Dear forum members,
We are delighted to inform you that our forums are back online. Every single topic and post are now on their places and everything works just as before, only better. Welcome back!


Discussion on Article:
The Fastest Graphics Cards of 2004: Ultimate Testing of 27 GPUs. Part II: Performance Analysis

Started by: sorienor | Date 02/16/05 04:28:52 PM
Comments: 213 | Last Comment:  11/29/16 08:05:47 AM

Expand all threads | Collapse all threads


I propose that your article shows heavy bias twords ATI.

1. You test the X700XT. This card is pure Vaporware and no consumer will ever have one.

2. Comparing a 6800GT to a X800XT you state both are priced at $470, based on Pricewatch. This is totally not true. My local Best Buy has a stack of BFG 6800GT OC's for $399.99 and the lowest price on Newegg is $370

3. Stating the value of a X800XL, you mention the MSRP of the X800XL at $299, and make a conclution of it's value based on that. However right now a X800XL can barely be had for $400.

I am not a peticular fan of either, but I felt compelled to point out these inconsistancies.
0 0 [Posted by:  | Date: 02/16/05 04:28:53 PM]

Also note the pricing of 299 is MSRP which of cours is SUGGESTED by the manufacturer and the pricing was from pricwatch, they didn't cliam it was the end all be all.
0 0 [Posted by:  | Date: 02/16/05 07:22:10 PM]
- collapse thread


There is more to this the "x card wins y benchmark" but price is also a major factor. When you conclude one card is better then another at a certain price point, and totally get the price points WRONG, then the only thing you can say is that the conclution is wrong.

Including the x700xt was pointless. Based on the other wording in the article, I can only conclude it was included to make sure ATI had the "Best of class" video card.

I didn't see any out-of-production nVidia cards like, say, a Ti4600. What was the point of including the 9500Pro?

I want to make it clear I am not a nVidia fanboy but only wish to point out blatent bias in what should be "impartial" reviewing.
0 0 [Posted by:  | Date: 02/16/05 11:22:00 PM]

The analysis of the dawn of war results is a bit odd

"The SLI mode quite unsurprisingly gives no speed boost and sometimes even worsens the performance."

The 6600GT/6800 ultra sli systems show huge leads in avg performance over their single card counter parts. Yes the 6800 ultra SLI did have a low min fps so your second bit sort of makes sense

"FSAA and anisotropic filtering enabled, ATI’s cards feel more confident and increase their advantage over their competitors from NVIDIA"

I think you have ATI and Nvidia the wrong way around

Other than that great review, shame you didnt include chronicles of riddick as I would of been interested in results from that
0 0 [Posted by:  | Date: 02/17/05 01:42:07 AM]

I was wondering whats the point in putting all the other PCI express cards on a P4 system and run the SLI version on a 4000+ [ as i see it, FAR superior CPU ].

Easy solution would have been to use the SAME nF4 system for the all the PCI express tests. Atleast then the damn values would have been comparable, now you can only compare the SLI versions against the AGP cards.

All that work and 40% of the article benchmarks are non-comparable... what a waste of your testing time.

Allso i wouldnt have even included the 6800 ULTRA SLI to the ultra-high end comparision.. OK, X850 XT PE costs a LOT, but 6800 ULTRA SLI is atleast in here around 2.4x more expensive than a X850 XT PE [yes, thats right]. SLI is very nice when it works, but as this test shows the list of the games you can actually get ANY value of the money you spent is quite small.
0 0 [Posted by:  | Date: 02/17/05 02:20:36 AM]

Another couple of things that members at a certain forum I frequent bring up is :

1."they don't have a PCI Express 6800 Ultra, and comparing benchmark results against two totally different plattforms is not exactly a great idea."

2."showing the "effective" memory speed made sense when not every video card had DDR. hey, guess what? they do now.

3."all the "effective" bullshit means is a nice marketing reminder that DDR works on the rising and falling edges of the clock."

4."Indeed. I usually tend to mention effective clock speed maybe once in a review, biut beyond that stick to the real clock speeds."

