Perl 6 - the future is here, just unevenly distributed

IRC log for #moarvm, 2015-03-20

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

All times shown according to UTC.

Time Nick Message
02:28 zakharyas joined #moarvm
04:02 FROGGS_ joined #moarvm
05:26 zakharyas joined #moarvm
07:49 Ven joined #moarvm
07:50 FROGGS joined #moarvm
08:18 Ven joined #moarvm
08:30 rurban_ joined #moarvm
08:54 brrt joined #moarvm
09:13 kjs_ joined #moarvm
10:11 kjs_ joined #moarvm
10:30 * lizmat sees daylight again for the second time today
10:46 brrt :-)
10:46 brrt i see cloudlight mostly
10:49 lizmat well, it got a bit darker here, like an early onset of the setting of the sun
10:56 colomon joined #moarvm
11:32 kjs_ joined #moarvm
11:50 Ven joined #moarvm
12:21 Ven joined #moarvm
12:35 [Coke] new moarvm is visible in the port list. whee.
12:35 [Coke] hopefully will make some progress on nqp this weekend.
12:35 brrt yay
12:42 FROGGS \o/
13:33 timotimo brrt: you think perf could tell us what's wrong with the performance?
13:34 kjs_ joined #moarvm
13:40 nebuchadnezzar joined #moarvm
13:56 kjs_ joined #moarvm
14:10 retupmoca joined #moarvm
15:09 FROGGS[mobile] joined #moarvm
16:13 Ven joined #moarvm
16:37 jnthn timotimo: You may get more insight from callgrind or so
16:40 timotimo mhm
16:53 hoelzro [Coke]++
16:58 dalek MoarVM: aa450df | jnthn++ | src/6model/serialization.c:
16:58 dalek MoarVM: Re-enable lazy deserialization.
16:58 dalek MoarVM:
16:58 dalek MoarVM: Provide a #define for easily turning it on/off also. We'll see what
16:58 dalek MoarVM: the ecosystem fall-out is; even if there is any, we really want to
16:58 dalek MoarVM: have this on for the startup and memory use, so we should discover
16:58 dalek MoarVM: and debug the issues ahead of the next release. The pre-comp tests
16:58 dalek MoarVM: in the spectest suite are no worse with this turned on.
16:58 dalek MoarVM: review: https://github.com/MoarVM/MoarVM/commit/aa450df4ad
16:59 FROGGS[mobile] I can do a ecosystem smoke if you want
17:00 jnthn Feel free.
17:00 FROGGS[mobile] and we can easily diff to the 2015.03 run
17:00 jnthn Right
17:01 jnthn I'm going to have more tuits starting soon also, so I'll be in a better position to debug any fallout
17:05 FROGGS joined #moarvm
17:28 hoelzro jnthn: I tested my test for RT #122773 with your lazy deserialization change, and it doesn't break  :thumbsup:
17:28 synopsebot Link: https://rt.perl.org/rt3//Public/Bug/Display.html?id=122773
17:29 jnthn Yeah, I can the test file that's in and it seemed happy
17:29 jnthn There were other pre-comp fixes in the meantime
17:29 hoelzro jnthn++
17:29 jnthn So I guess one of those nailed the real issue.
17:32 FROGGS "This has been fixed on MoarVM; the issue is still outstanding on JVM."
17:33 hoelzro FROGGS: I'm trying it on the JVM now =)
17:33 * hoelzro wonders if the corresponding roast test is skipped on JVM
17:34 colomon joined #moarvm
17:56 japhb jnthn: What are your 2015.04 goals?  Which is to say, what's up next on the docket?
17:57 jnthn Next major goal is NFG
17:57 jnthn (In Moar)
17:58 jnthn Also thinking lots about sized/shaped array stuff
17:58 jnthn In the backbrane at least. :)
17:59 jnthn Will keep on with the native array stuff a bit too, since that's not perfect yet
18:00 jnthn Wanting to land the various box/unbox and ref/deref opts in spesh I've started in on too
18:01 jnthn In May, concurrency and I/O things probably will rise to the top of the pile.
18:11 hoelzro FROGGS: it's still outstanding on JVM =(
18:12 hoelzro I was hoping that NQP or Rakudo had a change that fixed the issue for everyone
18:26 FROGGS :/
18:27 hoelzro although
18:28 hoelzro there could be something in NQP that was largely copied over from JVM -> Moar, and was fixed in the latter, but not the former
18:28 hoelzro (or Rakudo, for that matter)
