I was in a similar position when I got Crossfire, and I had to do a complete reinstall for it to work properly... I'm just sayin' don't rule it out.
This makes no sense. Just ran the benches, I get literally twice the score and fps with the crossfire test. I went straight into a game of Dirt3 afterwards, and was getting the crappy 40-odd-fps again. here's my results:
Very odd. Take us through exactly what you've done driver wise since installing the second card. Also, Dirt 3 can be a bit funny with Crossfire, so try Dirt 2 if you have it.
Sure; So first time around I just uninstalled and re-installed without rebooting, the other times I have done the trick that True_Gamer said about - Uninstall - reboot - sweep - reboot - reinstall - reboot. I've tried loads of games, not just dirt, and they all get lower FPS. Weird.
Run afterburner in the background and see how much VRAM your cards are using in CF in Dirt 3. Although this shouldn't be a VRAM limitation as I run it maxed with 8xq msaa. And I only have 1.5GB cards. Another test is to run dirt 3, and to only use 1 monitor and see what FPS is with a single card, and then in CF. I do believe that the limitation is the x4 slot as you are using a hell of allot of bandwidth to run that resolution. So a further test is to see what your heaven benchmark scores are at 5760x1080 with normal setting, and test with a single card and them in CF. This will tell you if the problem lies in the x4 slot bottleneck.
I'm not sure about the x16 x8 argument. I have a x16 x8 setup and the crossfire almost doubles my fps on everything that is crossfire enabled. I know x16 x4 might seem like it may be a problem as it is now slowing everything down to the x4 speed which may be why you get less fps. I would try a fresh win 7 install as a last resort before you purchase a mobo. You have nothing to lose and you will have to do a fresh install anyway on your new board!! But as I say you can buy a x16 x8 setup as it does not seem to slow me down any on my system. But considering how much you have spent on 2 6970 cards you may as well get a mobo with 2 x16 slots, or even 3 or 4 x16 slots!!
I beg to differ as we ran that test Here see for yourself, and the more you OC the more bandwidth you use. Also when your running at our resolution, your better of with x16 x16. As this is where you will see the biggest performance lost. Edit: updated link
For whatever reason your link isn't working. It drops the : after the http when clicked? In addition it merely shows synthetic benchmarks not real world gaming differences (the only thing that actually matters). Your own link from earlier in this thread showed a minor drop from 100 to 93 going from 16x/16x to 16x/4x. Now I am not saying there won't be a difference, there will be especially at eyefinity like resolutions but there is no way it should be causing negative scaling as in the ops case. It should still be scaling, just not as much.
The first link I posted does not give any indication to what resolution was run, what games was tested. Just merely a pencentage figure. Where as the heaven benchmark is showing a 10% lost at 1080p, so you could well see upto a 30% loss at 5760x1080. And if it's at x4 could therefore show you 120% lost, which is what it looks like is happening.
Right, just ran the benchmark at full res accross the three monitors, and then ran Dirt3 at just 1080p in one monitor to see how that ran. The bench results are below in an image, but the dirt 3 results are: Dirt3 @ 1080p, Single card = 122.1702128 average fps Dirt3 @ 1080p, CrossfireX = 182.3529412 average fps = around 50% increase in fps. And the high-res bench results were quite interesting. The single card had a lower score and average, but a (slightly) higher max fps, where as the dual cards had a slightly higher average, but lower max. They are pretty close though So, I think True_Gamer might well be onto something. New board should be with me on Monday, so I should be able to get it all re-installed and setup for more benching and testing on Monday evening. Fun, sort-of.
It was a bit of a fight to prove this, and now you have proved me to be correct. It was very long winded, but I had to prove to people it was your x4 slot causing you all the grief at your native resolution. I'm glad we sorted it, and I look forward to your results on Monday. Don't forget the +Rep...Hint hint
Here's some benchmarks of what you should be expecting. Although it's on a i7 920. They don't state what motherboard or whether it is running x16/x16 or x16/x8...But in some games its 100% Scaling @ 5760x1080
Right then, all back together, and I've not run the bench yet, but did play dirt3. 3-monitor (~6000x1080) resolution. Average on old motherboard with single card was: 59 Average on old motherboard (16x/4x) in crossfire was: 46.6 Average on new motherboard (16x/16x) is: 83 A marked improvement. That's before any overclocking of the CPU or graphics cards.
I don't think that's too bad for a semi-current game at mega-res really. I bet I would now get over 200fps at a 1080p. True_Gamer will need to do more than this to get man love from me..
I'm glad you got the improvement you wanted. It will be very interesting to see what your heaven benchmark results are. Sorry I'm not into man love lol.
Makes me question other sites findings about x16 x4 not being that big a hit. does beg the question that this could of been solved quicker if bit-tech done its own CF/SLI testing.*hint hint*
