Stream: t-compiler

Topic: perf nll measurements


simulacrum (Apr 24 2019 at 21:45, on Zulip):

@nikomatsakis fyi: nll dashboard on perf is now going to be empty; we merged NLL into 2015 edition so there's no "base case" anymore. The page hasn't been removed yet because that's more work than I had time for, but we are no longer collecting data.

simulacrum (Apr 24 2019 at 21:46, on Zulip):

This was also partially brought on by a desire to reduce the benchmark time (we already have a long queue)

nikomatsakis (Apr 25 2019 at 09:46, on Zulip):

OK

pnkfelix (Apr 29 2019 at 10:47, on Zulip):

we may want to repurpose the NLL dashboard to measure non-migrate mode, perhaps

pnkfelix (Apr 29 2019 at 10:47, on Zulip):

(though I suspect there aren't going to be the same sort of performance issues there that we saw for the initial NLL development)

nikomatsakis (Apr 30 2019 at 20:05, on Zulip):

At some point I'd like to start measuring polonius :)

nikomatsakis (Apr 30 2019 at 20:05, on Zulip):

But we're not there yet

Last update: Nov 22 2019 at 05:20UTC