github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
nwc10 good *, #moarvm 07:02
MasterDuke dogbert17: fwiw, i can't repro at moarvm and nqp HEAD 08:45
nine me neither 08:57
dogbert17 oops 09:02
Welcome to Rakudo(tm) v2020.12-39-g53da2061f. Implementing the Raku(tm) programming language v6.d. Built on MoarVM version 2020.12-56-g1e09c82b5. 09:03
dogbert17 nukes rakudo install and retries 09:11
Geth MoarVM/master: 27 commits pushed by (Stefan Seifert)++, niner++
review: github.com/MoarVM/MoarVM/compare/1...b8a815f97f
dogbert11 the errors remain, how odd 09:25
nine dogbert11: do you have an idea about what goes wrong? 09:26
dogbert11 weren't there changes made to spawnprocasync recently 09:28
Altai-man what's the matter, is it not safe to bump now? 09:29
nine commit 22f1038a952a89d93dd5389afe9b2fcd88397e54
dogbert11 prove -j0 -r -v --exec nqp t/nqp/111-spawnprocasync.t 09:30
not ok 2 - got the correct output on stdout
fails on my RPi 4 as well, the plot thickens 09:36
nine, MasterDuke: are you running 'This is nqp version 2020.12-4-gd65a8c109 built on MoarVM version 2020.12-83-ga5b8a815f' ? 09:37
MasterDuke This is nqp version 2020.12-7-gf7a0081b2 built on MoarVM version 2020.12-83-ga5b8a815f 09:38
dogbert11 ah, you're running a newer nqp 09:40
MasterDuke i think you're missing the merges of PRs 689 and 690
dogbert11 perhaps Altai-man should bump then :)
MasterDuke: it was the missing merges so false alarm 09:43
dogbert11 hello brrt 10:10
nine: still satisfied with your new monster machine? 10:11
brrt hello dogbert11 10:22
nine dogbert11: absolutely! Works like a charm. And as a bonus I finally have reliable sleep mode on my desktop 11:11
jnthn nine: A threadripper, or something else? 11:56
tellable6 2021-01-10T20:51:01Z #raku-dev <Xliff> jnthn When NativeCall loads a dynamic library, will it also load its prerequisites?
2021-01-10T20:59:46Z #raku-dev <nine> jnthn: I answered Xliff's question about dynamic libraries
nine jnthn: Ryzen 9 5950X, 16 cores/32 threads. Gets a spectest done in ~77s
jnthn nine: Wowser. How is its single-core performance? 11:57
nine CORE.c stage parse in about 30s
jnthn (For example, building CORE.setting)
hah, we thought of the same benchmark :D
nine it's what matters :D
jnthn Given how the pandemic seems to last for darn ever, I'm thinking to get something better for working at home. 11:58
Though that beats my machine at the office handily too
I want to get back to RakuAST hacking soon but it's quite painful on my current home box :/ 12:00
nine It's even slowed down a bit because I only have DDR3000 memory. Though I haven't tested yet how much this impacts the important workloads
sena_kun jnthn, latest ryzen plus 36000Mhz memory and you're ready. 12:01
jnthn Is there one too many 0 in that number? :) 12:02
MasterDuke unfortunately www.phoronix.com/scan.php?page=art...&num=1 doesn't have any code compilation benchmarks
sena_kun jnthn, yesh
nine well, an order of magnitude faster has never hurt anyone ;)
MasterDuke openbenchmarking.org/result/200706...&sgm=1 doesn't show much of a difference in compiling the kernel or llvm 12:05
but it's zen2, not zen3 12:06
nine Also the more cores you have, the more contention is around those to memory channels. And a C compiler may still have different memory usage than rakudo 12:07
sena_kun "much" - no, but there is a difference between 3000 and 3600 in most tasks and if you can throw money at 5950X, getting some ram is probably not an issue.
Xliff Getting SEGV from rakudo, but stacktrace points at MoarVM 17:26
MVM_spesh_log_invoke_target (tc=0x55555555a160, invoke_target=0x0, was_multi=0) at src/spesh/log.c:194
194 if (REPR(invoke_target)->ID == MVM_REPR_ID_MVMCode && IS_CONCRETE(invoke_target)) {
There is NO golf, as yet.
timotimo but is there a corsa? 17:59
lizmat you mean in ID.3 ? 18:11
timotimo hm? 18:13
Xliff: that smells like memory corruption of some kind; MVM_spesh_log_invoke_target would record what some invoke instruction actually invoked 18:14
i haven't checked, but i imagine invoking a literal null should have caused an exception earlier than that 18:15
lizmat timotimo: the modern successor of the VW Golf :-) 18:16
timotimo aah 18:18
Xliff timotimo: I think it has something to do with traits and dynamic variables. You will have to wait for the gist. I still have things to do for $dayJob before I can even start on that. 18:30
Thanks for the comment, though.
lizmat and the first Rakudo Weekly News of 2021 hits the Net: rakudoweekly.blog/2021/01/11/2021-...new-stuff/ 20:08
nine lizmat++ # wow, there's really much going on 20:13
lizmat yeah, I was pleasantly surprised 20:14
nine It's easy to miss when one just hangs around here. Funny, initially I read the weekly (way back then) to follow what happens in core development. Now I follow it to read what happens outside the core :) 20:15
jnthn lizmat++, and yes, really lots going on 20:37
brrt` lizmat++ :-) 20:38

Powered by Cro and the Raku® Programming Language.

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

Thank you!