This utility intended to automatically run multiple versions of blender and render some models for benchmarking purpose.
Runnable files list:
autotest.py
- Run tests, collects logs in/log
folder and produce results in/out
folder:.csv
,.log
files and.zip
archive collecting all this stuff together.analyzer.py
- Parse existing log files from/log
folder and generate summary.csv
file in/out
folderplotter.py
- Parse summary.csv
files from/out
folder draw some diagrams and store it in.png
files next to them
For minimal testing you need just pure python without any dependencies.
Given requirements.txt
suitable if you want to get
CPU frequency statistics and hardware/OS information in output log.
requirements-plot.txt
suitable for plotting.