***Official*** NV40 Benches (Updated as they go live) ANANDTECH Review Added

[JBT]

I think I will wait for some different reviews these ones are interesting to say the least. Even so I bet with a new driver it will play a bit better.

 

[apoppin]

Originally posted by: flexy
uhhh..apopping.....gimme a break with "ATI unlocking the secret hidden pipes for our pleasure"

We all know these pipes were there from the start and (IMHO) there is not ONE logical reason why they (OTHERWISE) should have kept 4 pipes secretley hidden and locked. Just doesnt make any sense to me...

ATI's R420 Increases Number of Pipelines
Argue with THEM, not me.

 

[Dug]

Originally posted by: NFS4

Originally posted by: mchammer187
anyone know if Anand will have benches @ 12:00 am?

looks in NFS4's direction

nueve

AM or PM?
13th or 14th
2004 or 2005

 

[DeeSlanger]

Originally posted by: MemberSince97
under NDA till 4/14 09:00 E.S.T.

 

[RussianSensation]

Ok guys I decided to bring a different view of this "supposed review"

What I did was look at each screenshot taken and simply compared the framerate in the left corner as it is without analyzing the graphs and here are the results:

Bear with me as I did not include the resolution, but you already know NV40 ran at higher resolution at all times.

1. Call of Duty
9800xt 4AA/16AF = 112 FPS
5950U 4AA/8AF = 107FPS
6800Ultra 4AA/16AF = 214FPS or 91% faster than the fastest

2. Battlefield: Vietnam
9800xt 4AA/8AF = 80FPS
5950U 4AA/8AF = 70FPS
6800Ultra 4AA/8AF = 93FPS or 16% faster

3. Splinter Cell
9800xt 0AA/8AF = 35FPS
5950U 0AA/8AF = 38FPS
6800Ultra 0AA/8AF = 63FPS or 66% faster

4. Far Cry
9800xt 4AA/8AF = 33FPS
5950U 4AA/8AF = 22FPS
6800U 4AA/8AF = 46FPS or 39% faster

5. UT2004
9800xt 4AA/8AF = 40FPS
5950U 4AA/8AF = 29FPS
6800U 4AA/8AF = 53FPS or 33% faster

Well I realize that taking a screenshot at one instance in time is less indicative than an average but it still gives you an indication how fast the cards were running at each exact (as much as possible) location. These results are not that bad now that I think about it considering either the card ran at higher resolution while still performing faster or it has higher visuals enabled, or both. Of course these are not 100% improvements but then again this review is not comparing the cards at equal settings.

What I do begin to realize is that it might take a while before we can run these new powerful games like far cry, doom 3 and hl2 at 1600x1200 with 4AA/8AF enabled at smooth frame rates. hmm......

 

[ViRGE]

Heh, I love the amount of unsubstantiated rumors and "information" in this thread. Just wait until tomorrow, boys, you'd be surprised what's right and what's wrong.

PS If the HardOCP numbers are right, I know someone who isn't getting an NV45 later this year

 

[UlricT]

uhhh?? ViRGE... where did you get the NV45 specs from anyway?

 

[GTaudiophile]

Originally posted by: ViRGE
Heh, I love the amount of unsubstantiated rumors and "information" in this thread. Just wait until tomorrow, boys, you'd be surprised what's right and what's wrong.

PS If the HardOCP numbers are right, I know someone who isn't getting an NV45 later this year

Can you blame kyle? Maybe he was just h4x0red like Gabe?

 

[ViRGE]

UlricT: the NV45 doesn't even exist yet(case in point: even Xbit doesn't have some of their "theoretical" specs for it), so how could I have the specs for it?

GT: Nvidia can be as vindictive as hell; they may very well go after Kyle if he left himself open

 

[UlricT]

This is why I asked about NV45...

Originally posted by: ViRGE
Heh, I love the amount of unsubstantiated rumors and "information" in this thread. Just wait until tomorrow, boys, you'd be surprised what's right and what's wrong.

