Inside the Xbox 360
QT writes "Ars Technica has an in-depth look at the Xenon processor
that will power the Xbox 360. It's the first technical look at the CPU itself, its design goals, and some of the differences between it and IBM's Cell processor. The Xbox
360's procedural synthesis capabilities look quite impressive, and I'm not as convinced as I was before that the PS3 would spank the Xbox 360."
The computer/electronics industry changes extremely quickly from a historical point of view. We should expect alliances to be made and undone all the time between competitors.
Apple isn't switching their CPUs to Intel, they are using Intel chipsets for other components of their computers...
Why would it?
You're refering to Intel's Xeon.
The XBox 360 has IBM's Xenon CPU.
Hope this clears it up.
^_^
There are many approaches. A big problem has been avoiding "popping", when an area suddenly is rendered with more detail.
Again, this is well understood. It just takes plenty of computer power to drive it.
Microsoft seems to be paving the way for game developers to use their specialized hardware. When the PS2 first came out, the development tools were weak, and it took about two years for developers to get the tools in place to use it effectively. The original XBox is basically a PC; you can develop, test, and debug on Win2K, then rebuild for the XBox target. The new XBox won't be like that; the target is drastically different from the development environment. So Microsoft has to do and promote more middleware development.
Because every developer on the floor knew that the most impressive demos for the PS3 were totally prerendered.
:
5 -23&res=l
While I generally frown on blatant karma whores who post links to Penny Arcade strips in the vain hope that some of the humour will rub off on themselves, I think that this one is particularly appropriate
http://www.penny-arcade.com/view.php3?date=2005-0
I will now spend the rest of the evening frantically flagellating myself with barbed wire in an attempt to atone for this act of rank hypocrisy.
1) Yes it is better. The designer doesn't "fit" a mathamatical curve -- the software handles that under the hood for them. They choose a circle and manipulate it, or they choose points to deform on a line and manipulate it.
... one "shape" may have more data than a single vertex, but it has far less data than a similar shape made out of vertexes. This reduces the amount of data that must be transfered over the system bus, which is critical in this context.
It is definately less information
The other advantage to this methodology is that the rendering software can dynamically vary the number of vertexes it generates based on the rendering load.
2) They applied for one and got it. Remember, patents don't cover ideas -- they cover how you implement an idea.