Leaked 68xx benchmarks

straxusii · 18 Oct 2010 at 16:21

http://vr-zone.com/articles/amd-radeon-hd-6800-dirt-2-battleforge-benchmarks-leaked/10091.html

Also indicates 6870 will have 1120 shaders and 6850 960

"Dirt 2 and Battleforge (both DX11 titles) benchmarks have been leaked at PCOnline Forums (originating from PCZilla BBS), comparing the upcoming Radeon HD 6850/6870 cards to their likely competitors - the GTX 460. Both titles were previously believed to favour the Nvidia cards, thanks to its stronger tessellation performance. However, it is clear that AMD have improved DX11 performance, as the HD 6800 cards end up well ahead of both GTX 460s. "

Scores are 56.7/52.3/47.2/39 for 6870/6850/4601GB/460768MB

Edit : 6850 and 6870 preview here : http://www.legitreviews.com/article/1444/2/

Ladiesman417 · 18 Oct 2010 at 16:23

RavenXXX2 · 18 Oct 2010 at 16:25

lol they are judging tessellation performance on Dirt 2 performance, I'll wait for some heaven tests thank you very much.

Weazle · 18 Oct 2010 at 16:27

Ohhhh I smells an potential Nvidia fanboi annoyance

... just j/k with ya mate

straxusii · 18 Oct 2010 at 16:31

There's been a few leaked heaven extreme tessellation shots bandied around and they show, what I guess will be 6970, destroying a 480.

Guess we will find out in 4 days

RavenXXX2 · 18 Oct 2010 at 16:32

Confirmed as fake.

Rroff · 18 Oct 2010 at 16:32

straxusii said:
There's been a few leaked heaven extreme tessellation shots bandied around and they show, what I guess will be 6970, destroying a 480.

Guess we will find out in 4 days

Going by those numbers and a little guess work it would put the 6970 at 36-37% faster than the GTX480 in dirt2 and tucking in right behind GTX470 SLI.

straxusii · 18 Oct 2010 at 16:37

RavenXXX2 said:
Confirmed as fake.

Yeah probably, but my guess is it will own the 480 in that bench.

RavenXXX2 · 18 Oct 2010 at 16:39

Time will tell, personally I think it will be on par or marginally slower.

RobotGirl · 18 Oct 2010 at 16:44

I'm interested to see how they perform in tessellation too.There isn't much point in high fps if you can't have all the visual effects maxed out.

GJ02 · 18 Oct 2010 at 16:49

AMD know what they need to do, and i reckon they will not fail
this is their time again i feel
just hurry up with it !!

straxusii · 18 Oct 2010 at 16:55

RavenXXX2 said:
Time will tell, personally I think it will be on par or marginally slower.

I think that will be a big fail for the 6970 if it turns out that way.

RichardH754 · 18 Oct 2010 at 17:05

What about a high OC 460 against it, As ATI cards don`t OC as well as Nvidia

RichardH754 · 18 Oct 2010 at 17:05

Well the OC does seem to do more fps on Nvidia cards

kylew · 18 Oct 2010 at 17:12

RichardH754 said:
Well the OC does seem to do more fps on Nvidia cards

This isn't true, and I don't know why people keep on acting like it is. At the very least, it's true in the sense of 1-2%, but that's not something you'd notice.

RavenXXX2 · 18 Oct 2010 at 17:15

Yeah overclockng on fermi nets more performance than on the 5*** cards with same percentage overclocks.

Bodom · 18 Oct 2010 at 17:36

Carefull with VR Zone guys my browser defender went nuts when I was on there this morning

kylew · 18 Oct 2010 at 17:54

RavenXXX2 said:
Yeah overclockng on fermi nets more performance than on the 5*** cards with same percentage overclocks.

We've been over this before Raven. Do you forget how you used to claim you got more performance in percentages than your overclock?

An Exception · 18 Oct 2010 at 18:09

kylew said:
We've been over this before Raven. Do you forget how you used to claim you got more performance in percentages than your overclock?

Raven is right (If a little exaggerated), GF100 does scale a little better with percentage clock increases which is likely down to the fact that Fermi's front end was probably designed to feed higher performing shaders...

Not sure about GTX460 1Gb Vs 5850 though, I would think this would be closer...

kylew · 18 Oct 2010 at 18:25

Ejizz said:
Raven is right (If a little exaggerated), GF100 does scale a little better with percentage clock increases which is likely down to the fact that Fermi's front end was probably designed to feed higher performing shaders...

Not sure about GTX460 1Gb Vs 5850 though, I would think this would be closer...

Exaggerated a lot more like. Every bit of "proof" he posted about it wasn't proof of his claims at all. You forget that he was claiming something like 19% more performance for like a 15% overclock.

Anyway, my point was more to the fact that the amount was so minimal that you'd never notice it in games, it was continuously sub 2% more than evergreen. At that level it's not worth claiming.