(Cracking nuckles)
I'm afraid you are extreamly wrong with that comment. That's like saying that the original PSX GPU was nothing better than A Matrox Millinum card. One is an apple, one is an orange
I'll give you a light overview on how the PS2 renders stuff, and how amazing this system really is. But before I do, let's backtrack into the 2D world for a second.
You know that with a 2D surface, kind of like a blackboard, you draw shapes, squares, circles, lines, and dots. Nothing really exciting, but you can, with time and effert, draw the Mona Lisa. But it takes a while and is very hard to do.
This is what a computer has to do without 3d acceleration
Now you can stream a 2d video onto the blackboard via a projecter get wonderful graphics and the work is nill and no memory is being used. This is because everything's been done before. It's just putting stuff on the board.
This is how streaming movies are done on a computer. Like the PS1's MDEC. The CPU doen't know or care what's on the screen. It can do something else.
Now the PS2 can do something rather extraordinary. It can stream polygons There are two sub-CPUs (VDU #1 and #2)with 64k ram each. Thier job is to render polygons and toss them to the emotion engine (a third CPU) which take the rotation data from the vector coprosser (a fourth CPU) puts it all to togeather and put it on the screen. All this without the primary CPU doing anything. When the PS2 is running peak, you have 5 CPUs running in tandom. (6 if you add the new MDEC and you want to stream polygons ontop of video, which Square has been know to do) This is why it's so difficult to program for.
Af far as the Anti-alising goes. The PS2 can do AA, but at a cost of dropping the framerate by 3/4. It's a bug with the system. (The 3D clipper was broken in the PS1 and had to be worked around as well) Oh well, what are you going to do.....