|
02:47
ilbot3 joined
06:14
JimmyZ joined
|
|||
| JimmyZ | www.cs.purdue.edu/homes/gkrichar/p...ruffle.pdf # Something about VM optimization, hopes someone may like it . | 06:20 | |
|
07:09
FROGGS[mobile] joined
07:23
FROGGS joined
|
|||
| jnthn | timotimo: Which benchmark is it? | 07:27 | |
| Having a lot of static frames in the nursery sounds very odd, yes | 07:28 | ||
| Is that on first collection, though? | |||
| Or a later one? | |||
| FROGGS | morning | 07:29 | |
| jnthn: we really need to do something about 32bit machines :o) - testers.perl6.org/dist/NativeCall | 07:30 | ||
| (where 'we' really is you *and* me hopefully) | |||
| jnthn | cpan - da testers! :D | 07:31 | |
| FROGGS | cpan date sters! | 07:32 | |
| :P | |||
| the page even would be fast if I'd precompute more things | |||
| jnthn | And yeah, agree... | 07:35 | |
| FROGGS | but... IIRC the NSA will be involved here? | 07:36 | |
| jnthn | No, I don't think it needs to be | ||
| FROGGS | cool | ||
| I guess it is just a tiny change that'd make x86 work... | 07:37 | ||
| but I already tried and failed twice :o) | |||
| jnthn | OK. I need to go to do some urgent but boring paperwork at the office...I'll have some tuits later in the day, and can take a look. | 07:38 | |
| FROGGS | ahh, no hurry | 07:39 | |
| just wanted to do some conversation :D | |||
|
07:39
JimmyZ joined
|
|||
| jnthn | .oO( ā¬ A little less conversation, a little moar action please ā¬ ) |
07:40 | |
| FROGGS | hehe | ||
| jnthn | hah, that's the same song that uses the word "satisfactioning" :) | 07:41 | |
| NSA is something I plan to dig into soon, though. Hopefully starting this week. :) | 07:42 | ||
| commute & | |||
| nwc10 | jnthn: NSA before NFG? | 08:00 | |
| TimToady | there's some chance NSA will be useful to NFG, and no chance of the reverse | 08:11 | |
| FROGGS | *g* | ||
| TimToady: I always imagine you speaking these sentences with a slow dark voice coming from the off :o) | 08:12 | ||
| (except when I don't) | |||
| timotimo | jnthn: that's the "for native postinc" | 10:37 | |
| and that was in one of the later collections | |||
|
11:04
ggoebel111111115 joined
|
|||
| jnthn | timotimo: hm, wtf... | 11:34 | |
| timotimo | yes, my thoughts exactly :) | 11:35 | |
|
11:38
daxim joined
|
|||
| FROGGS | jnthn: before I forget: there is something to review in case you are in the right mood this week: github.com/MoarVM/MoarVM/compare/sepsh | 11:46 | |
| jnthn: I don't consider it mergeable btw | 11:47 | ||
| I just wanna talk about it to get insights | |||
|
12:06
JimmyZ joined
12:18
brrt joined
|
|||
| timotimo | hmm. perhaps the fact that the code got OSR'd is somehow related to the massive build-up of MVMStaticFrames? | 12:26 | |
| brrt | \o | ||
| timotimo | perhaps a frame that results from OSRing doesn't get cached? instead it gets recreated each time the osrpoint gets hit? | ||
| brrt | that'd be a bug | 12:27 | |
| timotimo | yeah | ||
| just a theory i made up | |||
| i don't actually know how the OSR mechanism is implemented exactly | 12:28 | ||
| brrt | basicaylly, by sp ops iirc | ||
| which are compiled into labels, iirc | |||
| timotimo | well... that's pretty much a no-explanation :P | ||
| JimmyZ | I know how it triggers | ||
| brrt | i'm sorry timotimo :-) | 12:33 | |
| timotimo | brrt: how about that blog post? :3 | 12:35 | |
| brrt | :-$ | ||
| timotimo | i hope you've been all right | ||
| brrt | yes... just incredibly busy and barely finished my last trimester | ||
| timotimo | OK | ||
| brrt | :-) | 12:36 | |
| timotimo | you want me to mention anything you'd like to do this week? | ||
| brrt | (motivation to study is diminishing, though) | ||
| timotimo | in the weekly post, i mean | ||
| brrt | no, not yet :-) | ||
| timotimo | aaw | ||
| brrt | there will be enough weeklies | ||
|
13:26
woolfy1 joined
13:53
zakharyas joined
17:05
FROGGS[mobile] joined
18:12
FROGGS joined
18:40
zakharyas joined
19:14
japhb joined
19:50
camelia joined
20:34
zakharyas joined
|
|||