This thread is intended to point out that the review listed above is in error and has been a determining factor for my decision making which has now become a glaring error on my part. The content is not accurate and in some places is a blatant fabrication. It is a false statement and is a NVIDIA moronic fanboi's fantasy. It is a representation of OC3D.

And now you are walking on wafer thin ice. Reviewers here install the hardware, run the benchmarks/timedemos/fraps runs or what ever is needed and then report on the results they are presented with from such tests. I know, I am one. Attacking our work with nothing to back it up in any way will earn you nothing but a one way ticket out of here.

Now, please, explain yourself in clear, concise words. With evidence supplied where necessary.

Your point would be a valid one were it not for the fact that the 9800GX2's were provided by XFX. I didn;t buy them nor do I own them. The 2x ATI 3870x2's the GX2's were pitched against in the review were my own, in my own rig at the time of writing the review, paid for by myself and at the time of writing were the highest performing GPU's, hence why I bought them. The two 3870x2's replaced the 3870 I had before that and the x1900 before that so exactly how does that make me a moronic NVidia fanboi? If I was a fanboi surely I would be classed as an ATI fanboi as that is what I owned at the time.

With regards to Quad SLI. The results don't lie. Scaling is quite evidently there, SLI (QUAD) obviously works as the screenshots and results from the review show. I have two friends who have bought a Quad SLI GX2 setup on the back of my review who have not encountered the problems you have. Bizarrely, it worked for them too!

If you feel the review has been misleading in some way then I offer no apologies. I cannot be blamed for enthusiasts incapability to setup a PC and cannot be held responsible for the poor performance thereafter.

It is in no way in the reviewers interest to fabricate results, we do not purchase items such as that for review, they are supplied. There are control methods in place to ensure results gained are consistent, accurate and reliable.

Having read the review carefully I can see nothing misleading, simply facts that do not match your personal results. I'm going to be the first and call user error on this one. Get some hard evidence and prove yourself not to be at fault and maybe then can you call the results into question.

i have read the review through many times and i remeber the reviewer talking to me about the tests he was going to conduct on them. I am however a nvidia fan boy and owned many cards from them. I started with a 7900 GT xxx edition on a 939 sli board. At this time sli was very much a new technology and when nvidia launched the 8 series card i thought it wise to invest in another 7900GT .

Funny enough as soon as i plugged the second card in i was shocked at how much raw gfx power i had to play with . Since then i have owned a 8800gtx, a 3870x2 and now im on a 9800gx2. For my latest upgrade i have recently purchased a 790i board so i can get another 9800gx2 as they are now cheap for some more sli. When i have enough funds for the second card i cant wait to see the results. I am expecting great results and will post them when i have the card.

I also believe that the review was conducted fairly and the results certainly grabbed my attention and i am looking forward to having this setup 24/7

I can honestly say you've lost me there. SLI is the linking of two Nvidia GPUs to perform the necessary tasks. Combining two dual GPU cards gives four GPUs, making Quad SLI. The argument could be made that because they use single PCI-E lanes, that it's not 4 cards in use, but it is not the cards that count, its the cores.



Once again this statement is beyond me. They quad and tri SLI clearly were enabled as results on not only his website but thousands of others have shown. Along with the likes of HWBot which shows the tri and quad SLIs do blatantly give a performance boost at very least with benchmarks.



And now you are walking on wafer thin ice. Reviewers here install the hardware, run the benchmarks/timedemos/fraps runs or what ever is needed and then report on the results they are presented with from such tests. I know, I am one. Attacking our work with nothing to back it up in any way will earn you nothing but a one way ticket out of here.

Now, please, explain yourself in clear, concise words. With evidence supplied where necessary.

I lost you on Quad SLI. +1

Ok first off, multi quote or edit button. It's not hard.

We are inviting you to back up your claim of 'False statements'. So far all you have done is repeat yourself.

So, present your evidence. We're sat with open ears.

I can honestly say you've lost me there. SLI is the linking of two Nvidia GPUs to perform the necessary tasks. Combining two dual GPU cards gives four GPUs, making Quad SLI. The argument could be made that because they use single PCI-E lanes, that it's not 4 cards in use, but it is not the cards that count, its the cores.

Umm no. I'm sorry Ham as much as I am your side for everything else in this thread Quad SLi (if using situations in other SLi setups) does not exist. Quad SLi as you point out refers to the 4 cores being operated over 2x 16x PCI-e lanes however, this does not match triple or normal SLi setups.
In dual SLi you have two cores linked directly for glorious scaling.
In triple SLi has three cores directly linked together using twice the connectors to insure bandwidth is maintained and scaling is optimal.
In "Quad" SLi you have a dual SLi setup in each card and the information from one CARD is shared (not both cores on each card - that is the difference). Thereby making the cards act as really fast single core cards so they CAN share the information over the single slot bridge used during quad SLi. If quad SLi was actual quad SLi there would presumably be 3 SLi connectors to insure bandwidth between the respective cores or perhaps SLi connectors above each core to tandem SLi. This however is not the case.

That last comment was not necessary.

What you are describing is besides the point, as well as just affirming the last sentence you quoted from me. What you have written is how it works. The fact of the matter is there are 4 GPUs cooperating with each other, via what ever means, to produce your graphics.

If you wish to debate it further, feel free to start a new thread. This one is now awaiting y eye to explain himself.

Quad = four cores, SLI = linked cards. 4 cores + link = Quad SLI

and rrjwilson, you stated "Thereby making the cards act as really fast single core cards". ACTS. A quad core CPU ACTS like one CPU (because it is) but you still call it quad core do you not?


It's just a tag at the end of the day. The ASUS commando doesnt don face paint and drop from helicopters does it.

Quad = four cores, SLI = linked cards. 4 cores + link = Quad SLI

and rrjwilson, you stated "Thereby making the cards act as really fast single core cards". ACTS. A quad core CPU ACTS like one CPU (because it is) but you still call it quad core do you not?


It's just a tag at the end of the day. The ASUS commando doesnt don face paint and drop from helicopters does it.

Wow I don't see how this thread turned into such a flamefest! There are probably limitations to quad SLI, but I doubt the GPU's have the bandwidth to make sharing 16 lanes an issue.. Theres a couple of extra chips involved with quad gpu setups, which would mean there are overheads, but the extra GPU means that it'd still spank 3 cards and definitely whip a straight dual card setup assuming all cards are based on the same GPU.

I think my commando might don face paint and drop from a helicopter though. I'm sure i've already had that nightmare lol my mobo wants to kill me!
Thanks for the idea though, I'll have to get some UV paint going on :D