-
Use
matlab-benchmarks.py
to generatematlab-benchmarks.json
:$ python matlab_benchmarks.py -o matlab-benchmarks.json
-
Use
qutip-benchmarks.py
to generatequtip-benchmarks.json
:$ python qutip_benchmarks.py -o qutip-benchmarks.json
-
Use
benchmark_comparison.py
to generatebenchmark-data.json
which is used by the d3 script in benchmark.html. It should take two arguments specifying which benchmark runs to compare::$ python benchmark_comparison.py -i qutip-benchmarks.json -r matlab-benchmarks.json -o benchmark_data.json
-
The html file and d3 scripts uses benchmark_data.json to render the comparison graphics.
Now we can also compare different versions of QuTiP, like this:
# install qutip 2.2.0
$ python qutip_benchmarks.py -o qutip-benchmarks-2.2.0.json
# install qutip 2.3.0-dev
$ python qutip_benchmarks.py -o qutip-benchmarks-2.3.0-dev.json
# generate the comparison data, with 2.2.0 as reference
$ python benchmark_comparison.py -i qutip-benchmarks-2.3.0-dev.json -r qutip-benchmarks-2.2.0.json -o benchmark_data.json