculate_first_density_matrix 1 7.0 0.000 0.000 5.369 5.373 make_full_inverse_cholesky 11 9.9 0.000 0.000 5.092 5.247 acc_transpose_blocks 24660 15.4 0.116 0.118 4.823 4.875 make_m2s 4110 13.4 0.057 0.060 4.205 4.686 qs_ot_p2m_diag 48 11.0 0.028 0.044 4.663 4.681 make_images 4110 14.4 0.409 0.465 4.097 4.575 cp_dbcsr_syevd 48 12.0 0.003 0.003 4.194 4.194 acc_transpose_blocks_kernels 24660 16.4 0.323 0.332 3.742 3.786 pw_transfer 1331 11.6 0.066 0.074 3.557 3.692 cp_fm_diag_elpa 48 13.0 0.000 0.000 3.622 3.633 cp_fm_diag_elpa_base 48 14.0 3.573 3.594 3.619 3.629 fft_wrap_pw1pw2 1111 12.6 0.008 0.008 3.450 3.587 wfi_extrapolate 11 7.9 0.001 0.001 3.526 3.526 jit_kernel_transpose 5 15.6 3.419 3.459 3.419 3.459 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 3.309 3.378 grid_integrate_task_list 110 12.3 3.164 3.319 3.164 3.319 density_rs2pw 110 9.6 0.004 0.005 3.149 3.303 fft_wrap_pw1pw2_140 451 13.1 0.459 0.473 2.940 3.077 calculate_dm_sparse 110 9.5 0.001 0.001 3.042 3.074 hybrid_alltoall_any 4261 16.3 0.107 0.456 2.042 2.846 make_images_data 4110 15.4 0.048 0.052 2.280 2.777 fft3d_ps 1111 14.6 1.109 1.338 2.548 2.695 cp_fm_cholesky_invert 11 10.9 2.543 2.550 2.543 2.550 mp_sum_l 6594 12.7 1.801 2.502 1.801 2.502 grid_collocate_task_list 110 9.6 2.169 2.325 2.169 2.325 potential_pw2rs 110 12.3 0.008 0.009 2.283 2.296 jit_kernel_multiply 10 16.3 1.018 2.003 1.018 2.003 mp_alltoall_d11v 2046 13.8 1.803 2.001 1.803 2.001 qs_ot_get_orbitals 99 10.5 0.001 0.001 1.968 1.987 qs_energies_init_hamiltonians 11 5.9 0.000 0.001 1.848 1.849 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.789 1.802 multiply_cannon_metrocomm4 22605 15.4 0.077 0.082 0.782 1.647 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.001 0.001 1.610 1.619 transfer_rs2pw 451 10.6 0.006 0.007 1.445 1.616 mp_allgather_i34 2055 14.4 0.577 1.585 0.577 1.585 mp_sum_dm 438 4.9 1.483 1.533 1.483 1.533 mp_irecv_dv 57340 16.2 0.656 1.530 0.656 1.530 build_core_hamiltonian_matrix_ 11 4.9 0.001 0.002 1.390 1.492 md_write_output 11 3.9 0.031 1.447 0.032 1.465 md_output 10 3.0 0.000 0.000 0.034 1.465 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="202", plot="h2o_128_md", label="(8n/6r/2t)", y=73.214000, yerr=0.000000 PlotPoint: name="203", plot="h2o_128_md_mem", label="(8n/6r/2t)", y=560.181818, yerr=6.087516 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 665.776128E+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 3473 66421. MP_Allreduce 9774 562. MP_Sync 52 MP_Alltoall 1496 4511006. MP_SendRecv 6820 27424. MP_ISendRecv 6820 27424. MP_Wait 25498 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 0.089 0.195 66.677 66.678 qs_mol_dyn_low 1 2.0 0.004 0.006 66.187 66.197 qs_forces 11 3.9 0.003 0.005 65.647 65.650 qs_energies 11 4.9 0.005 0.017 62.366 62.375 scf_env_do_scf 11 5.9 0.001 0.001 50.238 50.239 scf_env_do_scf_inner_loop 99 6.5 0.003 0.007 41.367 41.367 velocity_verlet 10 3.0 0.002 0.002 34.042 34.059 dbcsr_multiply_generic 2055 12.4 0.107 0.112 33.416 33.702 multiply_cannon 2055 13.4 0.213 0.223 26.739 28.047 multiply_cannon_loop 2055 14.4 0.816 0.843 25.583 26.759 qs_scf_new_mos 99 7.5 0.001 0.001 26.288 26.406 qs_scf_loop_do_ot 99 8.5 0.001 0.001 26.288 26.405 ot_scf_mini 99 9.5 0.003 0.003 24.983 25.120 ot_mini 99 10.5 0.001 0.001 14.257 14.400 rebuild_ks_matrix 110 8.3 0.000 0.000 12.443 12.603 qs_ks_build_kohn_sham_matrix 110 9.3 0.013 0.019 12.442 12.602 mp_waitall_1 139946 16.5 8.113 11.369 8.113 11.369 qs_ks_update_qs_env 110 7.6 0.001 0.001 10.894 11.036 multiply_cannon_multrec 16440 15.4 3.499 4.250 9.831 10.428 qs_ot_get_derivative 99 11.5 0.001 0.002 9.686 9.822 init_scf_run 11 5.9 0.000 0.001 9.589 9.590 scf_env_initial_rho_setup 11 6.9 0.001 0.003 9.589 9.590 init_scf_loop 11 6.9 0.002 0.007 8.833 8.834 multiply_cannon_metrocomm3 16440 15.4 0.046 0.047 5.062 7.906 prepare_preconditioner 11 7.9 0.000 0.002 6.997 7.019 make_preconditioner 11 8.9 0.000 0.001 6.997 7.019 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 6.807 6.809 make_full_inverse_cholesky 11 9.9 0.000 0.001 6.268 6.623 dbcsr_mm_accdrv_process 34862 16.1 5.502 6.005 6.178 6.520 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 6.433 6.492 sum_up_and_integrate 110 10.3 0.061 0.064 6.444 6.463 integrate_v_rspace 110 11.3 0.003 0.003 6.382 6.399 calculate_first_density_matrix 1 7.0 0.001 0.002 6.378 6.380 qs_rho_update_rho_low 110 7.6 0.001 0.001 6.040 6.049 calculate_rho_elec 110 8.6 0.059 0.059 6.040 6.048 qs_ot_get_p 110 10.4 0.002 0.002 5.778 5.933 acc_transpose_blocks 16440 15.4 0.078 0.079 4.852 5.587 apply_preconditioner_dbcsr 110 12.6 0.000 0.001 5.017 5.557 apply_single 110 13.6 0.000 0.000 5.017 5.557 acc_transpose_blocks_kernels 16440 16.4 0.219 0.224 4.066 4.775 make_m2s 4110 13.4 0.050 0.052 4.358 4.681 make_images 4110 14.4 0.400 0.521 4.243 4.567 jit_kernel_transpose 5 15.6 3.846 4.554 3.846 4.554 ot_diis_step 99 11.5 0.011 0.013 4.515 4.516 qs_ot_p2m_diag 48 11.0 0.043 0.053 4.033 4.041 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 3.269 3.884 multiply_cannon_sync_h2d 16440 15.4 3.178 3.777 3.178 3.777 pw_transfer 1331 11.6 0.065 0.073 3.725 3.737 cp_dbcsr_syevd 48 12.0 0.003 0.003 3.649 3.651 fft_wrap_pw1pw2 1111 12.6 0.008 0.008 3.619 3.634 density_rs2pw 110 9.6 0.004 0.005 3.227 3.513 grid_integrate_task_list 110 12.3 3.175 3.390 3.175 3.390 fft_wrap_pw1pw2_140 451 13.1 0.576 0.585 3.118 3.133 wfi_extrapolate 11 7.9 0.004 0.016 3.116 3.116 cp_fm_diag_elpa 48 13.0 0.000 0.000 3.077 3.085 cp_fm_diag_elpa_base 48 14.0 3.007 3.036 3.074 3.083 make_images_data 4110 15.4 0.045 0.049 2.475 2.865 hybrid_alltoall_any 4261 16.3 0.109 0.384 2.191 2.821 cp_fm_cholesky_invert 11 10.9 2.770 2.776 2.770 2.776 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 2.624 2.705 calculate_dm_sparse 110 9.5 0.001 0.001 2.589 2.621 fft3d_ps 1111 14.6 1.089 1.101 2.563 2.574 multiply_cannon_metrocomm4 14385 15.4 0.048 0.053 0.939 2.554 mp_sum_l 6594 12.7 1.727 2.502 1.727 2.502 mp_irecv_dv 48980 15.7 0.864 2.418 0.864 2.418 grid_collocate_task_list 110 9.6 2.216 2.337 2.216 2.337 mp_alltoall_d11v 2046 13.8 2.030 2.302 2.030 2.302 potential_pw2rs 110 12.3 0.011 0.011 2.230 2.238 dbcsr_complete_redistribute 325 12.2 0.328 0.377 1.631 2.084 qs_energies_init_hamiltonians 11 5.9 0.001 0.004 2.007 2.011 transfer_rs2pw 451 10.6 0.005 0.005 1.487 1.767 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.746 1.766 cp_fm_upper_to_full 70 14.2 1.374 1.755 1.374 1.755 cp_fm_cholesky_decompose 22 10.9 1.607 1.625 1.607 1.625 mp_waitany 17072 13.8 1.297 1.575 1.297 1.575 mp_allgather_i34 2055 14.4 0.467 1.567 0.467 1.567 copy_fm_to_dbcsr 174 11.2 0.001 0.001 1.098 1.539 multiply_cannon_metrocomm1 16440 15.4 0.029 0.030 0.606 1.534 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.001 0.001 1.486 1.502 build_core_hamiltonian_matrix_ 11 4.9 0.001 0.001 1.373 1.473 qs_ot_get_orbitals 99 10.5 0.000 0.001 1.362 1.374 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="204", plot="h2o_128_md", label="(8n/4r/3t)", y=66.678000, yerr=0.000000 PlotPoint: name="205", plot="h2o_128_md_mem", label="(8n/4r/3t)", y=630.818182, yerr=7.813952 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 739.098624E+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 3473 66419. MP_Allreduce 9774 603. MP_Sync 52 MP_Alltoall 1496 5863162. MP_SendRecv 5060 43184. MP_ISendRecv 5060 43184. MP_Wait 20042 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 0.023 0.062 72.147 72.148 qs_mol_dyn_low 1 2.0 0.003 0.004 71.856 71.874 qs_forces 11 3.9 0.003 0.003 71.791 71.793 qs_energies 11 4.9 0.002 0.007 68.361 68.366 scf_env_do_scf 11 5.9 0.001 0.002 55.273 55.276 scf_env_do_scf_inner_loop 99 6.5 0.013 0.050 43.509 43.509 velocity_verlet 10 3.0 0.002 0.002 37.968 37.970 dbcsr_multiply_generic 2055 12.4 0.116 0.128 35.782 36.006 multiply_cannon 2055 13.4 0.236 0.253 28.353 29.539 qs_scf_new_mos 99 7.5 0.001 0.001 28.519 28.617 qs_scf_loop_do_ot 99 8.5 0.001 0.001 28.518 28.617 multiply_cannon_loop 2055 14.4 1.404 1.500 26.937 27.591 ot_scf_mini 99 9.5 0.003 0.003 26.751 26.857 ot_mini 99 10.5 0.001 0.001 15.275 15.396 multiply_cannon_multrec 24660 15.4 4.083 6.772 13.085 14.115 rebuild_ks_matrix 110 8.3 0.000 0.000 12.168 12.271 qs_ks_build_kohn_sham_matrix 110 9.3 0.013 0.016 12.168 12.271 init_scf_loop 11 6.9 0.001 0.004 11.722 11.723 qs_ot_get_derivative 99 11.5 0.001 0.001 10.918 11.019 qs_ks_update_qs_env 110 7.6 0.001 0.001 10.759 10.851 dbcsr_mm_accdrv_process 52304 16.0 7.616 9.031 8.845 10.608 init_scf_run 11 5.9 0.000 0.001 10.167 10.167 scf_env_initial_rho_setup 11 6.9 0.001 0.002 10.166 10.167 prepare_preconditioner 11 7.9 0.000 0.000 9.987 10.004 make_preconditioner 11 8.9 0.000 0.001 9.987 10.004 make_full_inverse_cholesky 11 9.9 0.000 0.001 8.245 9.678 mp_waitall_1 121746 16.5 5.858 8.022 5.858 8.022 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 7.431 7.433 acc_transpose_blocks 24660 15.4 0.114 0.118 5.642 7.143 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 7.055 7.101 calculate_first_density_matrix 1 7.0 0.000 0.002 7.059 7.061 qs_ot_get_p 110 10.4 0.002 0.002 6.298 6.459 sum_up_and_integrate 110 10.3 0.067 0.072 6.406 6.420 integrate_v_rspace 110 11.3 0.003 0.003 6.339 6.353 qs_rho_update_rho_low 110 7.6 0.001 0.001 6.007 6.018 calculate_rho_elec 110 8.6 0.078 0.081 6.006 6.018 make_m2s 4110 13.4 0.061 0.063 5.586 5.869 make_images 4110 14.4 0.580 0.699 5.443 5.723 acc_transpose_blocks_kernels 24660 16.4 0.311 0.316 4.112 5.387 jit_kernel_transpose 5 15.6 3.801 5.074 3.801 5.074 multiply_cannon_metrocomm3 24660 15.4 0.038 0.039 2.123 4.840 cp_fm_upper_to_full 70 14.2 3.321 4.810 3.321 4.810 qs_ot_p2m_diag 48 11.0 0.055 0.064 4.356 4.378 ot_diis_step 99 11.5 0.011 0.011 4.307 4.307 apply_preconditioner_dbcsr 110 12.6 0.000 0.000 4.174 4.304 apply_single 110 13.6 0.000 0.000 4.174 4.304 dbcsr_complete_redistribute 325 12.2 0.421 