options

gmx_mpi - 2023-08-08 09:18:53 - MAQAO 2.17.8

Help is available by moving the cursor above any symbol or by checking MAQAO website.

Global Metrics

Total Time (s)50.31
Profiled Time (s)46.33
Time in analyzed loops (%)89.0
Time in analyzed innermost loops (%)79.0
Time in user code (%)91.3
Compilation Options Score (%)100
Perfect Flow Complexity1.02
Array Access Efficiency (%)Not Available
GFLOPS0.0
Perfect OpenMP + MPI + Pthread1.03
Perfect OpenMP + MPI + Pthread + Perfect Load Distribution1.15
No Scalar IntegerPotential Speedup1.18
Nb Loops to get 80%17
FP VectorisedPotential Speedup1.59
Nb Loops to get 80%20
Fully VectorisedPotential Speedup1.67
Nb Loops to get 80%23
FP Arithmetic OnlyPotential Speedup1.76
Nb Loops to get 80%23

CQA Potential Speedups Summary

Loop Based Profile

Innermost Loop Based Profile

Application Categorization

Compilation Options

Source ObjectIssue
libgromacs_mpi.so.7
fft5d.cpp
threaded_force_buffer.cpp
stl_vector.h
pme_gather.cpp
listed_forces.cpp
partition.cpp
manage_threading.cpp
kernel_prune.cpp
pairs.cpp
vec.h
update.cpp
md_support.cpp
pme.cpp
kernel_common.cpp
mdatoms.cpp
lincs.cpp
pbc.cpp
constr.cpp
pme_grid.cpp
localtopology.cpp
kerneldispatch.cpp
pme_solve.cpp
pme_spread.cpp
calc_verletbuf.cpp
fft_fftw3.cpp
bonded.cpp
pairlist.cpp
sim_util.cpp
grid.cpp
settle.cpp
kernel_outer.h
domdec_constraints.cpp
redistribute.cpp
impl_arm_sve_util_float.h
atomdata.cpp

Loop Path Count Profile

Cumulated Speedup If No Scalar Integer

Cumulated Speedup If FP Vectorized

Cumulated Speedup If Fully Vectorized

Cumulated Speedup If FP Arithmetic Only

Experiment Summary

Application/home/eoseret/GROMACS/build/gcc_2/bin/gmx_mpi
Timestamp2023-08-08 09:18:53 Universal Timestamp1691486333
Number of processes observed1 Number of threads observed52
Experiment TypeMPI; OpenMP;
Machineip-172-31-47-199
Architectureaarch64 Micro ArchitectureARM_NEOVERSE_V1
OS VersionLinux 5.15.0-1039-aws #44~20.04.1-Ubuntu SMP Thu Jun 22 12:21:08 UTC 2023
Architecture used during static analysisaarch64 Micro Architecture used during static analysisARM_NEOVERSE_V1
Frequency DriverNA Frequency GovernorNA
Huge Pagesmadvise Hyperthreadingoff
Number of sockets1 Number of cores per socket64
Compilation Options
libgromacs_mpi.so.7: GNU C++17 11.1.0 -march=armv8.2-a+sve -msve-vector-bits=256 -mlittle-endian -mabi=lp64 -g -O3 -O3 -std=c++17 -fno-omit-frame-pointer -fcf-protection=none -fPIC -fexcess-precision=fast -funroll-all-loops -fopenmp -fasynchronous-unwind-tables -fstack-protector-strong -fstack-clash-protection
CommentsGNU g++ 12.2.0 (SIMD=SVE), AWS G3 (Neoverse V1), 10000 steps, 52 cores

Configuration Summary

Dataset
Run Command<executable> mdrun -s ion_channel.tpr -nsteps 10000 -pin on -deffnm gcc
MPI Commandmpirun -n <number_processes> --bind-to core
Number Processes1
Number Nodes1
Number Processes per Nodes1
FilterNot Used
Profile StartNot Used
Maximal Path Number4
×