Originally Posted by w3bbo

Cheers for the feedback guys. I'd like to say it was a pleasure reviewing the two setups but in all honesty it was bloody hard work. I toasted a windows install thanks to the 1200W PSU giving up during overclocking. Couldn't for the life of me figure out why SLIx4 wouldn't work (4th card was set to PhysX!!) and to top it all off both AMD and NVIDIA drivers are very sketchy and unpredicatable when all cards are installed making getting predictable results very difficult indeed.

I wanted to do a line graph to show scaling but whicever way I did it, it just didn't look right. I see what Helios is saying though and will throw some samples up with the other editors too see what they think.

I'm working on getting some more screens including 3D and Eyefinity but these don't come cheap.

I expect criticism by using 'older' NVIDIA drivers but I was in a no win situation here. Use the latest drivers and people will moan that PCIe 3.0 is not enabled for NVIDIA and claim bias. Use older ones (with PCIe 3.0 enabled) and people moan that they are out of date. The only way you can correctly review two product lines is to use the same setups. While the differences between PCIe 2.0 and 3.0 are minute at present, using PCIe 3 one one card and PCIe 2 on another you may as well use two different motherboards which is not the way it should be done.

Originally Posted by Helios1234

Perhaps a couple of tests to see whether new drivers have more of a beneficial impact on performance to compensate for the loss of PCIe 3 would be a good way to quieten the skeptics.

Originally Posted by BurningFeetMan

what do you mean you can't afford 3 30" 3d monitors

Originally Posted by w3bbo

You still have the problem of driver enhancements though. What PCIe 2.0 has lost may be claimed back by being a better driver so you wouldn't truly know if it were PCIe gen that was responsible or the driver itself. The only way to truly tell the difference would be to pitch PCIe 3 vs 2 from the same driver which would mean testing on a board that has PCIe 2 and PCIe 3. Some boards have this capability though the lanes are different. E.G the Big Bang uses pcie 3 and 2 but PCIe 3 is 16x and pice 2 via PLX is 8x. so that will also have an effect.

Like I say - no win . I'm certainly not about to go testing all the configurations in all the games with different drivers! A few sample graphs might work better though as you say to back up what was said in the review. I ran a few benchies to test and the differences were, as expected small. Some games benifitted slightly, others lost out. In the end I decided to go with the drivers on the CD which had PCIe 3.0 enabled as this matched the AMD setup better.

Originally Posted by Essexraptor

A lot of hard work Webbo - thanks

Unfortunately ..... I think your own conclusion summed it up... a waste of time and far to costly for very little gain.

It's the domain of idiots with very deep pockets quite frankly or a stack of freebies sitting around in a review office.

I'm still not convined about setups beyond a single powerful GPU give significant % benefits to the end user. AMD GPU's have the edge in the ability to run more than 2 screens on a single unit. Why the hell Nvidia still have not gone this route is beyond me?

Keep up the good works