Cool cool, but hasn't this been achived by multiple projects many many times before?
I'd say that cost is a significant factor for doing this, and if you can leverage the usage of those fancy kubernetes GPU-cluster providers it'd come out even cheaper. Sure, it wouldn't be as "sophisticated", but it'd perform more or less identical and cost far far less.
All the above used cloud-providers are expensive, if compared to something like vast.ai or runpod.
You're right cost is an issue, that's why Cloudy Pad is mostly suitable for occasional players (up to 30h or 40h / months) in which case it's still cost efficient to use such Cloud providers.
> more or less identical and cost far far less.
I wish it would, for I have tried: the problem is GPU availability which is often low or highly elastic. Considering data are persisted in a given region, there's high chance your RTX 4090 will be unavailable when you want it.
These platforms are great for AI and inference since you can easily hop onto another GPU type and it's not too much an issue if a certain GPU type is not available at a given time. Not so much for gaming purpose :(
Major Cloud providers have much more availability on their infra.
Yes, they build off other open source projects and list them in the README. I actually do run moonlight + sunshine myself, and have for more than a year. It's not too difficult to setup, depends on what you want to achieve.
I hadn't heard of wolf [0], but it checks a lot of boxes that sunshine does not. Namely, it supports multiple clients at once, multiple streams, and virtual displays out of the box (Linux + container first is almost neat). Sunshine is more for allowing your gaming desktop to be used as for game streaming. There's also a fork of sunshine, Apollo [1], that's more similar to wolf.
I'd say that cost is a significant factor for doing this, and if you can leverage the usage of those fancy kubernetes GPU-cluster providers it'd come out even cheaper. Sure, it wouldn't be as "sophisticated", but it'd perform more or less identical and cost far far less.
All the above used cloud-providers are expensive, if compared to something like vast.ai or runpod.
Check this out; https://github.com/selkies-project/selkies
> more or less identical and cost far far less.
I wish it would, for I have tried: the problem is GPU availability which is often low or highly elastic. Considering data are persisted in a given region, there's high chance your RTX 4090 will be unavailable when you want it.
These platforms are great for AI and inference since you can easily hop onto another GPU type and it's not too much an issue if a certain GPU type is not available at a given time. Not so much for gaming purpose :(
Major Cloud providers have much more availability on their infra.
> Check this out; https://github.com/selkies-project/selkies
Thanks, this is one of the project I had in mind beside Wolf and Sunshine but haven't gone as far. A good reminder to try it out again :)
I hadn't heard of wolf [0], but it checks a lot of boxes that sunshine does not. Namely, it supports multiple clients at once, multiple streams, and virtual displays out of the box (Linux + container first is almost neat). Sunshine is more for allowing your gaming desktop to be used as for game streaming. There's also a fork of sunshine, Apollo [1], that's more similar to wolf.
[0] https://github.com/games-on-whales/wolf [1] https://github.com/ClassicOldSong/Apollo