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
{{ message }}
This repository was archived by the owner on Nov 28, 2020. It is now read-only.
Repository navigation
This repository was archived by the owner on Nov 28, 2020. It is now read-only.
The delta between v6.x and master has gotten non trivial. I think it would make sense to Audit the 6.x branch and see what needs to be done to get things up to date
The benchmark tooling in the core was rewritten during v6.x as semver-major so it is not available in v6.x (nodejs/node#7094), therefore almost every update in the benchmark folder does not apply to v6.x.
I think we should skip backporting benchmark commits, and ask people updating benchmarks in their PR to split those changes into separate commits?
Uh...wait, if you are talking about bringing the existing benchmark folder in v6.x branch up-to-date, then yes, it needs to be audited because there could probably be some commits that depend on the new tooling landed. Although...do we still run the benchmarks in the v6.x branch? Is it necessary? If we are comparing the performance differnce between v6.x and other later release lines, since the tooling in v6.x does not generate the same output as others, we probably can't use them as-is anyway.
Activity
joyeecheung commentedon Oct 19, 2017
Are you talking about the benchmarks on https://benchmarking.nodejs.org/ or the core micro-benchmarks?
mhdawson commentedon Oct 19, 2017
Agreed, not really sure what you are asking for. Can you add some more detail.
MylesBorins commentedon Oct 19, 2017
I was referring to the benchmark folder in the 6.x release line. The tooling is quite out of date
joyeecheung commentedon Oct 20, 2017
The benchmark tooling in the core was rewritten during v6.x as semver-major so it is not available in v6.x (nodejs/node#7094), therefore almost every update in the benchmark folder does not apply to v6.x.
I think we should skip backporting benchmark commits, and ask people updating benchmarks in their PR to split those changes into separate commits?
joyeecheung commentedon Oct 20, 2017
Uh...wait, if you are talking about bringing the existing benchmark folder in v6.x branch up-to-date, then yes, it needs to be audited because there could probably be some commits that depend on the new tooling landed. Although...do we still run the benchmarks in the v6.x branch? Is it necessary? If we are comparing the performance differnce between v6.x and other later release lines, since the tooling in v6.x does not generate the same output as others, we probably can't use them as-is anyway.