Slashdot Mirror
2000x GPU Performance Needed To Reach Anatomical Graphics Limits For Gaming?
Vigile writes "In a talk earlier this year at DICE, Epic Games' Tim Sweeney discussed the state of computing hardware as it relates to gaming. While there is a rising sentiment in the gaming world that the current generation consoles are 'good enough' and that the next generation of consoles might be the last, Sweeney thinks that is way off base. He debates the claim with some interesting numbers, including the amount of processing and triangle power required to match human anatomical peaks. While we are only a factor of 50x from the necessary level of triangle processing, there is 2000x increase required to meet the 5000 TFLOPS Sweeney thinks will be needed for the 8000x4000 resolution screens of the future. It would seem that the 'good enough' sentiment is still a long way off for developers."
My question is this: how much more will games have to cost to support the development to this level of detail?
The soylentnews experiment has been a dismal failure.
better looking "anatomical peaks"!
"I say we take off, nuke the site from orbit. It's the only way to be sure."
there is a rising sentiment in the gaming world that the current generation consoles are 'good enough' and that the next generation of consoles might be the last
If developers can't find a way to improve games beyond the next generation, it's not because we've reached some peak of gaming possibilities, it's just because those particular developers have reached the peak of their imaginations.
Somewhere right now their is a young guy sitting somewhere who has an idea in the back of his head which will become the next great innovation in gaming. It will require a lot more computing power than the current generation of PC's, much less consoles. If he were to pitch it at EA, he would be laughed at. If he tried to explain it at a Game Developers Conference, he would be greeted by blank stares and derision. He's probably already used to hearing responses like "That can't be done", "Who would want THAT?", "That could never be done on a console", etc. But one day people will look back and say "Wow, how could they *not* have seen that that was the future?" and "How could they have been so arrogant as to think that gaming had peaked with the millionth variation of the FPS?".
What's more, I suspect that even Sweeney is off-base. The next real innovation won't be about improving resolution or framerates to some theoretical max, or making an even prettier FPS. It will be some whole new way of thinking about gaming that is just in the mind of that weird guy right now. Most of us can no more imagine it now than some guy playing Pacman could have foreseen Half-Life 2. But it's coming.
Every generation thinks it's special. But never be so arrogant as to think your generation has somehow reached the pinnacle of achievement in ANY area.
SJW: Someone who has run out of real oppression, and has to fake it.
When the article's authors have shoehorned a word so obviously not related to the subject matter into the subject line, and then go on to repeat it over and over again, only one of two things can be true:
1. There were no better words in the dictionary, and rather than taking the sensible approach of creating a new one, they opened to a page at random, stuck their finger on it, and started using whatever their finger touched.
2. Author was trying to sound trendy and interesting.
As a footnote, salahamada is a made-up word waiting patiently for its debut. Give it a little love?
#fuckbeta #iamslashdot #dicemustdie
Think of how glorious the reflective spheres and checkerboards will be bro!
Mod me down, my New Earth Global Warmingist friends!
I came here to read this. I leave satisfied.
Tim's explanations of first- and second- and third-order approximations are somewhat bizarro. Unreal doesn't use second-bounce in its lighting. All game engines are first-bounce only unless they contain some realtime radiosity simulation, and very very few do. This has been true since Wolf 3D and is true today.
And once you have a system for second-bounce, third- and fourth-bounce can be trivially computed (over multiple frames if need be), and the results are hardly different to second-bounce.
I wish I knew what he meant by these levels of approximation.
And he's only talking about rasterization. Expect a switch to raytracing somewhere in the not so near future.
But that won't really matter either. The problem at this point isn't the number of pixels, or the number of polygons, or the depth or resolution of the textures. It's the fact that the image is being projected on a rectangle with a strip of plastic around it. In the end, what we really are shooting for is what literature people call "Suspension of disblief". You can only get so far looking into a glowing rectangle. The wrap-around screens of eyefinity help some, and 3d glasses have the potential to help a little bit.
