Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Shadow: Rent a beasty machine accessible from anywere (shadow.tech)
11 points by BiteCode_dev on Jan 26, 2022 | hide | past | favorite | 9 comments


I was a customer of theirs for the better part of a year, up until I finally got my hands on an xbox series x. The good is that it generally performs as advertised: really smooth streaming, very low latency, never had an issue with installing any games I wanted to.

Downsides were that it became much pricier over time, there was a long waitlist to get access to it, and I did have some trouble getting my xbox controllers to work consistently with it from a Mac host. (Some of those may be worked out now.)

I think in general its an excellent solution to a problem most people don't have. It didn't really outperform stadia or Nvidia GFN by a wide enough margin, and the only real advantage was having the whole of a windows box to work with when you needed it (I never did, and probably would have used a virtual machine if I did.)


Is the latency really acceptable for gaming?


"Gaming" isn't a single activity with a single threshold above which latency is unacceptable. Competitive FPS players will probably never find streaming latency acceptable. A child playing roblox probably always will. The rest of us land everywhere in between.

If you have a specific use and requirement you'll pretty much need to try a streaming game service and see how it goes. There are several of them these days and they often have free trials so it's not too ridiculous to just sign up with one and see.


i found it pretty good playing destiny, yeah, but i’m kind of terrible at pvp so, you know, ymmv.


I've been following this company for a while now, and others like it, and I'm really wondering how the economics of these rent-a-big-GPU services shake out. The specs say you get a GeForce GTX 1080 (or equivalent, so maybe they're licensing NVIDIA vGPUs or something)... do they actually have one big GPU per subscription, or do I just have access to a big GPU whenever I want? You can't slice up a GPU as effectively as you can slice up CPU cores... Filling up a datacenter with big GPUs seems like a very expensive proposition.


When I originally used it at the beginning, if I recall, it was essentially a QEMU/KVM server with, indeed, a physical GPU passed to the VM, and each server had 2-4 GPUs assigned to it. When you logged in, you had to wait for one of the available servers to have an available GPU, in which point your virtual machine was mounted to that server and booted up, and their custom software on the VM would send the graphics back to you.


FYI, Linus Tech Tips showcased one of their servers a year ago: https://www.youtube.com/watch?v=DS1FcT93Qy8

Tl;dr 2-3 NVIDIA Quadros split up over hypervisor


Wonder if you could leverage Bitcoin farms for this.


I wondered about using it for ML type stuff, but figured they have application or time limit restrictions.




Consider applying for YC's Summer 2026 batch! Applications are open till May 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: