Perl 6 - the future is here, just unevenly distributed

IRC log for #moarvm, 2017-10-22

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

All times shown according to UTC.

Time Nick Message
01:55 ilbot3 joined #moarvm
01:55 Topic for #moarvm is now https://github.com/moarvm/moarvm | IRC logs at  http://irclog.perlgeek.de/moarvm/today
02:46 greppable6 joined #moarvm
02:46 evalable6 joined #moarvm
02:46 quotable6 joined #moarvm
02:46 nativecallable6 joined #moarvm
02:46 benchable6 joined #moarvm
02:46 releasable6 joined #moarvm
02:46 unicodable6 joined #moarvm
02:46 coverable6 joined #moarvm
02:46 bloatable6 joined #moarvm
02:46 committable6 joined #moarvm
02:46 bisectable6 joined #moarvm
02:46 squashable6 joined #moarvm
02:46 statisfiable6 joined #moarvm
03:14 evalable6 joined #moarvm
03:44 ggoebel joined #moarvm
04:07 bloatable6 joined #moarvm
04:20 committable6 joined #moarvm
07:18 evalable6 joined #moarvm
07:21 domidumont joined #moarvm
07:29 domidumont joined #moarvm
08:46 evalable6 joined #moarvm
09:03 nwc10 joined #moarvm
11:09 greppable6 joined #moarvm
11:15 committable6 joined #moarvm
12:41 MasterDuke joined #moarvm
16:03 buggable joined #moarvm
17:09 timotimo whoops, gotta put a little fix in for my debug name changes
17:11 Geth ¦ MoarVM: 19a2001bac | (Timo Paulssen)++ | src/spesh/dump.c
17:11 Geth ¦ MoarVM: fix wrong access to debug_name in spesh dump
17:11 Geth ¦ MoarVM: review: https://github.com/MoarVM/MoarVM/commit/19a2001bac
17:11 timotimo AlexDaniel`: this should make it into the release
17:28 zakharyas joined #moarvm
18:24 brrt joined #moarvm
18:28 brrt good * #moarvm
18:30 timotimo good * brrt
18:30 brrt hmm interesting...
18:30 timotimo whatcha got?
18:30 brrt waitaminute
18:31 brrt oh, i need to set MVM_SPESH_BLOCKING i think
18:31 timotimo that often helps
18:31 brrt i'm debugging an issue in the optimizer
18:32 timotimo mhm
18:32 timotimo the spesh optimizer or the jit optimizer? :)
18:32 brrt jit optimizer
18:32 brrt i'm pondering generalizing the jit bisect logic
18:33 brrt well, we can do that partially with MVM_SPESH_LIMIT, can't we?
18:33 timotimo ?
18:34 brrt i'm pondering generalizing the debug log as well
18:34 brrt and optimizing it away for the common case
18:34 brrt so rather than MVM_jit_log(tc, ''…")
18:34 brrt i'd have
18:41 timotimo if it takes that long to type, maybe it's not a good idea
18:53 brrt hehehe
18:54 brrt no
18:54 brrt i mean
18:54 brrt i want MVM_JIT_DEBUG(msg, args)
18:54 brrt but, i want that to do two things
18:54 brrt - compile-time switchable
18:54 brrt - sensitive to env var
18:55 brrt - speciifc to frame / block
18:55 brrt - specific to compilation phase
18:55 brrt so you could have MVM_JIT_DEBUG=ALL and it would dump everything
18:56 brrt or you could have MVM_JIT_DEBUG=REGISTER and dump only register allocation
18:56 brrt or you could have MVM_JIT_DEBUG=REGISTER/21/17:ALL/102/3:OPTIMIZER/29/3
18:56 brrt something like that
18:57 brrt tuneable
18:57 zakharyas joined #moarvm
18:57 timotimo oh
18:57 brrt and if it'd need compile-time support, who cares if it was 'expensive' to do
19:08 brrt joined #moarvm
19:12 greppable6 joined #moarvm
19:12 committable6 joined #moarvm
19:18 brrt i found the optimizer bug, yay
19:19 brrt actually
19:19 timotimo cool!
19:19 brrt it is kind of scary
19:20 brrt it is a sp_p6obind_o
19:20 brrt issue
19:20 brrt and it comes to front by the introduction of a few copy nodes
19:22 timotimo interaction with object registers and GC?
19:25 brrt no, more innocent than that
19:26 brrt but also a bit more scary
19:26 brrt the tl;dr
19:26 * timotimo listens intently
19:27 brrt sp_p6obind_o uses a macro write_barrier
19:27 brrt this macro conditonally loads the value to be assigned
19:27 brrt anyway, it uses a direct reference to the $2 (load) sequence, and evaluates that conditonally
19:29 brrt the insertion of the COPY forces the evaluation and makes it the same
19:29 brrt but
19:29 brrt the condition is not always kept
19:29 timotimo oooh
19:29 brrt so the two worries i have
19:29 brrt - this bug is always there in this case
19:30 brrt because the expr jit is allowed to tile them as one and the same value
19:31 brrt - i'm not sure whether to fix this in this specific template (and all other instances) by declaring a let:, or whether to somehow make the optimizer aware of inadvertent conditionally-evaluated references to values that are used in a block that isn't consequent to it
19:31 timotimo oof, good question
19:32 brrt obviously the first is much simpler than the second
19:32 timotimo yeah, but much less thorough
19:32 brrt true
19:33 brrt and strictly speaking, this *is* a template bug
19:33 timotimo i.e. the problem could reappear again and again
19:33 brrt because the value being referenced - $2 - well, it's being loaded conditionally first
19:33 brrt in the code, and that's not legal, since the expr JIT has no obligation to 'reload' the value
19:34 brrt the other way to fix it is to have the analyzer decouple them
19:34 Ven joined #moarvm
19:34 brrt but that's even less than ideal
19:41 AlexDaniel joined #moarvm
19:42 brrt the other fix is not to do this kind of optimization just yet
19:42 brrt but that's not where i want to go with this
19:46 * timotimo BBIAB
19:56 bartolin joined #moarvm
19:57 zakharyas joined #moarvm
20:07 ilmari[m] joined #moarvm
20:28 brrt joined #moarvm
21:19 btyler joined #moarvm
21:52 lizmat joined #moarvm
23:26 Geth ¦ MoarVM: 833803b6a5 | (Samantha McVey)++ | docs/ChangeLog
23:26 Geth ¦ MoarVM: Add changelog entries for the next release
23:26 Geth ¦ MoarVM: review: https://github.com/MoarVM/MoarVM/commit/833803b6a5
23:27 samcv .tell brrt can you add what you would like to the changelog for the expression engine? You'd be the best one to know what should be added to the changelog
23:27 yoleaux samcv: I'll pass your message to brrt.
23:38 timotimo samcv: i think the "check caller chain before promotion" is only active when GC_DEBUG (or something) is on? maybe that detail wants to be pointed out?
23:42 timotimo samcv: "seakable" wants to be "seekable"
23:47 samcv here's a seaking https://cdn.bulbagarden.net/upload/thumb/6/6a/119Seaking.png/250px-119Seaking.png
23:47 samcv will change :p
23:47 timotimo i like seadra more

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