5."As regards this article, I am kind of surprised they didn't use one of the PCI Express 6800 Ultras in isolation, or at least use an AGP part capable of 425MHz core. I guess they simply didn't have one."
0 0 [Posted by:  | Date: 02/17/05 03:33:04 AM]
- collapse thread

Truly speaking, I don't like to "respond" to posts from third party forums :) because there are too many forums and discussions of our articles everywhere.

I wonder why the didn't ask we we didn't include performance scores of the RADEON X850 XT, RADEON X850 XT PE, RADEON X850 PRO, RADEON X700? Still, the GeForce 6800 Ultra PCI Express results will be available in our future articles - that's for sure.

I am also curious why Intel uses 533MHz, 800MHz, 1066MHz terms instead of real clocks? They are less educated than visitors of certain forums?
0 0 [Posted by:  | Date: 02/17/05 04:17:41 AM]
"Truly speaking, I don't like to "respond" to posts from third party forums :) because there are too many forums and discussions of our articles everywhere."

Understood Anton, and thank-you for replying to us, your readers. It is important to ask questions and get answers:)

Positive criticism is just that, Positive.
0 0 [Posted by:  | Date: 02/17/05 06:20:31 AM]

Some of the graphs are incorrect.

Several of the bargraps are dots, connected by lines.
See pages 44, 50.
In a graph, this means that the value can be interpolated.

This is not the case.
If the score for a card in CSS Chateau is 90, and in CSS Piranesi it scores 40, then a line between these two scores implies that there is a game in between Chateau and Piranesi for whic h the card's score is 65.
This is not so, hence these graphs should be barrgraphs.

0 0 [Posted by:  | Date: 02/17/05 05:20:13 AM]

Thanx for the review but...
I would like to ask : How do u compare AMD AGP platform to intel PciX platform.?????!!!
0 0 [Posted by:  | Date: 02/19/05 01:58:16 PM]

Counter Strike Source.... and Nvidia PCI-E... what the hell were you guys doing?

I just bought an MSI K8N Platinum SLI.... and put a 6800GT in it (will upgrade to the second one later)... and I'm not having any problems with CSS. At 1280x1024 8xAniso 4xAA I get around 95 to 100 average in the visual stress test.... and it seems to run around that while I'm playing the game as well... I play all the time and have the newest version of the game and everything.... you guys must have really borked something.

0 0 [Posted by:  | Date: 02/21/05 01:07:07 PM]

How come in games tests, the 6800GT agp card beats the 6800GT PCI:E card? Im planning on getting a 6800GT PCI:E and im concerned that it only gets better results when run through simulation benchmarks... is this because of an incompatability with the games and PCI:E which could be sorted with software updates?
0 0 [Posted by:  | Date: 02/27/05 11:41:35 AM]
- collapse thread

That's because the testing platforms differ and the PCI-E is a new technology and graphic card makers hadn't took the biggest advantage of it(the speed of data transfer) Although if you read about Nvidia's Turbo Cache you will understand what I'm speaking about.. If you're buying a new system, a PCI-E is your best choise, cause AGP is dying and new graphic cards will be made for PCI-E.
0 0 [Posted by:  | Date: 03/01/05 02:28:40 AM]

How does it come that the 6800gt AGP in CS source has over 100 fps

And the6800gt PCI-e Just about 60?

And the 6800ultra SLI performing less as the 6600GT SLI?

I want to play CS source and i'm wondering why The AGP card runs twice so fast as the PCI-e?
Only CS source so far i saw ?

(sorry for my bad english)

0 0 [Posted by:  | Date: 03/09/05 10:18:47 AM]
- collapse thread

Im not sure mate... but i asked virtually the same question and that guy talked about not making the most of newer technologies... im not sure why this is tho... im still planning on getting a 6800GT PCI:E and am going to go religious and start praying that it performs well....
0 0 [Posted by:  | Date: 03/10/05 12:44:22 AM]

I loved this article!
0 0 [Posted by: Stefanos Pritskas  | Date: 01/20/15 02:56:58 PM]


Back to the Article

Add your Comment