Perl 6 - the future is here, just unevenly distributed

IRC log for #moarvm, 2015-04-21

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

All times shown according to UTC.

Time Nick Message
02:40 vendethiel joined #moarvm
02:59 dalek MoarVM/no-moar-linenoise: 1d37511 | hoelzro++ | / (13 files):
02:59 dalek MoarVM/no-moar-linenoise: Remove linenoise/readline
02:59 dalek MoarVM/no-moar-linenoise:
02:59 dalek MoarVM/no-moar-linenoise: Now that NativeCall is included with Rakudo, instead of having
02:59 dalek MoarVM/no-moar-linenoise: linenoise/readline functionality be bundled in as a VM-level op,
02:59 dalek MoarVM/no-moar-linenoise: let's create a NativeCall-based module for this functionality
02:59 dalek MoarVM/no-moar-linenoise: review: https://github.com/MoarVM/MoarVM/commit/1d37511726
03:07 xiaomiao joined #moarvm
05:06 timotimo joined #moarvm
05:25 vendethiel joined #moarvm
06:38 brrt joined #moarvm
06:38 vendethiel joined #moarvm
06:51 brrt \o
06:54 FROGGS joined #moarvm
06:56 brrt with regards to the tracing discussion, i see no other plausible way than to add a check in MVM_frame_invoke
06:57 brrt something like if (tc->is_tracing) { if (!code->tracing_frame) code->tracing_frame = MVM_spesh_add_tracing(tc, frame); invoke(tc, code->tracing_frame) } } else { ... }
06:57 jnthn That's certainly an easy way to do it, yes
06:58 brrt that is not too expensive for the general, non-tracing form, and it catches all mvm level invocations
06:58 jnthn The other is to tweak all the invokish ops to some other op that takes the invokish op in question as a parameter, but that's getting a bit awkward.
06:58 rurban joined #moarvm
06:59 brrt it can be done.. but sometimes the invokish bit is hidden rather far (istrue/isfalse come to mind)
07:01 jnthn *nod*
07:11 zakharyas joined #moarvm
07:36 * brrt wishes he had more time :-(
07:37 jnthn All being well, you'll have a good amount of time to make a nice difference with the JIT. :)
07:39 brrt that is true
07:40 brrt the last year has been so incredibly hectic :-o
07:44 vendethiel joined #moarvm
08:37 vendethiel joined #moarvm
10:07 * brrt lunch &
10:10 vendethiel joined #moarvm
10:30 Ven joined #moarvm
10:36 vendethiel joined #moarvm
10:51 Ven joined #moarvm
11:25 vendethiel joined #moarvm
11:35 brrt joined #moarvm
11:49 Ven joined #moarvm
11:52 rurban joined #moarvm
12:08 ggoebel joined #moarvm
12:42 vendethiel joined #moarvm
13:03 Ven joined #moarvm
13:20 vendethiel joined #moarvm
13:45 vendethiel joined #moarvm
13:47 ggoebel joined #moarvm
13:53 Ven joined #moarvm
13:57 brrt` joined #moarvm
14:20 timotimo i wonder at what point the trace jit will throw in guards against execution that'd differ?
14:21 Ven joined #moarvm
14:25 vendethiel joined #moarvm
14:26 rurban joined #moarvm
14:51 vendethiel joined #moarvm
15:01 Ven joined #moarvm
15:56 vendethiel joined #moarvm
16:05 ggoebel joined #moarvm
16:35 FROGGS[mobile] joined #moarvm
17:23 vendethiel joined #moarvm
17:39 lizmat joined #moarvm
17:50 FROGGS joined #moarvm
18:29 vendethiel joined #moarvm
18:42 lizmat_ joined #moarvm
19:22 synbot6 joined #moarvm
19:37 brrt joined #moarvm
19:40 brrt timotimo: long story really short
19:40 brrt in the trace you add only the bbs that are actually visited, and *all* the bbs that are visited below a level
19:41 brrt once you'd branch out of any of these bbs you deopt
19:42 brrt the assumption is that some paths within and through a routine are really hot, and some paths are not, even if they are in the same routine
19:44 brrt so if you'd have something like: sub foo { if $condition { bar(); } else { something-else(); } }; sub bar() { if $unlikely { branch-not-taken(); } 42; }
19:45 brrt then a trace starting at foo, given that condition would always yield true, would simply jump all the way to the end of bar
19:46 brrt sidenote; trace jit would require a generalised sp_guard_istrue
19:56 zakharyas joined #moarvm
20:04 brrt tracing spesh, i mean
20:18 jnthn Something like that, yeah
20:18 jnthn I hope we can mostly re-use the deopt mechanism we already have to power the deopt too
20:18 jnthn You can also do side-traces for commonly taken other branches
20:19 brrt right, but you'd only do that if they appear in the trace
20:19 brrt a trace isn't necessarily linear
22:59 flussence joined #moarvm
23:41 LLamaRider joined #moarvm

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