Radeon HD 2900 XT vs. 320MB 8800 GTS

[xlink]

Also, I can run cs source, with everything maxed with in game settings, thats 8XMSAA , etc @ 1900X1080, and get 150-200+FPS in all the cs source servers I have tried.

Kind of makes me wonder, where all these ty numbers in reviews are coming from, cause I just don't see them.

8800GTS

CSS 1680*1050
16xQ

so the point is moot.

[XS2K]

I'm sorry but 1 driver release doesn't mean 100% increase in performance. Nor is the XT intended to compete against the GTX.

Look at what v_rr said

Stalker with Cat. 7.5 improved 100% it´s performance

that was what I was talking about....
I know XT isn't intended to compete with GTX,although ~5months delay should put it in the position to compete with GTX but nevermind that,when I said that I hoped it will beat the crap of 8800GTS/X it was figurative....but I sure do hope so...


P.S I just read your sig that explains it....
For me that 5 month delay isn't getting old,I really hoped that this video card will worth the wait but all we got is something that sometimes is better than 8800GTS...
Anyway that's just me I am sure no one else cares about that 5 month delay.

[Eastcoasthandle]

Look at what v_rr said that was what I was talking about....
I know XT isn't intended to compete with GTX,although ~5months delay should put it in the position to compete with GTX but nevermind that,when I said that I hoped it will beat the crap of 8800GTS/X it was figurative....but I sure do hope so...

-driver improvements are still needed, bottom line
-there has only been 1 driver release since the product was available for sale
-the "X month delay" is getting old. The card is officially release. Make that statement 5 months from now if nothing improves.

[Plywood99]

-driver improvements are still needed, bottom line
-there has only been 1 driver release since the product was available for sale
-the "X month delay" is getting old. The card is officially release. Make that statement 5 months from now if nothing improves.

QFT. If this card had been out for six months then fine. But 1 MONTH, come on people, give it some time...


ply

[v_rr]

Boosted stalker 100%? Let's see a link from a notable site on that one, because ironically enough I haven't heard a word about that, and I've been personally watching the whole driver situation pan out.

E6600@3GHZ
HD 2900XT
Windows vista ultimate

1680*1050

drivers 8.36
STALKER




drivers 8.37




drivers 8.38






4xAA






Also some improvements in Call of Juarez DX_10, Colin Mcrae and much better image quality with Cat 7.5.
Globaly other games uped performance a litle bit, but there is more to come with Cat 7.6 and 7.7

I know that in paper (release notes) ATI didn´t said anything, but the real world is what interests, not on paper.

As for [H] using this driver, perhaps they had the article already written, and the editor just took that long to finish checking it to put it up? The driver did only come out less than 2 weeks ago, now consider how long it takes to run all of these tests, write up the article, get the editor to check it, then get it posted. It's possible the driver came out in the middle of testing as well.

Sorry but I think that makes no sense releasing something like that. The review is very bad. So bad that even in their foruns there are a lot of people laughing. And that argument that there are no diferences to newer drivers doesn´t stick. Is too bad....

[AbelJemka]

Using TUbench from TechPowerup http://www.techpowerup.com/tpubench/
With [H] settings i otain this :
********** Batch run started at 00:43:54 **********

===== STALKER =====
Average: 46.6 FPS
Min: 5.1 FPS
Max: 146.2 FPS
Middle: 51.0 FPS
Benchmark completed.
********** Batch run started at 00:50:38 **********

===== STALKER =====
Average: 59.9 FPS
Min: 5.3 FPS
Max: 214.3 FPS
Middle: 62.3 FPS

First with Full Dynamic Lighting, second with Object Dynamic Lighting.
If an GTS owner can do the bench will be good. I use Cat 7.5a.

[gdogg]

8800GTS

CSS 1680*1050
16xQ

so the point is moot.

Oh thanks for providing your fps, and using a resolution I didn't. Not using MSAA and forcing your aa levels.

MOOT that point. ThX.


"You're playing CS:S and bragging about frame-rates? The x1950xtx could match you on that one. Well, other than 8xMSAA, but could do it with 6x.

You're playing a 3 and a half year old game that was easy on videocards when it released. I know, I played it with a 6800GT just fine at 1680x1050.

Try loading up stalker yourself and you'll see the exact performance problems they're talking about."

Guess a x1950 pro user says otherwise.

Note the higher resolution 1900X1080, and 150-200 fps next time, I don't like stalker, if it was a good game, then maybe i'd care about its performance, but to me, its doom3. Dark, and this ... http://www.gamershell.com/pc/stalker...ts.html?id=363

I can also play a new game, civ 4 @ 1900X1080 maxed out, 16X AA (in game setting) without any lag.

