Memory regressions in daily benchmark

[KristofferC]

Putting it here so it doesn't get forgotten.

3a31155#commitcomment-19063411

https://github.com/JuliaCI/BaseBenchmarkReports/blob/c446686d038e02992f7b0a60490e04577069720f/daily_2016_9_17/report.md

Noticed by @Sacha0

[TotalVerb]

The ["array","cat",("catnd",5)] regression seems to have been introduced by #18467 and has been fixed by #18572.

Local benchmarking has indicated that all other regressions are likely caused by a different PR.

[KristofferC]

Anyone has an idea about this?

[tkelman]

Did we trigger nanosoldier on all the intermediate commits to see which one it was?

[KristofferC]

Commit range is ec3d29e...3a31155

[KristofferC]

@jrevels is there anyway to run nanosoldier on two arbitrary commits?

[nalimilan]

Looks like you should be able to do that by commenting on a specific commit, and giving the reference commit via vs. See https://github.com/JuliaCI/Nanosoldier.jl#comment-examples

[KristofferC]

Yeah, just thought if I could do it here for example.

[KristofferC]

f30c949#commitcomment-19082011 seems to suggest that up to that commit things are ok?

[TotalVerb]

Has this been figured out yet?