|
06:23
woolfy joined
07:20
woolfy joined
07:23
woolfy left
07:24
woolfy1 joined,
woolfy1 left
07:59
lizmat joined
08:08
FROGGS joined
08:29
vendethiel joined
08:33
woolfy joined
08:38
brrt joined
|
|||
| dalek | arVM: 7785f67 | nicholas++ | src/6model/sc.h: Merge the bodies of MVM_sc_get_{obj,stable}_sc and *_set_* The two *_get_* functions call MVM_sc_get_collectable_sc, and the two *_set_* functions call MVM_sc_set_collectable_sc. 4befa95 | jnthn++ | src/6model/serialization.c: Back out index-setting in deserialize again. This breaks the Panda build on some platforms. |
10:00 | |
|
10:01
dalek joined
|
|||
| nwc10 | :-) | 10:03 | |
| FROGGS | nwc10++ # killing dalek | 10:05 | |
| nwc10 | ammo doesn't kill bots, people kill bots | 10:06 | |
| so | |||
| jnthn++ | |||
| jnthn | I let the patch restore the thing I took out 'cus it SEGV'd the Panda build on Linux | 10:07 | |
| It still does. | |||
| nwc10 | :-/ | 10:09 | |
| jnthn: I'm still playing with the 24:8 split - right now it seems to be slower | 10:11 | ||
| which is wrong! as it should reduce memory allocation by about 256K | 10:12 | ||
| so going to look at the generated code | |||
| nwc10 | but also planning to use pahole to find all the gaps in structures, and where possible re-order to reduce memory usage | ||
| and see if (a) that doesn't look crazy stupid unmaintainable | 10:13 | ||
| (b) helps | |||
| second-order thing - pahole neatly points out cache lines | |||
| jnthn hopes for a hole lot of little improvements from it :) | |||
| nwc10 | and one structure has a bunch of mutexes which are close enough together that some share cache lines | ||
| nwc10 groans | 10:14 | ||
| jnthn | ugh, yes | ||
| nwc10 | reply pun failure | ||
|
10:47
brrt joined
|
|||
| masak .oO( soundcloud.com/pomdeterrific/pomde...-me-a-hole ) | 11:19 | ||
| jnthn | .oO( you're a hole ) |
11:20 | |
|
12:19
brrt joined
|
|||
| dalek | arVM: 9dacb45 | jonathan++ | src/ (2 files): Specialize getattr_[ins] on P6opaque. |
12:45 | |
| jnthn | With that, time spent in get_attribute in CORE.setting is done from 4.23% down to 3.96% | 12:47 | |
| Now for the object case... | |||
|
13:27
brrt joined
13:28
woolfy1 joined
13:51
btyler joined
|
|||
| btyler | jnthn: not sure if this is helpful information, but Moar at 4befa95 was ok with panda prior to rebuilding nqp and rakudo | 14:05 | |
| jnthn | btyler: Yeah, decided we'll put that off | 14:07 | |
| btyler: Until after next release | |||
| btyler | jnthn: alrighty, sounds good | 14:09 | |
|
14:25
zakharyas joined
|
|||
| dalek | arVM: b31ac96 | (Tobias Leich)++ | src/math/bigintops.c: a bigint is big when it does not fit in On 64bit platforms the DIGIT_BIT is 60, so when we want to store something bigger, we need a bigint. The condition before this patch was always false. |
14:28 | |
|
14:30
ggoebel111115 joined
|
|||
| masak | nwc10++ # speeding up moar by aligning bits | 15:15 | |
| dalek | arVM: 187eedf | jonathan++ | src/ (2 files): Optimize object attribute accesses in spesh. |
15:47 | |
| jnthn | Used to spend 4.23% of CORE.setting build in get_attribute; now it's 1.85%. | 15:48 | |
| And think there's some more win to be had yet. | |||
| tadzik | :o | 15:52 | |
| jnthn++ | |||
|
16:02
lizmat joined
16:03
woolfy joined
16:07
tgt joined
|
|||
| brrt | plz checkout my blog again :-) brrt-to-the-future.blogspot.nl/2014...chine.html | 16:13 | |
| jnthn | ooh, new post | 16:14 | |
|
16:14
FROGGS joined
|
|||
| brrt | yes, new post | 16:15 | |
| :-) | |||
| brrt dinner | |||
| moritz | brrt++ | ||
|
16:15
brrt left
|
|||
| moritz | brrt: iirc one of the reasons for making moarvm a stack machine is that it makes some codegen easier | 16:15 | |
| ... to slow | |||
| masak | ironic that the blog post requires a JavaScript VM to read ;) | ||
| moritz | aye, blogspot-- | 16:16 | |
| jnthn | uh, moarvam ain't a stack machine | ||
| moritz | s/stack/register/ | ||
| argl | |||
| tadzik | br<TAB><TAB> | 16:17 | |
| moritz | 18:15 -!- brrt [~bart@195-241-164-190.ip.telfort.nl] has left #moarvm [] | ||
| tadzik | yeah | ||
| is there a .tell bot here? | |||
| no. Oh well, I'll wait :) | 16:18 | ||
| lizmat | .tell | 16:20 | |
| :-( | |||
| masak | brrt++ # nice post | 16:26 | |
| jnthn | One other reason cheap deopt is desirable is it makes it uncostly to speculate lots | 16:27 | |
| dalek | arVM: ef5657e | jonathan++ | src/ (2 files): Spesh indirect form of attr ops with known names. |
16:39 | |
|
16:40
woolfy left
17:10
lizmat joined,
colomon joined
|
|||
| nwc10 | jnthn: I can haz SEGVs | 17:34 | |
| eg ./perl6-m t/spec/S05-transliteration/with-closure.rakudo.moar | 17:35 | ||
| goes away with export MVM_SPESH_DISABLE=1 | |||
|
17:55
brrt joined
|
|||
| brrt is back | 17:58 | ||
| did is say moarvm is a stack machine? | |||
| :-o | |||
| tadzik? | 18:01 | ||
| tadzik | brrt: oh | 18:29 | |
| freqeently in your blog post :) | |||
| brrt | hi :-) | ||
| really? that is embarassing | |||
| where | |||
| tadzik | it's a hyperlink | ||
| somewhere at the beginning iirc | 18:30 | ||
| also, I really liked it :) | 18:31 | ||
| brrt | thanks :-) | 18:32 | |
| oh i've found it | |||
| nwc10 | jnthn: 4 spectests SEGV with spesh. Probably all the same bug.paste.scsys.co.uk/377252 | 18:39 | |
| jnthn: culprit is ef5657e713e8c7a92753614cf1ea9e2bbcf3e0e5 -- Spesh indirect form of attr ops with known names. | 18:45 | ||
| jnthn: may not be obvious from that nopaste - 4 failures are t/spec/S05-transliteration/with-closure.rakudo.moar t/spec/S05-transliteration/trans.rakudo.moar t/spec/integration/advent2009-day09.t t/spec/integration/real-strings.t | 18:49 | ||
| brrt is off | 20:12 | ||
|
20:12
brrt left
|
|||
| timotimo | brrt: if you read this in the backlog, you may want to change the wording of the sentence about add_i to indicate that the first operand is the target, as is the case with all operations i can think of right nw | 20:29 | |
|
22:03
lizmat joined
22:14
camelia joined
22:18
lizmat joined
22:19
sergot_ joined,
nwc10_ joined,
dalek joined,
cognominal joined,
Util joined,
tadzik joined,
64MAAITRZ joined,
TimToady joined,
tokuhirom joined,
harrow joined,
flussence joined,
rurban_ joined,
timotimo joined,
bcode joined,
bonsaikitten joined,
zacts joined,
17SAADFTQ joined,
nwc10 joined
22:20
zacts_ joined
22:21
FROGGS joined
22:24
woolfy joined
23:01
lizmat joined
23:08
woolfy joined
23:11
woolfy joined
23:12
woolfy joined
23:21
xiaomiao joined
23:27
woolfy joined
23:47
woolfy left
|
|||