**** BEGIN LOGGING AT Sat Oct 24 10:59:57 2020
Oct 24 11:03:45 <Herrie> real    84m34.792s
Oct 24 11:04:05 <Herrie> This is "time" when building Gatesgarth for Pinephone from scratch
Oct 24 11:11:41 <Herrie> Now doing Mido for comparision, but with sstate
Oct 24 11:11:59 <Herrie> Mido has 7927 tasks
Oct 24 13:23:27 <Herrie> System locked up, not veyr helpful
Oct 24 13:23:32 <Herrie> Now got for Mido: real    41m7.447s
Oct 24 13:53:35 <Herrie> Tissot after Mido, only 8m36s, mainly kernel I guess
Oct 24 14:04:15 <JaMa> with gatesgarth it seems to use more ram, probably because of gcc-10, with test-oe-build-time with 128G ram I got 2 OOMK hits, while with dunfell it passed ok on the same system
Oct 24 14:04:44 <JaMa> bonaire still didn't came back, so cannot compare with that
Oct 24 17:19:30 <JaMa> test-oe-build-time is 12min (13%) slower in gatesgarth than dunfell
Oct 24 17:19:34 <JaMa> ==> 4-build-all-cores.log <==
Oct 24 17:19:34 <JaMa> 38.87user 5.03system 1:45:06elapsed 0%CPU (0avgtext+0avgdata 30296maxresident)k
Oct 24 17:19:37 <JaMa> 34536inputs+6008outputs (2364major+46264minor)pagefaults 0swaps
Oct 24 17:19:40 <JaMa> ==> threadripper-3970x-128gb-dunfell/4-build-all-cores.log <==
Oct 24 17:19:42 <JaMa> 36.15user 4.53system 1:33:11elapsed 0%CPU (0avgtext+0avgdata 29992maxresident)k
Oct 24 17:19:45 <JaMa> 29304inputs+6064outputs (1850major+17267minor)pagefaults 0swaps
Oct 24 17:20:55 <JaMa> but it's not clear where the extra time was spent, I'll have a look after all builds are finished, it's possible that e.g. new rust takes 10 extra minutes and causes all of this
**** ENDING LOGGING AT Sun Oct 25 02:59:57 2020