• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

SirJamesDTech - Asus Maximus IV Extreme Review

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
I'll probably check out the video later, can't get to it at work; the one you posted in the General Hardware forum as a sort of build log was pretty good.

Like I mentioned in that other thread, it would be nice to see some pictures and analysis here on the forums also instead of just a link to youtube. It would help foster more discussion here.

In anycase its always fun to see the highend hardware reviewed, keep up the good work.
 
The specs on Asus' website are a bit hazy, but its probably got an NF200 and its actually running in x16/x16 not x8/x8? Its hard to tell without being able to see the video/what they are talking about. It probably also depends on the video cards used, only the higher end ones are slightly limited by x8/x8 right?

Edit: wait, he said doing Better not scoring Equal to, yeah how is that possible?
 
x16/x16 via the NF200 isn't as good as native x8/x8 due to the latency to the NF200, "splitting" on the chip, and then sending the data from the NF200.

For instance, Asus set up the Maximus 4 Extreme to bypass the NF200 when using 1 or 2 GPUs.
 
Ahh ha! I thought that board had that chip and was perplexed a bit. Thanks for the explanation!
 
Yeah, they only use the NF200 for 3+ GPUs. Here's a quote from Asus that I've seen in a few reviews regarding the NF200:

Asus said:
Though there is unfortunately only x16 lanes of PCI Express Lanes available from the Intel Sandy Bridge CPU, resulting in limited expansion for high-end applications, there is a cure via the use of PCI Express expansion bridges such as the NVIDIA NF200. However, this also leads into a performance bottleneck in PCI express transfer due to the additional latency introduced. To overcome this unwanted trade-off, the PCI Express lanes has been uniquely arranged to go through such bridge only when absolutely necessary, ensuring both expansion and performance are not to be compensated by each other.
 
So there must be a mode on the board to run 16x/16x through the NF200. He only used 2 GPU's Im guessing by hank's 16x/16x response... otherwise, he misposted, or it can be used with 2 GPUs?

sorry if this info is in the video...
 
So there must be a mode on the board to run 16x/16x through the NF200. He only used 2 GPU's Im guessing by hank's 16x/16x response... otherwise, he misposted, or it can be used with 2 GPUs?

sorry if this info is in the video...

I don't see the option to enable x16/x16 for dual GPUs in the manual, maybe I'm missing it or it's not listed. But, here is the slot setup:

untitled.JPG
 
^^ That's precisely the slot configuration I went by.
PCIe slot 1 occupied, with a GPU each in 2 and 4.
According to GPU-Z and other monitors, they are running at x16.
The benchmarks certainly didn't impress me to that reading, though.
 
Can't see the video here at work... So you used a non-GPU PCIe card in the first slot and a GPU in the 2 and 4 slots?
 
Yep...as long as that slot is occupied then the 580s supposedly are at x16

580s.jpg
 
Can't see the video, but it sounds like this:

x16/x16 via the NF200 isn't as good as native x8/x8 due to the latency to the NF200, "splitting" on the chip, and then sending the data from the NF200.

For instance, Asus set up the Maximus 4 Extreme to bypass the NF200 when using 1 or 2 GPUs.

Plus these:
Can't see the video here at work... So you used a non-GPU PCIe card in the first slot and a GPU in the 2 and 4 slots?
Yep...as long as that slot is occupied then the 580s supposedly are at x16

Are the reason for x8/x8 performing better than x16/x16.
 
With 3 cards without the NF200 the 3rd card is limited to 4x, and those 4 lanes are also shared with everything else running on the PCIx bus I beleive, which can really bottleneck that 3rd card.
 
Back