0.460 2.659 3.777 cp_dbcsr_syevd 48 12.0 0.003 0.003 3.776 3.777 pw_transfer 1331 11.6 0.065 0.075 3.719 3.753 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 3.638 3.683 fft_wrap_pw1pw2 1111 12.6 0.008 0.008 3.613 3.651 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 3.465 3.521 grid_integrate_task_list 110 12.3 3.291 3.437 3.291 3.437 density_rs2pw 110 9.6 0.004 0.004 3.114 3.357 copy_fm_to_dbcsr 174 11.2 0.001 0.001 2.120 3.226 cp_fm_diag_elpa 48 13.0 0.000 0.000 3.214 3.224 cp_fm_diag_elpa_base 48 14.0 3.060 3.122 3.211 3.221 fft_wrap_pw1pw2_140 451 13.1 0.607 0.628 3.081 3.122 calculate_dm_sparse 110 9.5 0.001 0.001 3.085 3.110 make_images_data 4110 15.4 0.048 0.053 2.762 3.065 wfi_extrapolate 11 7.9 0.001 0.001 3.011 3.011 hybrid_alltoall_any 4261 16.3 0.123 0.463 2.376 2.952 transfer_fm_to_dbcsr 11 9.9 0.000 0.000 1.733 2.822 mp_alltoall_i22 605 13.7 1.671 2.815 1.671 2.815 cp_fm_cholesky_invert 11 10.9 2.805 2.814 2.805 2.814 fft3d_ps 1111 14.6 1.083 1.110 2.497 2.513 multiply_cannon_sync_h2d 24660 15.4 2.385 2.503 2.385 2.503 multiply_cannon_metrocomm1 24660 15.4 0.035 0.036 1.024 2.441 grid_collocate_task_list 110 9.6 2.262 2.419 2.262 2.419 qs_energies_init_hamiltonians 11 5.9 0.001 0.005 2.387 2.389 potential_pw2rs 110 12.3 0.013 0.013 2.158 2.168 mp_alltoall_d11v 2046 13.8 1.803 2.039 1.803 2.039 qs_ot_get_orbitals 99 10.5 0.001 0.001 1.869 1.963 mp_sum_l 6594 12.7 1.165 1.781 1.165 1.781 cp_fm_cholesky_decompose 22 10.9 1.724 1.770 1.724 1.770 build_core_hamiltonian_matrix_ 11 4.9 0.001 0.001 1.615 1.712 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.655 1.667 transfer_rs2pw 451 10.6 0.005 0.006 1.426 1.662 jit_kernel_multiply 9 15.9 0.878 1.604 0.878 1.604 acc_transpose_blocks_sync 73980 16.4 1.374 1.594 1.374 1.594 multiply_cannon_metrocomm4 20550 15.4 0.062 0.066 0.849 1.588 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.001 0.001 1.568 1.586 mp_allgather_i34 2055 14.4 0.498 1.550 0.498 1.550 mp_waitany 13376 13.8 1.230 1.527 1.230 1.527 mp_irecv_dv 62702 16.1 0.745 1.505 0.745 1.505 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="206", plot="h2o_128_md", label="(8n/3r/4t)", y=72.148000, yerr=0.000000 PlotPoint: name="207", plot="h2o_128_md_mem", label="(8n/3r/4t)", y=698.636364, yerr=9.246844 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 859.725824E+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 3473 66417. MP_Allreduce 9774 644. MP_Sync 52 MP_Alltoall 1496 8504061. MP_SendRecv 3300 54848. MP_ISendRecv 3300 54848. MP_Wait 13926 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 0.014 0.029 59.300 59.300 qs_mol_dyn_low 1 2.0 0.003 0.003 59.065 59.074 qs_forces 11 3.9 0.003 0.003 59.001 59.002 qs_energies 11 4.9 0.001 0.002 55.336 55.340 scf_env_do_scf 11 5.9 0.000 0.001 44.253 44.253 scf_env_do_scf_inner_loop 99 6.5 0.007 0.031 36.393 36.394 velocity_verlet 10 3.0 0.002 0.002 31.598 31.601 dbcsr_multiply_generic 2055 12.4 0.105 0.109 26.529 26.723 multiply_cannon 2055 13.4 0.239 0.259 20.773 22.089 qs_scf_new_mos 99 7.5 0.001 0.001 21.643 21.708 qs_scf_loop_do_ot 99 8.5 0.001 0.001 21.643 21.708 ot_scf_mini 99 9.5 0.002 0.002 20.351 20.391 multiply_cannon_loop 2055 14.4 0.607 0.636 19.485 19.913 rebuild_ks_matrix 110 8.3 0.000 0.000 11.779 11.814 qs_ks_build_kohn_sham_matrix 110 9.3 0.013 0.013 11.778 11.814 ot_mini 99 10.5 0.001 0.001 11.156 11.189 qs_ks_update_qs_env 110 7.6 0.001 0.001 10.498 10.529 multiply_cannon_multrec 8220 15.4 3.174 4.701 7.619 8.742 init_scf_run 11 5.9 0.000 0.001 7.944 7.945 scf_env_initial_rho_setup 11 6.9 0.001 0.001 7.944 7.945 init_scf_loop 11 6.9 0.000 0.000 7.813 7.814 mp_waitall_1 103326 16.6 6.198 7.737 6.198 7.737 qs_ot_get_derivative 99 11.5 0.001 0.001 7.341 7.381 sum_up_and_integrate 110 10.3 0.080 0.082 6.356 6.367 integrate_v_rspace 110 11.3 0.003 0.004 6.276 6.286 prepare_preconditioner 11 7.9 0.000 0.000 6.112 6.117 make_preconditioner 11 8.9 0.000 0.000 6.112 6.117 qs_rho_update_rho_low 110 7.6 0.001 0.001 6.026 6.041 calculate_rho_elec 110 8.6 0.114 0.114 6.025 6.041 make_full_inverse_cholesky 11 9.9 0.000 0.000 5.717 5.790 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 5.328 5.329 qs_ot_get_p 110 10.4 0.011 0.045 5.217 5.259 dbcsr_mm_accdrv_process 17442 15.9 3.058 4.172 4.304 5.230 calculate_first_density_matrix 1 7.0 0.000 0.000 5.049 5.050 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 4.933 4.955 multiply_cannon_metrocomm3 8220 15.4 0.019 0.020 3.264 4.693 make_m2s 4110 13.4 0.039 0.040 4.294 4.490 make_images 4110 14.4 0.649 0.713 4.163 4.356 acc_transpose_blocks 8220 15.4 0.040 0.041 4.162 4.203 pw_transfer 1331 11.6 0.065 0.071 3.938 3.947 apply_preconditioner_dbcsr 110 12.6 0.000 0.000 3.827 3.891 apply_single 110 13.6 0.000 0.000 3.827 3.890 fft_wrap_pw1pw2 1111 12.6 0.008 0.008 3.831 3.843 ot_diis_step 99 11.5 0.012 0.012 3.790 3.790 qs_ot_p2m_diag 48 11.0 0.081 0.084 3.710 3.714 grid_integrate_task_list 110 12.3 3.369 3.488 3.369 3.488 acc_transpose_blocks_kernels 8220 16.4 0.115 0.117 3.435 3.441 cp_dbcsr_syevd 48 12.0 0.003 0.003 3.357 3.359 jit_kernel_transpose 5 15.6 3.320 3.325 3.320 3.325 fft_wrap_pw1pw2_140 451 13.1 0.776 0.790 3.283 3.296 density_rs2pw 110 9.6 0.004 0.004 2.974 3.097 cp_fm_cholesky_invert 11 10.9 2.963 2.966 2.963 2.966 wfi_extrapolate 11 7.9 0.001 0.001 2.798 2.799 cp_fm_diag_elpa 48 13.0 0.000 0.000 2.756 2.763 cp_fm_diag_elpa_base 48 14.0 2.700 2.725 2.755 2.762 hybrid_alltoall_any 4261 16.3 0.201 0.855 2.277 2.674 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 2.662 2.663 make_images_data 4110 15.4 0.040 0.046 2.279 2.579 calculate_dm_sparse 110 9.5 0.001 0.001 2.529 2.571 grid_collocate_task_list 110 9.6 2.364 2.554 2.364 2.554 fft3d_ps 1111 14.6 1.134 1.160 2.495 2.511 multiply_cannon_sync_h2d 8220 15.4 2.342 2.493 2.342 2.493 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 2.077 2.113 potential_pw2rs 110 12.3 0.015 0.016 2.068 2.071 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 2.047 2.065 mp_alltoall_d11v 2046 13.8 1.849 2.056 1.849 2.056 build_core_hamiltonian_matrix_ 11 4.9 0.001 0.001 1.790 2.023 jit_kernel_multiply 10 15.5 0.929 1.793 0.929 1.793 cp_fm_cholesky_decompose 22 10.9 1.700 1.721 1.700 1.721 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.611 1.617 qs_env_update_s_mstruct 11 6.9 0.001 0.001 1.490 1.603 dbcsr_complete_redistribute 325 12.2 0.558 0.586 1.470 1.561 mp_allgather_i34 2055 14.4 0.456 1.537 0.456 1.537 cp_dbcsr_plus_fm_fm_t_native 22 8.9 0.001 0.001 1.447 1.458 multiply_cannon_metrocomm4 6165 15.4 0.019 0.020 0.478 1.329 qs_create_task_list 11 7.9 0.001 0.001 1.212 1.312 generate_qs_task_list 11 8.9 0.372 0.440 1.212 1.311 transfer_rs2pw 451 10.6 0.005 0.005 1.141 1.293 copy_dbcsr_to_fm 151 11.3 0.003 0.003 1.264 1.290 mp_irecv_dv 24056 15.7 0.452 1.286 0.452 1.286 mp_waitany 9240 13.8 1.069 1.202 1.069 1.202 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="208", plot="h2o_128_md", label="(8n/2r/6t)", y=59.300000, yerr=0.000000 PlotPoint: name="209", plot="h2o_128_md_mem", label="(8n/2r/6t)", y=808.272727, yerr=13.267122 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 1.393201E+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 67098. MP_Allreduce 9752 812. 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 0.018 0.047 91.435 91.435 qs_mol_dyn_low 1 2.0 0.003 0.003 91.093 91.121 qs_forces 11 3.9 0.003 0.003 90.960 90.962 qs_energies 11 4.9 0.001 0.002 86.834 86.837 scf_env_do_scf 11 5.9 0.001 0.001 73.914 73.914 velocity_verlet 10 3.0 0.002 0.002 56.383 56.397 scf_env_do_scf_inner_loop 99 6.5 0.003 0.007 45.437 45.439 dbcsr_multiply_generic 2055 12.4 0.120 0.125 33.000 33.095 init_scf_loop 11 6.9 0.000 0.000 28.403 28.404 qs_scf_new_mos 99 7.5 0.001 0.001 27.767 27.795 qs_scf_loop_do_ot 99 8.5 0.001 0.001 27.766 27.794 prepare_preconditioner 11 7.9 0.000 0.000 26.323 26.332 make_preconditioner 11 8.9 0.000 0.000 26.323 26.332 multiply_cannon 2055 13.4 0.333 0.354 25.528 26.200 ot_scf_mini 99 9.5 0.002 0.002 25.866 25.903 make_full_inverse_cholesky 11 9.9 0.000 0.000 20.568 25.793 multiply_cannon_loop 2055 14.4 0.826 0.845 23.672 23.969 cp_fm_upper_to_full 70 14.2 12.529 17.997 12.529 17.997 ot_mini 99 10.5 0.001 0.001 14.467 14.495 rebuild_ks_matrix 110 8.3 0.000 0.000 13.951 14.003 qs_ks_build_kohn_sham_matrix 110 9.3 0.013 0.014 13.951 14.003 qs_ks_update_qs_env 110 7.6 0.001 0.001 12.648 12.693 dbcsr_complete_redistribute 325 12.2 1.020 1.051 7.354 10.547 multiply_cannon_multrec 8220 15.4 4.053 4.231 9.718 9.869 qs_ot_get_derivative 99 11.5 0.001 0.001 9.692 9.729 copy_fm_to_dbcsr 174 11.2 0.001 0.001 6.323 9.510 transfer_fm_to_dbcsr 11 9.9 0.000 0.000 5.739 8.885 mp_waitall_1 84994 16.7 7.993 8.864 7.993 8.864 init_scf_run 11 5.9 0.000 0.001 8.633 8.633 scf_env_initial_rho_setup 11 6.9 0.001 0.001 8.633 8.633 mp_alltoall_i22 605 13.7 5.372 8.585 5.372 8.585 qs_rho_update_rho_low 110 7.6 0.001 0.001 7.247 7.278 calculate_rho_elec 110 8.6 0.224 0.225 7.246 7.277 sum_up_and_integrate 110 10.3 0.149 0.152 7.008 7.039 integrate_v_rspace 110 11.3 0.004 0.005 6.858 6.889 make_m2s 4110 13.4 0.044 0.045 5.659 6.198 qs_ot_get_p 110 10.4 0.001 0.002 6.002 6.037 make_images 4110 14.4 0.885 0.926 5.469 6.008 cp_dbcsr_sm_fm_multiply 37 9.5 0.001 0.001 5.821 5.825 dbcsr_mm_accdrv_process 11614 15.7 3.974 4.808 5.518 5.755 cp_fm_cholesky_invert 11 10.9 5.472 5.476 5.472 5.476 multiply_cannon_metrocomm3 8220 15.4 0.019 0.020 5.023 5.371 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 5.297 5.317 apply_preconditioner_dbcsr 110 12.6 0.000 0.000 4.883 5.311 apply_single 110 13.6 0.000 0.000 4.883 5.310 calculate_first_density_matrix 1 7.0 0.000 0.000 5.145 5.146 pw_transfer 1331 11.6 0.074 0.074 5.074 5.095 fft_wrap_pw1pw2 1111 12.6 0.009 0.009 4.958 4.978 ot_diis_step 99 11.5 0.015 0.015 4.752 4.752 acc_transpose_blocks 8220 15.4 0.041 0.041 4.614 4.647 qs_ot_p2m_diag 48 11.0 