X-bit labs

Started by: Niels | Date 05/24/07

Comments: 27 | Last Comment:  07/17/07

[1-19]

1. What a waiste of time to read this,still with 8.37.4 drivers....
A missed opportunity if you ask me to be the first with a decent 8.38 driver review

They were using older Nvidia drivers as well, 158.19 only, when the 158.22 so both sides are using old drivers sets...

The 8.38 BETA drivers are reported to have flakey or broken support for Crossfire, OpenGL, video acceleration under Vista and degraded image quality under some filtering. Most respectable sites benchmarking atm are using the STABLE 8.37.4.3 driver.

If you wanted the 8.38 BETA drivers used simply to give great 3DMark benches, tough.

I'd rather see the card correctly properly under stable drivers.

I'd rather see the card tested properly using stable drivers.

2. Excellent article that focuses on performance and not on power usage or GPU temperature like some other review articles. That's what should matter the most.

I was hoping that ATI would have released updated drivers with serious performance improvement in every game where it loses to X1950XTX in either average or minimum frame rates... before Xbitlabs posted this review. Too late, ATI... Maybe there's going to be a surprise where the updated drivers secretly enables another 16 TMU's on this GPU which would go to explain insane power consumption, even worse than 8800GTX in both idle and load modes.

I just do not understand the higher 3DMark05 scores against the 8800GTX when performs half as fast as the 8800 in all the 3 games, averaged together. So does the difference come from the CPU tests in 3dMark05? Well, since the default benchmark only requires those tests, Futuremark must be allowing something funny to happen with ATI's scores. The same goes for 3DMark06--when you look at each of the game test scores, we can only point to the CPU tests for making up the difference but we know that it is impossible for the slightly varying CPU tests to bring HD2900's score up to nearly equal of 8800GTX's score! 3DMark06's default benchmarking does not require the other miscellaneous tests either--the scores should be roughly the same without or with the rest of optional tests, regardless.

It is very commendable in those few games where the X2900XT beats the 8800GTX hands-down without any visible rendering issues (or visible cheats)--and goes a long way to show 2900's potential with future driver optimizations.

Perhaps ATI will shortly release the XTX version with 1 GB memory, with "Golden" drivers that unleashes full (perhaps hidden) performance of those 700-million transistor chips running at 215 watts (load wattage in normal temperature condition as stated by AMD)! This is perhaps the biggest video card mystery ever--more than Matrox's Parhelia was (but unfortunately never improved much as claimed with drivers) or even 3dfx's Voodoo6!

The 3dmark 05 and 06 scores are for default setting, and the individual scores are using FSAA 4x + Aniso 16x.

Do you all simply still believe that a piece of "software", a "magic" driver could somehow eliminate the serious "hardware" bottlenecks ??
Drivers can optimize performance a little bit, that's I have no doubt about, but that would be in certain cases in certain games, it won't allow this card to match or overtake the 8800 GTX in general case, maybe the R650/R670/R700 or whatever it will be called can do what the R600 was originally intended to do -presuming that these BIG hardware shorts are fixed-.
This battle is over for me, nVIDIA wins hands down, plain and simple.

> Excellent article that focuses on performance and not on power usage or GPU temperature like some other review articles. That's what should matter the most.

No kidding!

This is a not a notebook part, nor is is a hybrid motor car. Performance per watt/degree are irrelevant and meaningless metrics, unless they're so bad they pose a potential problem, which is not the case. This article is well balanced an informative, vs. the "other" website's somewhat anal perspective (IMHO)

This isn't the idea. You know, I do agree with you that performance per watt metric is not what matters with most of the gamers, still, compared to the targeted competator -the GTS- the R600 is highly inefficient in that metric. It's just one card being more economical than the other, since both can deliver an equal performance then why should I go to the R600 when the GTS can deliver the same performance with less price and lower electricity bills on the long run ?? keeping in mind that it's built on a bigger process -90nm vs. 80 nm- and still being more efficient, this is a big ponus for me.
If the R600 was leaving the GTS in behind then I might think about it, but even that is not the case, sadly.

3. well done good article

4. I would like to see an atricle on how the game play actually is between the cards and not just FPS. Feedback from else where says that ATI's image quality is suffering at the expense of FPS.

5. There's no UVD in the HD 2900XT, the HD Decoding capabilities are the same as X1950 cards...

6. well , we really dont know if ATI bring out the latest and stable driver can make it a lot of different...

I'm ATI fans really disappointed about HD2900XT performance this card should at least have same performance with 8800GTX due to 6 month late

and now nvidia are one step forward compare with ATI...

This make Nvidia are market dominant for Video Card segment and control the price for high end card...

Overall Xbitlabs make balance article good job and keep it up

7. "Will the novelty beat its two main competitors – Nvidia’s GeForce 8800 GTS and GeForce 8800 GTS?"

I think one of those should be gtx ??

8. Why can't you run FSAA + HDR in Oblivion? It works for all X1900 series cards, and I thought the 8800s as well.

I'm wondering the same thing
post it here when you find out something

9. "... and that the developer hasn’t yet enabled FSAA for the dynamic lighting (the reason why we have to test S.T.A.L.K.E.R. with anisotropic filtering alone), we can suspect some defects in the game itself, not only in ATI’s driver.."

This game uses deferred rendering therefore msaa in hardware does not work, wich is why most developers won't use this technique.
So this is not a defect but rather a design choice.
I've also read somewhere that it uses hardware accelleration for shadows on nVidia hardware.

10. very good article. although i recently upgraded to and 8800 gts 640mb, i still find this to be a really good choice. in most of the tests, the new hd2900 only out performed a few more frames a second more... which in real-world gaming doesn;t make too much or noticable difference in game play.

i would probably consider going SLI with another 8800 gts or upgrade my card when directx 11 comes out...

11. How about an article comparing the video playback performance of the HD 2900XT and the GeForce 8800 GTS/X?

I would especially be interested on what is the minimum CPU speed you could use with either of these cards and still have no dropped frames on 1080P content (HD-DVD, BluRay and WM-HD) - telling me their CPU usage with a quad-core Intel is meaningless as I'll never put that in a HTPC.

Once you find the minimum CPU these cards will play smooth 1080P video with, give us some 1920x1080 game benchies using that minimum config.

I'm sure I'm not the only one looking to use a HD 2900XT or 8800GTS for a home theatre PC :)

12. i plan to buy one video card to play starcraftII, is it one suitable?
_____________
Best Video Converter
http://www.bestvideoconverter.net/

13. An excellent card for 400$

14. will we get a revised bench with the new drivers?

15. ok

16. Interesting, this is cool, nearly made my day. :)

guide pocket zune
http://www.giftdownload.net

17. Sounds like a shit to me
this ati/amd/whatever product is light years far far away from success and it' shouldn't show up at all especially after such a long long time delay

18. Where are the power consumption charts for your ATI video card reviews? This site used to be the best place on the net i could find for reliable video card power consumption and now you dont even do it anymore? I can get benchmarks anywhere.

19. Yea,,,ATi HD2900XT it's the best Video Card for that Price...
Good Video Card and bad drivers...So bad for AMD...
Works more for drivers,,,because i can't beleave that 512bits loses against 320bits... =O ...
Anyway I use ATi HD2900XT ,,,noisy cooler,,,but good performance and games runs smothly...

[1-19]