So in all honesty, its driver issue, plus problems the stalker devs will have to fix, since there as much to blame.

[Clairvoyant129]

Review was pretty bad.

[Clairvoyant129]

this is as bad as anandtech selling out to intel.

ROFL, yea Anand, Toms, Xbit, Hexus etc etc all Intel paid pumpers./sarcasm

[Ron 61]

You got to remember that the H likes to stir the pot. When they stir up sh*t they get more hits on their site which means more advertising money.

[xlink]

Oh thanks for providing your fps, and using a resolution I didn't. Not using MSAA and forcing your aa levels.

MOOT that point. ThX.


"You're playing CS:S and bragging about frame-rates? The x1950xtx could match you on that one. Well, other than 8xMSAA, but could do it with 6x.

You're playing a 3 and a half year old game that was easy on videocards when it released. I know, I played it with a 6800GT just fine at 1680x1050.

Try loading up stalker yourself and you'll see the exact performance problems they're talking about."

Guess a x1950 pro user says otherwise.

Note the higher resolution 1900X1080, and 150-200 fps next time, I don't like stalker, if it was a good game, then maybe i'd care about its performance, but to me, its doom3. Dark, and this ... http://www.gamershell.com/pc/stalker...ts.html?id=363

I can also play a new game, civ 4 @ 1900X1080 maxed out, 16X AA (in game setting) without any lag.

So in all honesty, its driver issue, plus problems the stalker devs will have to fix, since there as much to blame.

ehh.
highest MSAA setting I have is 8xQ

16x and 16xQ both look better to me. CSAA seems to be the way to go in CSS

and frame rate is 75 solid. I have vsynch enabled.

all I was saying is that showing CSS as an example isn't all that wise since it's now easily maxed out.

[eXceeded]

Get mine soon

[gdogg]

ehh.
highest MSAA setting I have is 8xQ

16x and 16xQ both look better to me. CSAA seems to be the way to go in CSS

and frame rate is 75 solid. I have vsynch enabled.

all I was saying is that showing CSS as an example isn't all that wise since it's now easily maxed out.

Ya, I know, it was just only game I had tried out at that point. Not trying to be arguementative.

I have only had card 2 days today.

[grimREEFER]

my fx5900 played cs source well. im not even joking. it also owned in doom 3. perhaps we should stop benching with these games lol. and yea, the 2900xt is pretty trashy, idk whats the problem with the article, did it come down too hard or not hard enough on the amd card?

[DilTech]

E6600@3GHZ
HD 2900XT
Windows vista ultimate

1680*1050

drivers 8.36
STALKER...

Sorry but I think that makes no sense releasing something like that. The review is very bad. So bad that even in their foruns there are a lot of people laughing. And that argument that there are no diferences to newer drivers doesn´t stick. Is too bad....

I'm calling you out on that one, the 8.38 images aren't really running full dynamic lighting....

http://img513.imageshack.us/img513/4...0280339dc7.jpg
http://img514.imageshack.us/img514/2...0443564pv4.jpg
Look behind the ladder...See the shadow? The 8.37 one is different because with Dynamic lighting comes dynamic shadows, the shadow in the 8.38 one is a static shadow, which is obvious by it's lack of definition.

Furthermore, to prove my point...

http://img481.imageshack.us/img481/9...0482999nu5.jpg

Take note here... Anyone who plays stalker knows the engine doesn't support AA with Dynamic Lighting. You can turn it on, but it does absolutely nothing. Turn off Full Dynamic lighting and then AA will work. This is because the game engine uses deffered rendering for dynamic lighting, which under the DX9 structure cannot use AA in any way, shape, or form(and the reason why UE3.0 games will only support AA under DX10). Now, notice that AA worked fine for those 8.38 shots?... Now you know why. Because the 8.38 shots aren't using full dynamic lighting.

You should look a bit more closely at images before bringing them here. It's clear as day that the 8.38 images aren't using the same lighting model as the 8.37 shots. If you'd like, I can go to those exact locations and post screenshots with dynamic lighting when I get off work to show you the difference.

Whoever you got those shots from just lost all credibility for their performance numbers btw.

[bluep3ace]

my 9600xt plays cs:s well.
(on low)

[theteamaqua]

sigh ...

u released a card 6 months later than competitor, and its not even faster than competitors 3rd best card ... there is a reason for the $399 price tag .. btw my 8800GTS 640mb costs $330 back in February

i was waiting for this card... but well we all know how well this card performs, and it will be stupid for me to spend $410 + 8% tax for something thats not even faster

[gdogg]

If the next gen nvidia cards are late, and perform worse . I completely except all of you who bash the hd2900xt to act the same.

