Tom's DX10 performance analysis

trinibwoy

2[H]4U
Joined
Apr 25, 2005
Messages
2,090
When the Canadian company introduced this chip, it insisted that its graphics processors used the better and faster DX10 design. Its cards were looking so slow because they simply weren’t being challenged enough under Windows XP and were being tested in the wrong manner anyway.

ATI fans rallied around this statement, waiting for real DX 10 titles and always holding out hope for faster Windows XP drivers. At the time, THG predicted the performance difference between the chip families would not change much between DirectX 9 and DirectX 10, even when native DX 10 titles appeared. At the same time, message boards all over the Internet were rife with speculation. The common thread, if you will, was that ATI surely had some untapped potential in its cards that just needed to be teased out. ATI simply needed more time – after all, Nvidia had six months to tweak its graphics drivers. Given enough time, ATI’s drivers would be bound to improve, giving the Radeon 2900 XT the much-anticipated performance boost.

http://www.tomshardware.co.uk/DirectX10-Geforce-Radeon,review-29666.html
 
I remember ATI fan saying ATI's drivers would be ready, out of the gate, unlike nVidia. That's what they get for having the "wait-and-see" mentality.
 
That wasn't pretty for the HD2900XT.


Isn't that the truth. Ouch. :eek:

Let's hope that ATI's next gen product can compete much better with Nvidia's next gen product. Consumers want competition. Keeps the industry honest.
 
Anyone know what caused the 8800gts problems in CoH? Their test has it below the 8600gts :p
 
Man ATI got no love in that article. But I even find it funny that in WiC Test1, the 8600 GTS 512mb beats 8800 GTS 640's.
 
ATi and Nvida trade blows all the time so I fail to see what the big deal is anyways. One only has to go back to the NV30 to see how badly NVidia screwed the pooch.......
 
Oh you mean we have to go back 4 years to 2003 to find evidence of them "trading blows all the time"? ;)
 
what a great article, I am so glad I went with the MSI 8800 Ultra OC 768MB - I can't wait to get my computer!!
 
Man ATI got no love in that article. But I even find it funny that in WiC Test1, the 8600 GTS 512mb beats 8800 GTS 640's.



It just so happens that both cards are memory limited in the final test, so that they are both texturing via PCIe. Only the 768MB cards have enough memory to keep everything in local ram. Considering what a memory hog DX10 is turning out to be, that's really no surprise.
 
Man ATI got no love in that article. But I even find it funny that in WiC Test1, the 8600 GTS 512mb beats 8800 GTS 640's.

In CoH the 8600 256mb supposedly beats the 8800GTS 640mb also at 1280x1024 4x AA. :rolleyes:
 
In CoH the 8600 256mb supposedly beats the 8800GTS 640mb also at 1280x1024 4x AA. :rolleyes:

Sometimes tests are won because of character. Maybe the 8800 was cocky being 320bit and all :D hahaha
 
Does it really matter what's the best card, really?
I mean it doesn't matter. I have the 8800 and it's best that's all. :rolleyes:
 
why was the 8800gts 320 beating the 8800gts 640mb a lot of the time at high resolutions?
 
And why did the 2900 beat the 640 in company of heroes, it doesn't make any sense at all.
 
You have to give them a break on some of the tests. There was 170 different benchmarks run on just WIC. Thats a shit ton. I'm sure they used multiple machines and numerous geeks. :D
 
I know Toms isnt the most reliable source, but many of their articles are well done and informative.
This one paints a bleek picture for ATi users.

Sure glad my 8800GTS 320mb OC is right up there.
 
Anyone know what caused the 8800gts problems in CoH? Their test has it below the 8600gts :p

I recall seeing this type of anomaly at least once before. It may have been during the GF7 series that i saw a ridiculous benchmark (akin to a 7800 getting whooped by a mid-level ATI card in Half Life 2). I forget the actual benchmark but i know it was from Tom's. That was almost 2 years ago (the last time i visited that site). Here's to two more years of avoiding Tom's.
 
Any site can post informative articles, having reliable benchmarks is not something Tom's been puting out!
 
Back
Top