Official 7900 reviews thread

Considering NV's historic modus operandi, it's a pretty relevant inquiry. But maybe not for this particular forum.

Most posts here are speculative. To call Jawed's notion unfounded and irrelevant is a bit strong, especially in light of Xenos, R520, RV530, and RV515 for sale 3-4 months ago (and the fact that a purported library error caused a multi-month delay). Surely, given G71's die size and the fact that NV is selling what amounts to a 7800GTX at $300, NV would have preferred to move to 90m ASAP?

It doesn't really bother me, but I don't think it's a ridiculous notion. No need to scoff at it b/c you fear it's an attack on NV, which is what I'm sensing. NV's been pretty immune to reproach since the 6800, as you said.

 

I think it's simple. They had great margins on the G70, so they could afford to sit back and accumulate stock for a hard launch, rather than launch earlier, with worse supply and cannibalize their G70 sales as well. I don't think the G71 was 'delayed' in the way that the R520 was 'delayed' if that's what you're implying.

The G71 is more like this years 'R480'

 

Ati's problem wasn't a result of going to 90nm.

 

I found this techreport quote interesting :

"There is, however, one intriguing point that I'd like to note: the 7600 GT came out ahead in our timedemo tests and in 3DMark, but the X1800 GTO was faster in games we measured with FRAPS. I wish we'd had time to test with a broader range of games in order to establish whether this is a trend. We'll have to keep our eyes on that question."

Edit : Oops maybe this should be in the 7600 thread . But maybe the issue , if any at all may extend across the archietecture .

 

It's as usual funny to see how companies adapt depending on what they have on the market:

And Nvidias performance with HDR (FP16) on the NV40 was/is great then ?

Doesn't mean that i don't agree with Nvidia in a way (business wise). But it's not exactly an exciting launch for those who like additional features.

 

That's not what I heard. NVIDIA simply said that it's a hardware trade off and that at this time the trade off didn't make sense for them.

 

they don't call it tom's biasware just for fun

 

You know. I keep hearing this. But I have found Nvidias FP performance has for the "Most" part scaled with its fillrate. 6800GT SLI HDR is similar to a 7800GTX HDR and so on. While I agree that there have been improvements to the overall speed of HDR over the NV4x -- G70 --> G71. They havent been that dramatic either. I may test this later since I have some 6800GTS and can do some HDR comparisons with the 7800GTX @ 6800 clocks and pipeline setup. Btw this isnt an attack at your comment. Just my general curiosity of HDR scaling in relation to the 7800/6800 line.

 

I was only speculating on what Anandtech reported about what NVidia said.. obviously this part wasn't meant to have any significant changes... it's primarily a tweak of the existing architecture... The only thing I'm aware of is that they will allow MSAA with their FP16 framebuffer (? or whatever) in their next gen hardwar.. but as to anything else.. that's still up in the air.

 

lol

Isn't that supposed to be 287Million?



US

 

Why is trilinear being used(and again in HL2 settings, haven't checked the others yet) over anisotropic filtering? Is it just a preference of the reviewer? Isn't trilinear filtering worse than anisotropic?

US

 

So more or less Nvidia kicks butt today. 7900GT and 7600GT own their price classes, with extremely small die sizes.

However, the 7900GTX is not the clear performance leader. In newer Direct 3D games X1900 will continue to pull away.

BTW, it's clear to me the internet press is VERY pro-Nvidia, so I wouldn't expect fair criticism of them from most qaurters. The utter lack of criticism of them over the non-availability of the 512 GTX should tell you that.

 

They're not using one over another at all - the settings box shot is simply there to show what graphics details are being used. Other than that, they're using trilinear filtering for the 0xAF results and then trilinear-anisotropic filtering for the 8xAF results; same as we do it here.

 

If not for the power issues .. the 1900XTX/XT holds it's own very nicely indeed against the G71, even beating it in some tests(non OGL ok). I was expecting the G71 to romp the X1900 .. but it just ain't so.

US

 

From a business standpoint, I think what they`re doing is impressive. Does not any business seek to attain maximum outcome with minimum effort? That`s what NV have been doing. I imagine they`ve been constantly spending less in the past few years since the NV40s introduction in terms of R&D, fully exploiting what seems to have been a very very solid architecture. And I imagine it`s cheaper to produce a G71 rather then a R580, which means:
a)if they sell at the same pricepoint with ATI they`ll make more money;
b)they can afford pushing their prices lower and still be able to have at worst the same ammount of profit per card as ATI.
I can respect the way they`ve been doing business(please no sad sad stories of presentations, misleading the public, driver cheats etc. as I`m not talking about that (sadly necessary) aspect of their business conduit). For an enthusiast drooling over scientifically named features, this is a "move-along, nothing to see here card". But those enthusiasts account for quite a small percentage of the target user-base, so ignoring them may not be so bad

 

Is it irrelevent? Was 7800 GTX 512MB really supposed to be the product that lasted between November and now?