The cloud is just over the horizon [2018-2019]

We both agree that they did disappoint on delivering exactly what shown on the first demo, and simultaneously did do something beyond what the modern standard.
The rest is semantics. Let's not waste time there. I think the same as you.

There is no standard because the last game that used heavy physics was released a console generation earlier.
 
There is no standard because the last game that used heavy physics was released a console generation earlier.
I'm not impressed by the physics. I'm impressed by the amount of moving objects they are synching with multiple clients online with relatively low lag. I made that pretty clear.
 
No it didn't.

It's like saying that game X will support 4K 60fps ray traced graphics and then you release a 720p 25fps game that uses raytracing only for low resolution reflections.

Of course the "tech" is there but is nowhere near to what was supposed to do, that was the whole point of the claim.

milk said the tech came through. I agree. You just agreed.

...?
 
milk said the tech came through. I agree. You just agreed.

...?

There is a reason tech is between quotes.
most if not all the games that uses a servers client architecture have been processing physics on the server (the cloud).

This is why I used raytraced as comparison, raytracing as "tech" has been around for decades.

I'm not impressed by the physics. I'm impressed by the amount of moving objects they are synching with multiple clients online with relatively low lag. I made that pretty clear.

I am not, what they promised was vastly superior in any possible way to what we got.
 
It's one of the most important downgrade this gen on what was promised. If Crackdown delivers the technology expected, then the Killzone 2 trailer was fine, same for the Uncharted 4 teaser, etc.
 
There is a reason tech is between quotes.
most if not all the games that uses a servers client architecture have been processing physics on the server (the cloud).

This is why I used raytraced as comparison, raytracing as "tech" has been around for decades.

"The cloud" doesn't just mean "server". And to suggest that because servers have existed, no "tech" development has been required to implement crackdown is absurd.

Ray tracing has been around for decades. Doesn't mean Nvidia had no "tech" to develop for RTX.

I am not, what they promised was vastly superior in any possible way to what we got.

And yet still more impressive then anything comparable. Which seems difficult for some to comprehend in the presence of a downgraded product.

Both can be, and are, true.
 
Last edited:
If I understand lefantome correctly, he's saying that though the absolute nature of the tech is as intended - in that games would be hosted, physics calculated on the servers, and synchronised across players - because the scale of that achievement is far below what was intended and promised, and because it appears no different to what one would expect from a normal evolution of server-based processing although we have no examples to compare, it doesn't represent a success for cloud processing.

Breaking that down: (I have no idea on the validity on points 1 & 2, but am presenting the arguments)
  1. Game physics calculated on the server and sent out to clients is nothing new.
  2. The simple advance of processing power and internet tech means this sort of thing is possible now without any particular new tech developments.
  3. The destruction in C3 is nothing new itself.
Everyone is agreement with point 3. That is actually a noteworthy downgrade. The first demos of the power of the cloud had procedural destruction, from tiny bullet-holes to epic explosions shattering large structures. This was only possible by using huge amounts of processing power, which the cloud brought. In the end, the PhysX style preconstructed damage (predeconstruction!) is a significant step backwards and doesn't require anything fancy on the server side. The Cloud isn't necessary in this case.

That leaves the argument over whether the physics synchronisation is something new and impressive, or whether the entire network side of the game is a bust. But even then, whatever C3 is doing over the network doesn't require cloud computing to achieve and can be achieved with ordinary servers.

So, is Crackdown 3 a flop in its promises? Debatable. But is the Cloud a flop in its promises? I think that's a categorical 'yes'. What is C3 doing that requires a huge pool of networked processing that can quickly spin up jobs and distribute workloads? What is C3 doing that can't be achieved by a dedicated i7 PC running GPU PhysX? The only thing I can think of there is server proximity, with Azure being placed as locally as possible, as Iroboto will tell us. This keeps latency down and network performance more consistent for localised players. Which isn't a win for algorithmic advances (remember how we discussed what would be necessary to sync all that original data of procedurally created geometry? Alternative data representations etc.), but for MS's investment in networked computing structure.
 
Let me explain more.

I believe that the developers worked hard to develop the tech I assume they brought many important improvements to it.

but!

As far as I know this is the first and only game this generation trying involving heavy physics processing in online multiplayer.
Time and money have been spent to do this, starting with Cloudgine that was the company responsible for the cloud tech (and it was bought by Epic last years, this means they had definitely something to offer)

So how can we measure the progress made with the tech if there is no other instance it can be compared with?