The reality is that hte most immersive gaming experience I've had was in the mid to late 90's when i was hooked up to a real VR system with a helmet, and held a gun with approximately wii-controller input capability. The ability of that system, despite its craptacular by today's standard rendering capability, to be immersive was much higher, because the ability to see my entire environment by moving my neck and body was more important the the quality of the environment itself.
68% + 42% = 100% eh? Maybe quitting video games would be a good thing for you. It would give you more time to study math.
While describing the layer and textures, it is going to be offset by what is known as "uncanny valley". There is a point at which the reality is flawed because it looks too real for the context.
I'm even starting to see uncanny valley on magazine covergirls after they've been photoshopped till they are almost unrecognizable. There is a point where you stop fixing flaws and start making them.
Agent K: A *person* is smart. People are dumb, stupid, panicky animals, and you know it.
Everyone talks about how far we can push graphics.
But what about pushing the AI?
What about procedural generation of the game?
What about vastly improved physics including a destrucable world?
I'd rather see these things pushing hardware development than how many polygons you can crunch in a second.
I'm god, but it's a bit of a drag really...
I think 2000x GPU power is very much underestimating the potential for a number of reasons:
1: Raytracing / global illumination. In comparison to games with true global illumination, current technology 3D worlds with only direct illumination (or scanline rendering) look crude and unconvincing. Objects appear 'cookie-cutter' like and colours tend not to gel with the overall 3D landscape.
Toy Story 3 took around 7 hours to render each frame. To render in real-time for a video game (say 60 FPS), you would need a processor that was around 1 million times faster than what we have today. And AFAIK, that's mostly using Reyes rendering (which incorporates mostly rasterization techniques with only minimal ray tracing.
2: Worlds made of atoms, voxels or points. This makes a world of difference for both the user and the designer. Walls can be broken through realistically, water can flow properly, and explosions will eat away at the scenery.
2000x? Pah, try 2 TRILLION as a starting point.
Why OpalCalc is the best Windows calc
Developers used to start off with high resolution models and have to pare down the triangle count and adjust textures to meet memory and processing requirements. In the future, they won't have to do all of that tweaking and will be able to use full resolution models, so it will probably be cheaper.
Also, not all games aim for realistic depictions, many (most?) are stylized, and won't necessarily need to be highly detailed. The extra processing power could go to effects, deformation, physics, etc.
Twinstiq, game news
Nono, totally unrelated: 68% of the people won't repost. 42% who have vision will. So 32% of the people who repost are of the 76% that can see, meaning he considers 24% of the populace to be blind. Apparently there is a high level of head injury in his area resulting in eye trauma.
Exponential improvement in technology is the historical norm, yet it can still be difficult to fathom.
2000X should be achievable by 2024, at 2x improvement per year; or by 2029 at 2X every 18 months.
Some of us should see 2 trillion-fold improvement in about 40+ years at 2X per year; or by 2075 at 2X every 18 months.
Barring the occurrence of any variety of manmade and natural disasters, of course.
To me the real problem is focusing on the wrong details. Take Skyrim for example. Is it really a big thing if they, say, tripled the detail on the existing characters? Do the NPCs need pores or drops of sweat?
Or would it be more interesting to walk into Whiterun, and there's a 100 NPCs walking around, or you assault a fort with the Stormcloaks and there's 100 other soldiers at your side attacking the 100 Imperials in the fort, and clouds of arrows raining down [nice knowing ya, shieldless dual wielders :-) ]? It's a "more detailed objects" versus "more objects in the world" sort of argument, I guess. I'd rather see the power applied to "more objects" at this point, IMHO.
Most of us can no more imagine it now than some guy playing Pacman could have foreseen Half-Life 2. But it's coming.
The guy playing pacman (released in 1980) only had to move a couple cabinets over to play Battlezone (also released in 1980) to foresee Half Life 2 and FPS's in general.
