kelvie

According to nvtop, on both my nvidia and AMD computers, kscreenlocker_greet uses 200-400MB of VRAM while the screen is locked -- doesn't that feel excessive for a simple screen locker (I do realize that it's QML and thus in theory can use as much VRAM as say plasmashell).

This is kind of annoying as I was trying to set up a chatbot using my main desktop while it's idle, and would like that extra 400MB back for a higher context length.

Wasn't sure if this was a bug or just how software is nowadays so I opted to start a discussion rather than finishing filing a bug at bugs.kde.org.

Anyway, anyone know of an alternative screenlocker for kwin_wayland?

I thought I would disable kscreenlocker completely (by setting the screen to never lock?) and use something like swayidle and swaylock, but it doesn't look like kwin supports the wayland extensions required to use swaylock.

Yes, yes, I know, buy AMD, but I already have nVvdia to use CUDA, but this new patch on the nightly branch (on arch, you can use sunshine-git but with my patch here: https://aur.archlinux.org/packages/sunshine-git) finally makes it so that I don't have to "dual boot" into X11 to get game streaming at full performance.

Prior to this, wayland-based streamers had to make a round-trip through CPU ram, and now it stays within GPU ram and thus we can stream 4k on nvidia/Wayland!

... check that the breeding ranch is fully within your base's blue circle.

The building is massive, and the game lets you have part of it outside, but I ran into a ton of bugs with pals going into a weird state where they can't be assigned after they un-assign themselves to the pen (usually when fast travelling).

(Edit: this is before patch 1.3.0)

Anyone heard of this? I've been following it since the first few trailers looked fake, but now I'm more convinced this is going to be a real game (and actually looks kinda good).

Is it just memory bandwidth? Or is it that AMD is not well supported by pytorch well enough for most products? Or some combination of those?

It causes a bunch of weird controller errors, such as when a controller disconnects (or if the deck goes to sleep), it can't reconnect again.

Just let it use the default version (which I believe should be Proton Hotfix, or whatever Valve decides it should be since BG3 is an AAA headliner, they'll want this to work)

I use primarily 8bitdo controllers (in xbox emulation mode, so they show up as xbox controllers), and so far I've ran into several major bugs; wondering if others have experienced the same (i.e. it's something wrong with my setup) or found any workarounds

1. The controllers show up as 2 controllers when steam input is turned on -- to workaround this, I need to go into Steam's controller settings and turn off "Steam Input for XBox controllers", and they show up as one again.

2. It seems to have controllers working, they need to be connected before the game starts.

3. When loading a game with 2 controllers connected to the same computer, if you go into the session manager to move characters around (hit start -> LB I think), your first player's screen disappears forever and is unusable.

4. When a controller disconnects, you have to restart the entire game to reconnect them -- I can't figure out how to rejoin if a controller goes to sleep.

Edit: forgot to mention this is on (stock) steam deck. Turns out the problem was Proton Experimental, it works a lot better when I turn off "Force use of specific compatibility version".

I just had a second Samsung 980 PRO start dying, and had several more in my homelab cluster, so I was dreading doing this on my desktop computer (I have a small form factor), but it was a breeze on the framework.

I used a gparted liveUSB with the firmware on it (the tool is actually just a x86_64 binary), unscrewed the back screws (the captive screws makes it so I don't even need a screw tray), then took off the keyboard, and popped in the SSD, pop the keyboard back in its magnet and go and upgrade.

The brilliant part is that the keyboard is attached by magnets, and is easy to pop on and off to replace the SSDs, and before I knew it, all 5 SSDs were upgraded.

Yes, yes, who needs speeds this fast, but it was cheaper than my 1Gbps plan before this.

Anyway, we switched for 3 weeks and it's been down twice for us now (both on weekends). It's like I'm beta testing their new backbone or something -- any other early adopters want to share their experience?

2023-07-19 Edit: since posting this it went down 3 more times. The TELUS on-hold music is starting to give me PTSD. There was something wrong with their backend where my network access hub kept getting un-registered (and my account getting unregistered), multiple times, but it's been okay for 24 hours now (knock on wood).

A tech came by monday morning to look at it, and all they did was call their backend team, but he gave me specific instructions to give the frontline support, which was useful, but still frustrating.

I'm at I think about a 50% rate on "if we get disconnected, we'll call you back".

I've been using gparted live for the most part to repair all sorts of stuff, but I'm wondering if anyone else has any other more modern recommendations, preferably even ones with Wifi or more graphics card support!

I also find installing deb packages to be way slower than they should be on a modern system (what are deb packages doing that alpine apk and arch packages don't??)

Bonus if they boot fast, too.

Does anyone have a recommendation for someone with a domain name and a k8s cluster where I can set up a simple image and/or video host?

I don't want to have to use Imgur anymore, and am more than capable of setting up cloudflared, and just want to share a screenshot or two, or perhaps show some friends something cool in a video game without having to go through YouTube or Twitch.

Just wondering how many of us use ipv6 for our local hosts, as with my router upgrade, my ISP only allows me to have 253 IP ipv4 addresses (and I don't want to have to buy a new router/gateway, a 10gbe router/gateway is expensive).

Anyway, do you guys use statically assigned ULA addresses? Statically assigned global addresses? DHCPv6? SLAAC? What do you guys do for DNS resolution, avahi/mdns everywhere (given that ipv6 addresses seem to change all the time).

I've currently mostly gotten ipv6 working (dual stack) on machines I touch, my my k3s cluster is out of commission until I can figure out a way to not have them consume any precious ipv4 addresses.

I'm not even sure what prefix I want to choose for the cluster / service CIDR, should I be using a ULA or the one specified https://docs.k3s.io/installation/network-options#dual-stack-ipv4--ipv6-networking, 2001:cafe:42::