Perl 6 - the future is here, just unevenly distributed

IRC log for #moarvm, 2014-06-19

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

All times shown according to UTC.

Time Nick Message
00:30 japhb_ Dangit, I missed jnthn.  :-(
00:30 japhb_ Well, in any case ...
00:31 japhb_ .tell jnthn I was referring to problems with the output of https://github.com/japhb/perl6-bench/blob/master/nqp/rc-forest-fire
00:48 lizmat joined #moarvm
01:06 FROGGS_ joined #moarvm
01:26 woolfy joined #moarvm
01:35 btyler joined #moarvm
02:00 raiph joined #moarvm
02:29 carlin joined #moarvm
04:41 woosley joined #moarvm
05:48 woosley joined #moarvm
06:20 lizmat joined #moarvm
06:23 nebuchadnezzar joined #moarvm
06:36 woolfy joined #moarvm
06:58 brrt joined #moarvm
06:58 brrt hey, i think i've found the source of the bug in moar-jit
06:58 brrt it's the combination with inline
07:00 brrt i.e - inline 'inlines' a frame that has a instruction of sp_enter_jit
07:00 brrt sp_enter_jit looks as the (now changed) frame and decides there isn't any jitcode
07:00 brrt sp_enter_jit freaks out :-)
07:01 brrt (because the bigger frame, even if fully compatible with jit otherwise, cannot compile sp_enter_jit yet)
07:04 lizmat_ joined #moarvm
07:07 hoelzro joined #moarvm
07:10 zakharyas joined #moarvm
07:12 lizmat joined #moarvm
07:16 FROGGS joined #moarvm
07:21 woolfy joined #moarvm
07:25 bonsaikitten joined #moarvm
07:25 brrt tl;dr - i need some form of register for the jitcode and refer to that in the enter_jit opcode, otherwise inlining breaks jit
07:29 tadzik joined #moarvm
07:29 brrt the only question left is
07:29 brrt what register
07:30 brrt locally (hanging of the StaticFrame) or globally (hanging of the CompUnit or even the MVMInstance)?
07:31 FROGGS is enter_jit coupled to any of these three?
07:33 brrt ehm, no
07:33 brrt it used to be that the jit compiler compiled a whole frame or nothing
07:34 brrt the inliner obviously changes that, as i've seen
07:34 brrt the jitcode /used/ to hang off of the spesh candidate of the static frame
07:34 brrt or, does now, actually
07:35 brrt the register is just there to answer the question 'what jitcode do you want to call?'
07:38 brrt i think the core of it is this: if i register jit code globally, then during inlining, i don't need to change anything (yay), because the sp_enter_jit can still refer to the same entry
07:38 brrt if i register jit code on the static frame - in any sense, either on the spesh candidate or directly - then i need to take care of that in inlining
07:38 brrt i.e. merge the registers of inlined frames
07:39 brrt and update the sp_ente_jit opcodes to refer to the right register entries
07:39 brrt i actually think that is the better option, but i can't readily explain why :-)
07:43 FROGGS hmmm, well, jnthn's opinion will shed light on this then :o)
07:50 brrt he wouldn't be here today, would he?
07:51 FROGGS I have no idea
07:51 brrt ok, i know why a local register is better
07:52 brrt basically, it allows to keep the index small
07:53 timotimo It seems like it's a travel day for jnthn
08:59 * brrt is off for a few hours
08:59 brrt left #moarvm
11:16 cognominal joined #moarvm
11:36 woolfy1 joined #moarvm
11:57 LLamaRider joined #moarvm
12:10 cognominal joined #moarvm
12:14 woolfy joined #moarvm
13:07 woolfy left #moarvm
13:09 woolfy joined #moarvm
13:55 btyler joined #moarvm
14:15 FROGGS joined #moarvm
14:25 btyler joined #moarvm
14:40 woolfy1 joined #moarvm
15:47 raiph joined #moarvm
16:12 btyler_ joined #moarvm
16:14 jnap joined #moarvm
17:02 colomon joined #moarvm
18:10 btyler joined #moarvm
18:10 jnap joined #moarvm
18:10 raiph joined #moarvm
18:10 woolfy1 joined #moarvm
18:10 FROGGS joined #moarvm
18:10 tadzik joined #moarvm
18:10 bonsaikitten joined #moarvm
18:10 lizmat joined #moarvm
18:10 hoelzro joined #moarvm
18:10 nebuchadnezzar joined #moarvm
18:10 woosley joined #moarvm
18:10 japhb joined #moarvm
18:10 bcode joined #moarvm
18:10 japhb_ joined #moarvm
18:10 vendethiel joined #moarvm
18:10 Util joined #moarvm
18:10 oetiker joined #moarvm
18:10 lue joined #moarvm
18:10 BinGOs joined #moarvm
18:10 moritz joined #moarvm
18:10 [Coke] joined #moarvm
18:10 avar joined #moarvm
18:10 ChanServ joined #moarvm
18:10 ingy joined #moarvm
