You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently we use the changepoint to add changepoint data to our Krun JSON. We spent a long time running our example data through the changepoint library with different parameters, until we found something that seemed to work well for most of our VM / benchmark combinations.
In order to produce useful changepoints for instrumentation data, we would need to look again at how to parameterise the R library functions that we're using.
However, since we first wrote the changepoint code, @rkillick has developed EnvCpt which is a new version of the R changepoint code, which does not need to be parameterised. This would be a more robust option for the instrumentation data, but it would need to be well tested. It would also be worth moving the core changepoint code here to use EnvCpt, but again that would need careful testing over a large dataset.
The intention here is to make it easier for the user to see correlations between changes in timing data and changes in instrumentation data.
The text was updated successfully, but these errors were encountered: