Perl 6 - the future is here, just unevenly distributed

IRC log for #moarvm, 2015-10-22

| Channels | #moarvm index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
01:19 tokuhiro_ joined #moarvm
01:29 colomon joined #moarvm
01:34 colomon joined #moarvm
01:45 tokuhirom_ joined #moarvm
02:15 tokuhiro_ joined #moarvm
02:43 tokuhirom_ joined #moarvm
02:44 colomon joined #moarvm
02:57 colomon joined #moarvm
03:13 dalek joined #moarvm
05:29 tokuhirom_ joined #moarvm
06:42 Ven joined #moarvm
06:48 FROGGS joined #moarvm
07:23 dalek MoarVM/even-moar-jit: 473f005 | brrt++ | 3rdparty/dynasm:
07:23 dalek MoarVM/even-moar-jit: Fix another miscompile
07:23 dalek MoarVM/even-moar-jit:
07:23 dalek MoarVM/even-moar-jit: Of OPTREX bytes there should also be just one.
07:23 dalek MoarVM/even-moar-jit: review: https://github.com/MoarVM/MoarVM/commit/473f005da5
07:51 ShimmerFairy joined #moarvm
07:57 kjs_ joined #moarvm
08:06 Ven joined #moarvm
08:07 kjs_ joined #moarvm
08:17 brrt joined #moarvm
08:17 brrt good *
08:17 brrt the previous commits should almost fix autocasting, but we still get a segv
08:17 brrt also
08:17 timotimo hey brrt
08:17 timotimo i've recently had big integer arithmetic on my mind a little bit
08:17 brrt i don't really think we need to autocast consts; we may just propagate sizes
08:18 brrt tell me about it :-)
08:19 timotimo how's our plan currently looking?
08:19 timotimo i guess we can sort-of easily get to "only actually allocate an Int object at the very end" from where we are now
08:19 timotimo just using spesh
08:20 timotimo can we generate good fast-path code that just holds the smallbigints directly in integer registers and does computations on them with a lot of (probably cheap) checks to make it jump into a full-blown bigint computation path?
08:20 timotimo does that even seem worthwhile?
08:21 timotimo i may be a bit incoherent; sickness is keeping me on very little sleep - none in today's case :(
08:27 brrt that seems very worthwile
08:27 brrt worthwhile
08:28 timotimo also, i really shouldn't have dropped work on deoptimization bridges instantly like i did when i last thought about that
08:28 brrt however, the thing that makes it a jitty-thing, is that the c-languages don't export overflow flags
08:28 timotimo yeah
08:28 brrt i.e. there is no way to do a calculation and have it say 'oops this overflowed, need to switch to bigint semantics'
08:28 timotimo that's where the checks come in
08:29 brrt well, those checks are jit checks, is the thing
08:29 timotimo we'd have to look-before-you-leap
08:29 timotimo last i checked our smallbigints are 32bit integers, right?
08:29 timotimo so addition and subtraction will give us easy overflow checks if we do the arithmetic itself in 64bit ints
08:30 timotimo division shouldn't be able to underflow on int-int cases
08:30 timotimo and for multiplication i think all you have to do is add up the most significant bit's position of both operands?
08:30 timotimo modulus trivially works fine
08:31 timotimo shifts are easy to check, too
08:31 timotimo a good benchmark for this stuff would be interesting; there's also still the stepping-stone of directly emitting the smallbigint calculation code into the jitted code rather than calling to the C functions all the time
08:32 timotimo we'd been planning to get that anyway
08:40 timotimo i'll try to sleep again, hopefully getting a bit of rest now
08:43 timotimo i also feel a bit frustrated from the even-moar-jit effort not really letting me contribute much any more :|
08:52 brrt sorry, was off for a bit
08:52 brrt oh, well, *ahem*
08:52 brrt that is my fault, of course
08:52 brrt i've left it in a 'almost-working-but-broken-state'
08:53 brrt can you tell me how that would work, add up the msb of both operands.....
08:53 Ven joined #moarvm
08:53 brrt hmmmm
08:53 brrt ok, i see what you mean
09:01 brrt i see how that is simpler than my naive plan
09:02 brrt (fwiw, my naive plan would be: clone the graph and rewrite bigint ops into int ops, and have them check for overflow *after* the calculations
09:07 brrt timotimo: if you want, i can probably figure out a way for you to contribute :-)