18:10 brother joined #moarvm
18:10 lee__ joined #moarvm
18:10 ggoebel111116 joined #moarvm
18:10 nwc10 joined #moarvm
18:10 cxreg joined #moarvm
18:10 masak joined #moarvm
18:10 sergot joined #moarvm
18:10 synopsebot joined #moarvm
18:10 _sri joined #moarvm
18:10 betterworld joined #moarvm
18:10 daxim joined #moarvm
18:10 krunen joined #moarvm
18:10 jnthn joined #moarvm
18:10 dalek joined #moarvm
18:10 ashleydev joined #moarvm
18:10 harrow joined #moarvm
18:10 rurban_ joined #moarvm
18:10 timotimo joined #moarvm
18:10 flussence joined #moarvm
18:10 tokuhirom joined #moarvm
18:16 japhb_ joined #moarvm
18:16 btyler joined #moarvm
18:16 jnap joined #moarvm
18:16 raiph joined #moarvm
18:16 woolfy1 joined #moarvm
18:16 FROGGS joined #moarvm
18:16 tadzik joined #moarvm
18:16 bonsaikitten joined #moarvm
18:16 lizmat joined #moarvm
18:16 hoelzro joined #moarvm
18:16 nebuchadnezzar joined #moarvm
18:16 woosley joined #moarvm
18:16 japhb joined #moarvm
18:16 bcode joined #moarvm
18:16 vendethiel joined #moarvm
18:16 Util joined #moarvm
18:16 oetiker joined #moarvm
18:16 lue joined #moarvm
18:16 BinGOs joined #moarvm
18:16 moritz joined #moarvm
18:16 [Coke] joined #moarvm
18:16 avar joined #moarvm
18:16 ChanServ joined #moarvm
18:16 ingy joined #moarvm
18:16 brother joined #moarvm
18:16 lee__ joined #moarvm
18:16 ggoebel111116 joined #moarvm
18:16 nwc10 joined #moarvm
18:16 cxreg joined #moarvm
18:16 masak joined #moarvm
18:16 sergot joined #moarvm
18:16 synopsebot joined #moarvm
18:16 _sri joined #moarvm
18:16 betterworld joined #moarvm
18:16 daxim joined #moarvm
18:16 krunen joined #moarvm
18:16 jnthn joined #moarvm
18:16 dalek joined #moarvm
18:16 ashleydev joined #moarvm
18:16 harrow joined #moarvm
18:16 rurban_ joined #moarvm
18:16 timotimo joined #moarvm
18:16 flussence joined #moarvm
18:16 tokuhirom joined #moarvm
18:27 brrt joined #moarvm
19:09 zakharyas joined #moarvm
19:14 vendethiel joined #moarvm
20:02 dalek MoarVM/moar-jit: 22773f2 | jnthn++ | src/spesh/args.c:
20:02 dalek MoarVM/moar-jit: Don't refuse to spesh if we've a slurpy positional
20:02 dalek MoarVM/moar-jit: review: https://github.com/MoarVM/MoarVM/commit/22773f2330
20:02 dalek MoarVM/moar-jit: dd80dbf | (Timo Paulssen)++ | src/spesh/optimize.c:
20:02 dalek MoarVM/moar-jit: put in a missing break
20:03 dalek joined #moarvm
20:06 [Coke] we lost the review line there.
20:16 brrt joined #moarvm
20:17 brrt that should've been a merge line, though
20:40 timotimo yes, those are mostly old-ish commits
20:43 raiph joined #moarvm
21:33 dalek MoarVM: bc1bbc3 | Carlin++ | src/io/fileops.c:
21:33 dalek MoarVM: Change default mode to 0x01FF
21:33 dalek MoarVM:
21:33 dalek MoarVM: 0x0FFF is 7777, not 0777
21:33 dalek MoarVM: review: https://github.com/MoarVM/MoarVM/commit/bc1bbc3bab
21:33 dalek MoarVM: 8309dc2 | Carlin++ | src/io/syncfile.c:
21:33 dalek MoarVM: Change default mode to 0x01FF
21:33 dalek MoarVM: review: https://github.com/MoarVM/MoarVM/commit/8309dc225b
21:33 dalek MoarVM: 5cdc366 | Carlin++ | src/io/ (2 files):
21:33 dalek MoarVM: make default file mode 0644
21:33 dalek MoarVM: review: https://github.com/MoarVM/MoarVM/commit/5cdc366b15
21:33 dalek MoarVM: f418844 | (Tobias Leich)++ | src/io/ (2 files):
21:33 dalek MoarVM: Merge pull request #105 from carbin/master
21:33 dalek MoarVM:
22:06 pmichaud joined #moarvm
22:13 pmichaud regarding pull request / issue #105 in MoarVM, I think that file permissions in C programs should be specified as 0666... letting the umask setting take care of turning off group/other write permissions as appropriate.
22:13 pmichaud I left a similar comment in the ticket.
22:49 jnthn +1, please somebody make it as pmichaud++ said
22:51 jnthn .tell brrt spesh for inlining needs to have access to the specialized bytecode; I suggest the jitcode is hung off just the ->effective_bytecode of the frame, but the bytecode slot of spesh_cand is left as it is (I'm assuming since the bug exists, it's being hung off spesh_cand).
22:51 jnthn Thta is, hvae MVM_spesh_invoke make the "enter jitcode" decision.
22:51 jnthn uh, MVM_frame_invoke
22:52 jnthn And thus the inliner still finds the bytecode itself.
22:52 jnthn sleep &
23:30 benabik joined #moarvm

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