How is CrossfireX vs SLi scaling these days?

[Grooveriding]

There is a thing called a ignore post feature, if you dont like my posts,points,context ,use it. I'm not arguing with you in every thread, so you can personally attack me with nothing to back it up but your word.

I don't consider pointing out that you make multiple needless threads in this forum daily as personal. You do, and they are pointless. As free as you are to make them, so am I to say they're of no point.

If you're referring to the focus-group jab, it was that, a jab, but really one aimed at what is the point of the endless nvidia threads you constantly create, most of little to no merit.

I'll continue to point out inaccuracies where I see them. You were proven wrong in this thread multiple times, and as usual made assertions based on nothing concrete.

Please leave the focus group comments out of discussions. Thank you, Admin DrPizza

 

[happy medium]

"This message is hidden because Grooveriding is on your ignore list."

See it works.

And making posts like this in the technical forums just to antagonize people isn't allowed. -Admin DrPizza

 

[Wreckage]

I just got that monitor. It's fantastic!

That's what I like to hear! :thumbsup:

Yeah, I'm getting it. Now my 260 will really have to work.

 

[evolucion8]

I think that scaling difference between both isn't that big, is just that Fermi's scaling is more consistent and predictable, here for example, scaling with Crossfire is almost 100%

http://www.xbitlabs.com/articles/video/display/starcraft2-wings-of-liberty_6.html#sect1

 

[happy medium]

I think that scaling difference between both isn't that big, is just that Fermi's scaling is more consistent and predictable, here for example, scaling with Crossfire is almost 100%

http://www.xbitlabs.com/articles/video/display/starcraft2-wings-of-liberty_6.html#sect1

I think thats the wrong link bud.

 

[evolucion8]

I think thats the wrong link bud.

It worked for me, compare the HD 5970 performance vs the HD 5850 performance and see what I mean.

 

[happy medium]

It worked for me, compare the HD 5970 performance vs the HD 5850 performance and see what I mean.

ok, now I see what you mean. Thats kinda tough to do ,I thought the 5970 was faster then the crossfired 5850's? 3200 sp's vs 2880 and the same clocks right?

 

[evolucion8]

ok, now I see what you mean. Thats kinda tough to do ,I thought the 5970 was faster then the crossfired 5850's? 3200 sp's vs 2880 and the same clocks right?

They're about the same speed with the edge going to the HD 5850

 

[happy medium]

Learn something new everyday , looks like they need to work on the 5970's drivers. It should be faster.

 

[Wreckage]

Learn something new everyday , looks like they need to work on the 5970's drivers. It should be faster.

It's getting knocked around by a pair of old 285's.

 

[evolucion8]

It's getting knocked around by a pair of old 285's.

Yeah, and considering that the GTX 480 and the GTX 460 can't come close to the HD 5970, the GTX 285 knocked them down too!!

But back to reality, with recent driver optimizations, look!!

Suddenly the GTX 285 looked more attractive than the nuclear GTX 480. The poor GTX 260 SLI can't even outperform the GTX 285 SLI convincingly. I couldn't find newer benchmarks using Resident Evil 5, seems that it isn't demanding enough, so I just wiped out the smile of your face again Wreck, back to reality...

 

[happy medium]

I like these even better.
Love them 5850 and 5870 crossfire minimums! A single gtx 275 beats them.

The gtx 460 1gb in sli for 420$ with 99 percent of the performance of the $880 5870 crossfire setup

And remember thats stock gtx 460 performance. You can get 2 gtx 460 at 800 core for 420$ shipped. You have been drinking to much of that Puerto Rican rum!
They will woop the old ass 5870 crossfire setup!

all in good fun Please no fighting guys we are just kidding around.
At least we are on topic? HA HA

 

[evolucion8]

I like these even better.
Love them 5850 and 5870 crossfire minimums! A single gtx 275 beats them.

The gtx 460 1gb in sli for 420$ with 99 percent of the performance of the $880 5870 crossfire setup

And remember thats stock gtx 460 performance. You can get 2 gtx 460 at 800 core for 420$ shipped. You have been drinking to much of that Puerto Rican rum!
They will woop the old ass 5870 crossfire setup!

all in good fun Please no fighting guys we are just kidding around.
At least we are on topic? HA HA

What a troll you had become lately happy medium, at such high resolution of 2560x1600, is irrelevant because besides of the GTX 480/470 SLI, nothing there is playable, they're VRAM limited obviously, if in other games and other scenarions the GTX 460 SLI trails behind the HD 5870 CF, why it suddenly became faster at 2560x1600

Now let me use your mentality and sensationalist trolling; "But at 1920x1080, heck, an HD 4870X2 is outperforming even the GTX 470 in min frame rate and is nipping the heels of the GTX 480 for half the price!!!, it whoop the ass of nVidia's brand new architecture!! You must be on soviet crack!!!"