--> PS If the HardOCP numbers are right, I know someone who isn't getting an NV45 later this year

 

[Dean]

If those benchmarks are displaying an overall fair picture of the 6800ultra, its going to be in tough against even the 12 pipeline version of the r420.

I'm guessing here, but I bet the NV45 is actually an 8x2 architecture and not 16x1 like what was rumored.

 

[GTaudiophile]

Originally posted by: Dean
If those benchmarks are displaying an overall fair picture of the 6800ultra, its going to be in tough against even the 12 pipeline version of the r420.

I'm guessing here, but I bet the NV45 is actually an 8x2 architecture and not 16x1 like what was rumored.

Dave's 25+ page review at B3D will clear that up...

 

[Dean]

Originally posted by: GTaudiophile

Originally posted by: Dean
If those benchmarks are displaying an overall fair picture of the 6800ultra, its going to be in tough against even the 12 pipeline version of the r420.

I'm guessing here, but I bet the NV45 is actually an 8x2 architecture and not 16x1 like what was rumored.

Dave's 25+ page review at B3D will clear that up...

Yeah Wavey will get to the bottom of it. I still think the NV45 has more juice under its belt even though it does seperate itself from current cards fairly well now. The speed i was fairly impressed by(just not overwhelmed), the rendering quality is still a notch behind ATI's current offering though.

 

[UlricT]

gahhh.... whats with all the NV45 posts? Its NV40 ppl!!!

 

[Corporate Thug]

Originally posted by: RussianSensation
Ok guys I decided to bring a different view of this "supposed review"

What I did was look at each screenshot taken and simply compared the framerate in the left corner as it is without analyzing the graphs and here are the results:

Bear with me as I did not include the resolution, but you already know NV40 ran at higher resolution at all times.

1. Call of Duty
9800xt 4AA/16AF = 112 FPS
5950U 4AA/8AF = 107FPS
6800Ultra 4AA/16AF = 214FPS or 91% faster than the fastest

2. Battlefield: Vietnam
9800xt 4AA/8AF = 80FPS
5950U 4AA/8AF = 70FPS
6800Ultra 4AA/8AF = 93FPS or 16% faster

3. Splinter Cell
9800xt 0AA/8AF = 35FPS
5950U 0AA/8AF = 38FPS
6800Ultra 0AA/8AF = 63FPS or 66% faster

4. Far Cry
9800xt 4AA/8AF = 33FPS
5950U 4AA/8AF = 22FPS
6800U 4AA/8AF = 46FPS or 39% faster

5. UT2004
9800xt 4AA/8AF = 40FPS
5950U 4AA/8AF = 29FPS
6800U 4AA/8AF = 53FPS or 33% faster

Well I realize that taking a screenshot at one instance in time is less indicative than an average but it still gives you an indication how fast the cards were running at each exact (as much as possible) location. These results are not that bad now that I think about it considering either the card ran at higher resolution while still performing faster or it has higher visuals enabled, or both. Of course these are not 100% improvements but then again this review is not comparing the cards at equal settings.

What I do begin to realize is that it might take a while before we can run these new powerful games like far cry, doom 3 and hl2 at 1600x1200 with 4AA/8AF enabled at smooth frame rates. hmm......

if we omit the COD results, 6800U is about 39% faster than the fastest thing available today. good, not great.

 

[reever]

Originally posted by: UlricT
gahhh.... whats with all the NV45 posts? Its NV40 ppl!!!

We don't know that yet. This might be the real NV40 and the NV45 will be an actual refresh, but then again this might be the NV45 rebadged as NV40 and the refresh NV45 is just going to have higher clocks, pci-E and nothing else

 

[UlricT]

Originally posted by: reever

Originally posted by: UlricT
gahhh.... whats with all the NV45 posts? Its NV40 ppl!!!

We don't know that yet. This might be the real NV40 and the NV45 will be an actual refresh, but then again this might be the NV45 rebadged as NV40 and the refresh NV45 is just going to have higher clocks, pci-E and nothing else

Really!!!??? OK!

 

[AIWGuru]

Originally posted by: shady06

Originally posted by: RussianSensation
Ok guys I decided to bring a different view of this "supposed review"

What I did was look at each screenshot taken and simply compared the framerate in the left corner as it is without analyzing the graphs and here are the results:

Bear with me as I did not include the resolution, but you already know NV40 ran at higher resolution at all times.

1. Call of Duty
9800xt 4AA/16AF = 112 FPS
5950U 4AA/8AF = 107FPS
6800Ultra 4AA/16AF = 214FPS or 91% faster than the fastest

2. Battlefield: Vietnam
9800xt 4AA/8AF = 80FPS
5950U 4AA/8AF = 70FPS
6800Ultra 4AA/8AF = 93FPS or 16% faster

3. Splinter Cell
9800xt 0AA/8AF = 35FPS
5950U 0AA/8AF = 38FPS
6800Ultra 0AA/8AF = 63FPS or 66% faster

4. Far Cry
9800xt 4AA/8AF = 33FPS
5950U 4AA/8AF = 22FPS
6800U 4AA/8AF = 46FPS or 39% faster

5. UT2004
9800xt 4AA/8AF = 40FPS
5950U 4AA/8AF = 29FPS
6800U 4AA/8AF = 53FPS or 33% faster

Well I realize that taking a screenshot at one instance in time is less indicative than an average but it still gives you an indication how fast the cards were running at each exact (as much as possible) location. These results are not that bad now that I think about it considering either the card ran at higher resolution while still performing faster or it has higher visuals enabled, or both. Of course these are not 100% improvements but then again this review is not comparing the cards at equal settings.

What I do begin to realize is that it might take a while before we can run these new powerful games like far cry, doom 3 and hl2 at 1600x1200 with 4AA/8AF enabled at smooth frame rates. hmm......

if we omit the COD results, 6800U is about 39% faster than the fastest thing available today. good, not great.

Yes, let's remove the highest benchmark, ignore the higher settings, and call it 40%

Or we could leave in the highest benchmark, STILL ignore the higher settings, and call it 50%

 

[Corporate Thug]

Originally posted by: AIWGuru

Originally posted by: shady06

Originally posted by: RussianSensation
Ok guys I decided to bring a different view of this "supposed review"

What I did was look at each screenshot taken and simply compared the framerate in the left corner as it is without analyzing the graphs and here are the results:

Bear with me as I did not include the resolution, but you already know NV40 ran at higher resolution at all times.

1. Call of Duty
9800xt 4AA/16AF = 112 FPS
5950U 4AA/8AF = 107FPS
6800Ultra 4AA/16AF = 214FPS or 91% faster than the fastest

2. Battlefield: Vietnam
9800xt 4AA/8AF = 80FPS
5950U 4AA/8AF = 70FPS
6800Ultra 4AA/8AF = 93FPS or 16% faster

3. Splinter Cell
9800xt 0AA/8AF = 35FPS
5950U 0AA/8AF = 38FPS
6800Ultra 0AA/8AF = 63FPS or 66% faster

4. Far Cry
9800xt 4AA/8AF = 33FPS
5950U 4AA/8AF = 22FPS
6800U 4AA/8AF = 46FPS or 39% faster

5. UT2004
9800xt 4AA/8AF = 40FPS
5950U 4AA/8AF = 29FPS
6800U 4AA/8AF = 53FPS or 33% faster

Well I realize that taking a screenshot at one instance in time is less indicative than an average but it still gives you an indication how fast the cards were running at each exact (as much as possible) location. These results are not that bad now that I think about it considering either the card ran at higher resolution while still performing faster or it has higher visuals enabled, or both. Of course these are not 100% improvements but then again this review is not comparing the cards at equal settings.

What I do begin to realize is that it might take a while before we can run these new powerful games like far cry, doom 3 and hl2 at 1600x1200 with 4AA/8AF enabled at smooth frame rates. hmm......

if we omit the COD results, 6800U is about 39% faster than the fastest thing available today. good, not great.

