A repository for long running ROOT benchmark workflows
Run workflows which last a significant amount of time (tens of seconds, minutes) and measure the runtime. This allows to check for performance regressions.
The package is written for Python3.
Benchmarks are expressed in directories which are python packages.
These packages need to provide two submodules, preparation and benchmark, each providing a method: run.
A useful tool to run a benchmark, print to screen the runtime and save a file with the result of the benchmark in the
work directory is the utils.Benchmarking.runBenchmark function. In order to get a logger for your messages, use the function
utils.log.getLogger.
Don't forget to add the benchmark name (aka the directory name) to the list of directories in the run.py driver script.
The driver creates a workdir for each benchmark. It prepares the benchmark and runs it.
- HiggsDiscoveryRDF
- YOUR NICE CONTRIBUTION!