View Full Version : 3DMark2001 - Another NVidiaMark?

14th February 2001, 12:31
There are screen-shots of 3dMark2001 popping up - seems it's going to be released in the nearer future.

Not so long ago the webmaster of 3D Center wrote a few words about the strange "delay" in the 2001 version of 3DMark (the 2000 version was released even in 1999), suspecting MadOnion is waiting for the NV20 to be available before releasing this benchmark (probably to tweak the bench and make the NV20 look great, which, of course, is not that easy when the benchmark is released earlier than the chip).

When I look at this "strange coincidence" of 3DMark2001 being finished magically at about the same time the NV20 is finally hitting the shelves, I think he was correct.

Just wonder how much they get paid for this crap by NVidia (and perhaps Intel too).

[This message has been edited by Indiana (edited 18 February 2001).]

15th February 2001, 21:47

16th February 2001, 15:57
"Benchmarking has NOTHING to do with the "popularity" of a certain chip."

The selling of a Benchmarking prog has nothing to do with the cards that will be benchmarked with it??!! http://forums.murc.ws/ubb/eek.gif http://forums.murc.ws/ubb/eek.gif

Kosh Naranek
17th February 2001, 07:57
When The K7 came out 3DMark 99/Max didn't include special code to handle this new cpu, but AMD in cooperation with MadOnion supplied an extra .dll so that they (AMD) could show of their shiny new cpu.
In 3DMark 2000 support for the K7 was there from the beginning.
So i'm wondering how this makes AMD look to you ?
I'm sure that, if Intel had supplied a special .dll then everybody would have started screaming FAKE FAKE FAKE etc. but now that it's AMD everything is a ok. !

Kosh Naranek
17th February 2001, 07:58

[This message has been edited by Kosh Naranek (edited 17 February 2001).]

17th February 2001, 14:05
Kosh, actually it's vice-versa. I'm sure you know that AMD did indeed get flamed for this 3DNow dll. And while 3DMark99Max "overrated" AMDs 3DNow, the 2000 version does make the Intels SSE look too good. The rules that applied for AMD also apply for Intel.

If two benchmarking programs of the same team suddenly change the speed-order of two CPUs that basicly means that at least one of the benchmarks is crap - well, in FutureMarks/MadOnions case both of them are crap, one overrating AMD and the other overrating Intel. You could say MadOnion are willing and happy to support everyone that pays enough....

19th February 2001, 05:42
I don't see why the reason a Intel CPU gives better performance in 3DMark is because MadOnion is paid by Intel.

Maybe SSE gives more performance boost than 3DNow!/3DNow!2 ? Of course, if it's general knowledge that this is not the case, then please enlighten me http://forums.murc.ws/ubb/smile.gif

19th February 2001, 07:22
Sigh. Which card is more popular? Geforce or G400. Think about it.

19th February 2001, 08:19
And we all know how accurate 3DMark is.

forums.murc.ws/ubb/Forum13/HTML/000015.html (http://forums.murc.ws/ubb/Forum13/HTML/000015.html)


19th February 2001, 09:03
got a link to those screenshots ?

19th February 2001, 12:47
If you compare the 3DMark99/Max (this was at a time when FutureMark/MadOnion apparently was not bought out by NVidia) CPU results of the same CPUs - i.e. Athlon600 vs. P-III 600 - to those given by 3DMark2000, you have to wonder about the sudden speed change of those same CPUs. Besides the 3DMark results (and here not only the CPU-ones) don't compare to reality. MadOnion is supporting Intel nearly as much with their "benchmark" as NVidia.

When you develop a benchmark and specifically wait until the card of a certain developer is available in silicon so you can sport features specific to this card and make sure it runs the benchmark faster than the competition, what do you call this?
Benchmarking has NOTHING to do with the "popularity" of a certain chip.
Think about it.

Hmm, not a very good example as you can apply this hack to nearly any benchmarking software. It's not the vulnerability for hacks that makes 3DMark a bad benchmark, it's the internal code-tweaking to make "certain" (we all know whose) cards look good - those FSB/clock-hacks can only affect single results, but writing a "benchmark" in favor of certain vendors does have an impact in EVERY hardware test out there in the great internet-world.

Look at Madonions homepage. Another screenshot can be found on e.g. on http://www.rivastation.de .

[This message has been edited by Indiana (edited 19 February 2001).]

[This message has been edited by Indiana (edited 19 February 2001).]

19th February 2001, 12:59
To which I might add, when MadnVidiot added EMBM to their test suite, they added it at the end and if your card didn't support this feature it sure the heck didn't alert you, unlike the bogus cpu scores if you have a GeFarce installed.

It's my hope that if they do indeed impose some sort of anything promoting nV that every single MatroxUser will scream at them about it! http://forums.murc.ws/ubb/smile.gif

all is fair in Love and War