PS: I'm a guy who goes to a gym, so I don't drink,

 

[NoQuarter]

Dunno why everyone jumps on the VRAM when it's very easy to check by comparing the 5870 E6 2GB to 5870 1GB:

http://www.anandtech.com/show/3621/amds-radeon-hd-5870-eyefinity-6-edition-reviewed/6

These video cards are not really limited by their 1GB of VRAM in most games even at 2560x1600 4xAA. They probably have slightly higher minimum framerates during the rare texture loads though. But minimum frame rates get overanalyzed without a time graph to show the breakdown.

 

[PingviN]

I like these even better.
Love them 5850 and 5870 crossfire minimums! A single gtx 275 beats them.

Without a graph displaying fps over time, minimum fps tells you nothing. All it shows is that, at some point, fps dropped. It doesn't say when, it doesn't say for how long and it doesn't say how often it happens. A single drop to 10fps would be preferable over 50 drops down to 20, wouldn't you agree?

 

[Voo]

They probably have slightly higher minimum framerates during the rare texture loads though. But minimum frame rates get overanalyzed without a time graph to show the breakdown.

I agree that minimum frame rates without a time graph are mostly useless, but on the other hand VRAM limitations should mostly be seen exactly there - the average frame rates shouldn't suffer that much (as long as we don't have to swap every other frame, but it's not that bad with 1gb)

But yeah we really need more reviews with time graphs - just posting minimums is great for the fanboys who just search for the benchmark where there are the biggest differences, but for people who're interested in the overall picture that's rather useless..

 

[happy medium]

What a troll you had become lately happy medium, at such high resolution of 2560x1600, is irrelevant because besides of the GTX 480/470 SLI, nothing there is playable, they're VRAM limited obviously, if in other games and other scenarions the GTX 460 SLI trails behind the HD 5870 CF, why it suddenly became faster at 2560x1600

Now let me use your mentality and sensationalist trolling; "But at 1920x1080, heck, an HD 4870X2 is outperforming even the GTX 470 in min frame rate and is nipping the heels of the GTX 480 for half the price!!!, it whoop the ass of nVidia's brand new architecture!! You must be on soviet crack!!!"

PS: I'm a guy who goes to a gym, so I don't drink,

Hey man , I was just jokin with ya. Jeeez I thought I made that clear. No need to call me a troll.

 

[Grooveriding]

Dunno why everyone jumps on the VRAM when it's very easy to check by comparing the 5870 E6 2GB to 5870 1GB:

http://www.anandtech.com/show/3621/amds-radeon-hd-5870-eyefinity-6-edition-reviewed/6


These video cards are not really limited by their 1GB of VRAM in most games even at 2560x1600 4xAA. They probably have slightly higher minimum framerates during the rare texture loads though. But minimum frame rates get overanalyzed without a time graph to show the breakdown.

The caveat I would give to your post, is that AT tests Crysis always on Gamer setting with Enthusiast shaders.

If you put everything on Enthusiast and turn on 4xAA at high resolutions it changes the vram demands of the game and 1GB becomes insufficient. I ran into this situation with my own system.

I've seen some reviews of these settings showing the same results with 1GB cards, will try to dig up the websites with them.

A 2GB 5870 would be able to deliver under these conditions though, whereas a 1GB would not, it certainly has the gpu power, just needs the extra VRAM.

 

[Voo]

The caveat I would give to your post, is that AT tests Crysis always on Gamer setting with Enthusiast shaders.

If you put everything on Enthusiast and turn on 4xAA at high resolutions it changes the vram demands of the game and 1GB becomes insufficient. I ran into this situation with my own system.

The problem is: How does the VRAM limitation show itself? Low dips when it has to load some texture. Obviously as long as that doesn't happen often enough, the average won't be that strongly affected, but you'll get lots of lows where the game will lag, which makes it unenjoyable effectly although you may very well get average fps in the 30s.

That's why a timegraph is so important to really see how a card performs for a specific game, all those averages and minimums just don't show the whole picture

 

[Grooveriding]

The problem is: How does the VRAM limitation show itself? Low dips when it has to load some texture. Obviously as long as that doesn't happen often enough, the average won't be that strongly affected, but you'll get lots of lows where the game will lag, which makes it unenjoyable effectly although you may very well get average fps in the 30s.

That's why a timegraph is so important to really see how a card performs for a specific game, all those averages and minimums just don't show the whole picture

I can't offer that data, I no longer own my 5870s. For me the issue manifested its self with a lot of hitching when panning my view around in the game environment.

I agree with the comment on minimums, it could just be a hard drive burp, which has given me ridiculously low minimums when I've benched my system before.

 

[happy medium]

That's why a timegraph is so important to really see how a card performs for a specific game, all those averages and minimums just don't show the whole picture

I totally agree, some sites do use them,they are little hard to read, but do paint a better picture of actual performance.