It's like somebody developing today the best VHS ever. Of course it's an improvement, but nobody has been developing that tech for almost 20 years, a timespan that has seen many tech improvements that may indirectly benefit it. If you are the only player you are always a winner.
Is it something to be impressed of? it depends on the actual improvement. *

The only thing that the public has is Crackdown 3 and the downgrade from the previous demos is crystal clear while the user side improvements from one similar game of the previous generation (made with much bigger constraints) is not that huge.



* another example may be SpaceX or something like that
If nobody else is developing space travel and they promise to bring people to Mars by 2020 and then by 2030 they are only able to bring millionaires to a space station in Earth Orbit, of course it's an improvement but not that much considering money and time spent and nowhere near what they promised.

"The cloud" doesn't just mean "server". And to suggest that because servers have existed, no "tech" development has been required to implement crackdown is absurd.

Ray tracing has been around for decades. Doesn't mean Nvidia had no "tech" to develop for RTX.



And yet still more impressive then anything comparable. Which seems difficult for some to comprehend in the presence of a downgraded product.

Both can be, and are, true.

The "cloud" is a marketing term for a series of services of which online gaming with dedicated servers per match/world/whatever is a perfect example.

I am not implying that no tech development would have been required.
I am saying that there was no paradigm shift and that given time, money and more power, we have seen no groundbreaking improvements as consumers.


Yes the whole point of me citing Ray Tracing is that while the tech is still based on the same concepts, new improvements are making it be much faster.
I don't see the same in Crackdown 3.

It's more impressive than anything comparable because the last comparable game is 10 years old.

 
Last edited by a moderator:
If nobody else is developing space travel and they promise to bring people to Mars by 2020 and then by 2030 they are only able to bring millionaires to a space station in Earth Orbit, of course it's an improvement but not that much considering money and time spent and nowhere near what they promised.
This would be a terrible way of looking at it. You're too focused on the promise and you aren't focusing on the experience or educational value of developing the solution.
I can get into many analogies of how a microcosm of success, that can result in fundamental ways in which we approach problems.
You can look at AI and see that today.

Many times in the world, we have spent time developing a lot of technologies that are nothing more than parlour tricks, that eventually became something later. Ie. Electricity was largely viewed as being nothing more than a hobby back in the day.

Experience is the difference between failure and success. And what they learned from this experience can be applied to the next and the one after that until it becomes someting substantial. And no other company can just come in and just replicate their effort overnight because they've never had to work through all the problems that this team did.

We pay people's salaries for their experience.
 
This would be a terrible way of looking at it. You're too focused on the promise and you aren't focusing on the experience or educational value of developing the solution.
I can get into many analogies of how a microcosm of success, that can result in fundamental ways in which we approach problems.
You can look at AI and see that today.

Many times in the world, we have spent time developing a lot of technologies that are nothing more than parlour tricks, that eventually became something later. Ie. Electricity was largely viewed as being nothing more than a hobby back in the day.

Experience is the difference between failure and success. And what they learned from this experience can be applied to the next and the one after that until it becomes someting substantial. And no other company can just come in and just replicate their effort overnight because they've never had to work through all the problems that this team did.

We pay people's salaries for their experience.

I agree but in this case lots of money was spent on this tech to develop a game that is bombing hard.

Even Microsoft that seemed so much interested into this has let Epic buy Cloudgine, the owners of the tech.
In 5 years, this is the first and only game that has used it.

While a failure may lead to success, most of the time it doesn't.

I am sure that the tech is and is going to be useful, but not in the way it was presented, aka to power massive multiplayer distruction in games that are supposed to have free multiplayer.
The problem is here is economic and the tech cannot do miracles.

In my opinion Crackdown 3 made sense from an economic point of view because Microsoft wanted to promote Azure and make people believe that Xbox One was not massively underpowered compared to PS4, then it was not canceled because canceling another exclusive would have been a PR disaster.
 
I agree but in this case lots of money was spent on this tech to develop a game that is bombing hard.

Even Microsoft that seemed so much interested into this has let Epic buy Cloudgine, the owners of the tech.
In 5 years, this is the first and only game that has used it.

While a failure may lead to success, most of the time it doesn't.

I am sure that the tech is and is going to be useful, but not in the way it was presented, aka to power massive multiplayer distruction in games that are supposed to have free multiplayer.
The problem is here is economic and the tech cannot do miracles.

In my opinion Crackdown 3 made sense from an economic point of view because Microsoft wanted to promote Azure and make people believe that Xbox One was not massively underpowered compared to PS4, then it was not canceled because canceling another exclusive would have been a PR disaster.
Well cloudgine left, I think they rebuilt the cloudtech to own it themselves. Thats unfortunately an asterisk I cannot prove or disprove yet.

