github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
nwc10 good *, #moarvm 07:45
nine TGIF, #moarvm 07:59
Geth MoarVM: f7b4e490f6 | (Patrick Böker)++ | 6 files
Switch spawnprocasync to use a separate arg for the program name

That's a lot more obvious than putting the program name at the start of the args list.		 08:32
MoarVM: 2d9899c726 | (Patrick Böker)++ (committed using GitHub Web editor) | 6 files
Merge pull request #1416 from patrickbkr/spawnprocasync-prog-arg

Switch spawnprocasync to use a separate arg for the program name
travis-ci MoarVM build failed. Patrick Böker 'Merge pull request #1416 from patrickbkr/spawnprocasync-prog-arg 08:57
travis-ci.org/MoarVM/MoarVM/builds/754595860 github.com/MoarVM/MoarVM/compare/a...9899c726a8
jnthn moarning o/ 10:34
tellable6 2021-01-12T23:50:00Z #raku <Xliff> jnthn Is there any way to introspect the symbol table of a template during render()?
jnthn .tell Xliff I assume you're talking about Cro::WebApp? In which case a) no, b) there's a #cro that's good for such questions :) 10:38
tellable6 jnthn, I'll pass your message to Xliff
nwc10 \o 10:42
nine For those interested in these topics: I tried finding out if getting faster memory to accompany my 16 core CPU was worth it. Since I don't have faster memory to test with, instead I disabled XMP in the BIOS, turning my memory down from 3000MHz to 2150MHz. 14:39
The effect on CORE.c stage parse is....nonexistent. In a best-of-3-runs comparison we're talking 28.681 vs. 28.731s, which is very well within noise. 14:40
Spectest however clearly profits from faster memory. Here we are talking about 76 vs. 89s (again best-of-3-runs) 14:41
So buying faster memory will in all likelyhood not reduce rakudo's build time. I could get faster spectests though (maybe ~70s?).
vrurg nine: I think this is could be explained by spectesting using concurrency heavily while compilation is done mostly single-thread. 14:42
nine As to why there is such a large discrepancy between single- and multithreaded results, I think the huge 64 MiB of L3 cache makes memory speed a non-issue during the build, but with 35 processes competing for that cache, memory speed gains in importance. 14:43
jnthn nine: That's interesting to know. I'm getting a new machine soon, and was debating 3600 MHz vs 4000 MHz RAM, and all the reviews I could find suggested it wasn't really worth it. I guess it's about half the difference you tried to observe (and there improvement is non-linear anyway) though... 15:48

Powered by Cro and the Raku® Programming Language.

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

Thank you!