Jim: I predict a magical, "out of the blue", agreement between nVidia and Intel when push comes to shove when it comes to whose chip gets put in the next X-box version. AMD will likely be left out in the cold again or in the warm if you consider the profit on these X-box chips...then again...
Rumor has it MS is seriously considering not going with NVIDIA for the next XBox.
Btw, maybe you can explain something to me…
I don't understand the whole XBox concept. Here's why:
PC's today are extremely flexible, both in design and performance. They are great a wide variety of applications and this is their greatest strength, IMHO. A PC can be great at everything from playing games to serving web pages, from database mining to complex scientific calculations.
However, if you look at something like the XBox, it only needs to perform a small subset of these applications. Namely games - and the odd video encoding / decoding, plus perhaps some word processing, email and browsing. However, except for games, none of this really stresses the system to any appreciable degree, so it seems logical to design the XBox (or any console) precisely for this application.
Now, why on earth would you chose to go with what is essentially a PC, rather than design the whole box around the graphics core? Clearly, the two approaches do not yield the same end-product.
As far as I can tell, there are only a few arguments in favor of going with a PC:
1) existing software base
2) off-the-shelf hardware
3) lower system testing and verification costs
#1 is clearly out the window (pun intended), with MS changing the OS to be incompatible with standard PCs. You could then argue "ease of porting to a slightly modified OS" instead, but this argument doesn't hold water anyway. If you write your game in C (which everyone does) and target OpenGL, porting to virtually any other platform requires little more than a recompilation (some of my best friends work for a company that specializes in porting games - no, not Loki ;-) - so I know what I'm talking about here).
That leaves #2 and #3, which basically both rely on the argument that mass-production is good. And it is, but since this isn't some niche product, but rather a console sold in the millions, I just don't buy it. Sure, you can save some on R&D, but your overall component costs will likely still be higher, since the system is basically a PC - and capable of all the things a PC is capable of (and for which a console has no need).
To put it briefly, since the XBox won't run any of the applications that require a CPU-centric design, why base it on a PC?
-fyo |
