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
Despite putting in a lot of work, the regression tests (#475) are not working reliably in the CI. I have done some reading and found a few hints to why this is. Mainly the github CI is not very reliably. See here and ħere, one can expect large deviations (up to 50%) when running comparing the same code against itself in the CI. .
This means we might want to check for regressions differently:
only do it on demand (i.e. via github comment)
compute baseline at test time
I have also been looking into asv, which looks really cool and would be something to implement during a hackathon maybe.
The text was updated successfully, but these errors were encountered:
Despite putting in a lot of work, the regression tests (#475) are not working reliably in the CI. I have done some reading and found a few hints to why this is. Mainly the github CI is not very reliably. See here and ħere, one can expect large deviations (up to 50%) when running comparing the same code against itself in the CI. .
This means we might want to check for regressions differently:
I have also been looking into asv, which looks really cool and would be something to implement during a hackathon maybe.
The text was updated successfully, but these errors were encountered: