github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
brrt \o 15:43
[Coke] o/ 16:23
brrt everyone still healthy around here? 16:25
MasterDuke doing ok so far 16:26
jnthn I'm better from my throat infection that made me cancel one of my GPRW talks :) 16:27
But I think the unreasonably warm February has made something blossom early...
MasterDuke that somehow manages to sound both nice and ominous 16:32
jnthn The nice outweighs the ominous; I don't suffer too badly from those allergies. 16:39
I used to think it was odd that I always seemed to have a mild cold through a load of spring. Was only in the last years I realized...d'oh, it's not a mild cold... :P 16:41
jnthn (Or rather, had it pointed out to me...) 16:42
MasterDuke i never had any allergies in the US, but the entire (almost) year we've been UK i've had some mild allergies 16:43
jnthn: btw, just noticed that loop() problem is related to protos/multis 16:44
m: class FOO { our proto sub foo($) {*}; multi sub foo(Int $n) { my @a = ^$n; my $b; loop (my $i = 0; $i < $n; $i++) { $b = @a[$i] }; $b }; }; say FOO::foo(1_000_000); say now - INIT now
camelia 999999
2.4714556
MasterDuke m: class FOO { our sub foo(Int $n) { my @a = ^$n; my $b; loop (my $i = 0; $i < $n; $i++) { $b = @a[$i] }; $b }; }; say FOO::foo(1_000_000); say now - INIT now 16:45
camelia 999999
0.63979531
jnthn Hm, intersting
*interesting
MasterDuke a spesh log of the proto/multi version has ~850 entries of 'Latest statistics for ' foo/<unit>/MAIN/a bunch more 16:47
a spesh log of the single version as 10 entries for them
MasterDuke cause any lightbulbs to go off? 16:49
timotimo jnthn: i've been doing a "monthly injection to lessen allergy symptoms" thing. can not say whether it helps or not, but it's fun 17:25
MasterDuke i thought at first the 'after' spesh of `foo` was different in the two examples, but when i set MVM_SPESH_BLOCKING=1 they turned out to be identical (not the entire logs of course, they still had the discrepancies in 'Latest statistics for' lines) 17:30
another data point. the proto/multi version had almost 1m 'OSR hits' by the end of the log. the single sub version had 300 17:37
from a profile of the no proto version: 5174 call frames were entered and exited by the profiled code. Inlining eliminated the need to create 10996665 call frames (that's 99.95%). 17:44
linkable6 (2017-08-06) github.com/rgmcrary/Train-Schedule...109966659e Finalize
MasterDuke from the proto version: 4003706 call frames were entered and exited by the profiled code. Inlining eliminated the need to create 6998169 call frames (that's 63.61%).
MasterDuke article about ocaml's new GC strategy www.ocamlpro.com/2020/03/23/ocaml-n...collector/ 19:48

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!