The Fermi Thread - Part 3

[Gilgamesh]

Not a bad showing in my opinion... question remains though, what driver version were those 5870 #'s taken with?

http://futuremark.yougamers.com/foru...70&postcount=1

"They used the 10.2 ATi cats for their comparison by the way."

[GoldenTiger]

[*img]http://img228.imageshack.us/img228/3264/480order.jpg[/img]

Nice... I opted for 2-day air, myself . Glad I sold my 5870 when I did... got a good amount for it !

[GoldenTiger]

http://futuremark.yougamers.com/foru...70&postcount=1

"They used the 10.2 ATi cats for their comparison by the way."

10.2's I have seen on forums claim are around 3%-5% slower average than 10.3's... so it's pretty accurate still on the 5870's performance I'd say.

[eric66]

Now write that down so when real non-biased reviews come out we will know how credible nvidia is with providing factual information on benchmarks.

well they probably tried there best to show how powerfull is gtx 480 and seems the reason benching of far cry 2 hundred times is clear as i suspected its clearly favoring nvidia

[annihilat0r]

well they probably tried there best to show how powerfull is gtx 480 and seems the reason benching of far cry 2 hundred times is clear as i suspected its clearly favoring nvidia

yeah, those evil people, always trying to put their cards on a good light!!!!!11

[eric66]

yeah, those evil people, always trying to put their cards on a good light!!!!!11

you are trying too hard to be funny you know that ?

[eric66]

http://futuremark.yougamers.com/foru...70&postcount=1

"They used the 10.2 ATi cats for their comparison by the way."

well major improvements in cat 10.3 are 20 percent in hawx and 30 in dirt 2 they also promised some AvsP improvements on upcoming drivers as far as i remember so add those and you'll get better pic

[GoldenTiger]

well major improvements in cat 10.3 are 20 percent in hawx and 30 in dirt 2 they also promised some AvsP improvements on upcoming drivers as far as i remember so add those and you'll get better pic

Incorrect according to even hardocp's tests:

http://www.hardocp.com/article/2010/...er_performance

They saw a moderate improvement (12% across min/max/avg) in Dirt 2 and negligible in everything else they tested.

[annihilat0r]

Fermi is much more likely to add on performance over time (with drivers) IMHO.

[pentium777]

From Nvidia:

My eyes opened wide as I looked at the SLI scaling of BFBC2 amongst other titles, 2x480 in SLI looks good to me.

84% increase in BFBC2..... NICE....

[eric66]

Incorrect according to even hardocp's tests:

http://www.hardocp.com/article/2010/...er_performance

They saw a minor improvement in Dirt 2 and negligible in everything else they tested.

dirt improvement is different on various res and there isn't hawx test as far as i can see

ps : funny thing is that hardocp has same fps with cat 10.2 so if you add those improvement 5870 will match up 480 performance

[pentium777]

Wow, looks like I won't be playing Metro 2033 for quite a while. Damn this 30" monitor.

I got three letters for you, S-L-I... 3008WFP here and I had 3007WFP before this one. I've been 30 inches for a while now and once you go there you just can't go back! Look at the increases with SLI, IMO the scaling is better than crossfire solution which is mostly why I'm still sticking with Nvidia.

[eric66]

only thing that can fix metro performance is patch from developer

[annihilat0r]

SLI scaling indeed looks good, especially on lower FPS's, which is why you get a SLI system anyway

[GoldenTiger]

My eyes opened wide as I looked at the SLI scaling of BFBC2 amongst other titles, 2x480 in SLI looks good to me.

84% increase in BFBC2..... NICE....

Yeah, the scaling is nuts . Shame I can't take advantage of it right now... can't/don't want to spend the money for two of these, one's already a lot .

[SimBy]

Has anyone even checked the FPS numbers? BC2 for example. HD5870 according to nV gets 62 FPS yet in [H] review with lower AA it gets 43.5

God knows what settings nV used for this benchmarks. I wouldn't be surprised if its not even DX11.

[Carfax]

Fermi is much more likely to add on performance over time (with drivers) IMHO.

Exactly. ATI has had what, 7 months or so to refine their drivers? Nvidia should easily add about 15 to 20% (possibly more) increase in performance from driver optimizations alone over the next few months..

I'm betting the 200.xx drivers will be the first to have real G100 optimizations..

[Arxagelos]

If this is true, we have a deference from 470 to 480 32sp witch seems to be a +20% advantage for GTX480, if fermi was with its whole sp512 then maybe we have seen xtra +20% advantage from 5870.
Hope GTX480 ultra or something hits the marker.

[SimBy]

Exactly. ATI has had what, 7 months or so to refine their drivers? Nvidia should easily add about 15 to 20% (possibly more) increase in performance from driver optimizations alone over the next few months..

I'm betting the 200.xx drivers will be the first to have real G100 optimizations..

So are you saying nV driver team was drinking and partying for the past 4 months since they got ES out of the oven?!

I expect the drivers to be very mature actually.

[Carfax]

Yeah, the scaling is nuts . Shame I can't take advantage of it right now... can't/don't want to spend the money for two of these, one's already a lot .

Yeah, I'm on a 30 incher myself and those SLI numbers if true look really good..

And I'm sure they'll get better as well in the future.

But it's hard to justify spending 1,000 bucks on graphics alone, especially when only a handful of games require that sort of juice to play at 2560x1600 maxed settings..

[eric66]

So are you saying nV driver team was drinking and partying for the past 4 months since they got ES out of the oven?!

I expect the drivers to be very mature actually.

yeah this one can be said also and don't forget nvidia probably picked best scenarios so i expect actual difference to be between 15 - 20 percent on avg at best

[GoldenTiger]

So are you saying nV driver team was drinking and partying for the past 4 months since they got ES out of the oven?!

I expect the drivers to be very mature actually.

I don't... they had to have been tweaking things a lot, I doubt they had the first chips ready several months ago and had the specs 100% nailed down at that point in time.

Yeah, I'm on a 30 incher myself and those SLI numbers if true look really good..

And I'm sure they'll get better as well in the future.

But it's hard to justify spending 1,000 bucks on graphics alone, especially when only a handful of games require that sort of juice to play at 2560x1600 maxed settings..

Exactly... it's very hard to justify spending $1k for GFX when half that will max out virtually everything even at 2560x1600, with the remainder being playable with slightly lower settings.

[Dami3n]

From Nvidia:

Two results of BF BC2 and the same fps in 2 different configurations?

[Carfax]

So are you saying nV driver team was drinking and partying for the past 4 months since they got ES out of the oven?!

I expect the drivers to be very mature actually.

Anyone thats a long time user of Nvidia knows what I'm talking about..

There are a number of occassions in the past where Nvidia has released drivers that increased performance by as much as 20 to 30% on a new architecture..

The best example of this was the GF3 series.....but even the G200 got a good 20% boost in performance from the 180xx and 190xx drivers, compared to the release 170xx drivers.

[zalbard]

I've made the calculations:

GTX 480 beats the HD5870 %24 average on 1920 + 4xAA, and 23% average on 2560 + 4xAA.

GTX 470 performs nearly completely equal to HD5870 on both settings.

if those benches are correct.

That's really good if it's true, time to cross fingers!