18:31 FROGGS yeah, the question is: what fixed it?
18:32 hoelzro sounds like a job for git-bisect!
18:32 jnthn I blame FROGGS!
18:32 jnthn :P
18:33 FROGGS ò.ó
18:33 * FROGGS thinks that it might help the he look angry
18:33 hoelzro jnthn: you blame FROGGS for fixing it? =)
18:33 jnthn Yes :D
18:33 jnthn FROGGS++ fixes lots of things :)
18:34 FROGGS hmmm, I cannot remember touching something in that area
18:36 rurban_ joined #moarvm
18:38 jnthn ah well :)
18:38 jnthn Dammit, why can drunk annoying people be rich enough to afford first class... :/
18:39 FROGGS :/
18:41 kjs_ joined #moarvm
19:03 FROGGS jnthn: that's the only ecosystem regression: http://testers.p6c.org/reports/37368.html
19:03 FROGGS jnthn: I just had to run it, right? I did not have to enable anything, right?
19:04 jnthn right
19:04 jnthn Just build with latest Moar
19:04 jnthn (I didn't bump MOAR_REVISION etc)
19:05 FROGGS jnthn: you see the used versions at the top of the report
19:05 FROGGS moar 2015.3.6.gaa.450.df - that looks good
19:07 jnthn yes
19:08 jnthn except the weird dots :P
19:11 FROGGS jnthn: I'm not to blame this time :o)
19:11 FROGGS but, you get used to the dots
19:16 nwc10 jnthn: I appreciate that the release ship has sailed, but I was a bit, um, unthinking.
19:17 nwc10 Could you correct "PPC32" and "PPC64" to "Power (32 and 64 bit)"
19:17 nwc10 I believe that that is correct (or nearly correct) and certainly far less incorrect than PPC
19:18 jnthn nwc10: Correct it where?
19:18 nwc10 the website
19:18 nwc10 and whatever file in the repo
19:18 nwc10 aha, not in the repo
19:19 jnthn The website is in a repo, but asking you to fix it does no good since you don't have a commit bit :(
19:19 nwc10 sorry, I'm not *that* useful
19:20 * jnthn isn't especially sure what was wrong with the original rendering :)
19:20 dalek moarvm.org: 1403e08 | jnthn++ | index.html:
19:20 dalek moarvm.org: Wording tweak.
19:20 dalek moarvm.org: review: https://github.com/MoarVM/moarvm.org/commit/1403e0805c
19:22 nwc10 if I understand it correctly, PPC is just one variant of the Power family
19:22 nwc10 from a decade ago
19:23 nwc10 so saying "PPC" is sort of "so great, and you're y2k compliant too?"
19:23 jnthn ah, ok
19:23 jnthn Turns out they marketed PPC too well so I thought that was the entire line :P
19:24 nwc10 I think this at times. Hence I made the mistake
19:24 jnthn Not sure if marketing win or fail :)
19:39 jnthn japhb: You've asked before about half-precision floating point. Any idea how portably available that is?
19:39 jnthn japhb: https://gcc.gnu.org/onlinedocs/gcc/Half-Precision.html suggests "not"
19:47 jnthn S09 mentions "16-bit floating-point is also considered optional in this sense."
19:57 japhb http://en.wikipedia.org/wiki/PowerPC has a little of the history of POWER/Power/PowerPC/PPC and such
19:58 japhb I found it interesting that IBM won the contracts for all three of CPUs for the previous generation of gaming consoles.  Apparently in the early days Sony/Microsoft/Nintendo didn't realize that IBM was essentially having the design teams sitting together, which is why mysteriously they had the same clock ratings and such.
19:59 japhb .oO( Gee, 4.2 GHz is awfully specific to have identical on different consoles ... )
20:00 japhb jnthn: It might be worth looking at the LLVM docs, since it's used for doing GPU cross compiles, IIRC.
20:01 japhb jnthn: 16-bit FP ("half") is pervasively supported on GPUs and in any graphics software that supports HDR rendering or intermediates; see http://www.openexr.com/ for the library that kick-started this.
20:02 japhb As you note, it's supported directly on ARM, but also see: https://software.intel.com/en-us/blogs/2013/09/30/intel-half-precision-floating-point-format-conversion-instructions
20:09 jnthn japhb: OK. I think I'm inclined to class this as something I'm happy to have go in, but don't personally feel it's a great use of tuits for me to focus on.
