github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
diginet where could I find some decent info on 6model? 04:28
moritz diginet: github.com/edumentab/rakudo-and-nq...es-day2.md maybe? 05:33
brrt good * #moarvm 08:01
lizmat brrt o/ 08:03
brrt ohai lizmat 08:06
how's quarantine?
lizmat tbh, for someone who has been mainly working from home for the past 20+ years, not a lot has changed 08:07
brrt :-D
lizmat but am glad we're still able to a bicycle ride every day, including in Germany
too bad we're not allowed into Belgium atm, so that approximately halves the area in which we can cycle 08:08
and no restaurants along the way :-(
yes, we're suffering here :-)
oh, with all the trees in full bloom, the pollen make us quite sniffy 08:09
nwc10 \o 09:04
timotimo when's the last time someone tried jemalloc on moar and raku? 09:17
i don't actually know what areas jemalloc excels in the most 09:46
and we've got multiple custom allocators for our Very Special Requirements™ 09:47
AlexDaniel timotimo: github.com/rakudo/rakudo/issues/1501 10:32
timotimo: github.com/MoarVM/MoarVM/pull/1072 10:33
timotimo pahole's got some holes for us to stuff 14:09
timotimo haha, my manual attempts at reordering made it worse actually 14:34
nope! i was just too tired to notice that i was diffing pahole output the wrong way around 14:41
making short-lived structs smaller isn't worth very much, tho, of course 14:46
or structs that only have few instances made
MVMNormalizer is one of the two, but not sure which. but it shrunk from 56 to 48 14:47
MVMDecodeStream from 144 to 136
Geth_ MoarVM: ec53e4dfca | (Timo Paulssen)++ | 3 files
Give MVMSpeshAnn an "order" field, use it for comments

I hope this can be insightful when something odd happens and the order of optimizations spesh does is important.
14:53