A small toolset for postprocess plumed benchmark and the plumed time report at the end of the simulations
as now it can extract a dict from a file like:
BENCH: Kernel: this
BENCH: Input: plumed.dat
BENCH: Comparative: 1.000 +- 0.000
BENCH: Cycles Total Average Minimum Maximum
BENCH: A Initialization 1 0.214297 0.214297 0.214297 0.214297
BENCH: B0 First step 1 0.062736 0.062736 0.062736 0.062736
BENCH: B1 Warm-up 199 12.618833 0.063411 0.055884 0.076860
BENCH: B2 Calculation part 1 400 25.567659 0.063919 0.054110 0.113234
BENCH: B3 Calculation part 2 400 25.594014 0.063985 0.059516 0.102646
PLUMED: Cycles Total Average Minimum Maximum
PLUMED: 1 64.054325 64.054325 64.054325 64.054325
PLUMED: 1 Prepare dependencies 1000 0.003443 0.000003 0.000001 0.000013
PLUMED: 2 Sharing data 1000 0.305915 0.000306 0.000015 0.037867
PLUMED: 3 Waiting for data 1000 0.003051 0.000003 0.000002 0.000013
PLUMED: 4 Calculating (forward loop) 1000 63.459357 0.063459 0.054012 0.091577
PLUMED: 5 Applying (backward loop) 1000 0.008520 0.000009 0.000005 0.000044
PLUMED: 6 Update 1000 0.043188 0.000043 0.000031 0.000080
BENCH:
BENCH: Kernel: ../../src/lib/install/libplumedKernel.so
BENCH: Input: plumed.dat
BENCH: Comparative: 0.941 +- 0.002
BENCH: Cycles Total Average Minimum Maximum
BENCH: A Initialization 1 0.216190 0.216190 0.216190 0.216190
BENCH: B0 First step 1 0.058967 0.058967 0.058967 0.058967
BENCH: B1 Warm-up 199 11.983512 0.060219 0.056412 0.102643
BENCH: B2 Calculation part 1 400 24.035510 0.060089 0.056539 0.113900
BENCH: B3 Calculation part 2 400 24.084369 0.060211 0.056866 0.097184
PLUMED: Cycles Total Average Minimum Maximum
PLUMED: 1 60.373083 60.373083 60.373083 60.373083
PLUMED: 1 Prepare dependencies 1000 0.003351 0.000003 0.000001 0.000014
PLUMED: 2 Sharing data 1000 0.329323 0.000329 0.000015 0.032672
PLUMED: 3 Waiting for data 1000 0.003078 0.000003 0.000001 0.000013
PLUMED: 4 Calculating (forward loop) 1000 59.752459 0.059752 0.056310 0.083841
PLUMED: 5 Applying (backward loop) 1000 0.008900 0.000009 0.000006 0.000034
PLUMED: 6 Update 1000 0.043015 0.000043 0.000032 0.000239
or
PLUMED: Cycles Total Average Minimum Maximum
PLUMED: 1 60.373083 60.373083 60.373083 60.373083
PLUMED: 1 Prepare dependencies 1000 0.003351 0.000003 0.000001 0.000014
PLUMED: 2 Sharing data 1000 0.329323 0.000329 0.000015 0.032672
PLUMED: 3 Waiting for data 1000 0.003078 0.000003 0.000001 0.000013
PLUMED: 4 Calculating (forward loop) 1000 59.752459 0.059752 0.056310 0.083841
PLUMED: 5 Applying (backward loop) 1000 0.008900 0.000009 0.000006 0.000034
PLUMED: 6 Update 1000 0.043015 0.000043 0.000032 0.000239
One way of extracting the timing from the ouput of the benchmark is:
awk '/BENCH: Kernel: /,EOF' benchmark.out > times_benchmark.out
pip install plumed-bench-ppplumed-bench-pp is distributed under the terms of the MIT license.