0.151 0.156 4.320 4.330 fft_wrap_pw1pw2_140 451 13.1 1.278 1.285 4.301 4.323 hybrid_alltoall_any 4261 16.3 0.264 0.562 3.050 3.849 cp_dbcsr_syevd 48 12.0 0.003 0.003 3.780 3.780 make_images_data 4110 15.4 0.045 0.048 3.071 3.765 grid_integrate_task_list 110 12.3 3.692 3.725 3.692 3.725 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 3.673 3.675 density_rs2pw 110 9.6 0.004 0.004 3.639 3.665 acc_transpose_blocks_kernels 8220 16.4 0.117 0.118 3.537 3.541 qs_ot_get_derivative_taylor 52 13.0 0.001 0.001 3.096 3.540 jit_kernel_transpose 5 15.6 3.421 3.423 3.421 3.423 wfi_extrapolate 11 7.9 0.001 0.001 3.367 3.367 calculate_dm_sparse 110 9.5 0.001 0.001 3.269 3.304 cp_fm_diag_elpa 48 13.0 0.000 0.000 3.159 3.160 cp_fm_diag_elpa_base 48 14.0 2.622 2.820 3.158 3.158 multiply_cannon_sync_h2d 8220 15.4 3.125 3.149 3.125 3.149 fft3d_ps 1111 14.6 1.307 1.327 3.010 3.030 qs_ot_get_derivative_diag 47 12.0 0.001 0.001 2.724 2.743 grid_collocate_task_list 110 9.6 2.682 2.719 2.682 2.719 potential_pw2rs 110 12.3 0.021 0.022 2.377 2.397 qs_env_update_s_mstruct 11 6.9 0.001 0.001 2.191 2.256 mp_alltoall_d11v 2046 13.8 2.168 2.214 2.168 2.214 build_core_hamiltonian_matrix_ 11 4.9 0.001 0.001 2.085 2.145 cp_fm_cholesky_decompose 22 10.9 2.041 2.059 2.041 2.059 qs_create_task_list 11 7.9 0.001 0.001 1.888 1.933 generate_qs_task_list 11 8.9 0.730 0.784 1.887 1.932 qs_ks_update_qs_env_forces 11 4.9 0.000 0.000 1.903 1.910 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="210", plot="h2o_128_md", label="(8n/1r/12t)", y=91.435000, yerr=0.000000 PlotPoint: name="211", plot="h2o_128_md_mem", label="(8n/1r/12t)", y=1256.909091, yerr=54.658368 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/15/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1420242647040 0.0% 0.0% 100.0% flops 32 x 32 x 32 1943472701440 0.0% 0.0% 100.0% flops 22 x 9 x 32 1972057190400 0.0% 0.0% 100.0% flops 9 x 22 x 32 1977770336256 0.0% 0.0% 100.0% flops 22 x 22 x 32 2734287699968 0.0% 0.0% 100.0% flops 32 x 32 x 9 4416300122112 0.0% 0.0% 100.0% flops 32 x 32 x 22 5397700149248 0.0% 0.0% 100.0% flops 9 x 32 x 32 5443971710976 0.0% 0.0% 100.0% flops 22 x 32 x 32 6653743202304 0.0% 0.0% 100.0% flops 9 x 32 x 9 11528903135232 0.0% 0.0% 100.0% flops 22 x 32 x 9 15129160814592 0.0% 0.0% 100.0% flops 9 x 32 x 22 15129160814592 0.0% 0.0% 100.0% flops 22 x 32 x 22 19767995056128 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 93.514766E+12 0.0% 0.0% 100.0% flops max/rank 1.094965E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 6755941440 0.0% 0.0% 100.0% number of processed stacks 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 631.279616E+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 3175955383376 4194304 < size <= 16777216 261888 1145794321408 16777216 < size 0 0 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 4020 57505. MP_Allreduce 11129 795. MP_Sync 87 MP_Alltoall 2226 3158756. MP_SendRecv 24320 18752. MP_ISendRecv 24320 18752. MP_Wait 42476 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 0.013 0.031 213.397 213.400 qs_mol_dyn_low 1 2.0 0.006 0.024 212.898 212.914 qs_forces 11 3.9 0.005 0.005 212.766 212.772 qs_energies 11 4.9 0.001 0.002 207.111 207.135 scf_env_do_scf 11 5.9 0.001 0.001 185.648 185.654 scf_env_do_scf_inner_loop 117 6.6 0.003 0.009 164.725 164.728 dbcsr_multiply_generic 2507 12.6 0.181 0.185 130.445 131.210 qs_scf_new_mos 117 7.6 0.001 0.001 124.745 125.012 qs_scf_loop_do_ot 117 8.6 0.001 0.001 124.744 125.011 velocity_verlet 10 3.0 0.004 0.014 124.724 124.726 ot_scf_mini 117 9.6 0.003 0.003 118.063 118.333 multiply_cannon 2507 13.6 0.241 0.252 106.404 108.149 multiply_cannon_loop 2507 14.6 2.388 2.469 104.155 105.984 ot_mini 117 10.6 0.001 0.001 66.195 66.474 multiply_cannon_multrec 60168 15.6 31.959 34.139 42.361 45.324 qs_ot_get_derivative 117 11.6 0.001 0.001 41.438 41.678 rebuild_ks_matrix 128 8.3 0.001 0.001 33.932 34.304 qs_ks_build_kohn_sham_matrix 128 9.3 0.016 0.024 33.931 34.304 mp_waitall_1 267128 16.5 29.881 33.610 29.881 33.610 qs_ks_update_qs_env 128 7.6 0.001 0.001 30.454 30.811 qs_ot_get_p 128 10.4 0.002 0.009 29.742 30.082 multiply_cannon_sync_h2d 60168 15.6 26.620 28.722 26.620 28.722 apply_preconditioner_dbcsr 128 12.6 0.000 0.001 24.420 25.077 apply_single 128 13.6 0.001 0.001 24.420 25.077 ot_diis_step 117 11.6 0.008 0.008 24.512 24.513 qs_ot_p2m_diag 83 11.4 0.079 0.092 23.083 23.136 init_scf_loop 11 6.9 0.000 0.001 20.847 20.849 cp_dbcsr_syevd 83 12.4 0.005 0.005 20.472 20.473 qs_ot_get_derivative_diag 77 12.4 0.002 0.002 19.844 20.047 multiply_cannon_metrocomm3 60168 15.6 0.121 0.128 15.938 18.523 cp_fm_diag_elpa 83 13.4 0.000 0.000 17.443 17.481 cp_fm_diag_elpa_base 83 14.4 17.369 17.408 17.439 17.476 init_scf_run 11 5.9 0.000 0.001 17.275 17.276 scf_env_initial_rho_setup 11 6.9 0.001 0.001 17.275 17.276 prepare_preconditioner 11 7.9 0.000 0.000 16.204 16.252 make_preconditioner 11 8.9 0.000 0.000 16.204 16.252 make_full_inverse_cholesky 11 9.9 0.000 0.000 15.437 15.622 make_m2s 5014 13.6 0.105 0.113 14.035 14.472 make_images 5014 14.6 0.396 0.414 13.853 14.300 sum_up_and_integrate 128 10.3 0.089 0.106 14.214 14.234 integrate_v_rspace 128 11.3 0.004 0.005 14.124 14.146 qs_rho_update_rho_low 128 7.7 0.001 0.001 13.426 13.562 calculate_rho_elec 128 8.7 0.045 0.064 13.426 13.561 dbcsr_mm_accdrv_process 124484 16.2 4.590 4.840 9.960 11.825 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.003 11.559 11.571 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 10.670 10.784 mp_sum_l 7950 12.9 8.744 10.243 8.744 10.243 multiply_cannon_metrocomm1 60168 15.6 0.100 0.104 7.323 9.830 cp_fm_cholesky_invert 11 10.9 9.354 9.363 9.354 9.363 wfi_extrapolate 11 7.9 0.001 0.001 9.151 9.151 calculate_dm_sparse 128 9.5 0.001 0.001 8.504 8.620 qs_ot_get_derivative_taylor 40 13.0 0.001 0.001 8.078 8.234 qs_ot_get_orbitals 117 10.6 0.001 0.001 7.962 8.020 calculate_first_density_matrix 1 7.0 0.000 0.001 7.941 7.946 pw_transfer 1547 11.6 0.074 0.088 7.734 7.906 make_images_data 5014 15.6 0.067 0.074 6.855 7.749 fft_wrap_pw1pw2 1291 12.7 0.010 0.012 7.532 7.697 density_rs2pw 128 9.7 0.006 0.007 7.034 7.649 grid_integrate_task_list 128 12.3 7.083 7.440 7.083 7.440 hybrid_alltoall_any 5200 16.5 0.296 2.279 5.994 7.239 fft_wrap_pw1pw2_140 523 13.2 1.133 1.181 6.573 6.742 acc_transpose_blocks 60168 15.6 0.290 0.306 6.295 6.482 fft3d_ps 1291 14.7 2.194 2.832 5.530 5.851 mp_alltoall_d11v 2415 14.1 4.426 5.591 4.426 5.591 grid_collocate_task_list 128 9.7 4.844 5.186 4.844 5.186 cp_fm_cholesky_decompose 22 10.9 4.666 4.679 4.666 4.679 potential_pw2rs 128 12.3 0.009 0.010 4.629 4.650 jit_kernel_multiply 13 15.9 3.290 4.598 3.290 4.598 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="400", plot="h2o_256_md", label="(8n/12r/1t)", y=213.400000, yerr=0.000000 PlotPoint: name="401", plot="h2o_256_md_mem", label="(8n/12r/1t)", y=596.545455, yerr=6.372721 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/16/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1420242647040 0.0% 0.0% 100.0% flops 32 x 32 x 32 1943472701440 0.0% 0.0% 100.0% flops 22 x 9 x 32 1972057190400 0.0% 0.0% 100.0% flops 9 x 22 x 32 1977770336256 0.0% 0.0% 100.0% flops 22 x 22 x 32 2734287699968 0.0% 0.0% 100.0% flops 32 x 32 x 9 4416300122112 0.0% 0.0% 100.0% flops 32 x 32 x 22 5397700149248 0.0% 0.0% 100.0% flops 9 x 32 x 32 5443971710976 0.0% 0.0% 100.0% flops 22 x 32 x 32 6653743202304 0.0% 0.0% 100.0% flops 9 x 32 x 9 11528903135232 0.0% 0.0% 100.0% flops 22 x 32 x 9 15129160814592 0.0% 0.0% 100.0% flops 9 x 32 x 22 15129160814592 0.0% 0.0% 100.0% flops 22 x 32 x 22 19767995056128 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 93.514766E+12 0.0% 0.0% 100.0% flops max/rank 2.183246E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 6755941440 0.0% 0.0% 100.0% number of processed stacks 5975232 0.0% 0.0% 100.0% average stack size 0.0 0.0 1130.7 marketing flops 144.580175E+12 ------------------------------------------------------------------------------- # multiplications 2507 max memory usage/rank 838.909952E+06 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 2406720 MPI messages size (bytes): total size 4.100943E+12 min size 0.000000E+00 max size 17.653760E+06 average size 1.703955E+06 MPI breakdown and total messages size (bytes): size <= 128 14916 0 128 < size <= 8192 0 0 8192 < size <= 32768 70860 2317615104 32768 < size <= 131072 722992 55511613440 131072 < size <= 4194304 1375664 1398181724160 4194304 < size <= 16777216 154704 1463834845264 16777216 < size 67584 1181116006400 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 4002 58214. MP_Allreduce 11083 960. MP_Sync 87 MP_Alltoall 1969 5295516. MP_SendRecv 12032 47072. MP_ISendRecv 12032 47072. MP_Wait 25916 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 0.026 0.052 196.160 196.162 qs_mol_dyn_low 1 2.0 0.003 0.003 195.721 195.735 qs_forces 11 3.9 0.015 0.086 195.626 195.628 qs_energies 11 4.9 0.009 0.038 188.609 188.636 scf_env_do_scf 11 5.9 0.001 0.001 167.770 167.780 scf_env_do_scf_inner_loop 117 6.6 0.003 0.008 134.617 134.620 velocity_verlet 10 3.0 0.001 0.002 120.102 120.104 dbcsr_multiply_generic 2507 12.6 0.188 0.193 101.864 103.240 qs_scf_new_mos 117 7.6 0.001 0.001 95.771 96.306 qs_scf_loop_do_ot 117 8.6 0.001 0.001 95.770 96.305 ot_scf_mini 117 9.6 0.004 0.005 90.929 91.526 multiply_cannon 2507 13.6 0.503 0.565 81.222 86.452 multiply_cannon_loop 2507 14.6 1.582 1.655 78.241 80.952 ot_mini 117 10.6 0.001 0.001 50.535 51.076 mp_waitall_1 214728 16.6 25.527 39.960 25.527 39.960 multiply_cannon_multrec 30084 15.6 20.769 25.228 32.004 37.273 rebuild_ks_matrix 128 8.3 0.001 0.001 32.517 33.149 qs_ks_build_kohn_sham_matrix 128 9.3 0.017 0.021 32.516 33.149 init_scf_loop 11 6.9 0.001 0.006 33.059 33.061 multiply_cannon_metrocomm3 30084 15.6 0.097 0.101 16.540 30.105 qs_ks_update_qs_env 128 7.6 0.001 0.001 29.176 29.764 qs_ot_get_derivative 117 11.6 0.001 0.002 28.412 29.020 prepare_preconditioner 11 7.9 0.000 0.001 28.638 28.719 make_preconditioner 11 8.9 0.000 0.002 28.638 28.719 make_full_inverse_cholesky 11 9.9 0.000 0.001 27.324 27.892 apply_preconditioner_dbcsr 128 12.6 0.000 0.000 22.198 23.373 apply_single 128 13.6 0.001 0.001 22.197 23.373 qs_ot_get_p 128 10.4 0.002 0.010 22.432 23.080 ot_diis_step 117 11.6 0.014 0.015 21.940 21.942 multiply_cannon_sync_h2d 30084 15.6 17.815 20.009 17.815 20.009 qs_ot_p2m_diag 83 11.4 0.188 0.216 17.338 17.380 cp_fm_cholesky_invert 11 10.9 16.700 16.713 16.700 16.713 cp_dbcsr_syevd 83 12.4 0.005 0.006 16.097 16.098 init_scf_run 11 5.9 0.001 0.003 15.917 15.918 scf_env_initial_rho_setup 11 6.9 0.001 0.002 15.916 15.918 make_m2s 5014 13.6 0.089 0.095 14.457 15.697 make_images 5014 14.6 1.176 1.374 14.250 15.486 sum_up_and_integrate 128 10.3 0.116 0.134 14.273 14.312 integrate_v_rspace 128 11.3 0.004 0.004 14.156 14.203 qs_rho_update_rho_low 128 7.7 0.001 0.001 13.577 13.619 calculate_rho_elec 128 8.7 0.087 0.104 13.577 13.618 cp_fm_diag_elpa 83 13.4 0.000 0.000 12.880 12.910 cp_fm_diag_elpa_base 83 14.4 12.626 12.717 12.874 12.899 dbcsr_mm_accdrv_process 62242 16.2 5.639 7.691 10.693 12.089 qs_ot_get_derivative_diag 77 12.4 0.002 0.002 11.302 11.730 multiply_cannon_metrocomm4 27577 15.6 0.105 0.118 3.893 11.090 mp_irecv_dv 69486 16.3 3.688 10.677 3.688 10.677 make_images_data 5014 15.6 0.067 0.075 8.534 10.059 hybrid_alltoall_any 5200 16.5 0.352 1.546 7.117 9.280 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 8.900 8.911 pw_transfer 1547 11.6 0.084 0.102 8.789 8.859 fft_wrap_pw1pw2 1291 12.7 0.010 0.011 8.566 8.637 wfi_extrapolate 11 7.9 0.001 0.001 8.443 8.443 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 7.952 8.152 calculate_dm_sparse 128 9.5 0.001 0.001 7.991 8.125 grid_integrate_task_list 128 12.3 7.159 7.673 7.159 7.673 fft_wrap_pw1pw2_140 523 13.2 1.224 1.254 7.549 7.635 density_rs2pw 128 9.7 0.006 0.007 7.234 7.633 calculate_first_density_matrix 1 7.0 0.001 0.003 7.273 7.276 qs_ot_get_derivative_taylor 40 13.0 0.001 0.001 6.306 7.120 cp_fm_cholesky_decompose 22 10.9 7.002 7.091 7.002 7.091 mp_sum_l 7950 12.9 4.398 6.511 4.398 6.511 fft3d_ps 1291 14.7 2.836 2.997 6.115 6.172 qs_ot_get_orbitals 117 10.6 0.001 0.001 5.523 5.592 grid_collocate_task_list 128 9.7 5.000 5.429 5.000 5.429 mp_alltoall_d11v 2415 14.1 4.303 5.294 4.303 5.294 acc_transpose_blocks 30084 15.6 0.150 0.154 5.187 5.234 potential_pw2rs 128 12.3 0.016 0.018 4.682 4.718 mp_allgather_i34 2507 14.6 1.502 4.683 1.502 4.683 mp_sum_d 4470 12.1 2.803 4.221 2.803 4.221 dbcsr_complete_redistribute 395 12.7 0.775 0.849 3.233 4.093 jit_kernel_multiply 13 16.0 1.992 3.965 1.992 3.965 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="402", plot="h2o_256_md", label="(8n/6r/2t)", y=196.162000, yerr=0.000000 PlotPoint: name="403", plot="h2o_256_md_mem", label="(8n/6r/2t)", y=799.454545, yerr=3.114880 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/17/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1420241154048 0.0% 0.0% 100.0% flops 32 x 32 x 32 1943472701440 0.0% 0.0% 100.0% flops 22 x 9 x 32 1972057190400 0.0% 0.0% 100.0% flops 9 x 22 x 32 1977770336256 0.0% 0.0% 100.0% flops 22 x 22 x 32 2734287699968 0.0% 0.0% 100.0% flops 32 x 32 x 9 4416300122112 0.0% 0.0% 100.0% flops 32 x 32 x 22 5397700149248 0.0% 0.0% 100.0% flops 9 x 32 x 32 5443971710976 0.0% 0.0% 100.0% flops 22 x 32 x 32 6653743202304 0.0% 0.0% 100.0% flops 9 x 32 x 9 11528896499712 0.0% 0.0% 100.0% flops 22 x 32 x 9 15129160814592 0.0% 0.0% 100.0% flops 9 x 32 x 22 15129160814592 0.0% 0.0% 100.0% flops 22 x 32 x 22 19767995056128 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 93.514757E+12 0.0% 0.0% 100.0% flops max/rank 2.928533E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 6755939872 0.0% 0.0% 100.0% number of processed stacks 3984192 0.0% 0.0% 100.0% average stack size 0.0 0.0 1695.7 marketing flops 144.579337E+12 ------------------------------------------------------------------------------- # multiplications 2507 max memory usage/rank 934.416384E+06 # max total images/rank 2 # max 3D layers 1 # MPI messages exchanged 1042912 MPI messages size (bytes): total size 2.716210E+12 min size 0.000000E+00 max size 26.214400E+06 average size 2.604448E+06 MPI breakdown and total messages size (bytes): size <= 128 6424 0 128 < size <= 8192 0 0 8192 < size <= 32768 264 8650752 32768 < size <= 131072 281856 36943429632 131072 < size <= 4194304 660064 996105256960 4194304 < size <= 16777216 65632 931531083728 16777216 < size 28672 751619276800 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 3992 58351. MP_Allreduce 11057 1000. MP_Sync 87 MP_Alltoall 1712 9388896. MP_SendRecv 7936 75008. MP_ISendRecv 7936 75008. MP_Wait 21820 MP_ISend 11748 275205. MP_IRecv 11748 275205. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.014 0.056 178.697 178.698 qs_mol_dyn_low 1 2.0 0.003 0.003 178.074 178.250 qs_forces 11 3.9 0.004 0.005 177.981 177.984 qs_energies 11 4.9 0.001 0.002 171.459 171.469 scf_env_do_scf 11 5.9 0.001 0.001 152.866 152.867 scf_env_do_scf_inner_loop 117 6.6 0.006 0.030 118.178 118.179 velocity_verlet 10 3.0 0.002 0.002 112.292 112.294 dbcsr_multiply_generic 2507 12.6 0.181 0.192 84.896 85.967 qs_scf_new_mos 117 7.6 0.001 0.001 81.071 81.493 qs_scf_loop_do_ot 117 8.6 0.001 0.001 81.070 81.493 ot_scf_mini 117 9.6 0.003 0.004 76.831 77.298 multiply_cannon 2507 13.6 0.518 0.544 64.961 68.903 multiply_cannon_loop 2507 14.6 1.132 1.206 62.219 64.864 ot_mini 117 10.6 0.001 0.001 42.671 43.137 init_scf_loop 11 6.9 0.000 0.000 34.589 34.590 mp_waitall_1 170520 16.6 25.040 34.093 25.040 34.093 prepare_preconditioner 11 7.9 0.000 0.000 30.611 30.660 make_preconditioner 11 8.9 0.000 0.000 30.611 30.660 rebuild_ks_matrix 128 8.3 0.001 0.001 30.183 30.654 qs_ks_build_kohn_sham_matrix 128 9.3 0.016 0.017 30.183 30.653 make_full_inverse_cholesky 11 9.9 0.000 0.000 28.304 29.670 qs_ks_update_qs_env 128 7.6 0.001 0.001 27.228 27.656 multiply_cannon_multrec 20056 15.6 12.973 16.639 22.310 25.898 multiply_cannon_metrocomm3 20056 15.6 0.062 0.066 15.623 24.650 qs_ot_get_derivative 117 11.6 0.002 0.002 22.965 23.445 apply_preconditioner_dbcsr 128 12.6 0.000 0.000 19.838 20.952 apply_single 128 13.6 0.001 0.001 19.837 20.952 qs_ot_get_p 128 10.4 0.002 0.002 19.531 20.055 ot_diis_step 117 11.6 0.018 0.019 19.593 19.594 make_m2s 5014 13.6 0.080 0.087 14.891 15.692 make_images 5014 14.6 1.159 1.246 14.659 15.464 multiply_cannon_sync_h2d 20056 15.6 13.548 15.462 13.548 15.462 qs_ot_p2m_diag 83 11.4 0.265 0.272 15.264 15.273 cp_fm_cholesky_invert 11 10.9 14.320 14.328 14.320 14.328 cp_dbcsr_syevd 83 12.4 0.005 0.006 14.199 14.200 sum_up_and_integrate 128 10.3 0.132 0.145 14.040 14.069 integrate_v_rspace 128 11.3 0.004 0.004 13.908 13.936 qs_rho_update_rho_low 128 7.7 0.001 0.001 13.570 13.604 calculate_rho_elec 128 8.7 0.130 0.145 13.569 13.604 init_scf_run 11 5.9 0.000 0.001 13.585 13.585 scf_env_initial_rho_setup 11 6.9 0.001 0.001 13.584 13.585 cp_fm_diag_elpa 83 13.4 0.000 0.001 10.935 10.953 cp_fm_diag_elpa_base 83 14.4 10.529 10.691 10.932 10.950 make_images_data 5014 15.6 0.062 0.071 8.899 10.266 multiply_cannon_metrocomm4 17549 15.6 0.067 0.077 3.531 9.572 qs_ot_get_derivative_diag 77 12.4 0.002 0.002 9.044 9.396 hybrid_alltoall_any 5200 16.5 0.450 2.041 7.693 9.373 mp_irecv_dv 50230 16.2 3.402 9.304 3.402 9.304 dbcsr_mm_accdrv_process 41502 16.2 5.623 5.967 8.801 8.960 pw_transfer 1547 11.6 0.084 0.100 8.691 8.793 fft_wrap_pw1pw2 1291 12.7 0.010 0.011 8.468 8.574 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 8.507 8.514 grid_integrate_task_list 128 12.3 7.301 7.874 7.301 7.874 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 7.599 7.749 fft_wrap_pw1pw2_140 523 13.2 1.312 1.360 7.508 7.626 wfi_extrapolate 11 7.9 0.001 0.001 7.431 7.432 cp_fm_cholesky_decompose 22 10.9 7.272 7.289 7.272 7.289 density_rs2pw 128 9.7 0.006 0.006 6.943 7.275 cp_fm_upper_to_full 105 14.8 5.622 7.105 5.622 7.105 dbcsr_complete_redistribute 395 12.7 1.171 1.198 4.558 6.235 calculate_first_density_matrix 1 7.0 0.000 0.001 5.964 5.966 calculate_dm_sparse 128 9.5 0.001 0.001 5.846 5.928 fft3d_ps 1291 14.7 2.757 2.972 5.832 5.893 grid_collocate_task_list 128 9.7 5.157 5.515 5.157 5.515 qs_ot_get_derivative_taylor 40 13.0 0.001 0.001 4.630 5.309 copy_fm_to_dbcsr 209 11.7 0.002 0.002 3.378 5.054 mp_alltoall_d11v 2415 14.1 4.279 4.971 4.279 4.971 acc_transpose_blocks 20056 15.6 0.100 0.104 4.806 4.921 mp_sum_l 7950 12.9 3.214 4.469 3.214 4.469 potential_pw2rs 128 12.3 0.021 0.023 4.429 4.442 mp_allgather_i34 2507 14.6 1.215 4.275 1.215 4.275 qs_ot_get_orbitals 117 10.6 0.001 0.001 3.982 4.013 transfer_fm_to_dbcsr 11 9.9 0.020 0.023 2.287 3.955 acc_transpose_blocks_kernels 20056 16.6 0.288 0.296 3.786 3.818 qs_energies_init_hamiltonians 11 5.9 0.000 0.001 3.734 3.735 mp_alltoall_i22 716 14.1 1.900 3.698 1.900 3.698 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="404", plot="h2o_256_md", label="(8n/4r/3t)", y=178.698000, yerr=0.000000 PlotPoint: name="405", plot="h2o_256_md_mem", label="(8n/4r/3t)", y=884.454545, yerr=7.165113 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/18/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1420242647040 0.0% 0.0% 100.0% flops 32 x 32 x 32 1943472701440 0.0% 0.0% 100.0% flops 22 x 9 x 32 1972057190400 0.0% 0.0% 100.0% flops 9 x 22 x 32 1977770336256 0.0% 0.0% 100.0% flops 22 x 22 x 32 2734287699968 0.0% 0.0% 100.0% flops 32 x 32 x 9 4416300122112 0.0% 0.0% 100.0% flops 32 x 32 x 22 5397700149248 0.0% 0.0% 100.0% flops 9 x 32 x 32 5443971710976 0.0% 0.0% 100.0% flops 22 x 32 x 32 6653743202304 0.0% 0.0% 100.0% flops 9 x 32 x 9 11528903135232 0.0% 0.0% 100.0% flops 22 x 32 x 9 15129160814592 0.0% 0.0% 100.0% flops 9 x 32 x 22 15129160814592 0.0% 0.0% 100.0% flops 22 x 32 x 22 19767995056128 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 93.514766E+12 0.0% 0.0% 100.0% flops max/rank 4.353791E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 6755941440 0.0% 0.0% 100.0% number of processed stacks 5977344 0.0% 0.0% 100.0% average stack size 0.0 0.0 1130.3 marketing flops 144.580175E+12 ------------------------------------------------------------------------------- # multiplications 2507 max memory usage/rank 1.151959E+09 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 1143192 MPI messages size (bytes): total size 2.023815E+12 min size 0.000000E+00 max size 17.653760E+06 average size 1.770320E+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 319024 36042702848 131072 < size <= 4194304 715736 785529176064 4194304 < size <= 16777216 70320 665379475120 16777216 < size 30720 536870912000 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 4002 58205. MP_Allreduce 11082 1082. MP_Sync 87 MP_Alltoall 1712 12503107. MP_SendRecv 5888 75008. MP_ISendRecv 5888 75008. MP_Wait 22442 MP_ISend 14952 244818. MP_IRecv 14952 244818. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.036 0.119 196.660 196.664 qs_mol_dyn_low 1 2.0 0.004 0.015 196.153 196.168 qs_forces 11 3.9 0.006 0.007 194.175 194.183 qs_energies 11 4.9 0.003 0.016 187.006 187.016 scf_env_do_scf 11 5.9 0.001 0.001 166.137 166.149 velocity_verlet 10 3.0 0.004 0.005 128.046 128.126 scf_env_do_scf_inner_loop 117 6.6 0.003 0.009 119.646 119.648 dbcsr_multiply_generic 2507 12.6 0.188 0.196 85.624 86.416 qs_scf_new_mos 117 7.6 0.001 0.001 83.326 83.614 qs_scf_loop_do_ot 117 8.6 0.001 0.001 83.325 83.613 ot_scf_mini 117 9.6 0.003 0.004 78.570 78.880 multiply_cannon 2507 13.6 0.552 0.584 59.840 62.707 multiply_cannon_loop 2507 14.6 1.795 1.938 56.088 58.030 init_scf_loop 11 6.9 0.000 0.000 46.365 46.366 ot_mini 117 10.6 0.001 0.001 43.785 44.085 prepare_preconditioner 11 7.9 0.000 0.000 42.241 42.265 make_preconditioner 11 8.9 0.000 0.001 42.241 42.265 make_full_inverse_cholesky 11 9.9 0.014 0.023 35.864 40.877 multiply_cannon_multrec 30084 15.6 13.577 18.919 26.525 31.659 rebuild_ks_matrix 128 8.3 0.001 0.001 29.340 29.626 qs_ks_build_kohn_sham_matrix 128 9.3 0.017 0.019 29.340 29.626 mp_waitall_1 147882 16.7 18.398 28.311 18.398 28.311 qs_ks_update_qs_env 128 7.6 0.001 0.001 26.459 26.711 qs_ot_get_derivative 117 11.6 0.002 0.002 23.919 24.230 make_m2s 5014 13.6 0.097 0.102 21.226 22.136 make_images 5014 14.6 2.001 2.504 20.916 21.825 qs_ot_get_p 128 10.4 0.002 0.002 20.800 21.176 apply_preconditioner_dbcsr 128 12.6 0.000 0.001 19.295 19.958 apply_single 128 13.6 0.001 0.001 19.294 19.958 ot_diis_step 117 11.6 0.018 0.018 19.748 19.750 cp_fm_upper_to_full 105 14.8 11.414 16.829 11.414 16.829 qs_ot_p2m_diag 83 11.4 0.342 0.389 16.423 16.472 cp_fm_cholesky_invert 11 10.9 16.161 16.170 16.161 16.170 multiply_cannon_metrocomm3 30084 15.6 0.049 0.052 6.832 15.560 cp_dbcsr_syevd 83 12.4 0.005 0.006 15.012 15.014 init_scf_run 11 5.9 0.000 0.001 14.915 14.917 scf_env_initial_rho_setup 11 6.9 0.001 0.002 14.915 14.916 sum_up_and_integrate 128 10.3 0.140 0.151 14.080 14.112 integrate_v_rspace 128 11.3 0.004 0.004 13.940 13.976 qs_rho_update_rho_low 128 7.7 0.001 0.001 13.803 13.859 calculate_rho_elec 128 8.7 0.173 0.189 13.802 13.858 dbcsr_mm_accdrv_process 62264 16.2 8.453 9.313 12.509 13.069 dbcsr_complete_redistribute 395 12.7 1.515 1.635 9.175 12.964 make_images_data 5014 15.6 0.066 0.072 11.199 12.663 cp_fm_diag_elpa 83 13.4 0.000 0.000 11.711 11.728 cp_fm_diag_elpa_base 83 14.4 10.746 11.076 11.705 11.720 copy_fm_to_dbcsr 209 11.7 0.001 0.002 7.777 11.557 multiply_cannon_sync_h2d 30084 15.6 10.495 11.438 10.495 11.438 hybrid_alltoall_any 5200 16.5 0.534 2.218 10.106 11.385 qs_ot_get_derivative_diag 77 12.4 0.002 0.002 10.184 10.417 transfer_fm_to_dbcsr 11 9.9 0.001 0.005 6.355 10.028 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 9.578 9.648 mp_alltoall_i22 716 14.1 5.636 9.334 5.636 9.334 pw_transfer 1547 11.6 0.085 0.099 8.958 9.044 fft_wrap_pw1pw2 1291 12.7 0.010 0.011 8.733 8.821 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 8.404 8.503 grid_integrate_task_list 128 12.3 7.504 7.872 7.504 7.872 fft_wrap_pw1pw2_140 523 13.2 1.453 1.483 7.766 7.870 wfi_extrapolate 11 7.9 0.001 0.001 7.837 7.837 cp_fm_cholesky_decompose 22 10.9 7.515 7.611 7.515 7.611 multiply_cannon_metrocomm4 25070 15.6 0.084 0.096 2.834 7.325 mp_irecv_dv 76098 16.2 2.679 7.039 2.679 7.039 acc_transpose_blocks 30084 15.6 0.150 0.155 5.864 7.032 density_rs2pw 128 9.7 0.006 0.006 6.767 7.026 calculate_first_density_matrix 1 7.0 0.000 0.000 6.881 6.883 calculate_dm_sparse 128 9.5 0.001 0.001 6.471 6.553 mp_alltoall_d11v 2415 14.1 5.175 6.082 5.175 6.082 fft3d_ps 1291 14.7 2.844 2.935 5.901 5.965 grid_collocate_task_list 128 9.7 5.291 5.758 5.291 5.758 acc_transpose_blocks_kernels 30084 16.6 0.404 0.412 3.926 4.880 qs_ot_get_derivative_taylor 40 13.0 0.001 0.001 4.605 4.706 jit_kernel_transpose 5 15.6 3.521 4.468 3.521 4.468 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 4.447 4.451 qs_ot_get_orbitals 117 10.6 0.001 0.001 4.299 4.382 potential_pw2rs 128 12.3 0.023 0.023 4.293 4.305 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="406", plot="h2o_256_md", label="(8n/3r/4t)", y=196.664000, yerr=0.000000 PlotPoint: name="407", plot="h2o_256_md_mem", label="(8n/3r/4t)", y=1086.090909, yerr=20.429277 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/19/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 9 x 9 x 32 1410024443904 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 11444712984576 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 92.796586E+12 0.0% 0.0% 100.0% flops max/rank 5.820059E+12 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 6705502176 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 1.543418E+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 716108700816 ------------------------------------------------------------------------------- - - - DBCSR MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Bcast 14 12. MP_Allreduce 8852 52. MP_Alltoall 9584 804353. MP_ISend 39716 2104723. MP_IRecv 39716 2103824. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 4003 58187. MP_Allreduce 11085 1167. MP_Sync 86 MP_Alltoall 1700 18828148. 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 0.027 0.056 175.250 175.250 qs_mol_dyn_low 1 2.0 0.003 0.003 174.520 174.843 qs_forces 11 3.9 0.004 0.004 174.406 174.407 qs_energies 11 4.9 0.001 0.002 167.032 167.040 scf_env_do_scf 11 5.9 0.001 0.001 145.459 145.472 velocity_verlet 10 3.0 0.002 0.002 111.315 111.319 scf_env_do_scf_inner_loop 116 6.6 0.003 0.008 109.803 109.805 dbcsr_multiply_generic 2485 12.5 0.179 0.190 76.497 76.945 qs_scf_new_mos 116 7.6 0.001 0.001 73.275 73.368 qs_scf_loop_do_ot 116 8.6 0.001 0.001 73.275 73.367 ot_scf_mini 116 9.6 0.003 0.004 68.855 68.937 multiply_cannon 2485 13.5 0.556 0.576 57.934 61.308 multiply_cannon_loop 2485 14.5 0.804 0.837 54.783 55.784 ot_mini 116 10.6 0.001 0.001 37.793 37.896 init_scf_loop 11 6.9 0.000 0.000 35.500 35.501 mp_waitall_1 124680 16.7 26.330 32.777 26.330 32.777 prepare_preconditioner 11 7.9 0.000 0.000 31.644 31.679 make_preconditioner 11 8.9 0.000 0.000 31.644 31.679 make_full_inverse_cholesky 11 9.9 0.014 0.026 29.552 29.822 rebuild_ks_matrix 127 8.3 0.001 0.001 28.876 29.018 qs_ks_build_kohn_sham_matrix 127 9.3 0.017 0.018 28.875 29.018 qs_ks_update_qs_env 127 7.6 0.001 0.001 26.332 26.462 multiply_cannon_multrec 9940 15.5 10.239 15.501 17.811 22.352 multiply_cannon_metrocomm3 9940 15.5 0.025 0.027 13.785 21.486 qs_ot_get_derivative 116 11.6 0.001 0.002 20.596 20.683 qs_ot_get_p 127 10.4 0.002 0.002 18.400 18.533 cp_fm_cholesky_invert 11 10.9 18.110 18.118 18.110 18.118 apply_preconditioner_dbcsr 127 12.6 0.000 0.000 17.285 17.495 apply_single 127 13.6 0.001 0.001 17.285 17.495 ot_diis_step 116 11.6 0.020 0.020 17.123 17.124 make_m2s 4970 13.5 0.066 0.069 14.634 15.532 make_images 4970 14.5 2.155 2.553 14.332 15.222 init_scf_run 11 5.9 0.000 0.001 14.741 14.742 scf_env_initial_rho_setup 11 6.9 0.001 0.001 14.741 14.741 qs_ot_p2m_diag 82 11.4 0.489 0.495 14.575 14.590 sum_up_and_integrate 127 10.3 0.180 0.189 14.126 14.173 integrate_v_rspace 127 11.3 0.004 0.004 13.946 13.998 qs_rho_update_rho_low 127 7.7 0.001 0.001 13.930 13.950 calculate_rho_elec 127 8.7 0.254 0.265 13.929 13.949 cp_dbcsr_syevd 82 12.4 0.005 0.006 13.405 13.406 multiply_cannon_sync_h2d 9940 15.5 10.637 11.242 10.637 11.242 cp_fm_diag_elpa 82 13.4 0.000 0.000 10.505 10.515 cp_fm_diag_elpa_base 82 14.4 10.269 10.344 10.500 10.511 make_images_data 4970 15.5 0.055 0.064 8.364 9.673 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 9.514 9.526 hybrid_alltoall_any 5155 16.4 0.838 3.767 8.234 9.414 pw_transfer 1535 11.6 0.084 0.094 9.217 9.244 fft_wrap_pw1pw2 1281 12.7 0.010 0.011 8.996 9.029 multiply_cannon_metrocomm1 9940 15.5 0.029 0.030 4.914 8.852 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 8.408 8.455 qs_ot_get_derivative_diag 76 12.4 0.002 0.003 8.053 8.129 grid_integrate_task_list 127 12.3 7.717 8.026 7.717 8.026 fft_wrap_pw1pw2_140 519 13.2 1.802 1.830 7.960 7.997 cp_fm_cholesky_decompose 22 10.9 7.877 7.974 7.877 7.974 dbcsr_mm_accdrv_process 20590 16.1 3.140 4.452 7.227 7.907 wfi_extrapolate 11 7.9 0.001 0.001 7.343 7.343 calculate_first_density_matrix 1 7.0 0.000 0.000 7.150 7.153 density_rs2pw 127 9.7 0.005 0.006 6.693 7.077 calculate_dm_sparse 127 9.5 0.001 0.001 6.102 6.187 acc_transpose_blocks 9940 15.5 0.052 0.053 4.657 5.995 grid_collocate_task_list 127 9.7 5.520 5.813 5.520 5.813 fft3d_ps 1281 14.7 2.738 2.802 5.733 5.774 dbcsr_complete_redistribute 393 12.7 2.114 2.163 5.244 5.574 multiply_cannon_metrocomm4 7455 15.5 0.026 0.028 1.854 5.416 mp_irecv_dv 28618 15.9 1.814 5.325 1.814 5.325 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 5.253 5.253 mp_alltoall_d11v 2401 14.1 4.610 5.048 4.610 5.048 acc_transpose_blocks_kernels 9940 16.5 0.149 0.155 3.694 5.005 jit_kernel_transpose 5 15.6 3.545 4.855 3.545 4.855 mp_allgather_i34 2485 14.5 1.338 4.612 1.338 4.612 potential_pw2rs 127 12.3 0.026 0.027 4.250 4.264 build_core_hamiltonian_matrix_ 11 4.9 0.001 0.001 3.591 3.876 qs_ot_get_derivative_taylor 40 13.0 0.001 0.001 3.644 3.730 copy_fm_to_dbcsr 208 11.6 0.001 0.002 3.382 3.715 qs_ot_get_orbitals 116 10.6 0.001 0.001 3.646 3.696 copy_dbcsr_to_fm 185 11.7 0.004 0.004 3.493 3.572 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="408", plot="h2o_256_md", label="(8n/2r/6t)", y=175.250000, yerr=0.000000 PlotPoint: name="409", plot="h2o_256_md_mem", label="(8n/2r/6t)", y=1444.454545, yerr=37.657219 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/20/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 11.606412E+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 1947808 0.0% 0.0% 100.0% average stack size 0.0 0.0 3442.6 marketing flops 143.507742E+12 ------------------------------------------------------------------------------- # multiplications 2485 max memory usage/rank 3.174506E+09 # max total images/rank 2 # max 3D layers 1 # MPI messages exchanged 99400 MPI messages size (bytes): total size 1.127422E+12 min size 0.000000E+00 max size 104.857600E+06 average size 11.342272E+06 MPI breakdown and total messages size (bytes): size <= 128 572 0 128 < size <= 8192 0 0 8192 < size <= 32768 0 0 32768 < size <= 131072 44 2883584 131072 < size <= 4194304 44768 34745614336 4194304 < size <= 16777216 43984 376564613120 16777216 < size 10032 716108490000 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 3991 59293. MP_Allreduce 11055 1504. MP_Sync 86 MP_Alltoall 1700 36954339. MP_SendRecv 1778 218624. MP_ISendRecv 1778 218624. MP_Wait 9728 MP_ISend 6360 1080477. MP_IRecv 6360 1080477. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.025 0.044 293.121 293.122 qs_mol_dyn_low 1 2.0 0.003 0.003 292.599 292.611 qs_forces 11 3.9 0.005 0.006 292.500 292.504 qs_energies 11 4.9 0.002 0.002 283.603 283.614 scf_env_do_scf 11 5.9 0.001 0.001 258.272 258.278 velocity_verlet 10 3.0 0.011 0.013 209.150 209.157 scf_env_do_scf_inner_loop 116 6.6 0.004 0.009 134.356 134.358 init_scf_loop 11 6.9 0.000 0.000 123.640 123.642 prepare_preconditioner 11 7.9 0.000 0.000 118.764 118.786 make_preconditioner 11 8.9 0.000 0.000 118.764 118.786 make_full_inverse_cholesky 11 9.9 0.036 0.039 95.117 115.832 qs_scf_new_mos 116 7.6 0.001 0.001 90.984 91.053 qs_scf_loop_do_ot 116 8.6 0.001 0.001 90.983 91.052 ot_scf_mini 116 9.6 0.004 0.004 86.095 86.136 dbcsr_multiply_generic 2485 12.5 0.211 0.223 85.017 85.539 cp_fm_upper_to_full 104 14.8 52.482 75.000 52.482 75.000 multiply_cannon 2485 13.5 0.679 0.744 61.489 62.889 multiply_cannon_loop 2485 14.5 1.047 1.057 57.356 58.756 ot_mini 116 10.6 0.001 0.001 44.783 44.848 dbcsr_complete_redistribute 393 12.7 4.034 4.122 29.790 42.497 copy_fm_to_dbcsr 208 11.6 0.001 0.002 26.203 38.916 transfer_fm_to_dbcsr 11 9.9 0.030 0.031 23.603 36.150 mp_alltoall_i22 712 14.1 21.392 34.340 21.392 34.340 rebuild_ks_matrix 127 8.3 0.001 0.001 34.097 34.137 qs_ks_build_kohn_sham_matrix 127 9.3 0.017 0.017 34.096 34.136 cp_fm_cholesky_invert 11 10.9 33.441 33.449 33.441 33.449 mp_waitall_1 102768 16.8 28.296 32.502 28.296 32.502 qs_ks_update_qs_env 127 7.6 0.001 0.001 31.607 31.652 qs_ot_get_p 127 10.4 0.002 0.002 26.436 26.528 qs_ot_get_derivative 116 11.6 0.002 0.002 24.803 24.848 qs_ot_p2m_diag 82 11.4 0.869 0.875 22.256 22.286 cp_dbcsr_syevd 82 12.4 0.006 0.006 20.448 20.450 make_m2s 4970 13.5 0.077 0.080 18.829 20.387 multiply_cannon_metrocomm3 9940 15.5 0.025 0.026 18.815 20.236 ot_diis_step 116 11.6 0.021 0.022 19.936 19.936 make_images 4970 14.5 3.050 3.257 18.354 19.913 apply_preconditioner_dbcsr 127 12.6 0.000 0.000 19.264 19.430 apply_single 127 13.6 0.001 0.001 19.264 19.430 multiply_cannon_multrec 9940 15.5 10.239 12.065 17.970 18.030 cp_fm_diag_elpa 82 13.4 0.000 0.000 17.249 17.249 cp_fm_diag_elpa_base 82 14.4 12.895 14.479 17.245 17.245 qs_rho_update_rho_low 127 7.7 0.001 0.001 16.425 16.432 calculate_rho_elec 127 8.7 0.477 0.478 16.424 16.431 sum_up_and_integrate 127 10.3 0.319 0.321 16.220 16.309 integrate_v_rspace 127 11.3 0.004 0.004 15.901 15.988 init_scf_run 11 5.9 0.000 0.001 14.978 14.979 scf_env_initial_rho_setup 11 6.9 0.003 0.003 14.977 14.978 multiply_cannon_sync_h2d 9940 15.5 14.182 14.209 14.182 14.209 hybrid_alltoall_any 5155 16.4 1.303 3.056 10.923 13.135 make_images_data 4970 15.5 0.064 0.069 10.731 13.001 pw_transfer 1535 11.6 0.091 0.092 11.355 11.372 fft_wrap_pw1pw2 1281 12.7 0.011 0.011 11.122 11.137 fft_wrap_pw1pw2_140 519 13.2 3.039 3.091 9.904 9.925 qs_ot_get_derivative_diag 76 12.4 0.002 0.002 9.789 9.817 dbcsr_mm_accdrv_process 20590 16.0 4.342 6.046 7.483 9.300 cp_dbcsr_sm_fm_multiply 37 9.5 0.002 0.002 9.194 9.252 cp_fm_cholesky_decompose 22 10.9 8.950 9.059 8.950 9.059 wfi_extrapolate 11 7.9 0.001 0.001 9.058 9.058 grid_integrate_task_list 127 12.3 8.489 8.688 8.489 8.688 qs_energies_init_hamiltonians 11 5.9 0.001 0.001 8.115 8.118 density_rs2pw 127 9.7 0.005 0.006 7.503 7.556 cp_dbcsr_sm_fm_multiply_core 37 10.5 0.000 0.000 7.131 7.179 mp_alltoall_d11v 2401 14.1 7.007 7.155 7.007 7.155 calculate_dm_sparse 127 9.5 0.001 0.001 6.775 6.831 fft3d_ps 1281 14.7 2.825 2.836 6.391 6.460 grid_collocate_task_list 127 9.7 6.364 6.437 6.364 6.437 copy_dbcsr_to_fm 185 11.7 0.004 0.004 6.276 6.373 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="410", plot="h2o_256_md", label="(8n/1r/12t)", y=293.122000, yerr=0.000000 PlotPoint: name="411", plot="h2o_256_md_mem", label="(8n/1r/12t)", y=2882.272727, yerr=152.792843 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 1.259401E+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 255646. MP_Allreduce 3139 6114. MP_Sync 4 MP_Alltoall 54 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 0.038 0.098 88.363 88.365 qs_energies 1 2.0 0.000 0.001 87.812 87.822 ls_scf 1 3.0 0.000 0.000 86.914 86.924 dbcsr_multiply_generic 111 6.7 0.014 0.015 75.950 76.112 multiply_cannon 111 7.7 0.018 0.020 58.812 60.205 multiply_cannon_loop 111 8.7 0.222 0.242 55.373 56.936 ls_scf_main 1 4.0 0.000 0.000 52.320 52.321 density_matrix_trs4 2 5.0 0.002 0.003 46.789 46.864 ls_scf_init_scf 1 4.0 0.000 0.000 31.579 31.580 ls_scf_init_matrix_S 1 5.0 0.000 0.000 30.452 30.503 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.001 28.357 28.381 mp_waitall_1 11031 10.9 22.549 26.544 22.549 26.544 multiply_cannon_multrec 2664 9.7 8.158 8.965 15.638 17.423 multiply_cannon_sync_h2d 2664 9.7 13.647 15.398 13.647 15.398 make_m2s 222 7.7 0.009 0.011 13.180 13.606 make_images 222 8.7 0.098 0.108 13.158 13.587 multiply_cannon_metrocomm1 2664 9.7 0.010 0.011 9.555 12.153 multiply_cannon_metrocomm3 2664 9.7 0.009 0.011 5.619 10.193 make_images_data 222 9.7 0.004 0.005 7.707 8.333 hybrid_alltoall_any 227 10.6 0.215 1.843 6.571 8.142 dbcsr_mm_accdrv_process 4760 10.4 0.586 0.695 7.099 8.060 dbcsr_mm_accdrv_process_sort 4760 11.4 6.312 7.204 6.312 7.204 calculate_norms 4752 9.8 5.529 6.161 5.529 6.161 apply_matrix_preconditioner 6 5.3 0.000 0.000 5.010 5.141 mp_sum_l 887 5.1 3.472 4.974 3.472 4.974 compute_matrix_preconditioner 1 6.0 0.000 0.001 4.257 4.261 dbcsr_multiply_generic_mpsum_f 86 7.8 0.000 0.000 2.649 3.879 make_images_sizes 222 9.7 0.000 0.000 0.788 3.651 mp_alltoall_i44 222 10.7 0.788 3.650 0.788 3.650 multiply_cannon_metrocomm4 2442 9.7 0.012 0.015 2.035 3.358 mp_irecv_dv 6231 10.9 2.019 3.331 2.019 3.331 acc_transpose_blocks 2664 9.7 0.018 0.021 3.018 3.317 arnoldi_extremal 4 6.8 0.000 0.000 3.237 3.274 arnoldi_normal_ev 4 7.8 0.001 0.003 3.237 3.274 acc_transpose_blocks_kernels 2664 10.7 0.035 0.040 2.884 3.175 build_subspace 16 8.4 0.009 0.012 3.140 3.143 jit_kernel_transpose 1 13.0 2.849 3.140 2.849 3.140 ls_scf_post 1 4.0 0.000 0.000 3.015 3.025 ls_scf_store_result 1 5.0 0.000 0.000 2.839 2.871 dbcsr_special_finalize 555 9.7 0.005 0.006 2.333 2.775 dbcsr_merge_single_wm 555 10.7 0.471 0.602 2.324 2.766 make_images_pack 222 9.7 2.206 2.620 2.208 2.622 dbcsr_matrix_vector_mult 304 9.0 0.006 0.014 2.316 2.555 dbcsr_sort_data 658 11.4 2.111 2.478 2.111 2.478 dbcsr_matrix_vector_mult_local 304 10.0 2.068 2.460 2.070 2.461 ls_scf_dm_to_ks 2 5.0 0.000 0.000 2.313 2.400 buffer_matrices_ensure_size 222 8.7 1.788 2.117 1.788 2.117 qs_ks_update_qs_env 3 6.3 0.000 0.000 1.772 1.773 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="500", plot="h2o_32_nrep3_ls", label="(8n/12r/1t)", y=88.365000, yerr=0.000000 PlotPoint: name="501", plot="h2o_32_nrep3_ls_mem", label="(8n/12r/1t)", y=1140.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 2.191454E+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 266673. MP_Allreduce 3138 10075. 