• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Breaking news, first benchmarks for the fermi are leaked.

Also, do people really think ATi will be sitting back? They will have SKUs to respond...I am pretty sure NV will have a slightly faster SKU than the 5870/5970, the usual suspects here will foam at the mouth because they can run Mirrors Edge + PhysX faster, then ATi will bring out a SKU to trump the Fermi.

And back to good times again for consumers, lower prices hopefully.
 
D.P. said:
Quite, simply, if those provided specs in thew OP are correct those are the numbers you would expect, .
Click to expand...

Are they though? Looking at single and double point and memory performance compared to G200 cards, they look very high. They show the GTX360 to be 88% faster than a GTX285 in Stalker yet here is the comparison:

GTX 285 vx GTX360

Performance: SP 1.063GFLOPS VS 1.152 GFLOPS = 8% GAIN

Pixel fillrate: 20.736 vs 26 = 25% GAIN

Texel Fillrate: 51.84 vs 62.4 = 20% GAIN

Memory Bandwidth: 159 vs 168 = 9% GAIN

I fail to see how that can equate to an 88% performance gain in actual games. The GTX 360 is certainly faster than a GTX285 but maybe 20%.

In fact if you compare the GTX360 specs of bandwidth, fillrates etc, they are all slower than a GTX 295 yet somehow in those charts it manages to beat a GTX 295.

SO my best guess unless there is some magic going on, is that the GTX360 is going to be head to head with a 5850 and the gtx380 slightly faster than a 5870. I still reckon a HD5970 will still be the fastest card you can buy albeit dual core.

So it will all be down to price really. If Nvidia can launch the GTX360 for £200 and the GTX380 for £300 then it will be good times for consumers. Somehow, I suspect they will be more money though.
 
Last edited:
markiejt said:
why does everyone think they are fake....

o wait these are ati forums.... Tbh if they were fake they would have atleast made them look een better.

I'm glad i waited
Click to expand...

There is an abundance of stupidity on these forums. :confused:

I guess you didn't read the link to where those graphs were posted by someone called 'successful_troll', who then provided a broken link to his supposed source, following which admin closed the topic.
 
"And back to good times again for consumers, lower prices hopefully".

Assuming that they'll be available to buy and this supply debacle will be over and everyone who wants one can get one. This supply prob is a bit of a joke atm.:)
 
FAKE FAKE FAKE wheres the troll repelant

Heres new benchmar there legit honest :D

benchmarkhtml.th.png
[/URL][/IMG]
 
I really hope they are an indication of the performance of the new nvidia cards i really do i need something to lower the price of 5870's so i can get another for xfire. Fact is though we have no solid spec on these cards yet we have nothing solid on release date or availability when they are released so until something that can be validated comes out i call fake on all these so called leaks which i think have more to do with desperate nvidia fanatics then anything we're likely to see soon.
 
engstrom0304 said:
Its fake! Cmon think ab out it a 5970 will not be beaten by a gtx 380.
Click to expand...

We can't know for sure til we see final clocks and just how well (or not) the shaders run on the 300 series... I'm pretty convinced personally that it will (somewhat dissapointingly) run right up behind the 5970 - beating it in a few cases but losing by around 15% overall... but if the shader setup is any good potentially it could be upto 40% faster... but doubt we will see that.
 
Greebo said:
GTX 285 vx GTX360

Performance: SP 1.063GFLOPS VS 1.152 GFLOPS = 8% GAIN
Click to expand...

I'm sure someone will come along and correct me shortly - as this is the bit I don't really understand - but I think the 200 series has an extra execution unit (or is that operator?) that is mostly un-used, that gives it its GFLOP figure but actually its only capable of 66% of that performance in real world... whereas the 300 series shaders are opptomised in that regard so the real gain is atleast 41%... add in other efficency enhancements and your probably looking at an extra 15-20% on that.
 
Last edited:
Rroff said:
We can't know for sure til we see final clocks and just how well (or not) the shaders run on the 300 series... I'm pretty convinced personally that it will (somewhat dissapointingly) run right up behind the 5970 - beating it in a few cases but losing by around 15% overall... but if the shader setup is any good potentially it could be upto 40% faster... but doubt we will see that.
Click to expand...

Do you serously, and genuinely believe that nVidia's new single GPU cards are going to be close to 4x faster than their previous generation? REALLY?
 
Last edited:
kylew said:
Do you serously, and genuinely believe that nVidia's new cards are going to be close to 4x faster than their previous generation? REALLY?
Click to expand...

Of course they won't be, so long as you don't put it into that context, Rroff doesn't have to think of his glaring mistake.
 
Hardly 4x - there are a small number of cases where 5970 cf does get massive gains close to 4x but in most cases its barely even 3x.

Taking the average across 20 random benchmarks from google:

285GTX: 100%
5870: 152%
5970: 210%

Pretty much all the info we have points to the 380GTX being about 40% faster than the 360GTX and the 360GTX being more than 40% faster than the 285GTX. The shaders have the potential to push that even more but I don't think we'll see that... unless nvidia gets really lucky with the final production silicon.
 
Rroff,

There is a lot of wasted silicon on the fermi that has nothing to do with gaming what so ever.

The cache might give it a performance boost, however this is nvidia's first try at this style of architecture they won't mximise the supposed benefits on their first try.
 
I don't think nVidia will care if the card isn't a great performer, as long as:

1) It's competitive and puts some money back in the bank.
2) The architecture shows promise and they can build on it.
 
You must log in or register to reply here.
Back
Top Bottom