5 Followers
22 Following
58 Posts
@pixelcluster i used numbers because of the limited text size, main idea is basically to set (in some way) at runtime that "process xyz should be at ultra priority, process abc should be at high priority, other processes are normal/default", etc. I guess it is fine for cgroups to be predefined (could be an install instruction), AFAICT the requirement is mainly to be able to assign the processes at runtime. I'll need to look into cgroups more though, never had an incentive about them before :-P
@pixelcluster hey, just found this. By reading the post i get the impression this is (or could be) a "VRAM priority" model where you can assign priorities to different processes for their VRAM residency. Say i do some VRAM heavy thing in the background but also want to play a light game, so i want to have game +2, bgtask +1, all else 0. Ignoring the gnarly details, could this be used to write an app like amdgpu_top that lets you pick an process and put it in a "ultra/high/normal" priority group?

Just released a new version of my View3D plugin for #krita that adds Krita 6 support, visibility masks to isolate parts of a 3D model, moat generation (to avoid issues with mipmaps/bilinear filtering) and a bunch of fixes.

https://codeberg.org/badsector/kritaview3d/releases/tag/v20260325

http://runtimeterror.com/tools/kritaview3d/

@logicalerror yeah, this thing - https://www.youtube.com/watch?v=lQsvTrRTBRs - is impressive, but AFAIK the real computation is done in a remote server, which makes it a big fat NO for me. I *hope* the existence of Taalas that prints NNs to silicon to run much much faster than software means there will be "robots with local AI only" without any networking but considering that people these days see smart TVs as a feature, i do not have much hope.
Introducing Helix 02

Last year, Helix showed that a single neural network could control a humanoid’s upper body from pixels. Today, Helix 02 extends that control to the entire ro...

YouTube

@RYStorm yeah, but... i keep using 2.79b since then because i never liked the 2.80 (and later) UI :-P.

The day it stops working will be the day i'll try to compile it myself :-P.

@RYStorm though personally i always have a "do bother me" take since often users try to fix issues themselves in suboptimal ways (i remember a case where a system needed an index map that for w/e reason was shown grayscale in the engine and the artists printed a cheatsheet with brightness ranges -- i replaced that with colors and editor UI for specifying exactly what each index does, but only did that because i was talking with an artist and the figures on the cheatsheet caught my attention)

@RYStorm I found this amusing: "While an engineer may be dangerous to interrupt"

engineers are known to bite when you disturb them in their coding đŸ˜›

The part about the "magic wand" is good but can be a dualsided blade if people treat it as plans (as happened in a recent job i had, though that place seemed to approach every aspect of toolmaking the wrong way).

@RYStorm ah yeah, recordings are also nice. Note that i'm talking from the "receiving" perspective when it comes to sharing the slides outside a talk :-). Personally i keep my own archive of all talks/slides/etc, i found interesting over the years and having these as PDFs (+video recording, if any) is much easier than any other alternative as it allows people to read them from any device and operating system (and there are many tools to work with PDFs too, e.g. for searching).
@RYStorm Actually please have PDFs anyway (+separate video files if needed) since they can be read anywhere :-P
@RYStorm (though i'm not sure about this applying to all sort of software, not only because of technical reasons but also because of incentives -- many websites would rather an ad susceptible human as a visitor instead of a robot - at least until they figure out how to serve ads to bots too :-P)