I think main challenge is the interaction between player and environment. On something like MW3 that is limited to blowing up the odd chicken, window or set piece designed into game.
I want to swish my ( virtual ) hand through a river and see ( and feel ) the water flow around it.
Any true physics model would require awesome cpu capacity. We have at one end mindcraft ( where the atoms are decidely blocky ) and second life ( where behavioural programs can be attached to objects).
My dream would be a virtual universe with atoms large enough to fit 1000 times current cpu capacity but behaves like a 'real' experience.
By 'atom' i mean the simplist construct in this universe which satisfies greatest potential complexity but still cpu possible.
In conclusion future 'games' will not be about fps or polygons per sec but model calculations.
By using eye tracking, we dont really need to render the whole screen at high resolution.
We only need to render the part the eyes are looking at at high resolution
The ability of the eye to percieve high resolution is only limited to a very small area, and the brain fakes it by moving the eyes around.
By superimposing a small image with high dpi on top of a larger image with low dpi, we get a high resolution window into the larger image.
If this high res window follows the eyes around, the brain will percieve a large high resolution image.
Naturally for this to work, the smaller image has to be updated to show the same part of the scene that it is replacing.
This can also be used to emulate a high resolution screen by keeping an area your screen black, and using a projector to project the smaller high-dpi image on the black area.
Oh, and by the way. Remember this post and use it as prior art in case some troll patents "A method of simulating high resolution images by combining multiple images of different scales and resolution"
Good enough does not mean that you have to match the anatomical limits of human perception. That is asking for perfection.
Unless the increase of graphics performance will lead to new radical ways of gaming, then the current GPU performance is good enough.
It is not like it was 10 years when 3D graphics opened the door to new types of gameplay, like the creation of Grand Theft Auto III. Now 3D gaming has matured and there isn't any more frontiers to discover other than just better graphics, which are just marginal improvements.
I am thinking innovation will come from input devices like touch screens, Kinects or another technology that no one has thought of yet.
Because raytracing uses so much less computing resources? Or, because you don't really know what it is?
For very complex scenes, yes it does use less resources. Raytracing grows logarithmically while Rasterization grows linearly. Intel estimates the scene complexity (whats visible) where Raytracing overtakes Rasterization is ~1 million polygons. After that, Rasterization has no chance to compete in efficiency.
"His name was James Damore."
Best test for gaming AI...
Some day Xbox or PS will sell a multiplayer game, but simulate all the other players. When no one notices, I will say we have achieved the pinnacle of gaming AI. Of course we may have to train an AI to chronically swear and make racist slurs, but if that is what progress takes so be it!
We passed a million polygons on screen over a decade ago. Your telephone can just about do that today (the iPhone 4S does 30 million per second), modern game consoles that came out 7 years ago will do about ten times that (500 million per second on a 360), and a modern high-end PC probably does ten times that again.
In other words, we're at the point where we're using rasterization to push 100 million polygons, and raytracing is still so much slower that it's not even remotely practical to duplicate the same quality. Intel's latest attempts to do so have produced low-resolution low-quality results that still require a massive array of hardware. They're basically throwing eight PCs worth of hardware at the problem. About all the demos do is demonstrate that it's easier to calculate accurate reflection and refraction with raytracing.
In other words, you either mis-remembered Intel's estimate, or their estimate was laughably inaccurate.
I want millions. I want the entire population of New York City, all eight million people, turned into shambling, flesh-eating monsters.
Sounds like you haven't been to New York lately.
Have you got your LWN subscription yet?
IDK about you, but i have a 22" 1080P samsung display and from arms length, I can see the pixels when the screen is white. Okay to be honest what i'm really seeing is the tiny spaces between the pixels leaving some sort of weird pattern.
I won't be happy with display tech, until I mistake them for windows. Same goes for your 3d volumetric display. I'd better be able to be fooled into thinking the things it is showing me are real.
All of the above was encrypted with a Quad ROT-13 method. Unauthorized decryption is in violation of the DMCA.