@lina@vt.social titelbild
@lina@vt.social avatar

lina

@lina@vt.social

Hello everyone, Asahi Lina here! I'm a Linux developer VTuber!
I also help run VT Social!

EN/日本語|she/her|🎨 #AsahiLinArt|My coding cat: @cyan 🩵
Model https://twitter.com/NananoNanase
Design https://twitter.com/shiranui_illust
PFP @mary

Dieses Profil is von einem föderierten Server und möglicherweise unvollständig. Auf der Original-Instanz anzeigen

karolherbst , an Random Englisch
@karolherbst@chaos.social avatar

I'm currently looking into what's the best way to support SVM/USM in #rusticl and one thing I'm wondering about is, if there are any drawbacks doing:

mmap(some_chosen_address, ram_size, PROT_NONE, MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED_NOREPLACE, 0, 0);

and reserve a lot of virtual memory, and then suballocate SVM allocations out of this region with "PROT_READ | PROT_WRITE" and "MAP_FIXED"?

Alternatively, I was considering allocating smaller heaps on demand.

lina ,
@lina@vt.social avatar

@karolherbst @jhwgh1968 I think what you really want is to logically detach the mapping from the backing (and go back to PROT_NONE) without leaving the address space momentarily available, right? So instead of unmapping you just do a mmap(addr, size, PROT_NONE, MAP_PRIVATE | MAP_ANONYMOUS, 0, 0); and that should work I think?

lina , an Random Englisch
@lina@vt.social avatar

Hi I took Lina's phone. Here's a secret about Lina:

She's a sucker for American chocolate chip cookies 🍪

video/mp4

lina OP ,
@lina@vt.social avatar

@technobaboo @cyan Yesssss!!!!

mary , an Random Englisch
@mary@icosahedron.website avatar

quick scribble #rkgk

lina ,
@lina@vt.social avatar

@mary That is sooooooo adorable omg!!!!!!! 💖💖💖💖

thephd , an Random Englisch
@thephd@pony.social avatar

Getting my m1 restored at the Apple Store to deal with not having a second Apple to restore the broken one.

Man I do not like this kind of vendor lock-in!!

(Yes I can restore from Linux; but the Apple device WAS my Linux device 😂😭)

lina ,
@lina@vt.social avatar

@thephd Since restores are done with Finder now without installing anything, I think you can probably just walk up to a demo Mac, plug in with a USB cable to your Mac, and do it yourself? Feels weird but if they're going to make asking for a restore that painful... and it would only take a few minutes.

https://support.apple.com/en-us/108900

lina ,
@lina@vt.social avatar

@millenomi @thephd There is no Network Recovery on Apple Silicon, if the System RecoveryOS doesn't work you need to DFU using another Mac (like with iPhones).

mary , an Random Englisch
@mary@icosahedron.website avatar
lina ,
@lina@vt.social avatar

@mary Waaaa so cute!!!!!!!! Thank you so much for all your amazing fanart!!! 💖💖💖💖

mary , an Random Englisch
@mary@icosahedron.website avatar

Maybe one day I will draw like a pro

lina ,
@lina@vt.social avatar

@mary Your art is amazing!!! Totally pro level ^^

lina , an Random Englisch
@lina@vt.social avatar

It's really silly that OBS doesn't support multichannel audio input (splitting it into multiple audio sources)... it makes the OBS mixer useless for mixing audio in sync! ;;

If you mix all audio outside of OBS (like I do) it works well, but then you can't record split tracks...

Most people will tell you to use something like VoiceMeeter (on Windows) to send audio via separate virtual devices to OBS... but that doesn't keep them in sync since OBS considers them completely independent so there is a random delay offset between them (I tested it).

lina OP ,
@lina@vt.social avatar

BTW, audio monitoring also has a random offset in OBS (since the monitoring path is completely separate from the stream path). The monitoring mix you hear is not the same mix that goes out to the stream!

lina OP ,
@lina@vt.social avatar

Basically I only use the OBS mixer for stuff like the intro/waiting screen BGM (where I don't care if it's in sync with the mic). Everything else including my stream BGM gets mixed in outside of OBS ^^

lina OP ,
@lina@vt.social avatar

@wuest Yeah, that doesn't guarantee sync between tracks but at least you can fix it in post if you need to... I did something similar for Cyan's stream just now (one track full mix, one track voice only).

lina OP ,
@lina@vt.social avatar

@scherzog That's the problem, if you have separate instances of stereo streams there is no guarantee that they are synced together... there needs to be something on the OBS side that takes a synchronized multichannel input natively and splits it into sources/tracks in a way that guarantees the sync is preserved, but I don't think that exists...

lina OP ,
@lina@vt.social avatar

@imikoy I've done that before ^^;; though for streams it's overkill, might as well just use jackrec and record a parallel multichannel wav for each stream.

lina OP ,
@lina@vt.social avatar

@scherzog OBS supports routing to separate tracks, so that is in some way a "multichannel" concept... but I don't know if it supports it at the source level. It might need some core code changes...

lina OP ,
@lina@vt.social avatar

@lanodan Depends on your setup, but at least on Windows definitely enough that I can tell listening to the mix when singing karaoke and stuff like that.

If the underlying audio device is the same at the end of the day then at least it should not drift over time during normal usage. However, if there are any glitches due to CPU overload or things like that, they could cause the sync to change suddenly if different sources skip different amounts of audio.

lina OP ,
@lina@vt.social avatar
  • Alle
  • Abonniert
  • Moderiert
  • Favoriten
  • random
  • haupteingang
  • Alle Magazine