I never said 4k at 144hz. I play 2560x1440 at 144hz and would take that over 4k@60 anyday.
But regardless, the new NVIDIA 3090 does 4k@144hz these days.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Amazon's Luna Gaming Service Available on Mac and iOS as Part of Early Access Rollout
- Thread starter MacRumors
- Start date
- Sort by reaction score
I never said 4k at 144hz. I play 2560x1440 at 144hz and would take that over 4k@60 anyday.
But regardless, the new NVIDIA 3090 does 4k@144hz these days.
Google doesn't only deploy Stadia servers at it's data centers, but also it's edge nodes:
“Stadia servers are deployed on Google’s Edge locations that are closest to partnered [Internet service providers] to further ensure a seamless and consistent gaming experience,”
Source: https://spectrum.ieee.org/tech-talk/telecom/internet/how-the-youtube-era-made-cloud-gaming-possible
Edge nodes don't contain the machines that power Stadia games. Edge nodes are just there to cache data (like when you start up Stadia and see your library quickly).
If you're talking about Edge PoP, then there are only 90-100 points in the world. They themselves don't have Stadia gaming machines, but rather Stadia machines that directly connects to Google data centers and relay the video stream in a more direct path instead of making many hops.
Last edited:
No, according to Google they actually do deploy Stadia servers at edge nodes. They have been quite specific about this.
Caching would do nothing for Stadia. Unlike video streaming, for example, game streaming has very little static, cacheable data. Everything must be rendered and transmitted in real time.
No. You are 100% wrong.
The nodes have a direct connection from the ISP (where the nodes are installed) and to Google's data center. So when the user starts playing, they connect directly to the ISP and then directly to the closest Google data center. This skips many hops, reducing latency.
If the games are processed at the node within the ISP, then there's no need for a data center as shown in the Stadia slide.
Each edge node cannot possibly provide enough processing power to every player in the area. I don't know why you're suggesting this. It's not 7,500 data centers, it's 7,500 nodes with direct fiber connections to the data centers. They don't render the graphics, they simply relay the connection on a faster and more direct line instead of jumping through 10-20 different points to get to the data center.
Do a search on what exactly an edge node does. You cannot possibly deploy 7,500 datacenters just to run up to 100-200 simultaneous video games at each node. That'll cost upwards of billions of dollars.
Read this reddit link:
I'm not suggesting they've deployed Stadia servers at 7,500 edge nodes or data centers. Clearly Stadia is not deployed at that scale and it is not launched worldwide.
However, it is a direct quote from Majd Bakar, Google's VP of Engineering for Stadia, that they have Stadia servers deployed at Edge locations:
“Stadia servers are deployed on Google’s Edge locations that are closest to partnered [Internet service providers]"
You're probably right that this means "Edge POPs" rather than "Edge nodes", but you are wrong to claim/assume that Stadia servers are only present at Google's main data centers.
Stadia servers mean they’re dedicated in relaying stadia video streams and other stadia functions.
For example, if you were playing on a 4k/60 screen and switch to a 720p cell phone, the edge node can transcode the 4k60 frame to a 720p resolution on the fly so your phone doesn’t need to download a 4k stream. This is also good for providing multiple resolutions for YouTube Live where people can select different resolutions to watch you play and they can “chat” and respond to your gameplay with low latency. IIRC they talked about this in the keynote
It doesn’t make sense to have these “stadia servers” process graphics since there could be potentially thousands of players in Los Angeles trying to use a single node. It’s not scalable that way.