sc94597 said:
Mifely said:
sc94597 said: @ Mifely I got 2306867 without anything else running. Are you sure you are using ALL of the systems memory. All 88 mbs. |
Not sure what you're referring to, but the pther posters are referring to an older GameCube game. I'm not going to post on this any further, since... well there's no point. And, for reference, 24 million / 40 = 600K single-textured, unlit polys at once, with no code, or game of any sort, assuming the 3MB of GC VRAM (they're talking about a GC game, as I said) is all used for textures and the framebuffer. The GPU performance of the GC is listed pretty clearly on wikipedia -- 337.5K polys/frame at 60 fps, and that's untextured, unlit, single-cycle fill which is probably about 10-20x faster than your typical game polygon. |
"The Wii is probably about 600K at 60 fps, again with tiny, untextured, unlit polys" I got 2000 polys/frame. Also I think they are talking about per second while you are talking about per frame. So you guys may both be correct. They are talking about second you are talking about frame. |
20M / 60 is only about 335K, which would mean that these games would have to be running untextured, unlit, <= 4 pixel wide/long polygons to achieve this. I don't recall Rogue Squadron or Rebel Strike being that boring. =) Even if they are talking per second, they're dreamin. Besides:
Lets take a realistic situation. 3 float for x,y,z + 2 float for 1 u,v tex coord (one texture is kinda weak though), plus 3 float for a vertex normal == 8 floats per vert (32 bytes). at an average of 2 verts per triangle, this is still 64 bytes per poly. If you actually had 335K polys, such that you could validly render 20M per second at 60 fps, that's:
335000 x 64 = 21.44M... almost all of the GameCube's RAM, just to hold the mesh data. That's ridiculous, to say the least, not to mention that the GC's realistic GPU performance amounted to about 20-40K polys/frame at 60fps (variable depending on all sorts of factors like # of textures, total fill, # of dynamic lights, etc), not 335K of the boring uselss polys they use to gauge max performance. I'm being pretty generous here, even. Most GC games had about 20-30K polys/frame, but didn't break 30 fps.
Back to the original topic, the point is that, even though the Wii exceeds the GC's performance numbers by a good margin, and even the original XBox to some (IMO minor) degree, its not even in the same class as the other consoles of this generation. There's not really any point in discussing it.
You could claim that "good programming" could make the Wii really shine... but that's not really unique to the Wii at all, is it? You might say its less so on the Wii, due to the much more detailed pipelines on the other consoles, even. Even on a good day, the Wii can't hold a candle to the other two platforms, in terms of graphics. That's why Nintendo games are all about gameplay, and not just a pretty shine.