Can this be true....475 for the 6800Ultra core now?

I agree with your statement, Joe, to the extent that both ATi's and nVidia's offerings can run Doom 3 at 1600x1200 without dropping below the frame rate cap. Then it does become a question solely of IQ. If they can both do this at max AF and AA (they presumably won't be able to), then it is a question of whose AF and AA implementation is better.
 
Joe DeFuria said:
Chalnoth said:
The benchmarks won't be capped, and a faster card will always allow higher resolution/FSAA/AF gameplay.
Click to expand...

Actually, it does matter a bit. Because if it's capped at 60 FPS, then it becomes an issue of image quality, of which resolution is a part...as is AA.
Click to expand...

And usable resolution is of course depending on speed. But if both cards can run at 1600*1200, 4X FSAA and max AF at >= 60 fps, then we're in pure FSAA land when it comes to Doom 3. I doubt that that's the case though.

Edit: damn, to late :)
 
Joe DeFuria said:
Actually, it does matter a bit. Because if it's capped at 60 FPS, then it becomes an issue of image quality, of which resolution is a part...as is AA.
Click to expand...
I really do not see how this changes anything at all. People have always run at higher resolutions (monitor permitting) when the performance was good, as long as the game scaled well with resolution (i.e. the UI scaled properly).

Only if the game is playable at the highest resolutions and settings by some specific competing graphics cards will performance mean less. I doubt this will be the case with the NV40 and R420.
 
Zod said:
I agree with your statement, Joe, to the extent that both ATi's and nVidia's offerings can run Doom 3 at 1600x1200 without dropping below the frame rate cap. Then it does become a question solely of IQ. If they can both do this at max AF and AA (they presumably won't be able to), then it is a question of whose AF and AA implementation is better.
Click to expand...

I don't think you understand what I'm saying. My statement is not based on some prerequisite that Doom3 run at 16x12 on both cards without any AA. BTW...is anyone actually expecting that to be achieved? I know I'm not....but you never know.

What if NV40 runs at 60 FPS at 1280x960 with 2X FSAA?

What if R420 runs at 60 FPS at 1024x768 with 4X or 6X FSAA?

Which is better "quality?"
 
Joe DeFuria said:
What if NV40 runs at 60 FPS at 1280x960 with 2X FSAA?

What if R420 runs at 60 FPS at 1024x768 with 4X or 6X FSAA?

Which is better "quality?"
Click to expand...

I think that'll get you different answers depending on who you ask.
 
Joe DeFuria said:
Bjorn said:
Joe DeFuria said:
What if NV40 runs at 60 FPS at 1280x960 with 2X FSAA?

What if R420 runs at 60 FPS at 1024x768 with 4X or 6X FSAA?

Which is better "quality?"
Click to expand...

I think that'll get you different answers depending on who you ask.
Click to expand...

Well, that's exactly my point. ;)
Click to expand...

Then we raise the bar a bit :)

Can the NV40 run 1600*1200 and 2X FSAA at >= 60 fps ?

Can the R420 run 1280*1024 and 4X MSAA at >= 60 fps ?

If the answer is no on both then we just see who is the faster (let's say that the reviewer thinks that those settings are similar in quality f.e) at those settings and voila, we have a "winner".

Look, what Doomtrooper implied was that we're not going to be able to choose a winner in Doom3 because it's capped at 60 fps and say yes, we can, as long as they both can't run at max settings at 1600*1200.
 
Sounds like the Radeon 8500 vs. GeForce 3 then..

Sure there's differences, but you really can't go wrong with either.
 
dksuiko said:
Sure there's differences, but you really can't go wrong with either.
Click to expand...

Agree 100%. That, in a nut-shell, should be the overriding take-home point in this round....with the only likey wrinkle to be how the cards are priced.

Unlike R3xx vs. NV3x, where you actually could go wrong with NV3x IMO.
 
Joe DeFuria said:
Agree 100%. That, in a nut-shell, should be the overriding take-home point in this round....with the only likey wrinkle to be how the cards are priced.
Click to expand...

That is, of course, assuming the R420 and the NV40 are evenly matched. For all we know, the R420 could smash the NV40... or vice-versa. The fun in waiting...

(Someone remind me of the rumored announce date of the R420, I can't seem to find it.)
 
So it is back to what resolution you play at, and what goodies you want to turn on. A Radeon 9700 Pro should be more than enough for Doom 3 with 2X FSAA and 8XAF at 1024 x 768.

In reality almost all games comes down to IQ enhancing features that seperate each other. I personally would not break the bank for the ability to up the FSAA by one mode...but that is just me.
 
Joe DeFuria said:
dksuiko said:
Sure there's differences, but you really can't go wrong with either.
Click to expand...

Agree 100%. That, in a nut-shell, should be the overriding take-home point in this round....with the only likey wrinkle to be how the cards are priced.

Unlike R3xx vs. NV3x, where you actually could go wrong with NV3x IMO.
Click to expand...

The NV3X is definitely a black sheep in the NVidia family. And it's pretty obvious that Nvidia think so too since they got rid of the FX part for the NV4X.
 
dksuiko said:
(Someone remind me of the rumored announce date of the R420, I can't seem to find it.)
Click to expand...

There are two dates floating around:

April 26th and May 4th. Your guess is as good as mine as to which one relates to NDA's lifting and p/reviews appearing.
 
Bjorn said:
The NV3X is definitely a black sheep in the NVidia family. And it's pretty obvious that Nvidia think so too since they got rid of the FX part for the NV4X.
Click to expand...

Yeah, all we can do is feel sorry for the folks who decided to buy it. :(
 
I think it would be a nice change if and doom3 timedemos were capped at 60 frames. This would force reviews to eventually only look at the bottom 10% or 25%. The crappy slowdowns are the only thing that bother me during gameplay. The average isn't really indicative of how slow it runs, nor is the frame minimum for a given sample. It would be nice if we started seeing performance bins as a percentage of the total run instead of min, average and max type comparisons.
 
You must log in or register to reply here.
Back
Top