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
#3084 seeks to introduce benchmarks on CI for local nodes in a Node.js environment. These are intended to be run per PR workflow and currently take ~ 12 mins.
We would like to support running these in the browser on CI but that would require extra setup which could take the workflow in excess of ~25 mins. I don't think this would be appropriate to be run per PR or even in a release process, but could be useful post-release (the downside of that is if we introduce a regression in the browser it would be caught after the release)
I do think though that there is a low probability one could introduce a performance regression that is vastly more significant in the browser than in node, but still possible.
The text was updated successfully, but these errors were encountered:
I elaborated a bit more here but TLDR If we can essentially cache the SDK builds + pretest, as well as the chrome setup ubuntu builds, we can take each benchmark time down to roughly 6 mins which would be 12 mins total.
#3084 seeks to introduce benchmarks on CI for local nodes in a Node.js environment. These are intended to be run per PR workflow and currently take ~ 12 mins.
We would like to support running these in the browser on CI but that would require extra setup which could take the workflow in excess of ~25 mins. I don't think this would be appropriate to be run per PR or even in a release process, but could be useful post-release (the downside of that is if we introduce a regression in the browser it would be caught after the release)
I do think though that there is a low probability one could introduce a performance regression that is vastly more significant in the browser than in node, but still possible.
The text was updated successfully, but these errors were encountered: