• 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.

1st Fermi review

Gerard said:
They seem pretty much inline with a lot of reports as of late. Waiting on hardocp and other sites to get their reviews online. Seeing as they will have had the cards for probably a week or so now id be suprised if newer drivers are included in the reviews.
Click to expand...

Yeah I didn't see half the charts (slow internet), edited my post a bit.
 
Geckovich said:
Performance aside I want heat/noise/power consumption results.
Click to expand...

335eaa6eb5b44158859d5c3.jpg
 
hominid said:
Also using resolution/aa/af that use more than 1gig video ram, but less that 1.3gig, is also disingenuous, on some of the tests.
Click to expand...

This is the problem, somewhere like Anand will hopefully use decent settings and a range of resolutions.

Pushing framerate way below 60fps by going for the highest res and AA option you can just to push AMD out of its memory safety net is going to screw with results.

But most people would want to play say Dirt 2 with less AA and most people will be playing on a lower res, how the cards compare there we might well not know.

We'll have to see how discriminating Nvidia were with who gets to review it, based on the GTS250 scenario and only giving cards to sites that agreed to only bench certain games in certain settings we'll have to see.

The results don't look bad, they way they did them, obviously the SLI and no xfire comparison was just so the charts all seem like theres a massive green bar and it destroys the 5870.

Now the strange thing, the others sites I've seen benchmark Metro 2033 with advanced physx on showed no difference in performance on both 5870's and the 285/295gtx, which seems to suggest it wasn't gpu based physx being used.

THe fact that now the 5870 is getting horrendous performance in Metro 2033 seems odd to say the least.

But this is the problem the "physx" titles are obviously heavily involved with Nvidia and will have all manner of crap in. As most people have seen DOF/Tesselation are worthless in Metro 2033 but kill performance for AMD cards, while I'm not certain Nvidia purposefully are crippling AMD performance, I'm very sure they've help code the tesselation/DOF effects to run well in a very specific way for their hardware. Thats fine IF it gives Nvidia an actual IQ improvement. If they've added in code that puts excess code complexity in for no actual improvement its just been done to screw with benchmarking results, that I would have a major issue with.

In all the somewhat "normal" benchmarks its pretty easy, 470gtx marginally ahead of the 5850, and behind the 5870. 5850 will continue to be the best card out of both companies new gen offerings. The question really should be how many 5850's you buy, not what other card do you get for more performance.


I really don't understand how, with all the maxed out settings in Metro, while I get crap performance, its no where near sub 10fps even with all dx11 options enabled, something dodgey going on with maybe a patch for a Metro benchmark that all of a sudden enables gpu based physx.
 
Last edited:
lol they used physx titles, resolutions/settings that are crippled by 1gb ati cards and old drivers.

Wonder if after the NDA falls it will allow none-nvida "recomended" games/settings with some reviews.
 
i love the fact that they used physx benchmakrs cause ati fanboys tears are sweet

Stop trolling, it's getting really tiresome now.
 
You must log in or register to reply here.
Back
Top Bottom