20:10 jnthn 'cus it'll need a bunch of Configure probes to do.
20:13 japhb jnthn: I'm fine with that.  Most important to me is that we not design it to break, like assuming a 16-bit value is a short int.
20:14 jnthn japhb: Given that'd be int/num confusion, I don't see it happening.
20:14 japhb (I wasn't expecting us to do that, I just meant that as an example.)
20:15 japhb I'm basically fine with not having it available soon, as long as it's not hard to add later, is all I'm saying.  :-)
20:17 jnthn ah, oK :)
20:18 dalek MoarVM: 0c95c4b | jnthn++ | src/6model/reprs/MVMArray.h:
20:18 dalek MoarVM: Add type codes for bit-packed array sizes.
20:18 dalek MoarVM: review: https://github.com/MoarVM/MoarVM/commit/0c95c4bc6e
20:20 japhb *Way* more useful to me is 32-bit and 64-bit int/float "reinterpretation", which these days I assume will come from CUnion support.
20:39 kjs_ joined #moarvm
20:42 TimToady joined #moarvm
20:45 FROGGS joined #moarvm
20:47 timotimo oh wow
20:47 timotimo no wonder devirting reprops made things break
20:47 timotimo i mean, performance wise
20:48 jnthn Oh?
20:49 timotimo i had "return jgb_consume_reprop(...)" in the jgb_consume_ins switch
20:49 timotimo wait
20:49 timotimo that's not wrong
20:50 timotimo mrrr
20:51 dalek MoarVM: 74c1982 | jnthn++ | src/6model/reprs/MVMArray.c:
20:51 dalek MoarVM: Fix a possible buffer-overrun in existspos.
20:51 dalek MoarVM: review: https://github.com/MoarVM/MoarVM/commit/74c198283b
20:51 dalek MoarVM: 01be8e3 | jnthn++ | src/6model/reprs/MVMArray.c:
20:51 dalek MoarVM: First steps to handling bit-packed arrays.
20:51 dalek MoarVM:
20:51 dalek MoarVM: This gets them recognized in compose/deserialize_repr. Also factor out
20:51 dalek MoarVM: some repeated code to make this change easier.
20:51 dalek MoarVM: review: https://github.com/MoarVM/MoarVM/commit/01be8e3548
20:54 timotimo and now configure.pl refuses to accept my moar because it's "too old"
20:55 dalek MoarVM/jit_devirtualize_reprops_2: cfff29d | timotimo++ | build/probe.pm:
20:55 dalek MoarVM/jit_devirtualize_reprops_2: Merge branch 'master' into jit_devirtualize_reprops
20:55 dalek MoarVM/jit_devirtualize_reprops_2: review: https://github.com/MoarVM/MoarVM/commit/cfff29da50
20:55 dalek MoarVM/jit_devirtualize_reprops_2: b5962a0 | timotimo++ | src/jit/graph.c:
20:55 dalek MoarVM/jit_devirtualize_reprops_2: when failing from consume_reprop, do it loudly
20:55 dalek MoarVM/jit_devirtualize_reprops_2: review: https://github.com/MoarVM/MoarVM/commit/b5962a05ea
20:55 timotimo er
20:55 timotimo did not mean to push
20:55 timotimo i wanted to rebase this
20:55 dalek Heuristic branch merge: pushed 22 commits to MoarVM/jit_devirtualize_reprops_2 by timo
20:56 timotimo my brane seems to still be napping
20:57 dalek Heuristic branch merge: pushed 22 commits to MoarVM/jit_devirtualize_reprops_2 by timo
21:07 timotimo ah, got it
21:08 dalek MoarVM/jit_devirtualize_reprops_2: ceeb4b5 | timotimo++ | src/jit/graph.c:
21:08 dalek MoarVM/jit_devirtualize_reprops_2: no longer bail from many unhandled repr ops
21:08 dalek MoarVM/jit_devirtualize_reprops_2: review: https://github.com/MoarVM/MoarVM/commit/ceeb4b567c
21:08 timotimo that was a dumb mistake :)
21:34 timotimo i'll walk by my desktop quickly and turn it on so that i can run benchmarks
21:34 timotimo now that i'm not accidentally bailing on a whole lot of jitted frames
21:50 dalek MoarVM/jit_devirtualize_reprops_2: ac1ca37 | timotimo++ | src/jit/graph.c:
21:50 dalek MoarVM/jit_devirtualize_reprops_2: fill out switch-case, revert pointer "hack"
21:50 dalek MoarVM/jit_devirtualize_reprops_2: review: https://github.com/MoarVM/MoarVM/commit/ac1ca372be
22:03 timotimo benchmarks will commence now
22:24 moritz joined #moarvm

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