The Fastest Graphics Cards of 2004

[Kocur]

Hi guys,

I just wanted to point out to you the article on www.xbitlabs.com/articles/video/display/2004-27gpu2.html. In my opinion it is very unprofessional. There are two issues:

(1) The comments regarding the results for each particular test are really bad.

The authors do not try to find out what is really behind the scenes, but use, interchangeably two or three sentences, depending on who the winner is.

Is that all what these guys are capable of?

(2) The reader cannot really compare the PCI-E results, since single card PCI-E results were obtained using P4 platform and SLI results were obtained using A64 platform.

Moreover, they seem to claim that FarCry does not work in SLI, which is, of course, not true (see, for axample, articles on Anandtech and Digit-life for reference).

Summing up, they have got nice test suite, but have to learn a lot yet. Do you agree with me?

 

[Stumps]

nah your just to picky

 

[thegimp03]

Haha.

 

[RobotOfHatred]

Your just sad that they choice the X850XT-PE as the fastest PCI-E card

 

[otispunkmeyer]

they not that bad, you shouldnt really take anything from the web as the absolute complete truth, more of a guide, anyway least i dont see any bias, even if they are unimaginative in their explanations, NV win doom3, ATI win HL2 (by a bigger margin than i thought too!!!, but its jus a guide) and then its a toss up between the rest

even the fact that nv owns doom and ATi own HL2, it doesnt really matter, both games are outclassed in terms of enjoyability by many other games, those two werent really all they were cracked up to be

 

[Matthias99]

Originally posted by: Kocur

Hi guys,

I just wanted to point out to you the article on www.xbitlabs.com/articles/video/display/2004-27gpu2.html. In my opinion it is very unprofessional. There are two issues:

(1) The comments regarding the results for each particular test are really bad.

The authors do not try to find out what is really behind the scenes, but use, interchangeably two or three sentences, depending on who the winner is.

Is that all what these guys are capable of?

I'm pretty sure that English isn't these guys' native language. I'm willing to cut them a little slack. Besides, most reviews have pretty lame commentary.

(2) The reader cannot really compare the PCI-E results, since single card PCI-E results were obtained using P4 platform and SLI results were obtained using A64 platform.

That did seem a little odd, but you can still compare all the single PCIe cards. It would obviously be nice if they had tested all the PCIe cards on the A64 platform, but considering the scope of this review, that might not have been feasable within a sane amount of time.

Moreover, they seem to claim that FarCry does not work in SLI, which is, of course, not true (see, for axample, articles on Anandtech and Digit-life for reference).

They seemed to have SLI issues with a number of games. I think it's fair to say not all the bugs have been worked out yet.

Summing up, they have got nice test suite, but have to learn a lot yet. Do you agree with me?

I think it was a pretty solid set of reviews. Not a lot of surprises, but always good to see stuff tested with the most recent drivers, etc., especially with such a large variety of benchmarks.

 

[otispunkmeyer]

hold on a sec, why do they not comment on the CS: S performance of the PCI cards? they perform half as good as the AGP ones, well the nvidia cards anyway, the ATI ones fair better, but they still cant match the AGP group, i know Intel arent the force in gaming, but surely the difference isnt THAT big

 

[Cheesetogo]

Originally posted by: otispunkmeyer
hold on a sec, why do they not comment on the CS: S performance of the PCI cards? they perform half as good as the AGP ones, well the nvidia cards anyway, the ATI ones fair better, but they still cant match the AGP group, i know Intel arent the force in gaming, but surely the difference isnt THAT big

look at this to see how big a difference it is. http://images.anandtech.com/graphs/half...2%20cpu%20scaling_01260560148/6035.png

 

[RobotOfHatred]

Edit: HAHAHA woops, wrong Post. Osrry about that