github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
00:45 vrurg left 00:46 vrurg joined 00:51 vrurg left 01:08 vrurg joined 06:44 vrurg left 06:45 vrurg joined
nine zostay: /win 10 06:46
06:49 vrurg left 06:55 harrow left 06:56 harrow joined 07:19 vrurg joined
nwc10 good *, #moarvm 07:23
07:24 zakharyas joined
nine *, nwc10 07:44
08:16 leont_ joined
nwc10 just when you thought that you knew what was going on, you find github.com/senderista/hashtable-benchmarks 08:27
(actually, I don't know what is going on, so I'm building with GH_DEBUG at 3)
08:42 patrickb joined 08:48 sena_kun joined 08:53 Altai-man_ joined 09:13 sena_kun left, sena_kun joined 09:14 Altai-man_ left
jnthn morning o/ 09:20
lizmat jnthn /o 09:21
nwc10 \o
09:27 sena_kun1 joined 09:57 sena_kun1 left
Geth_ MoarVM: ffontaine++ created pull request #1326:
Fix build with gcc 4.8
10:10
11:06 vrurg left 11:07 vrurg joined 11:11 Altai-man_ joined, vrurg left 11:13 zakharyas left 11:14 sena_kun left 11:42 sena_kun joined 11:45 vrurg joined 11:51 vrurg left, vrurg joined 12:03 brrt joined 12:12 brrt left 12:25 zakharyas joined 12:32 squashable6 left 12:36 squashable6 joined 13:14 sena_kun left, Altai-man_ left 13:15 sena_kun joined 13:19 MasterDuke left 13:28 sena_kun left 13:29 sena_kun joined 13:38 Altai-man_ joined 14:25 patrickb left 14:27 sena_kun left 15:13 sena_kun joined 15:14 Altai-man_ left
Kaiepi is it bad style to do for loops with pointer arithmetic? 15:45
i grepped for them to see if anyone else does them, but didn't see any
nine Kaiepi: well it is C :) 16:15
17:12 Altai-man_ joined 17:14 sena_kun left 17:22 zakharyas left 18:13 sena_kun joined 18:18 squashable6 left 18:21 squashable6 joined 18:43 sena_kun left 19:13 sena_kun joined 19:14 Altai-man_ left 19:23 squashable6 left 19:24 squashable6 joined 19:27 zakharyas joined 20:03 moon-child joined 20:07 moon-child left 20:38 MasterDuke joined 20:56 zakharyas left 21:12 Altai-man_ joined 21:14 sena_kun left
timotimo usually use index arithmetic and expect to see compiler to make the same code in the end 22:31
have not yet verified but expected to be just fine 22:35
jnthn I vaguely recall my compilers professor going through an example which was nice and clear using indexes, then one using pointer arith, and I think the punchline was that in that particular example the more readable code came out faster after optimization. 22:45
Of course this won't generalize to every case.
timotimo wonder if you should go look through the code for opportunities to use branchless programming 22:48
and surely theres something for the JIT optimizer to do at some point in the future 22:49
22:58 MasterDuke left 23:13 sena_kun joined 23:14 Altai-man_ left