rchapman45
It will show up 16x 16x in slot 2 and slot 4. But, it is all off the PLX plus those two alots put the carda very close together, very. If you water cool them you arw better off but K will tell you two fan cooled...say...780ti almost literally touch making the top card run much hotter, not a little, much. As in the top car qill want to thrpttlw down due to heat while the bottom one runs full song. It is really f'd up to me especially when they claim 16x 16 x with slot 1 and 3 like written in the manual..if you call it a manual. It took three weeks for us to be told it is a typo...really?! I'm not that dumb, they tested it and likely tried to fix it but cannot, so all who buys is misled. Needless to say mine went back to Amazon for a full refund and I got an Asus Z97 and it is flawless. As my last Asus was. Its your choice and money...$380 of your money!

zildjian75
Jesus Christ EVGA!!!  Get your crap together!!!
 
To EVGA...  Please stop with the mis-representations. The manual is wrong and y'all know it!! STOP!!!!  If you guys can't answer appropriate questions with accurate answers, please re-direct the questions to this thread so we forum users can give folks accurate information!!!  FOR PETE'S SAKE!!!!
 
 

AKToronto
I wouldnt get too hung up on this issue guys. Nothing has really changed about this issue for a few mobo itterations based on my research. See anandtech on the PLX chip here:
anandtech.  com/show/6170/four-multigpu-z77-boards-from-280350-plx-pex-8747-featuring-gigabyte-asrock-ecs-and-evga 
if you scroll down you will see discussion about EVGA's implementation, which is intended to provide better single and dual gpu performance by having a dedicated 8x to the CPU directly and not route all traffic through the plx.
 
From what i am hearing, the plx chip implementation is still looking like the following:          anandtech.com/doci/6170/EVGA%20Z77%20FTW%20Chipset.jpg
 
But no matter how you configure things on the EVGA board, you wont see much of a difference because  you will still be bottlenecked by the 16 lanes going to the CPU. For EVGA Either you are putting both graphics cards through the plx chip and then bottlenecked by the 8x going from the plx to the cpu, or you split it with the dedicated 8x from one card to the cpu and the 16x from the other card being routed through the plx which will still be bottlenecked by the other 8x going to the cpu. This is probably why people on this thread havent noticed a performance difference either way. My suggestion is to keep one GPU in slot 1 so that you can spread out the cards more and not overwork the plx chip needlessly.
 
Even with other mobo implementations that route all traffic through the PLX chip (and supposedly add 100 nanoseconds of latency, which is super small, but may add up over many many operations being performed) they are also providing some misinformation given the z-series limitation on pcie lanes, so again, I wouldnt cry over the spilt milk re evga's info. Inotherwords, there is no "true" 16/16, etc solution on the z97 platform, although perhaps you might consider the switching technology to be akin to something like intel's "hyper threading" technology on the core i7 cpus....
 
However, I wouldnt fault either solution so far. For AMD graphics cards that want to crossfire, it may be that a pure plx chip solution will help with x16 b/c from what i have read the PLX chip can facilitate with communication between the pcie lanes that do not involve data routing to the cpu. For NVIDEA cards that have SLI fingers, this point becomes a bit moot, which may be part of the reason EVGA chose the dedicated CPU lane implementation and only route 8 lanes to the PLX.
 
However, it seems to me that one of the main benefits of EVGA's implementation (and possibly of other plx methods too) is that you may be able to add more pcie cards and take advantage of some "hyperthreading"-like performance. For example, having two cards in SLI and also having a pcie SSD (as well as the more obvious 3/4 way SLI). So even if your not putting in 3/4 graphics cards, so long as your adding a bunch of pcie cards, the plx should help. In fact, with the EVGA implementation, when you are mixing gpu and non gpu pcie cards, you are still guarunteed solid graphics card performance with at least one graphics card talking directly to the cpu (and this will be more common now that the pcie bus is being used for SSD's too).
 
So for as far as my humble opinion goes, no matter how you slice it, once you start putting in multiple pcie cards into your rig (even if your not doing 3/4 way SLI) you can bet that the PLX chip will be helping you get the absolute most out of your 16 pcie lanes to the CPU. AND the EVGA implementation ensures you minimize any gpu latency when you only have one card in the rig (or when you are mixing various pcie cards). So while the evga manual may suck, don't hate on the evga board. The plx will help ensure this board lasts (a bit) longer than other non plx boards given the increased use of the pcie bus for newer computer components.