🚀 go-pugleaf

Jo Lux (lux@uni-bonn.de) wrote:
: 
: It is evident that you can't compare the pure fps performance of a
: game to the results of 3D Winbench 98 (3D Winmark), because it is a
: performance benchmark AND a quality benchmark.

So by what you're saying, since 3D Winbench 98 rates the Intel chip
as being nearly the equal of Voodoo2, and since it obviously is not
performing as such, we can therefore expect the Intel chip to
give us an image quality approximately twice as good as the Voodoo2?

...Heh.  I guess I can't wait ;)

: "Each WinMark test produces a result in frames per second. The 3D
: WinMark 98 is simply the sum of the results for all the tests. If a
: test cannot be completed because the required capabilities aren't
: supported or because the test failed, the test contributes zero to the
: overall 3D WinMark 98 score."

This is pretty enlightening; it basically means that 3D Winbench 98
was specifically designed to test for rhetorical features that
would only exist on a chipset if the designers were deliberately
trying to get good 3D Winbench 98 scores.

: Didn't Tom realize that? He wrote:
: "Read the shocking truth about the reliability of 3D Winbench 98."

Remember, he had real-world benchmarks (you know.. using the actual
applications one is bound to use with 3d hardware?) to support his
claims.  Of course, the Intel chip _is_ good for one thing: You can
spend hours drooling over your kick-ass 3D Winbench 98 score.

: Haha! I get the impression that he is not very competent as to that -
: or he wants to overhype the Voodoo2.

That was foolish.  But I think I'll let someone else respond to
this particular piece of flamebait.

-Maruku    Marc Brown    maruku@server-4.busprod.com