-
Ever wanted an RSS feed of all your favorite gaming news sites? Go check out our new Gaming Headlines feed! Read more about it here.
-
We have made minor adjustments to how the search bar works on ResetEra. You can read about the changes here.
Google: Stadia exclusives to have features “not possible” on home hardware
- Thread starter stormfire
- Start date
Loading is a linear task. If the PS5 has speed or bandwidth advantage, the engine is going to use it automatically, so the developers don't have to redo the engine to support this feature. CPU utilization is neither a linear, nor it is an automated task in most engines.
And on top of that, PS5 is going to be important. All of the engines will be retooled for PS5 to support the hardware, because nobody is going to bet that PS5 will be a failure and Stadia is going to take the most market share next-gen.
Netflix is a bad comparison. Streaming was an additional feature on certain DVD subscription tiers for a while so while their streaming service was coming together, they still had their subs for DVDa to support them.
While long term obviously matters more, launch is incredibly important to get word of mouth started. If the current consensus is that this isn't ready and people shouldn't buy games on Saltadia, that's going to be a sentiment that lasts for a long time, especially since they're not launching with the free versions of the service.
If you read what developers are saying, it's more about putting that hardware to work together, instead of assigning 200,000 instances to 100,000 players. An example of this could be an 8 player Call of Duty match, where the processing for everything that happens on the level is shared on the 8 instances that are playing the game together. This would be a lot better than having the 8 instances running the same repeated simulations. You could also dedicate a single or multiple instances to simulate something like the destruction or water physics on something like a 64 player Battle Royale Game. In that case it would be 64 player instances + the 1 or 2 dedicated instances for physics simulations. This can also be applied to single player games, imagine a GTA like game where every NPC can talk back to you and has their own life, can get married, have children and build their own houses. Just by the fact that all of this hardware is connected to each other allows for solutions that are a lot better than brute for solutions like assigning multiple instances per user.
Ubi? It seems to me they love to try everything (and they will have UPlay+ on Stadia)...tho i bet it will be only timed exclusive and after 6-12 months end on consoles/PC.
Remember when Microsoft claimed that for certain games? Pepperidge Farms remembers.
I believe it. Microsoft was a bit premature with their power of the cloud claims but in the next 2-3 years we'll definitely see things from Google, MS, Sony etc that are only capable with distributed computing. It's inevitable.
I just think we are at the point where Stadia needs to be more show than tell. I believe the technology could do some crazy things down the road but let it speak for itself for now.
"The power of the cloud!" was a bad joke 5 years ago, it's asinine for Google to make the same idiotic pitch with the same total lack of evidence.
Stop telling us about the awesome things your service can potentially do and start showing it actually doing them.
Stop telling us about the awesome things your service can potentially do and start showing it actually doing them.
This is really the best part:
"but those developers will have to figure out what setup actually makes that experience fun for players."
Unless some Google owned studio takes several years to make something like this exclusive for Stadia i can not imagine what 3rd party developer would even bother.
It was an interview, they plan on doing this and the service has not launched
They've had like 15 launch events for this fucking thing, if they had anything interesting to show they really should have shown it.
The most concrete thing they demonstrated was some neat fluid dynamics and lighting tech, but even that was decorative and additive rather than transformative.
Someone has to figure this stuff out eventually, right? And since there's no box, Google doesn't have much choice if they want their platform to stand out.
Problem is that it takes time and money, like anything else. The level of ambition for the things she's talking about is going to cost more and take longer to implement. Unlike normal games, it's not like you have ready to implement middleware you can purchase to do what you want to do.
And it's got to be fun. Like when talking about not having canned lines, an AI generated a dialogue isn't going to have the right emotion. Even if you could get to sound human, and not learn to be racist off the internet, it's still unlikely to be effective. There's a reason actors have scripts!
Hopefully the service works well enough for normal multiplatform titles to keep it going, as I do want to see some highly advanced simulation and other potentional advantages realized.
Problem is that it takes time and money, like anything else. The level of ambition for the things she's talking about is going to cost more and take longer to implement. Unlike normal games, it's not like you have ready to implement middleware you can purchase to do what you want to do.
And it's got to be fun. Like when talking about not having canned lines, an AI generated a dialogue isn't going to have the right emotion. Even if you could get to sound human, and not learn to be racist off the internet, it's still unlikely to be effective. There's a reason actors have scripts!
Hopefully the service works well enough for normal multiplatform titles to keep it going, as I do want to see some highly advanced simulation and other potentional advantages realized.
I haven't read through the thread, but the way this was worded left too much ambiguity. If the statement was "We are working on titles and experiences that just are simply impossible on home consoles" it would give a fleck of hope and optimism.
The way she worded it was way too hypothetical, and I'm sure some console manufacturer could say the same vs cloud. Just bleh.
The way she worded it was way too hypothetical, and I'm sure some console manufacturer could say the same vs cloud. Just bleh.
Strictly speaking not anymore, if you have a hw guarantee of nvme class ssd in every box and software stack that doesn't artificially limit I/O because it was designed around 30 year old storage paradigms.
To be clear, software designed For new consoles can completely remove the concept of 'loading' and accessing data in 'files'. But that level of integration won't come 'free' especially in 3rd party middleware that is dragged down by decades of legacy.
It's true and potentially interesting, like programming a game for one of the top500 supercomputer. But... nothing that couldn't be sided with local hardware. It's neither antinomic nor reserved to Stadia. If Nintendo wants to do " Yoshi Battleground " with hundreds of thousands Yoshies, they can use their servers to manage computing than inject the visual result into the player session, or in the client of a full streamed game.
Can mods actual step up and start banning these obvious trolls that come into every Stadia thread?
What you are describing there is using the power of the cloud to assist a local game. That is possible and it's already being done today. The big difference here is the amount of bandwidth available from one server blade to another, is several orders of magnitude larger than the regular internet connection from a server to a house. This will allow for developers to put all this hardware working together in ways that we haven't seen before.
I am so excited for the streaming bubble to burst. This is clearly just the next hot thing that companies are throwing themselves into with reckless abandon hoping that it sticks and game streaming will never be a suitable replacement for dedicated gaming hardware. It's pretty obvious that Google barely even knows how to market the thing which is telling in and of itself.
That's not true at all. What stadia has the potential for is completely different than what Microsoft tried to do at the beginning of the generation.
No idea why stadia threads attract so many trolls.
Google gave plenty of examples in their GDC presentation.
Why would they not be possible on home hardware? I'm really not trolling, genuine question. Some features we saw at the GDC were indeed unique but i can't see why they could not be implemented on PS5 or XBox Two?
I don't doubt it. In fact, I think they need that. Right now, I'm not their target audience, but the features that Stadia can theoretically provide certainly can result in very interesting game concepts. Let's wait and see.
At that point you are just describing a console using the power of the cloud. You can call the Chromecast Ultra or the smartphone used to play Stadia, a console that is using the power of the cloud if you like.
Not a troll. Stadia is streaming games that run on their own hardware. Playing a game on a traditional console doesn't mean a developer can't have their games work the same way.
Stadia next month is early access
Next year is a storefront
After that we will see first party results, the platform is not dependent on it like a traditional console.
Stadia hasn't invented or patented clustering or virtualization, just branded a possibility around it.
Nvidia knows how to virtualize GPU (Geforce Now runs on Tesla server), AMD too with Vega.
It wouldn't be impossible for Microsoft, Nintendo or Sony to offer a similar infrastructure for devs.
And for the client it has no importance on what the server run.
But it's true that Stadia could be the first to offer a game using that potential and I would like too see one!
Yeah this. I'm not interested in trying it now but the tech seems interesting. Look at what Flight Sim is doing with local hardware still a part of the equation. They could potentially do some wild shit being completely cloud based.
No one said Stadia is the only one that has access to the things that are possible with all that hardware connected in a server. We will surely see the same things on Xcloud, Geforce Now, PS Now and others. Stadia is just the only on that is specifically talking about it and seems to be the first that is making it available to developers.
The same reason why ps5 exclusives can't run on switch just in a different way
I am so excited for the streaming bubble to burst. This is clearly just the next hot thing that companies are throwing themselves into with reckless abandon hoping that it sticks and game streaming will never be a suitable replacement for dedicated gaming hardware. It's pretty obvious that Google barely even knows how to market the thing which is telling in and of itself.
This post will age really well. I imagine people like you describing polygon graphics during the N64/PS1 era, as the next hot thing that companies are throwing themselves into with reckless abandon. What you should start talking about in detail, is how it wont be possible for developers to use the hardware in a server to enhance the gameplay.
Last edited:
How would you implement multi GPU rendering on a closed box that is already defined as having a single GPU? The difference is the computational power available to developers, to run things as physics simulations or AI. On a PS5 or Xbox Two, they have a single CPU/GPU on each box and they cannot make it work with another PS5/XBox Two CPU/GPU, that is in a separate house. In a server environment they can make the connected hardware cooperate with each other. You can of course create hybrid solutions like the one below, but it would be a less efficient solution than the one that Stadia is implementing.
Well, there are SOME things. We have had RAM-intensive and CPU/thread-intensive games on PC for ages, games that didn't make it on console because they would run like shit if they are ported 1to1 without changing the motor under the hood.
A few examples:
- Dwarf Fortress very dependent on RAM and a little less on CPU. The size of the map and the amount of dwarfs are directly influenced on how much RAM and how good your CPU is
- Minecraft. VERY VERY dependent on RAM. Not so much in your standard Vanilla Minecraft, but with big structures and render-distance you could be strangled by RAM fast. Modded Minecraft is on another whole level. Some Modpacks are not recommended if you have below 10 gig of RAM and even then, it is quite possible that you couldn't finish the modpack because the endgame would be laggy as hell. Most modpacks even take 5 to 30 minutes just to load, dependent on your CPU and RAM
- Paradox 4X games are RAM and CPU dependent, determining how smooth and fast your game is running the days.
- Rimworld is CPU and RAM dependent determining how big your map can be and how many people can be on screen.
- Aurora 4X is entirely RAM dependent. The game will start running smoothly but will chug down immensly as the ingame years progress. a 5-day intervall would take a second at the start and would get longer and longer as the game progresses.
Now as for Stadia, could a game theoretically use 100Gb of RAM?
Don't tell BlueManifest I said positive things about Stadia.
Google, I'm sorry, but Microsoft telling us about the "power of the cloud" since 2013 and then releasing Crackdown 3 just this February has forever poisoned this well. Try something else.
Hahah you people have no chill. I love it.
Hahah you people have no chill. I love it.