4 Way SLI

WizZard

If nVidia enable this, wonder what will ATI do for the comeback?
Here (Chinese lang)

hairball

start selling chocolate bars instead??

po0k

Waste of time using 4 cards.
The GPU power might scale, but you're effectively wasting 75% of the ram you paid so much for.
Why they don't just release multi-core gpu packages with a unified mem controller is beyond me.

ATI are b0ned unless the physics effects acceleration they're pushing on the R5xx works out...which I think it just might - based on a few logical assumptions on performance trade-offs.

My x800xl runs Quake4 and BF2 acceptably. I could give a toss about 64x Deco-sample FSAA and 48-tap-gangbang Anistrophic filtering at 3200x1800 res.

WizZard

SyxPak wrote:

I could give a toss about 64x Deco-sample FSAA and 48-tap-gangbang Anistrophic filtering at 3200x1800 res.

Very, very few people will either. Until HD gaming/TV becomes more popular this side of the pond anyway.

gline

id love to see what kind of 3dmark scores they were getting with those 2 badboys. those cards cost about 800 each though, and optional external power source as far as i read about them before

po0k

3DMark is cack. The benchmark demos don't even look that good, and they've gradually removed more and more functionality from the free editions, cheap bastards.

A pure OpenGL benchmark is the only gauge.

gline

SyxPak wrote:

A pure OpenGL benchmark is the only gauge.

the only guage for open gl preformance yeh

po0k

3Dmark nor DirectX are open. In-game average performance running repeatable demos or open benchmarks are the two far superior performance comparison methods.

You cannot quantify subjective performance in a single rating with so many background factors that are hardly ever mentioned.

3DMarks are meaningless, I buy hardware to play games and do work comfortably, not to run endless benchmarks for fleeting bragging rights. I'll run timedemos to get a rough indicator of any positive or negative effect a given bit of fiddling has on my system.

mad m

SyxPak wrote:

3Dmark nor DirectX are open. In-game average performance running repeatable demos or open benchmarks are the two far superior performance comparison methods.

You cannot quantify subjective performance in a single rating with so many background factors that are hardly ever mentioned.

3DMarks are meaningless, I buy hardware to play games and do work comfortably, not to run endless benchmarks for fleeting bragging rights. I'll run timedemos to get a rough indicator of any positive or negative effect a given bit of fiddling has on my system.

Takes a couple of steps back,does the wild coyote look before making a run for the door!!!!

WizZard

Read about the next version of 3dMark here

3dMark is only good for benchmarking, and that's it. It gives no real indication of performance in real-world(tm) games

The Real B-man

if it plays the games without any lag & has acceptable graphics im happy