Send News. Want a reply? Read this. More in the FAQ.   News Forum - All Forums - Mobile - PDA - RSS Headlines  RSS Headlines   Twitter  Twitter
Customize
User Settings
Styles:
LAN Parties
Upcoming one-time events:

Regularly scheduled events

Report this Comment
Use this form to report the selected comment to the moderators. Reporting should generally be used only if the comment breaks forum rules.

62. Re: reviews Jul 24, 2004, 18:28 WaltC
 
"The benchmarking was conducted on-site, and the hardware vendors did not have access to the demo before hand, so we are confident that there is no egregious cheating going on, but it should be noted that some of the ATI cards did show a performance drop when colored mip levels were enabled, implying some fudging of the texture filtering"

Yes, Carmack is indeed quoted by [H] as having said exactly that. I find what he didn't say, however, to be far more interesting. What he doesn't talk about at all, even in passing, is the fact that nVidia drivers also provide trilinear filtering optimizations, and that they are enabled by default in the nVidia drivers. He doesn't mention such "fudging of texture filtering" by nVidia, nor does he even allude to the presence in the nVidia drivers of such "fudging" being enabled by default in the nVidia drivers...;)

I mean, it would have been one thing had Carmack said, "nVidia's drivers by default engage a similar kind of texture-filtering fudging, but we turned off such optimizations, and I personally guarantee that the nV fudging-defeat switch worked properly in D3, and all such fudging was turned off in the nVidia drivers when we ran these tests." Carmack simply makes no mention at all of nVidia's by-default "fudging of the texture filtering," let alone stating he turned nVidia's fudging off, or that he personally guarantees that the nVidia "fudging off" switch actually turned all of the fudging off as one might reasonably expect it to do..;) I wonder why, as Carmack is far too experienced for such things to escape him.

I found that glaring omission extremely odd. I mean, if you aren't going to specifically mention nVidia's texture-filter fudging, which is enabled by default in the nVidia drivers, then what is the point of specifically mentioning ATi's, and only characterizing ATi as "fudging"? I think he will need to explain that disparity in the coming weeks, as he will no doubt often be asked exactly that question.

Last, I found Carmack's framing of ATi's so-called "fudging of the texture filtering" to be odd in itself. He remarks that "ATI cards did show a performance drop when colored mip levels were enabled, implying some fudging of the texture filtering," which is precisely the behavior that ATi has been telling people to expect for weeks, ever since ATi's initial public disclosure of the nature of its trilinear filtering optimizations: in the presence of colored mipmap tests, and other conditions, the ATi drivers *automatically* turn OFF the trilinear optimizations (which is to say they turn OFF the "fudging," as Caramck puts it), which is why the performance drops. There is no reason on earth that Carmack needs to state that this behavior implies anything at all, since ATi has been publicly talking about that *very behavior* of its trilinear optimization algorithms for several weeks now. Nothing whatever may be "implied" by this behavior, because ATi for weeks has publicly verified it as fact. I'll take facts over "implications" any day of the week...:D

Again, the really strange and bizarre thing is that Carmack doesn't bother to tell *anyone* in his statement that if you are running the nVidia drivers in their *default* settings, then when you run colored mipmaps the performance *doesn't drop* because the nV driver *keeps on fudging the texture filtering* even in the presence of colored mipmaps, as there exists no automatic device in the nVidia driver to *turn the fudging off,* and as such it will result in *degraded image quality*, unlike for the ATi Catalysts which turn it OFF automatically under conditions where the optimization is likely to cause IQ degradation.

Heh...:D Just illustrates beautifully the old saw about how sometimes it's the things people *don't say* which is more important than what they say, doesn't it?...;)

 
Avatar 16008
 
It is well known that I do not make mistakes--so if you should happen across a mistake in anything I have written, be assured that I did not write it!
 
Subject
  
Optional
Message
 
Login Email   Password Remember Me
If you don't already have a Blue's News user account, you can sign up here.
Forgotten your password? Click here.
 




footer

Blue's News logo