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 0.025 0.057 92.633 92.634 qs_energies 1 2.0 0.000 0.000 92.189 92.199 ls_scf 1 3.0 0.000 0.000 90.859 90.869 dbcsr_multiply_generic 111 6.7 0.015 0.016 77.030 77.320 multiply_cannon 111 7.7 0.027 0.036 55.334 60.081 multiply_cannon_loop 111 8.7 0.135 0.145 52.564 56.579 ls_scf_main 1 4.0 0.000 0.000 54.448 54.455 density_matrix_trs4 2 5.0 0.002 0.003 48.814 49.006 ls_scf_init_scf 1 4.0 0.000 0.000 32.734 32.735 ls_scf_init_matrix_S 1 5.0 0.000 0.000 31.531 31.639 mp_waitall_1 9105 10.9 20.711 29.685 20.711 29.685 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.001 29.155 29.167 multiply_cannon_multrec 1332 9.7 13.346 16.923 22.762 27.457 multiply_cannon_metrocomm3 1332 9.7 0.007 0.008 11.410 21.074 make_m2s 222 7.7 0.006 0.008 15.021 15.547 make_images 222 8.7 1.368 1.696 14.990 15.518 dbcsr_mm_accdrv_process 4041 10.4 0.348 0.530 9.005 10.578 dbcsr_mm_accdrv_process_sort 4041 11.4 8.519 10.048 8.519 10.048 make_images_data 222 9.7 0.004 0.005 8.692 9.568 hybrid_alltoall_any 227 10.6 0.544 2.561 8.003 9.480 mp_sum_l 887 5.1 5.197 8.219 5.197 8.219 multiply_cannon_metrocomm4 1221 9.7 0.007 0.009 3.214 7.840 mp_irecv_dv 3311 11.0 3.194 7.781 3.194 7.781 calculate_norms 2376 9.8 6.075 6.772 6.075 6.772 dbcsr_multiply_generic_mpsum_f 86 7.8 0.000 0.000 3.973 6.684 multiply_cannon_sync_h2d 1332 9.7 4.926 6.647 4.926 6.647 apply_matrix_preconditioner 6 5.3 0.000 0.000 4.876 5.129 multiply_cannon_metrocomm1 1332 9.7 0.003 0.004 1.261 4.841 arnoldi_extremal 4 6.8 0.000 0.000 4.606 4.623 arnoldi_normal_ev 4 7.8 0.001 0.005 4.606 4.623 build_subspace 16 8.4 0.014 0.021 4.357 4.360 ls_scf_post 1 4.0 0.000 0.000 3.676 3.687 compute_matrix_preconditioner 1 6.0 0.000 0.001 3.527 3.533 ls_scf_store_result 1 5.0 0.000 0.000 3.386 3.487 dbcsr_matrix_vector_mult 304 9.0 0.010 0.023 3.129 3.354 dbcsr_matrix_vector_mult_local 304 10.0 2.747 3.225 2.749 3.227 acc_transpose_blocks 1332 9.7 0.008 0.009 2.651 3.007 mp_allgather_i34 111 8.7 0.756 2.947 0.756 2.947 acc_transpose_blocks_kernels 1332 10.7 0.020 0.022 2.574 2.926 jit_kernel_transpose 1 13.0 2.554 2.907 2.554 2.907 ls_scf_dm_to_ks 2 5.0 0.000 0.000 2.599 2.694 dbcsr_data_new 4174 10.1 2.112 2.402 2.112 2.402 make_images_pack 222 9.7 1.822 2.127 1.824 2.129 dbcsr_sort_data 436 11.2 1.828 2.083 1.828 2.083 qs_ks_update_qs_env 3 6.3 0.000 0.000 1.851 1.853 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="502", plot="h2o_32_nrep3_ls", label="(8n/6r/2t)", y=92.634000, yerr=0.000000 PlotPoint: name="503", plot="h2o_32_nrep3_ls_mem", label="(8n/6r/2t)", y=1774.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 2.866246E+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 265448. MP_Allreduce 3138 10896. 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 0.066 0.108 91.824 91.826 qs_energies 1 2.0 0.002 0.006 91.128 91.131 ls_scf 1 3.0 0.002 0.005 89.688 89.694 dbcsr_multiply_generic 111 6.7 0.016 0.021 74.346 74.608 ls_scf_main 1 4.0 0.000 0.006 55.637 55.643 multiply_cannon 111 7.7 0.036 0.096 51.699 55.607 multiply_cannon_loop 111 8.7 0.116 0.133 48.936 52.928 density_matrix_trs4 2 5.0 0.003 0.022 49.708 49.930 mp_waitall_1 7281 11.0 22.822 32.565 22.822 32.565 ls_scf_init_scf 1 4.0 0.000 0.002 30.316 30.319 ls_scf_init_matrix_S 1 5.0 0.000 0.000 29.085 29.185 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.002 26.736 26.753 multiply_cannon_multrec 888 9.7 12.662 15.532 21.339 24.850 multiply_cannon_metrocomm3 888 9.7 0.004 0.004 10.681 22.437 make_m2s 222 7.7 0.006 0.009 15.820 16.618 make_images 222 8.7 1.582 1.856 15.781 16.577 make_images_data 222 9.7 0.004 0.004 9.164 10.228 hybrid_alltoall_any 227 10.6 0.640 2.955 8.831 9.834 dbcsr_mm_accdrv_process 3754 10.4 0.293 0.490 8.219 9.498 dbcsr_mm_accdrv_process_sort 3754 11.4 7.783 9.008 7.783 9.008 mp_sum_l 887 5.1 5.159 8.787 5.159 8.787 multiply_cannon_sync_h2d 888 9.7 5.996 7.131 5.996 7.131 multiply_cannon_metrocomm4 777 9.7 0.004 0.005 2.447 7.037 dbcsr_multiply_generic_mpsum_f 86 7.8 0.000 0.000 4.007 7.014 mp_irecv_dv 2335 11.1 2.431 6.993 2.431 6.993 multiply_cannon_metrocomm1 888 9.7 0.003 0.003 3.621 6.875 arnoldi_extremal 4 6.8 0.000 0.000 5.068 5.089 arnoldi_normal_ev 4 7.8 0.001 0.006 5.067 5.089 apply_matrix_preconditioner 6 5.3 0.000 0.000 4.744 4.999 calculate_norms 1584 9.8 4.385 4.773 4.385 4.773 build_subspace 16 8.4 0.014 0.022 4.755 4.760 dbcsr_matrix_vector_mult 304 9.0 0.010 0.022 3.435 3.764 ls_scf_post 1 4.0 0.000 0.001 3.733 3.737 dbcsr_matrix_vector_mult_local 304 10.0 3.028 3.605 3.030 3.607 mp_allgather_i34 111 8.7 0.875 3.566 0.875 3.566 ls_scf_store_result 1 5.0 0.000 0.000 3.426 3.521 ls_scf_dm_to_ks 2 5.0 0.000 0.000 2.863 2.964 dbcsr_data_new 4116 9.9 2.109 2.486 2.109 2.486 dbcsr_sort_data 325 11.1 1.862 2.135 1.862 2.135 qs_ks_update_qs_env 3 6.3 0.000 0.000 1.903 1.905 rebuild_ks_matrix 3 7.3 0.000 0.000 1.885 1.887 qs_ks_build_kohn_sham_matrix 3 8.3 0.001 0.001 1.885 1.887 make_images_pack 222 9.7 1.626 1.872 1.628 1.875 make_images_sizes 222 9.7 0.000 0.000 0.828 1.863 mp_alltoall_i44 222 10.7 0.828 1.862 0.828 1.862 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="504", plot="h2o_32_nrep3_ls", label="(8n/4r/3t)", y=91.826000, yerr=0.000000 PlotPoint: name="505", plot="h2o_32_nrep3_ls_mem", label="(8n/4r/3t)", y=2202.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 3.354862E+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 266673. MP_Allreduce 3138 13030. 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 0.056 0.100 97.517 97.519 qs_energies 1 2.0 0.001 0.004 96.678 96.684 ls_scf 1 3.0 0.000 0.001 95.020 95.024 dbcsr_multiply_generic 111 6.7 0.017 0.018 78.772 79.059 ls_scf_main 1 4.0 0.003 0.018 58.881 58.882 multiply_cannon 111 7.7 0.039 0.076 51.948 56.874 density_matrix_trs4 2 5.0 0.004 0.005 52.762 52.928 multiply_cannon_loop 111 8.7 0.154 0.169 46.870 50.056 ls_scf_init_scf 1 4.0 0.000 0.002 32.892 32.894 ls_scf_init_matrix_S 1 5.0 0.000 0.000 31.627 31.706 mp_waitall_1 6369 11.0 22.737 30.078 22.737 30.078 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.002 29.009 29.021 multiply_cannon_multrec 1332 9.7 14.108 17.433 22.193 24.975 make_m2s 222 7.7 0.007 0.008 21.051 22.482 make_images 222 8.7 3.143 3.605 21.001 22.433 multiply_cannon_metrocomm3 1332 9.7 0.003 0.003 9.360 17.586 make_images_data 222 9.7 0.004 0.005 11.709 13.395 hybrid_alltoall_any 227 10.6 0.800 3.784 11.016 12.813 dbcsr_mm_accdrv_process 3641 10.4 0.286 0.474 7.726 9.313 dbcsr_mm_accdrv_process_sort 3641 11.4 7.221 8.731 7.221 8.731 mp_sum_l 887 5.1 4.217 7.471 4.217 7.471 multiply_cannon_sync_h2d 1332 9.7 5.465 6.079 5.465 6.079 multiply_cannon_metrocomm4 1110 9.7 0.005 0.007 2.079 5.945 mp_irecv_dv 3229 10.9 2.054 5.884 2.054 5.884 dbcsr_multiply_generic_mpsum_f 86 7.8 0.000 0.000 3.251 5.830 multiply_cannon_metrocomm1 1332 9.7 0.003 0.003 2.501 5.321 arnoldi_extremal 4 6.8 0.000 0.000 5.193 5.205 arnoldi_normal_ev 4 7.8 0.002 0.006 5.193 5.205 build_subspace 16 8.4 0.015 0.021 4.859 4.866 apply_matrix_preconditioner 6 5.3 0.000 0.000 4.682 4.840 mp_allgather_i34 111 8.7 2.186 4.565 2.186 4.565 calculate_norms 2376 9.8 4.171 4.519 4.171 4.519 dbcsr_matrix_vector_mult 304 9.0 0.011 0.022 3.571 3.872 dbcsr_matrix_vector_mult_local 304 10.0 3.198 3.695 3.200 3.696 dbcsr_sort_data 658 11.4 3.038 3.399 3.038 3.399 ls_scf_post 1 4.0 0.000 0.001 3.247 3.251 dbcsr_special_finalize 555 9.7 0.006 0.007 2.777 3.231 dbcsr_merge_single_wm 555 10.7 0.532 0.661 2.769 3.223 ls_scf_dm_to_ks 2 5.0 0.000 0.000 3.012 3.054 ls_scf_store_result 1 5.0 0.000 0.000 2.977 3.048 dbcsr_data_release 10477 10.7 1.582 2.395 1.582 2.395 dbcsr_finalize 304 7.8 0.049 0.061 1.806 1.992 qs_ks_update_qs_env 3 6.3 0.000 0.000 1.972 1.973 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="506", plot="h2o_32_nrep3_ls", label="(8n/3r/4t)", y=97.519000, yerr=0.000000 PlotPoint: name="507", plot="h2o_32_nrep3_ls_mem", label="(8n/3r/4t)", y=2771.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 4.745839E+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 265536. MP_Allreduce 3129 15263. 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 0.053 0.102 92.887 92.888 qs_energies 1 2.0 0.000 0.000 91.932 91.934 ls_scf 1 3.0 0.000 0.000 89.940 89.943 dbcsr_multiply_generic 111 6.7 0.017 0.019 71.513 71.717 ls_scf_main 1 4.0 0.000 0.000 57.139 57.139 multiply_cannon 111 7.7 0.096 0.177 52.994 56.010 multiply_cannon_loop 111 8.7 0.088 0.094 50.380 51.908 density_matrix_trs4 2 5.0 0.002 0.003 50.171 50.263 ls_scf_init_scf 1 4.0 0.000 0.000 29.316 29.317 mp_waitall_1 5436 11.0 24.796 29.007 24.796 29.007 ls_scf_init_matrix_S 1 5.0 0.000 0.000 28.063 28.101 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.001 25.970 25.981 multiply_cannon_multrec 444 9.7 13.762 16.204 20.950 22.410 multiply_cannon_metrocomm1 444 9.7 0.002 0.002 11.126 15.845 make_m2s 222 7.7 0.005 0.005 13.687 14.664 make_images 222 8.7 2.041 2.481 13.619 14.594 multiply_cannon_metrocomm3 444 9.7 0.001 0.002 6.034 14.519 hybrid_alltoall_any 227 10.6 0.801 3.822 8.253 9.965 make_images_data 222 9.7 0.003 0.004 8.453 9.767 multiply_cannon_sync_h2d 444 9.7 6.730 8.348 6.730 8.348 dbcsr_mm_accdrv_process 3003 10.4 0.353 0.400 6.868 7.978 dbcsr_mm_accdrv_process_sort 3003 11.4 6.502 7.581 6.502 7.581 arnoldi_extremal 4 6.8 0.000 0.000 5.840 5.851 arnoldi_normal_ev 4 7.8 0.002 0.006 5.839 5.851 build_subspace 16 8.4 0.015 0.020 5.447 5.455 apply_matrix_preconditioner 6 5.3 0.000 0.000 4.443 4.660 mp_sum_l 887 5.1 2.848 4.423 2.848 4.423 dbcsr_matrix_vector_mult 304 9.0 0.011 0.022 4.184 4.397 dbcsr_matrix_vector_mult_local 304 10.0 3.715 4.194 3.717 4.196 mp_allgather_i34 111 8.7 1.178 3.788 1.178 3.788 calculate_norms 792 9.8 3.618 3.737 3.618 3.737 multiply_cannon_metrocomm4 333 9.7 0.001 0.002 1.569 3.693 mp_irecv_dv 1241 11.2 1.553 3.646 1.553 3.646 ls_scf_dm_to_ks 2 5.0 0.000 0.000 3.526 3.597 dbcsr_multiply_generic_mpsum_f 86 7.8 0.000 0.000 2.000 3.538 ls_scf_post 1 4.0 0.000 0.000 3.485 3.487 make_images_sizes 222 9.7 0.000 0.000 0.845 3.325 mp_alltoall_i44 222 10.7 0.845 3.325 0.845 3.325 ls_scf_store_result 1 5.0 0.000 0.000 3.243 3.282 dbcsr_finalize 304 7.8 0.062 0.077 2.200 2.298 dbcsr_data_new 4608 9.7 1.791 2.152 1.791 2.152 dbcsr_merge_all 275 8.9 0.478 0.525 2.059 2.138 qs_ks_update_qs_env 3 6.3 0.000 0.000 2.030 2.031 rebuild_ks_matrix 3 7.3 0.000 0.000 1.998 1.999 qs_ks_build_kohn_sham_matrix 3 8.3 0.000 0.001 1.997 1.999 qs_energies_init_hamiltonians 1 3.0 0.000 0.001 1.976 1.976 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="508", plot="h2o_32_nrep3_ls", label="(8n/2r/6t)", y=92.888000, yerr=0.000000 PlotPoint: name="509", plot="h2o_32_nrep3_ls_mem", label="(8n/2r/6t)", y=3736.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_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 8.847688E+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 284089. MP_Allreduce 3123 21388. 