If not, then you just proved what you are.

[Eastcoasthandle]

I'm calling you out on that one, the 8.38 images aren't really running full dynamic lighting....

http://img513.imageshack.us/img513/4...0280339dc7.jpg
http://img514.imageshack.us/img514/2...0443564pv4.jpg
Look behind the ladder...See the shadow? The 8.37 one is different because with Dynamic lighting comes dynamic shadows, the shadow in the 8.38 one is a static shadow, which is obvious by it's lack of definition.

Furthermore, to prove my point...

http://img481.imageshack.us/img481/9...0482999nu5.jpg

Take note here... Anyone who plays stalker knows the engine doesn't support AA with Dynamic Lighting. You can turn it on, but it does absolutely nothing. Turn off Full Dynamic lighting and then AA will work. This is because the game engine uses deffered rendering for dynamic lighting, which under the DX9 structure cannot use AA in any way, shape, or form(and the reason why UE3.0 games will only support AA under DX10). Now, notice that AA worked fine for those 8.38 shots?... Now you know why. Because the 8.38 shots aren't using full dynamic lighting.

You should look a bit more closely at images before bringing them here. It's clear as day that the 8.38 images aren't using the same lighting model as the 8.37 shots. If you'd like, I can go to those exact locations and post screenshots with dynamic lighting when I get off work to show you the difference.

Whoever you got those shots from just lost all credibility for their performance numbers btw.

You cannot determine what's running "full dynamic lighting" unless you have another pic to show it at the same position. And, you are basing your opinion on HDR and dynamic shadow from ss that are taken from different angles. Why on earth would you expect the exact same lighting and shadow from 2 different angles? Still drawing straws

[DilTech]

If the next gen nvidia cards are late, and perform worse . I completely except all of you who bash the hd2900xt to act the same.

If not, then you just proved what you are.

Then I'll do exactly what I did when NVidia launched the FX... I'll buy an ATi card if I'm looking to upgrade.

If you stay to one brand, you'll end up with an under performing part at some point. NVidia fans had that problem when the FX came out, ATi fans have it right now, AMD fans have it right now, and Intel fans had it when the A64 launched.

It happens to all companies at some point in time, that's totally out of the debate at this point. The real question is, how long before ATi releases a new part to make us forget the short-comings of the R600 like the 6800GT/Ultra did for NVidia climbing out of the FX hole.

[gdogg]

Then I'll do exactly what I did when NVidia launched the FX... I'll buy an ATi card if I'm looking to upgrade.

If you stay to one brand, you'll end up with an under performing part at some point. NVidia fans had that problem when the FX came out, ATi fans have it right now, AMD fans have it right now, and Intel fans had it when the A64 launched.

It happens to all companies at some point in time, that's totally out of the debate at this point. The real question is, how long before ATi releases a new part to make us forget the short-comings of the R600 like the 6800GT/Ultra did for NVidia climbing out of the FX hole.

I got the best card for what I do at its price, after being screwd by the evga 680i, I will never got nV again, period, its amd and intel for me now. Plus I want to fold on my gpu while im not game.

[DilTech]

You cannot determine what's running "full dynamic lighting" unless you have another pic to show it at the same position. And, you are basing your opinion on HDR and dynamic shadow from ss that are taken from different angles. Why on earth would you expect the exact same lighting and shadow from 2 different angles? Still drawing straws

Allow me to explain a bit more while you laugh...

You see the trees in the 8.37 screenshots? See how they cast shadows on the ground? See how they're DEFINED shadows on the ground in the backround? See how the characters sitting down cast shadows as well?

Take a look at 8.38


No shadows at all from the trees. Furthermore, even the people sitting around casting shadows. That's because there's no dynamic lighting... Even if it is from a different angle, they're not casting a shadow at all.

If you really want to continue to call me out on this, I'll go to a few locations and turn off and on full dynamic lighting and personally show you that's the difference in the 8.37 and 8.38 scenarios.

[gdogg]

and for such a late card, 850/2100 for stock voltages and cooling on a hd2900xt is pretty nice, and I havn't even started overclocking yet, cause ati tool won't work for me under vista x64.

[DilTech]

and for such a late card, 850/2100 for stock voltages and cooling on a hd2900xt is pretty nice, and I havn't even started overclocking yet, cause ati tool won't work for me under vista x64.

Back the memory down to 999, once you hit 1000 latencies raise and performance actually goes down.

[FghtinIrshNvrDi]

You know, Catalyst and 7.6 will probably fix this, and everyone knows that this card will do much better in DX10.

Ryan