Yes, let's remove the highest benchmark, ignore the higher settings, and call it 40%

Or we could leave in the highest benchmark, STILL ignore the higher settings, and call it 50%

all the comparisons with the 9800XT use the same settings

in such a small smaple the highest (and for that matter the lowest) will sckew the overall representation, intro statistics

 

[AIWGuru]

Originally posted by: shady06

Originally posted by: AIWGuru

Originally posted by: shady06

Originally posted by: RussianSensation
Ok guys I decided to bring a different view of this "supposed review"

What I did was look at each screenshot taken and simply compared the framerate in the left corner as it is without analyzing the graphs and here are the results:

Bear with me as I did not include the resolution, but you already know NV40 ran at higher resolution at all times.

1. Call of Duty
9800xt 4AA/16AF = 112 FPS
5950U 4AA/8AF = 107FPS
6800Ultra 4AA/16AF = 214FPS or 91% faster than the fastest

2. Battlefield: Vietnam
9800xt 4AA/8AF = 80FPS
5950U 4AA/8AF = 70FPS
6800Ultra 4AA/8AF = 93FPS or 16% faster

3. Splinter Cell
9800xt 0AA/8AF = 35FPS
5950U 0AA/8AF = 38FPS
6800Ultra 0AA/8AF = 63FPS or 66% faster

4. Far Cry
9800xt 4AA/8AF = 33FPS
5950U 4AA/8AF = 22FPS
6800U 4AA/8AF = 46FPS or 39% faster

5. UT2004
9800xt 4AA/8AF = 40FPS
5950U 4AA/8AF = 29FPS
6800U 4AA/8AF = 53FPS or 33% faster

Well I realize that taking a screenshot at one instance in time is less indicative than an average but it still gives you an indication how fast the cards were running at each exact (as much as possible) location. These results are not that bad now that I think about it considering either the card ran at higher resolution while still performing faster or it has higher visuals enabled, or both. Of course these are not 100% improvements but then again this review is not comparing the cards at equal settings.

What I do begin to realize is that it might take a while before we can run these new powerful games like far cry, doom 3 and hl2 at 1600x1200 with 4AA/8AF enabled at smooth frame rates. hmm......

if we omit the COD results, 6800U is about 39% faster than the fastest thing available today. good, not great.

Yes, let's remove the highest benchmark, ignore the higher settings, and call it 40%

Or we could leave in the highest benchmark, STILL ignore the higher settings, and call it 50%

all the comparisons with the 9800XT use the same settings

in such a small smaple the highest (and for that matter the lowest) will sckew the overall representation, intro statistics

No they don't. He omitted the resolution settings. They run at different resolutions.
So...50% with skewed settings. I wonder what it will be with fair settings. We'll see tomorrow.

 

[Corporate Thug]

atleast UT is the same res...looks to be the only one

 

[flexy]

just as a sidenote (found time to gu thru all the pics)....i am convinced the pics/specs are genuine.
From that point of view i conclude the NDA as irrelevant...these are definetly genuine and real looking leaked pics/specs.

Also liked the reference to dx9.0c which is supposed to come out in summer.

And as some people stated/"calculated".....well 33% to 50% over 9800XT seems reasonable ! (My guess)

I just realy wisg the R420 would have shader 3.0 support. It's almost a given it wont (which, in real lfe, is irrelevant)....but i have to admit the features of the NV40 are impressive. HOWEVER, in raw FPS the R420 might be able to beat it. (Almost certain)


Edit: RGMS <----- good or bad ???? (Sorry, too lazy to look up now)

 

[Chad]

::resisting temptation to mock nVidia::

 

[flexy]

Originally posted by: Chad
::resisting temptation to mock nVidia::

Hey, i like ATI....but i dont think it would be adequate to 'mock' NV there. This card is still very fast and looks nice feature wise. (Ok, i consider me baned from rage3d now But - let's wait how R420 will look and we can start mocking

Edit: FPS is not anything...i am very looking f/w to the image quality comparisons, some shots there looked weird...subjective impression still that Radeon shots look 'smoother'...too early to tell.....also what perf. impact certain AA settings and trilinear-optimization will have...thats all important to know

 

[warcrow]

Originally posted by: Luthien
OH, and if you own Nvidea stock you better go sell them NOW!

OMGIGOD hi Margret!