Breaking news, first benchmarks for the fermi are leaked.

Tute · 10 Dec 2009 at 18:43

Nvidia logo in the corner, are these marketing benchmarks?

svs22 · 10 Dec 2009 at 18:46

RizlaKing said:
My farcry 2 score was a lot better then listed on that benchie hell that figure islower then the lowest fps i got so i am calling bs on those benchies. May well be right for their new cards but they got the 5 series bechies totally wrong.

http://www.tomshardware.co.uk/radeon-hd-5870,review-31685-14.html :rolleyes:

chaparral · 10 Dec 2009 at 18:50

HybridX said:
Any of those screaming things in this thread actually seen that the "benchmarks" are posted by someone called SUCCESSFUL_TROLL

I also notice that

psychas · 10 Dec 2009 at 18:51

Martini1991 said:
The graph states 37.. You're looking at farcry.
It's fake, nuff said.

soz my bad

Martini1991 · 10 Dec 2009 at 18:54

You can't use old benchies, the 9.11 drivers gave massive improvement, that with overclocking I'm averaging quite a happy amount of FPS with crysis, one min, time to bench.

Lightnix · 10 Dec 2009 at 18:55

chaparral said:
I also notice that

Instantly I am reminded of this.

chaparral · 10 Dec 2009 at 18:56

Martini1991 said:
Fake, I can average 60 FPS on Crysis at 1920x1080 with no AA.

Taken from 5870 review

jigger · 10 Dec 2009 at 19:00

With old drivers and at a higher res Tom's got 51 fps. 60 would seem doable would it not ?

Martini1991 · 10 Dec 2009 at 19:03

Luls, I cited when using high settings, my bad..
Running unpatched Crysis on very high I average around 40.

Owenb · 10 Dec 2009 at 19:32

Fake. This is what a real Nvidia Vs ATI graph looks like! A zero based graph- no chance!

Also the Crysis benchmark is done at 1900*1200.

chaparral · 10 Dec 2009 at 19:39

Martini1991 said:
Running unpatched Crysis on very high I average around 40.

Just ran the crysis benchmark with my SLI gtx 275 cards and it came out at 42fps....

Seems about right as that chart shows 40fps with a GTX 295

chaotic_uk · 10 Dec 2009 at 19:44

i did not think the gtx380 had 2 gig of memory ?

asoxedri · 10 Dec 2009 at 19:45

Martini1991 said:
Fake, I can average 60 FPS on Crysis at 1920x1080 with no AA.

what do you get, when you turn on the AA?

on topic though, i also think its fake figures.

Orangey · 10 Dec 2009 at 20:15

Rroff, I know you're desperate to talk up Fermi but this is a 100% fake troll thread.

svs22 = successful_troll

http://www.tomshardware.co.uk/forum/page-276100_15_0.html#t2034061

svs22 said:
Looking at those benchmarks the GTX 3xx seems to be more powerful than we expected, I am surprised.

successful_troll said:
Looking at those benchmarks the GTX 3xx seems to be more powerful than we expected, I am surprised.

Sorry green team, but your princess is in another castle... :cool:

TAC4U · 10 Dec 2009 at 20:28

Well, obviously if the benchmarks are true and reflect similar performance in other games then ATi has something to worry about. But ATi's drivers still have a way to go with the potential for some considerable improvements. And where are the DX11 benchmarks? The tessellation results will be interesting due to nVidia's different approach. And where are the 5870 Crossfire results? Afterall, the 5970 is designed for overclocking / overvolting and easily surpasses 5870 Crossfire performance.

We don't know the final pricing, what ATi has to offer with follow-up products and/or drivers, or when they're finally going to be released. But if we're talking about a single GPU nVidia card offering that sort of performance at a vaguely realistic price then things are going to get interesting.

straxusii · 10 Dec 2009 at 20:29

Martini1991 said:
Fake, I can average 60 FPS on Crysis at 1920x1080 with no AA.

Not very high DX10 you cant!

Martini1991 · 10 Dec 2009 at 20:40

TAC4U said:
Well, obviously if the benchmarks are true and reflect similar performance in other games then ATi has something to worry about. But ATi's drivers still have a way to go with the potential for some considerable improvements. And where are the DX11 benchmarks? The tessellation results will be interesting due to nVidia's different approach. And where are the 5870 Crossfire results? Afterall, the 5970 is designed for overclocking / overvolting and easily surpasses 5870 Crossfire performance.

We don't know the final pricing, what ATi has to offer with follow-up products and/or drivers, or when they're finally going to be released. But if we're talking about a single GPU nVidia card offering that sort of performance at a vaguely realistic price then things are going to get interesting.

Two 5870's at my clocks would beat your 5970 by a good 10%

Raikiri · 10 Dec 2009 at 20:46

jigger said:
With old drivers and at a higher res Tom's got 51 fps. 60 would seem doable would it not ?

What graph are you reading? It gets 35.1...

TAC4U · 10 Dec 2009 at 20:48

Martini1991 said:
Two 5870's at my clocks would beat your 5970 by a good 10%

For starters you have to pay more for such a setup and, more importantly, I can't run two dual-slot graphics cards without losing all my PCI slots (one of which I need for my soundcard, an 18 input / 20 output card I use for multitrack recording). I can clock my card considerably higher but the noise starts to become an issue. As soon as ATi gets some better drivers out I'll try to push it a bit further. If I were to grab a Firewire soundcard to replace my current one (around £400; my current one cost me £350) then I could add a 5850 or another 5970 in Crossfire at a later date, though I'm in no hurry to do so.

Rroff · 10 Dec 2009 at 20:49

Orangey said:
Rroff, I know you're desperate to talk up Fermi but this is a 100% fake troll thread.

svs22 = successful_troll

http://www.tomshardware.co.uk/forum/page-276100_15_0.html#t2034061

Sorry green team, but your princess is in another castle...

Troll thread it migh be - but the results aren't that far out for the quoted specs - tho the quoted specs seem to be well off... if you scale the figures by the difference compared to specs elsewhere they do show fairly believeable performance figures.