NV40 3DMark 2003 scores revealed -theinquirer

[Evildeus]

Remember the composition of GT2 & 3.

I've forgotten... I need to have a look at your reviews of 3DMark then

 

[Guden Oden]

Wouldn't this most likely be a violation of the EULA??? after all the drivers haven't been certified for the NV40.

Good luck trying to enforce that EULA in court, with the right of free speech and whatnot to fight against...

 

[Arun]

Yes, I mean to each pipeline, 991060.
Ah yes, I forgot about the VS optimizations Tim - what are those again? I kind of forgot. I think it was doing VS work on the CPU - am I right? If that's the case, then I'd once again have to say the impact would be reduced since the VS is obviously faster now too... But good point anyway

Uttar

 

[SlmDnk]

All the same.

All I care about is whether Far Cry is going ro run smoothly @ 1600x1200 with high level AF and with all the bells and whistles set to max 8)

 

[Ante P]

I know 3dmark03 is video card limited on todays cards. But on tomorrows cards that might not be the case...

the highest scores today are at 10k+ and at those scores overclocking the CPU a few 100 Mhz still does little help compared to say 5 Mhz on the graphics core

I'd say when we start reacing 18-19k+ then we'll see CPUs doing more of a difference

 

[Quitch]

All the same.

All I care about is whether Far Cry is going ro run smoothly @ 1600x1200 with high level AF and with all the bells and whistles set to max 8)

I'd be far more worried about whether the next generation of games will run at those settings After all, the nv35 was more than capable of doing just that on existing games of the time.

 

[Mark]

I remember the exact same kinda stuff before the NV30 release where some websites claimed overly "enthusiastic" 3DMark scores. Now we have the same kinds of "reports" (likely from the same "source") for the NV40 and we're supposed to take them as fact? Again? I don't think so.

edit: Don't get me wrong. I'm sure it'll be hella fast and break 10k without breaking too much of a sweat, and I wouldn't be all that surprised if what the Inq said is true and that it'll even break 12k, but after the NV30 launch fiasco I'm still just a little jaded when it comes to pre-launch benchmark scores.

 

[g__day]

Do folk expect NV40 will be balanced or biased shader model 2.0 or 3.0 optimised if these are conflicting design goals?

Can you design a card that is equally balanced regardless of whether you are talking DX8, DX9 or shader model 3.0?

 

[Arun]

All I care about is whether Far Cry is going ro run smoothly @ 1600x1200 with high level AF and with all the bells and whistles set to max 8)

I can confirm that Far Cry @ 1280x1024, 4x AA runs slightly smoother on the NV40 than on the Radeon 9800XT with no AA. But hey, with 1200Mhz effective memory, what did you expect?

Uttar

 

[KimB]

Do folk expect NV40 will be balanced or biased shader model 2.0 or 3.0 optimised if these are conflicting design goals?

From what I've seen, PS 2.0 and 3.0 aren't conflicting at all, and so if it's fast at one, it should certainly be fast at the other.

Can you design a card that is equally balanced regardless of whether you are talking DX8, DX9 or shader model 3.0?

Remember that "DX8" level pixel shaders require much less precision than "DX9" level pixel shaders. This is not true between PS 2.0 and PS 3.0, which have essentially the same precision requirements.

Anyway, the #1 problem with the NV3x was floating-point shader performance (this directly led to poor PS 2.0 performance). If nVidia has fixed this problem, then the NV4x generation will be great. Other bonuses, of course, would be improved FSAA and at least an option to not use "brilinear" filtering.

 

[Miksu]

But hey, with 1200Mhz effective memory, what did you expect?

How about R420's clockspeeds?

 

[Tim Murray]

But hey, with 1200Mhz effective memory, what did you expect?

How about R420's clockspeeds?

Nobody knows. 1200Mhz is false, it's slightly lower than that. I'm guessing that R420 will have much faster memory than NV40, probably in the range of 1300-1400Mhz. So, they'll have a distinct advantage there. Core speeds, of course, are TOTALLY UNKNOWN.

 

[Moose]

Full Trilinear is not a "bonus" option for me. If it can't do full Trilinear then I don't want it period.

I'm sure the performance between this next round of cards will be pretty close so my buying decision will be decided by other factors like image quality. AA modes/quality etc.


Brilinear is not an option.

 

[Evildeus]

At least we will see some benchmark, for sure, on the 13th

 

[991060]

I'd like to see a patch from futuremark adding VS/PS 3.0 support, maybe that's what they do with 3dmark04?

 

[Dave Baumann]

At least we will see some benchmark, for sure, on the 13th

Nope.

 

[Dave Baumann]

Core speeds, of course, are TOTALLY UNKNOWN.

Are they? 8)

 

[Evildeus]

At least we will see some benchmark, for sure, on the 13th

Nope.

Are you sure? :? According to Marc from HFR we will surely.

 

[Miksu]

At least we will see some benchmark, for sure, on the 13th

Nope.

That's disappointing

 

[Tim Murray]

Core speeds, of course, are TOTALLY UNKNOWN.

Are they? 8)

Oh Dave. Have I mentioned that I hate you for your teasing?