If it actually does what it says it does, then yes it will be able to run on a standard desktop. I'm surprised there hasn't been more hype about this. This sort of technology will end up giving game artists a nice boost in pay if they're good enough
Having to worry about polycount SUCKS ASS. I really want to know what the hell gives it this "unlimited" property...seeing as how a processor can process a few billion calculations a second, it seems as though the number of "atoms" would cap out at the same number, seeing as how it has to calculate each location every time the camera moves. I look forward to using this.
Nothing about this proves to me that they're using it in real time though...all of it could be pre-rendered.
Well it only has to calculate the numbers of pixels on your screen. This would be infinitely fast than anything they are currently doing to render polygons on your screen. There is really nothing inefficient about and as far as I can tell the only limitation is going to be how detailed game makers are willing to make their game. If they spent an infinite amount of time on the game and drew textures as detailed as atoms then it would still render really fast, but your game would be like 500gb's or larger.
Even if it's only calculating display data for pixels/voxels, that still leaves animation. It WILL have to calculate new data for anything moving, I don't see any way around that at all. So blades of moving grass = problem.
This is probably why nothing is animated in the video.
If you had something such as a cube moving up and down as your animation. The code would just apply a transformation matrix to the object which is easily done by the gpu and then it would just redo the unlimited detail algorithm over and over assumingly 60 times per second for each pixel on your screen.
From what I understand from these videos is that the only problem with standard polygon rendering is that it spends time rendering graphics that aren't even in line of sight. Why spend the time rendering the opposite side of the cube when you can't see it?
Comments like Izzy's make me happy I didn't get into Computer Science after all. Awesome stuff though, I assume these graphics although wouldn't be all that intensive on the CPU, the amount of data required for this kinda stuff would be gigantic. Would this be RAM intensive too?
Comments like Izzy's make me happy I didn't get into Computer Science after all. Awesome stuff though, I assume these graphics although wouldn't be all that intensive on the CPU, the amount of data required for this kinda stuff would be gigantic. Would this be RAM intensive too?
Comments like Izzy's make me happy that I'm about to go into Computer Science.
Какой идиот придумал Бутерброд с дикобраза? Он хулиган и бездельник.
Animating a cube moving is one thing...animating a full soft-body deformable mesh is another.
ZBrush is capable of creating a single mesh of up to 1 billion polygons, lets say this 1b polygon character is running through your scene. So every face of that mesh is moving independently of one another at different rates yet still a group object...quite a bit different that a box moving up and down.
Comments like Izzy's make me happy that I'm about to go into Computer Science.
I like knowing how a computer works in detail and having a general understanding about stuff like this, but knowing mathematical equations inside and out and making new ones makes me want to run for the hills. Computer Scientists are more like mathematicians than programmers, and even less like IT guys; although they exceed at programming, it's because of their math skills. The courses I was looking into had 2 years of math, mostly calculus before even touching a computer so I took the neckbeard IT route.
I suck at math, all of the mathematical formulas are generally going to be prewritten for you in library's of functions. You just have to know what they do and how to use them. I barely passed calculus myself. Programming is always going to be more about logic than math.
If it actually does what it says it does, then yes it will be able to run on a standard desktop.
It's going to be a very long time until we have exabyte+ HDDs.
Originally posted by Jewpinthethird
"Hey Keywii" Said Foil in a raspy voice.
"Hey Foil. What's that you got there?" inquired Keywii.
"Oh, just my cock." Replied Foil.
"That just will not do." was keywii's response as she lunged for the scissors, pulled the blades apart, and clamped them down on the base of foil's shaft. Blood start gushing out of the wound where his penis used to be.
"NOOOOOOOOOOOO!" Yelled Foil in horror.
"Don't worry. I'm a wizard" uttered Keywii. And with that, Foil's penis grew back.
It's going to be a very long time until we have exabyte+ HDDs.
You wouldn't need to have exabyte drives sheesh. There would obviously be tweaking and optimization to each game just like the current process. Things like those clumps of dirt could be procedurally generated, which would only take up maybe 1mb worth of space instead of including each clump as a piece of geometry. Same with pretty much every other type of fine detail, procedural is the way to go.
Originally posted by notch.tumblr
170 000 three-terrabyte harddrives full of information
I'm pretty sure they don't have 170,000 hard drives linked together to run this, even if each "block" is instanced and taking up far less space. In the video, when they converted that elephant statue to a pointcloud, there are DEFINITELY NOT MILLIONS OF POINTS IN THAT MODEL, sure I'm sure it has the capability to render those at that sort of definition, but it's pretty god damn pointless to model anything at a 1 billion polygon level, so each model placed in this engine isn't going to actually take up all that much space.
Comment