I’ve always found the documentation around virtio-GPU and virtgl very lacking, and have never gotten them working. Would love to get pointers if anyone has a good source.

I don’t see any performance differences with the vgpu actually. I have more performance bottlenecks with the CPU, and my RAM isn’t the fastest, so I think I’m more CPU limited. Benchmarks I have run that are GPU focused seem to show little to no difference from what the physical card would do.

Yeah unfortunately. 20xx is last generation supported so far via the patch, not sure if support for later cards is coming or not.

No, but I think you’d have some problems. Only the host has access to the actual DisplayPort outputs, all the vgpus have virtual displays, I don’t think there’s a way to make them use the physical out.

Sure, but you’ll get diminishing returns most likely as consumer hardware doesn’t really have the resources to scale that way very well if all the VMs are running demanding apps simultaneously.

Even for something like 4 VMs that just do NVenc, there are limits for how many streams the GPU can do. I think there’s another patch that lets you raise that, but at some point you’ll run out of resources quick. Even powerful consumer gear isn’t really designed to be used by more than one user/app and it starts to show the more you virtualize and split those resources.

I’ve been doing exactly that at home for a couple years now. First with Parsec, now Sunshine/Moonlight.

Host is Proxmox on Ryzen 5800x, 64gm RAM GPU is 2070 Super, with VGPU patched drivers from https://gitlab.com/polloloco/vgpu-proxmox

When I’m gaming I’ll dedicate the full 8Gb to my windows Vm, otherwise I split it in 2 or 4Gb chunks to Jellyfin or my home camera monitoring. 8gb can’t split very many ways, and most things require at least 2 to run.

Locally at home I can run 1440p 60fps rock solid over wifi on any device, from my phone/old laptop/apple tv/raspberry pi. Remote I can do 1080p60, but a bit more hit or miss depending on my network connection.

Experimenting with LLMs I’ve done through the same windows VM, or to a ubuntu dev VM. Works the same way. I’m thinking of transitioning my gaming VM to Linux too.

The amount of VRAM is the hard limitation to get past, the virtualization tech itself has been there for a while.

But to be perfectly honest……it really was just a “let’s see if I could do this” type task, direct GPU pass though is more straightforward and it’s not really worth splitting 8Gb these days. Unless you get a card with significantly more VRAM passthrough is much less work.

Yes. That’s what allows Unix legends like this: https://www.ee.torontomu.ca/~elf/hack/recovery.html

In most games not noticeable. Only game I have trouble with is emulating Wii, playing Mario Galaxy. The pointer on screen lags, but I think that’s more due to the bluetooth adapter compatibility than any latency added by the usb-> ip -> wifi link.

I’m not an FPS player, so can’t speak to sub second latency….but I do racing sims on this, and it has no trouble with controls and force feedback.

In the house, anywhere with wifi. Can run decently down to 10-15mbps at 1080p60.

Remotely, over Tailscale, my home uplink is too slow for anything more than 720p60, but its low latency enough I can play games like Mario RPG and get timed hits correct. Or Clone Hero. Games like rocket league tend to be too fast tho, and video breaks up badly.

so Long as you have fast enough uplink, I think I’d be fine anywhere. Sunshine and moonlight are amazing, I used to use Parsec extensively but now it’s just moonlight and sunshine.

Works great. It’s my portable gaming box. I use virtualhere usb over ip on the same Pi too so I can use multiple controllers like a wheel or joystick, pass a full bluetooth adapter directly to it for emulators.

You keep the user-changeable files on a separate filesystem. Whether that’s just a separate partition, or an external disk. Keep the system itself read only, and write-heavy directories like logs and caches in RAM.

Go for a vintage correct OS for a challenge, try Haiku!

I had a pi 1b running my hvac/humidifier/HRV unit at home for years. Only removed it when we moved out.

So I’m SUPPOSED to run a miner to keep mine from being overly idle??

The most annoying part I think is because I so rarely need them. All my Pis run headless, but the one time I do need direct console access I have to find the bloody adapters. Leaving them attached and unused is just asking them to get damaged.

Rather than using micro-hdmi (which hardly anything uses), stick a pair of usb-c DP ports instead if size is an issue. at least then I don’t need adapters that are ONLY needed for the Pi.

I loved Pi’s, but I hate the micro hdmi connectors

Depends on your system. Desktop have different requirements than servers.

On both at minimum, I’d keep /home and /var/log separate. Those usually see the most writes, are least controlled, and so long as they’re separate partitions they can fill up accidentally and your system should still remain functional. /tmp and /var/tmp should usually be mounted separately, for similar reasons.

/boot usually keep separate because bootloaders don’t always understand the every weird filesystem you might use elsewhere. It would also be the one unencrypted partition you need to boot off of.

On a server, /opt and /srv would usually be separate, usually separate volumes for each directory within those as well, depending how you want to isolate each application/data store location. You could just use quotas; but mounting separately would also allow you to specify different flags, i.e. noexec, nosuid for volumes that should only ever contain data.

/var/lib/docker and other stuff in /var/lib I usually like to keep on separate mounts. i.e. put /var/lib/mysql or other databases on a separate faster disk, use a different file system maybe, and again different mount options. In distant past, you’d mount /var/spool on a different filesystem with more inodes than usual.

Highly secure systems usually require /var/log/audit to be separate, and needs to have enough space guaranteed that it won’t ever run out of space and lock the system out due to inability to audit log.

Bottom line is its differnet depending on your requiremtns, but splitting unnecessarily is a good way to waste space and nothing else. Separate only if you need it on a different type of device, different mount options, different size guarantees etc, don’t do it for no reason.