I'm not disagreeing with your point mind you, it most certainly was marketed as a method to soften the power blow.

But now that ti's done and shipped, now the question becomes if anything comes from that learning. If something does, perhaps this wasn't such a bad lesson to learn.
 
Last edited:
It's more impressive than anything comparable because the last comparable game is 10 years old.


In Crackdown 3, if i'm not mistaken, you can see advanced physics only in the multiplayer mode and on the most simplistic maps...

And yet still more impressive then anything comparable. Which seems difficult for some to comprehend in the presence of a downgraded product.

Killzone 2 was one if not the best looking game at its release date. Yet, the trailer was total bullshit... even if the final product does more than other games, it doesn't mean the original tech was delivered.
 
Last edited:
Had forgotten about the cool physics in Just Cause 3 and 4.
 
Not sure if it belongs here but Flight Simulator is taking advantage of the cloud...

Real-world locations including Dubai, San Francisco, and Egypt are recreated in photorealistic 4K resolution, with geography and textures captured from satellite imagery. Players will be able to control numerous authentic planes from light aircrafts to passenger vehicles, set their own flight paths, and navigate through unpredictable scenarios thrown up by Microsoft’s Azure cloud computing platform, which leverages artificial intelligence and machine learning to dynamically shift flight conditions. This will keep players on their toes – get too comfortable running a commuter flight, and Azure might throw up sudden shifts in the weather, turning a sunny flight into a battle to keep aloft as storm conditions roll in.
...
If Flight Simulator looks ahead of its time, that’s because, in many ways, it is. Not only do the underlying AI and graphical technologies behind its creation push current gaming technology to its limits, but they begin to surpass them. The geographic data alone, needed to render the world players will be flying around, comes in at two petabytes – equivalent to 20,000 Blu-ray discs.

It will be far in excess of what can fit on a disc or be reasonably downloaded, and so Flight Simulator is at the vanguard of a new breed of game that can only live, perhaps fittingly, in the cloud. Microsoft servers will do a lot of the visual processing remotely and beam the results to a player’s screen, doing away with the need for a super-powerful PC at home (though you will need a fast internet connection). This is part of a wider shift that is already under way in video games: as technology improves and developers strive to create ever more realistic virtual worlds, games balloon in size, and the hardware needed to run them becomes more and more expensive. Offloading some of that data and processing to cloud servers is the solution that Microsoft – among other companies, such as Google – is betting on.

For games like Flight Simulator, the cloud provides “a system that kind of understands how much capability I need at any point,” Spencer explains. “It’s no different than what we’ve done in years past, streaming into available RAM that’s available on our consoles.” Flight Simulator will have an offline mode when it launches in 2020, but given how heavily it relies on data streamed from Microsoft servers, it’s likely to be severely truncated.

Always-online games are sometimes controversial – not everyone likes logging into a server every time they want to play (and indeed, some players aren’t able to). But in this case, the advantages are surely worth it. “We’re getting to the point with games where there’s the [question of] how much data can you actually download and store?” says Spencer. “[But] there’s also the matter of how much data do I need to have locally at any time? I’m not playing the beginning, middle and end of every game at every time … Not only do we have two petabytes of data behind [Flight Simulator], but we’ve got Azure AI running [so] you have different levels of fidelity in different parts of the planet. As you’re flying around, you’re seeing the highest buildings and the weather’s changing in real time. It is just unreal.”

https://www.theguardian.com/games/2...ouds-the-return-of-microsoft-flight-simulator

Tommy McClain
 
That probably the most optimal use of the cloud for a game. All of the scenery in a flight sim is basically non-interactive and usually a great distance away, so it doesn't matter too much if it takes multiple frames to get the data for it from the cloud while the plane and flight characteristics are modeled locally.

Regards,
SB
 
That sounds like it would require huge bandwidth. Which suggests to me the reality will be far short of what they claim.
 
Nah. It's hybrid game streaming from the sounds of it...
"Microsoft servers will do a lot of the visual processing remotely and beam the results to a player’s screen"
 
Nah. It's hybrid game streaming from the sounds of it...
"Microsoft servers will do a lot of the visual processing remotely and beam the results to a player’s screen"
It’s a slow moving game. The rate of change at those altitudes are relatively minor. Curious to see what happens if you try to Top Gun it.
 
It'll be fascinating to find out what extra layers of processing they're adding on to regular Bing 3d maps. I can easily imagine them adding AI layers to bake in additional detail where the photogrammetry detail is too sparse. Also, they're presumably flattening out texture detail so it can be relit?

The trailer looked stunning, but there were some trees and cars showing they typical bing/Gmaps artifacts.
 
Back
Top