• 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.

Leaked 68xx benchmarks

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.

Lol, yes I forgot about the poor maths, I was referring to a previous thread, and by the end of that thread I think Raven understood his maths was a little off...

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.



Mostly agreed, but if I remember there were some exceptions where GF100 scaled noticeably better.
 
Yeah overclockng on fermi nets more performance than on the 5*** cards with same percentage overclocks.

Not completely true and you never answered my detailed questions when I called you up on this. From memory you were claiming a bigger percentage performance increase from the nvidia overclock compared to the core clock speed overclock.

I claimed you were wrong and it was impossible as it would be extra performance for free but you never supplied proof and still claim it's true.

Something like a 19% gain from a 15% overclock from memory?

Either show the proof or stop claiming it tbh. This is the third time I have queried it.
 
5870 showed crap gains from 850 core to 1000, 480 rocks when overclocked. That's my take on things having experienced both cards, the 5870 gains steadily decline after 900 core, must be the architecture.
 
5870 showed crap gains from 850 core to 1000, 480 rocks when overclocked. That's my take on things having experienced both cards, the 5870 gains steadily decline after 900 core, must be the architecture.

When you had a 5870 you used to rave about the gains you had from 850 to 1000...
 
Not completely true and you never answered my detailed questions when I called you up on this. From memory you were claiming a bigger percentage performance increase from the nvidia overclock compared to the core clock speed overclock.

I claimed you were wrong and it was impossible as it would be extra performance for free but you never supplied proof and still claim it's true.

Something like a 19% gain from a 15% overclock from memory?

Either show the proof or stop claiming it tbh. This is the third time I have queried it.

He did show "proof" and it proved him wrong in the end.
 
Poor in comparison to a 480, for example in crysis I gained approximately 10 fps with 1000 core 5870, 480 is like 20 fps. It's much smoother when the **** hits the fan.

When you raved about it, it was always in comparison to a GTX480. I really don't think you gained an extra 20FPS from overclocking in Crysis though.

Don't forget we also determined that your GTX480 had had a greater percentage overclock than your 5870, which you seemed to struggle to grasp.
 
Last edited:
When you raved about it, it was always in comparison to a GTX480. I really don't think you gained an extra 20FPS from overclocking in Crysis though.

Don't forget we also determined that your GTX480 had had a greater percentage overclock than your 5870, which you seemed to struggle to grasp.

Oh man you took his bait, he stuck a bit fat juice worm on his hook and just like Robbie Coltrane at an all you can eat chips special buffet you just couldn't say no!
 
I'm not really sure why people are interested in the 6850/6870 anyway. Ok, obviously not everyone upgrades all the time, but is anyone with a 5850 through 5970, or a 460 through 480gtx going to upgrade to one, no.

THe 6970(again assume the naming is accurate) and 6950 will be the "performance" cards.

Even on the slim chance that Nvidia can make a lol 580gtx card which by latest rumours is a bigger GF104(which would probably actually suggest a 576 shader card and they are assuming 1 or 2 clusters disabled so shipping cards of 528 or 480 shaders, but slightly more efficient than a GF100). I don't think it will compete in speed, however, again that will end up a 500mm2 card, with yields in the toilet while a 6950 will offer great performance and superb value.

At the moment we don't have a clue what AMD did, assuming a 1120sp card is beating an old 1440 shader card, they've improved efficiency dramatically already. Did the 6970 stay around 1600 shaders, or is it much closer to 2000 shaders.

Would be nice if they launch the "series" together and give us an idea of 6970 performance straight off rather than wait a couple weeks.
 
They had a whole year to improve the 5xxx series and name them 6xxx series. No new arch or nm process. So the gains bound to time are well within the margin.
 
Can't believe I just read all this thread.
I'll be sticking to my 5850's anyway and skip this 6 series nonsense.
A weeks away in the sun instead for me and the wife will win me some brownie points and give me leverage for the next gen GPU's. ;)
I don't think i'll be missing out on much before then. In fact, I know I won't. (other than a few FPS gains that my ageing eyes won't even notice when everything is already running smooth)
A week away in the sun or half a dozen extra FPS. Hmmm
 
Back
Top Bottom