[ExpRev] G94 VS G92: NVIDIA SLI performance scaling test - Overclock.net

i_ame_killer_2 · 03-25-08

Quote:

Back to Nov 2007, we managed to run a HD 3870 CF VS 8800GT SLI test. At those time our conclusion is:

“In more than 90% items, Crossfire scales more better than SLI. In 60% test items CrossFire scale up more than 50% performance compare to single card. But in SLI’s side, it can only scale 40% items up to 50%.

More surprisingly, in 30% items there are more than 80% scale in Crossfire. But we can hardly find a 70% scale in SLI solutions. In one word to sum this up: ATi’s Crossfire is more efficient than SLI.”

And now it is 2008, after more than 4 months’ hardware development, 9600GT shows up, and its SLI scaling ratio is more better than 8800GT SLI ratio in our old article. Also, both AMD and NVIDIA have released lots of drivers, we think we should run a SLI scaling test.

So in this article we mainly focus on this questions:

　　●With the new Forceware, will 8800GT SLI scales better than it used to be?

　　● 9600 GT SLI scales better than ever? How good it is?

　　● 174.16 WHQL help 9600 GT SLI scales better?

　　● 8800GT SLI and 9600GT SLI, which have a better price/performance ratio?

Source

Pheatton · 03-25-08

Wow learn to speak engrish imo. Interesting anyway.

i_ame_killer_2 · 03-25-08

Quote:

Originally Posted by Pheatton

Wow learn to speak engrish imo. Interesting anyway.

The site is Chinese and they just use like google translate to translate it to English.

Here is the original site in Chinese if you want: http://www.expreview.com/

Ihatethedukes · 03-25-08

LMAO, it doesn't say they scale better at all. They have no single card control runs to compare to. And if those are calculated versus the single cards, then... who effing cares, we only care which set up is faster, who cares if the 9600's scale better if they are slower than the 8800s anyhow.

mtbiker033 · 03-25-08

nice article thanks! +1

I want a pair of the 9600GT now bad, cheap, effective performance.

Sylence · 03-25-08

Quote:

Originally Posted by Ihatethedukes

LMAO, it doesn't say they scale better at all. They have no single card control runs to compare to. And if those are calculated versus the single cards, then... who effing cares, we only care which set up is faster, who cares if the 9600's scale better if they are slower than the 8800s anyhow.

Because if 9600's scale better in SLI then it's possible that SLI'd 9600's might be better then SLI'd 8800's...might

.

trueg50 · 03-25-08

Quote:

Originally Posted by Sylence

Because if 9600's scale better in SLI then it's possible that SLI'd 9600's might be better then SLI'd 8800's...might

.

In "theory" it "might" is the key here.

+50-80% performance boost for 100% more cost? Not a good idea.

Just pick-up an 8800GTS 512 for now, or just an 8800GT / 9600GT, then pickup another card later down the road.

RussianGrimmReaper · 03-25-08

Quote:

Originally Posted by Sylence

Because if 9600's scale better in SLI then it's possible that SLI'd 9600's might be better then SLI'd 8800's...might

.

Theoretically they could be, but they aren't.

scottb75 · 03-25-08

Would this mean that if Nvidia stuck 112sps (instead of 64) in the 9600GT, that the 9600s in SLI would on average be 10% faster then the 8800GT in SLI?

RussianGrimmReaper · 03-25-08

Quote:

Originally Posted by scottb75

Would this mean that if Nvidia stuck 112sps (instead of 64) in the 9600GT, that the 9600s in SLI would on average be 10% faster then the 8800GT in SLI?

Most likely.

		Overclock.net - Overclocking.net > Industry News > Hardware News
[ExpRev] G94 VS G92: NVIDIA SLI performance scaling test

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)