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 0.120 0.169 115.361 115.362 qs_energies 1 2.0 0.000 0.000 113.936 113.940 ls_scf 1 3.0 0.000 0.000 111.009 111.013 dbcsr_multiply_generic 111 6.7 0.024 0.027 84.670 84.813 ls_scf_main 1 4.0 0.000 0.000 63.192 63.192 multiply_cannon 111 7.7 0.169 0.232 52.964 57.850 multiply_cannon_loop 111 8.7 0.097 0.099 49.969 55.969 density_matrix_trs4 2 5.0 0.002 0.003 54.256 54.306 ls_scf_init_scf 1 4.0 0.000 0.000 44.053 44.054 ls_scf_init_matrix_S 1 5.0 0.000 0.000 42.545 42.573 matrix_sqrt_Newton_Schulz 2 6.5 0.001 0.002 39.772 39.784 mp_waitall_1 4527 11.1 21.806 25.138 21.806 25.138 make_m2s 222 7.7 0.005 0.005 22.311 23.394 make_images 222 8.7 3.574 3.885 22.203 23.284 multiply_cannon_multrec 444 9.7 17.866 18.371 22.572 23.117 hybrid_alltoall_any 227 10.6 1.655 3.633 12.535 15.253 make_images_data 222 9.7 0.003 0.004 12.718 14.829 multiply_cannon_metrocomm3 444 9.7 0.001 0.001 10.501 13.993 compute_matrix_preconditioner 1 6.0 0.002 0.002 11.141 11.146 multiply_cannon_sync_h2d 444 9.7 8.798 8.851 8.798 8.851 mp_sum_l 887 5.1 6.989 8.364 6.989 8.364 arnoldi_extremal 4 6.8 0.000 0.000 7.443 7.455 arnoldi_normal_ev 4 7.8 0.003 0.009 7.443 7.455 dbcsr_multiply_generic_mpsum_f 86 7.8 0.000 0.000 6.399 7.442 build_subspace 16 8.4 0.026 0.036 6.895 6.906 acc_transpose_blocks 444 9.7 0.003 0.003 4.389 6.846 acc_transpose_blocks_kernels 444 10.7 0.006 0.006 4.307 6.766 jit_kernel_transpose 1 13.0 4.302 6.761 4.302 6.761 dbcsr_matrix_vector_mult 304 9.0 0.017 0.034 5.518 5.677 dbcsr_matrix_vector_mult_local 304 10.0 5.086 5.412 5.089 5.415 ls_scf_dm_to_ks 2 5.0 0.000 0.000 5.072 5.161 apply_matrix_preconditioner 6 5.3 0.000 0.000 4.854 5.099 dbcsr_mm_accdrv_process 1814 10.4 0.244 0.351 4.522 4.646 dbcsr_mm_accdrv_process_sort 1814 11.4 4.187 4.322 4.187 4.322 ls_scf_post 1 4.0 0.000 0.000 3.764 3.768 make_images_sizes 222 9.7 0.000 0.000 1.455 3.545 mp_alltoall_i44 222 10.7 1.454 3.545 1.454 3.545 ls_scf_store_result 1 5.0 0.000 0.000 3.478 3.515 mp_allgather_i34 111 8.7 1.104 3.482 1.104 3.482 calculate_norms 792 9.8 3.225 3.264 3.225 3.264 dbcsr_finalize 304 7.8 0.082 0.089 3.078 3.177 dbcsr_merge_all 275 8.9 0.885 0.915 2.863 2.956 qs_energies_init_hamiltonians 1 3.0 0.000 0.001 2.896 2.896 dbcsr_complete_redistribute 5 7.6 1.444 1.487 2.746 2.861 matrix_ls_to_qs 2 6.0 0.000 0.000 2.385 2.514 dbcsr_sort_data 325 11.1 2.437 2.506 2.437 2.506 qs_ks_update_qs_env 3 6.3 0.000 0.000 2.404 2.406 rebuild_ks_matrix 3 7.3 0.000 0.000 2.338 2.340 qs_ks_build_kohn_sham_matrix 3 8.3 0.000 0.001 2.338 2.340 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="510", plot="h2o_32_nrep3_ls", label="(8n/1r/12t)", y=115.362000, yerr=0.000000 PlotPoint: name="511", plot="h2o_32_nrep3_ls_mem", label="(8n/1r/12t)", y=7012.000000, yerr=0.000000 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ~~~~~~~~~ RESULT ~~~~~~~~~ RESULT file: /scratch/snx3000/mkrack/rt/../rt/CRAY-XC50-gnu/069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800_performance_tests/27/result.log @@@@@@@@@@ Run number: 1 @@@@@@@@@@ ------------------------------------------------------------------------------- - - - DBCSR STATISTICS - - - ------------------------------------------------------------------------------- COUNTER TOTAL BLAS SMM ACC flops 32 x 32 x 32 7009386627072 0.0% 0.0% 100.0% flops 9 x 9 x 32 7335108845568 0.0% 0.0% 100.0% flops 9 x 22 x 32 9866241589248 0.0% 0.0% 100.0% flops 22 x 9 x 32 9884108906496 0.0% 0.0% 100.0% flops 22 x 22 x 32 13354440523776 0.0% 0.0% 100.0% flops 32 x 32 x 9 20607185977344 0.0% 0.0% 100.0% flops 32 x 32 x 22 25186560638976 0.0% 0.0% 100.0% flops 9 x 32 x 32 28458319085568 0.0% 0.0% 100.0% flops 22 x 32 x 32 34782389993472 0.0% 0.0% 100.0% flops 9 x 32 x 9 42881542373376 0.0% 0.0% 100.0% flops 22 x 32 x 9 55680402235392 0.0% 0.0% 100.0% flops 9 x 32 x 22 55680402235392 0.0% 0.0% 100.0% flops 22 x 32 x 22 72328573419520 0.0% 0.0% 100.0% flops inhomo. stacks 0 0.0% 0.0% 0.0% flops total 383.054662E+12 0.0% 0.0% 100.0% flops max/rank 733.641090E+09 0.0% 0.0% 100.0% matmuls inhomo. stacks 0 0.0% 0.0% 0.0% matmuls total 26899403712 0.0% 0.0% 100.0% number of processed stacks 118860288 0.0% 0.0% 100.0% average stack size 0.0 0.0 226.3 marketing flops 780.439111E+12 ------------------------------------------------------------------------------- # multiplications 1445 max memory usage/rank 588.316672E+06 # max total images/rank 3 # max 3D layers 1 # MPI messages exchanged 102097920 MPI messages size (bytes): total size 37.227590E+12 min size 0.000000E+00 max size 4.551360E+06 average size 364.626312E+03 MPI breakdown and total messages size (bytes): size <= 128 731472 0 128 < size <= 8192 11922720 97670922240 8192 < size <= 32768 24718992 614677610496 32768 < size <= 131072 20000256 1970081366016 131072 < size <= 4194304 42515668 24886801223040 4194304 < size <= 16777216 2208812 9656099886720 16777216 < size 0 0 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - MESSAGE PASSING PERFORMANCE - - - ------------------------------------------------------------------------------- ROUTINE CALLS AVE VOLUME [Bytes] MP_Group 4 MP_Bcast 4640 78072. MP_Allreduce 13232 2081. MP_Sync 1064 MP_Alltoall 2588 4803606. MP_SendRecv 168740 11136. MP_ISendRecv 92040 11136. MP_Wait 102830 MP_comm_split 40 MP_ISend 26090 85106. MP_IRecv 37890 59644. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- - - - T I M I N G - - - ------------------------------------------------------------------------------- SUBROUTINE CALLS ASD SELF TIME TOTAL TIME MAXIMUM AVERAGE MAXIMUM AVERAGE MAXIMUM CP2K 1 1.0 0.055 0.187 226.576 226.580 qs_mol_dyn_low 1 2.0 0.005 0.039 224.872 224.895 qs_forces 5 3.8 0.008 0.084 224.748 224.778 qs_energies 5 4.8 0.007 0.047 221.668 221.692 scf_env_do_scf 5 5.8 0.001 0.008 206.607 206.610 scf_env_do_scf_inner_loop 105 6.6 0.010 0.047 180.448 180.449 qs_scf_new_mos 105 7.6 0.000 0.001 141.141 141.304 qs_scf_loop_do_ot 105 8.6 0.001 0.001 141.140 141.303 dbcsr_multiply_generic 1445 12.2 0.126 0.136 133.824 134.280 ot_scf_mini 105 9.6 0.003 0.007 131.216 131.333 multiply_cannon 1445 13.2 0.275 0.287 114.743 116.477 multiply_cannon_loop 1445 14.2 2.864 3.039 113.098 114.449 velocity_verlet 4 3.0 0.001 0.002 106.566 106.567 ot_mini 105 10.6 0.001 0.007 60.022 60.147 multiply_cannon_multrec 69360 15.2 29.650 34.567 39.566 44.208 mp_waitall_1 488190 16.1 35.415 44.039 35.415 44.039 qs_ot_get_p 112 10.4 0.001 0.002 41.119 41.403 qs_ot_get_derivative 55 11.6 0.001 0.001 38.371 38.514 multiply_cannon_metrocomm3 69360 15.2 0.202 0.219 26.100 35.101 multiply_cannon_sync_h2d 69360 15.2 29.057 33.054 29.057 33.054 qs_ot_p2m_diag 40 11.0 0.020 0.031 29.997 30.088 rebuild_ks_matrix 110 8.4 0.000 0.000 28.739 28.900 qs_ks_build_kohn_sham_matrix 110 9.4 0.012 0.020 28.739 28.900 cp_dbcsr_syevd 40 12.0 0.002 0.007 26.741 26.741 qs_ks_update_qs_env 112 7.6 0.001 0.001 26.317 26.459 init_scf_loop 7 6.6 0.001 0.007 26.130 26.134 apply_preconditioner_dbcsr 62 12.6 0.000 0.001 22.995 23.213 apply_single 62 13.6 0.000 0.000 22.995 23.213 cp_fm_syevd 40 13.0 0.000 0.002 21.722 21.872 prepare_preconditioner 7 7.6 0.000 0.000 21.314 21.350 make_preconditioner 7 8.6 0.000 0.003 21.314 21.350 ot_new_cg_direction 55 11.6 0.001 0.002 20.973 20.974 cp_fm_redistribute_end 40 14.0 8.555 17.071 8.559 17.072 cp_fm_syevd_base 40 14.0 8.506 17.024 8.506 17.024 qs_rho_update_rho_low 110 7.6 0.001 0.005 16.372 16.791 calculate_rho_elec 110 8.6 0.029 0.032 16.371 16.790 qs_ot_get_orbitals 105 10.6 0.001 0.001 14.811 14.979 make_full_inverse_cholesky 7 9.6 0.000 0.002 14.104 14.161 qs_ot_get_derivative_taylor 37 12.8 0.001 0.001 14.011 14.117 mp_sum_l 4764 12.2 12.435 13.193 12.435 13.193 init_scf_run 5 5.8 0.000 0.001 12.296 12.297 scf_env_initial_rho_setup 5 6.8 0.001 0.003 12.295 12.296 pw_transfer 1645 12.4 0.079 0.098 11.661 11.848 fft_wrap_pw1pw2 1425 13.5 0.013 0.016 11.522 11.713 calculate_dm_sparse 110 9.5 0.000 0.001 11.515 11.655 density_rs2pw 110 9.6 0.005 0.007 10.704 11.124 dbcsr_mm_accdrv_process 154766 15.8 6.179 6.390 9.785 10.726 qs_vxc_create 110 10.4 0.003 0.005 10.413 10.464 qs_ot_get_derivative_diag 18 12.0 0.000 0.001 10.376 10.434 fft_wrap_pw1pw2_240 915 15.0 1.150 1.237 10.171 10.346 cp_fm_cholesky_invert 7 10.6 9.862 9.870 9.862 9.870 check_diag 80 13.5 8.596 8.886 9.309 9.468 fft3d_pb 915 16.0 2.402 2.672 8.348 8.624 acc_transpose_blocks 69360 15.2 0.353 0.372 7.331 8.393 sum_up_and_integrate 60 10.3 0.028 0.033 8.326 8.343 integrate_v_rspace 60 11.3 0.002 0.002 8.298 8.314 multiply_cannon_metrocomm1 69360 15.2 0.101 0.107 5.038 8.016 transfer_rs2pw 445 10.6 0.007 0.009 7.480 7.871 cp_dbcsr_sm_fm_multiply 15 9.3 0.001 0.004 7.771 7.788 xc_rho_set_and_dset_create 110 12.4 0.076 0.098 7.319 7.566 cp_dbcsr_sm_fm_multiply_core 15 10.3 0.000 0.000 7.291 7.354 calculate_first_density_matrix 1 7.0 0.000 0.006 7.316 7.330 xc_vxc_pw_create 60 11.3 0.039 0.050 6.972 7.023 make_full_single_inverse 7 9.6 0.001 0.004 6.957 6.994 make_m2s 2890 13.2 0.078 0.087 6.213 6.749 make_images 2890 14.2 0.238 0.258 6.108 6.642 xc_pw_derive 510 13.4 0.005 0.007 6.170 6.244 mp_alltoall_z22v 2340 17.7 5.348 5.624 5.348 5.624 acc_transpose_blocks_kernels 69360 16.2 0.850 0.901 4.549 5.495 mp_waitany 7680 13.5 4.413 4.960 4.413 4.960 multiply_cannon_metrocomm4 67915 15.2 0.187 0.203 2.033 4.870 wfi_extrapolate 5 7.8 0.001 0.011 4.661 4.661 potential_pw2rs 60 12.3 0.003 0.003 4.590 4.614 jit_kernel_transpose 5 15.0 3.699 4.611 3.699 4.611 ------------------------------------------------------------------------------- ~ ~ ~ ~ DATA POINTS ~ ~ ~ ~ PlotPoint: name="601", plot="h2o_512_md", label="(64n/12r/1t)", y=226.580000, yerr=0.000000 PlotPoint: name="602", plot="h2o_512_md_mem", label="(64n/12r/1t)", y=560.600000, yerr=1.854724 ~ ~ ~ ~ END DATA POINTS ~ ~ ~ ~~~~~~ END RESULT ~~~~~~~~ ========= END RESULTS =========== CommitSHA: 069eaf4d8c9e1ef80555d84ef1ab26ee75fd1800 Summary: empty Status: OK