09:07 brrt have you seen the new tile-writing documentation
09:07 brrt i've been busy trying to rewrite / refactor the tiler-table-generator into Understandable Perl
09:17 brrt and one of the things i want to do is mark tiles with the 'parameters' that should be passed to the function
09:20 brrt the thing is, most of what needs to be done is currently plumbing
09:22 Ven joined #moarvm
09:24 kjs_ joined #moarvm
09:37 brrt and debugging
09:37 brrt (to the tune of ballmer: debuggers debuggers debuggers debuggers... xx 100)
09:39 edehont joined #moarvm
09:45 FROGGS joined #moarvm
09:53 * brrt wonders if dynasm has turned into more of a loss than a gain now
10:14 rarara_ joined #moarvm
10:21 kjs_ joined #moarvm
10:42 FROGGS joined #moarvm
11:15 FROGGS :S
11:16 brrt well, it's not that serious. it's just that we're using it in a pessimal case
11:19 Ven joined #moarvm
11:20 FROGGS well, but you can always improve on it no?
11:20 brrt no, that's not what i mean
11:20 brrt dynasm is basically ideal if you have long sequences and you fill in a few bits here and there
11:20 FROGGS I mean, sometimes it is important to get a good onramp effect before thinking about the most optimal solution
11:20 brrt we have all sorts of small, short sequences that have to be extensively modified in practice
11:20 FROGGS hmmm
11:21 brrt but yeah, i can see how we could create an interface-compatible clone some day in the feature
11:21 brrt future
11:21 FROGGS so we are not dynasms target audience
11:21 brrt haha
11:21 brrt no
11:21 brrt not anymore
11:21 FROGGS I see :o)
11:27 brrt anyway, i hope to have even-moar-jit mergeable by Christmas
11:28 brrt i'd hoped to have it mergable by halloween, but that seems unlikely
11:30 tokuhirom_ joined #moarvm
11:47 Ven joined #moarvm
12:13 kjs_ joined #moarvm
12:45 tokuhirom_ joined #moarvm
13:05 zakharyas joined #moarvm
13:13 Ven joined #moarvm
13:59 tokuhirom_ joined #moarvm
14:03 Ven joined #moarvm
14:32 hoelzro o/ #moarvm
14:33 nwc10 \o hoelzro
14:40 hoelzro o/ nwc10
14:57 ggoebel2 joined #moarvm
15:00 diakopte1 joined #moarvm
15:01 zakharyas joined #moarvm
15:14 _longines_ joined #moarvm
15:15 colomon joined #moarvm
15:26 [Coke] Hey. Um. is there a moarvm release this month yet?
15:26 [Coke] Can anyone other than jnthn++ do that?
15:39 kjs_ joined #moarvm
15:44 lizmat I think FROGGS can ?
15:46 tokuhiro_ joined #moarvm
15:48 [Coke] ok. that's a blocker on a birthday release.
15:50 dalek joined #moarvm
15:50 colomon joined #moarvm
15:57 lizmat if nobody can do a MoarVM release, then I guess we'll have to wait until next week  :-(
16:11 Ven joined #moarvm
16:47 synbot6 joined #moarvm
16:47 tokuhiro_ joined #moarvm
16:56 zakharyas joined #moarvm
17:03 tokuhirom_ joined #moarvm
17:21 ggoebel joined #moarvm
17:24 leont joined #moarvm
17:27 FROGGS joined #moarvm
18:17 vendethiel joined #moarvm
18:34 kjs_ joined #moarvm
18:34 [Coke] so, the instructions seem super easy; anyone have access to upload a tarball?
18:35 [Coke] diakopter, do you have access for that?
18:47 FROGGS ohh ups
18:48 FROGGS still nobody can AFAIK
18:48 dalek MoarVM: 4f9d381 | coke++ | docs/ChangeLog:
18:48 dalek MoarVM: Add changelog entries for last month
18:48 dalek MoarVM: review: https://github.com/MoarVM/MoarVM/commit/4f9d38178f
18:49 [Coke] SFAIK, the only thing that actually needs the uploaded tarball is macports.
18:49 tokuhiro_ joined #moarvm
18:49 [Coke] NQP just needs a tag.
18:49 [Coke] so, I'm just going to pull the trigger, we can get the tarball uploaded after the fact.
19:03 dalek MoarVM: 429ce74 | coke++ | VERSION:
19:03 dalek MoarVM: prep for 2015.10 release
19:03 dalek MoarVM: review: https://github.com/MoarVM/MoarVM/commit/429ce745fc
19:06 FROGGS [Coke]++
19:19 tokuhirom_ joined #moarvm
19:47 flussence joined #moarvm
19:53 zakharyas joined #moarvm
20:47 kjs_ joined #moarvm
20:50 tokuhiro_ joined #moarvm
21:50 tokuhirom_ joined #moarvm
21:57 zakharyas joined #moarvm
22:18 tokuhiro_ joined #moarvm
23:10 tokuhirom_ joined #moarvm

| Channels | #moarvm index | Today | | Search | Google Search | Plain-Text | summary