02-17-2010, 02:12 PM
The spheres would be plotted in 3D space as a composite object made up of smaller objects. They would all be positioned as an offset relative to the center of the packed group.
The whole area of computing power for the masses is fascinating to me. I hope carrie won't mind my having a digression here. She's probably better up to date on much of this than I am anyway and might be able to correct my mistakes.
3D graphics power available to consumers these days is astounding. For a few hundred dollars, a GeForce 8800 card can shade and overlap more than 3 million triangles a frame at 60 frames per second. With 1,000 spheres shown at once, this would be 3,000 triangles per sphere, each just a few pixels across. Each pixel would have whatever colors most beautifully represent a smoothly drawn surface. The 3D engine runtime would adjust the level of detail in a surface depending on how much it fills the screen.
With a card like that driving an HDTV, you'd never see any kind of angles or seams, just the impression of a thousand smooth spheres - with moving lights shining through fog and smoke, if you like.
High-end consoles like Playstation 3 and Xbox 360 have somewhat less powerful graphics engines. But the consoles include custom processors, system architectures and operating systems designed to easily manipulate massive amounts of realtime data for 3D simulations.
I haven't tried to run a benchmark, but I suspect that most ordinary desktop and laptop computers have a powerful enough graphics chip to give a pretty good rendering of a thousand close-packed spheres. They might only be able to do flat lighting, and show some seams and jagged edges. But you'd be able to figure out at a glance that you're looking at spheres, and probably spin and zoom them in realtime at 30+ frames per second.
For about a decade, high-end gamers have been the main source of improvements in consumer-level computers. If it wasn't for their desire to navigate photorealistic 3D virtual world models in realtime, there would be little reason to buy anything beyond a year-2000 level machine. That was when pretty much every machine for $500 and up became powerful enough to do all the office work that most people needed, with the machine mostly waiting around for the user to think of something to type.
Unless you're compiling a large program, crunching a very large spreadsheet, creating audio or video or photos, trying to turn off the dancing paper clip and search dog, or getting rid of spam, when was the last time you ever had to wait more than a few seconds for your computer to do anything you asked?
Within a few years, even netbooks, PDA's and tablets will have the same power as today's desktop machines.
I think within a decade, the question of whether a machine is powerful enough for 3D rendering will make as much sense as asking today if a computer is powerful enough to show different types of fonts. That was a serious issue when the 128KB, 8 MHz Mac was introduced, but is now something we can take for granted - usually, even on phones. As Moore's law and consumer electronics industry continue their magic, 3D will be more and more affordable until it's just thrown in for free to pretty much everything.
I think that Neil Stephenson's description from Snow Crash (a very trippy novel) is an astute prediction: Using any old computer, you'll be able to see a good enough version of a 3D virtual world to navigate and enjoy. A sign of luxurious wealth will be to add gratuitous photo-realastic rendering, especially of difficult to model materials such as flowing water and leaves. In the same way that wealthy entities show off with their buildings now, in the future they'll show off with spectacular renderings that they create and show to you that your own little machine couldn't create on its own.
When I have time to post my Smalltalk series here, I hope to trace that all the way up through Open Cobalt. As far as I know, this is the closest anyone's come to building Stephenson's "Metaverse" out of entirely free software. Unlike the novel, we have an opportunity to use free software tools that put the power of understanding and changing the system in the hands of ordinary users. Another generation or two and modifying this code will be an everyday skill for everyone, just as most of us already know how to fry an egg and to type. I'm almost done with my first Smalltalk post about how I believe its designers used Law of One principles in its foundation, and hope to post that within a week or so.
Sorry about hijacking the thread, back to the circle algorithm. As you can tell, this is stuff I've been thinking about a lot.
The whole area of computing power for the masses is fascinating to me. I hope carrie won't mind my having a digression here. She's probably better up to date on much of this than I am anyway and might be able to correct my mistakes.
3D graphics power available to consumers these days is astounding. For a few hundred dollars, a GeForce 8800 card can shade and overlap more than 3 million triangles a frame at 60 frames per second. With 1,000 spheres shown at once, this would be 3,000 triangles per sphere, each just a few pixels across. Each pixel would have whatever colors most beautifully represent a smoothly drawn surface. The 3D engine runtime would adjust the level of detail in a surface depending on how much it fills the screen.
With a card like that driving an HDTV, you'd never see any kind of angles or seams, just the impression of a thousand smooth spheres - with moving lights shining through fog and smoke, if you like.
High-end consoles like Playstation 3 and Xbox 360 have somewhat less powerful graphics engines. But the consoles include custom processors, system architectures and operating systems designed to easily manipulate massive amounts of realtime data for 3D simulations.
I haven't tried to run a benchmark, but I suspect that most ordinary desktop and laptop computers have a powerful enough graphics chip to give a pretty good rendering of a thousand close-packed spheres. They might only be able to do flat lighting, and show some seams and jagged edges. But you'd be able to figure out at a glance that you're looking at spheres, and probably spin and zoom them in realtime at 30+ frames per second.
For about a decade, high-end gamers have been the main source of improvements in consumer-level computers. If it wasn't for their desire to navigate photorealistic 3D virtual world models in realtime, there would be little reason to buy anything beyond a year-2000 level machine. That was when pretty much every machine for $500 and up became powerful enough to do all the office work that most people needed, with the machine mostly waiting around for the user to think of something to type.
Unless you're compiling a large program, crunching a very large spreadsheet, creating audio or video or photos, trying to turn off the dancing paper clip and search dog, or getting rid of spam, when was the last time you ever had to wait more than a few seconds for your computer to do anything you asked?
Within a few years, even netbooks, PDA's and tablets will have the same power as today's desktop machines.
I think within a decade, the question of whether a machine is powerful enough for 3D rendering will make as much sense as asking today if a computer is powerful enough to show different types of fonts. That was a serious issue when the 128KB, 8 MHz Mac was introduced, but is now something we can take for granted - usually, even on phones. As Moore's law and consumer electronics industry continue their magic, 3D will be more and more affordable until it's just thrown in for free to pretty much everything.
I think that Neil Stephenson's description from Snow Crash (a very trippy novel) is an astute prediction: Using any old computer, you'll be able to see a good enough version of a 3D virtual world to navigate and enjoy. A sign of luxurious wealth will be to add gratuitous photo-realastic rendering, especially of difficult to model materials such as flowing water and leaves. In the same way that wealthy entities show off with their buildings now, in the future they'll show off with spectacular renderings that they create and show to you that your own little machine couldn't create on its own.
When I have time to post my Smalltalk series here, I hope to trace that all the way up through Open Cobalt. As far as I know, this is the closest anyone's come to building Stephenson's "Metaverse" out of entirely free software. Unlike the novel, we have an opportunity to use free software tools that put the power of understanding and changing the system in the hands of ordinary users. Another generation or two and modifying this code will be an everyday skill for everyone, just as most of us already know how to fry an egg and to type. I'm almost done with my first Smalltalk post about how I believe its designers used Law of One principles in its foundation, and hope to post that within a week or so.
Sorry about hijacking the thread, back to the circle algorithm. As you can tell, this is stuff I've been thinking about a lot.