Sign-in / Sign-up
Your question

Asus Z77 motherboard Video card setup

Tags:
  • Asus
  • PCI Express
  • Components
Last response: in Components
September 9, 2012 8:04:34 PM

I am running an Intel I7 3770k on an Asus Z77 Sabertooth.

I am about to purchase a GTX 680 to replace my GTX560tiFTW. I considered running the GTX560 as the PhysX driver.

I am aware if I run the 680 in the primary pci-e 3.0 slot, with the 560 running in the 2nd pci-e 3.0 slot, it will bog down my primary card, as it will only run at x8. There is a 3rd PCI express lane marked as PCI-E 2.0x16 (x4 max).

Before I invest the money in to a bigger power supply to accomadate the 2 cards, I need to know if this means I am able to run the GTX 680 in the primary pci-e 3.0 slot at x16 with the gtx560 in the pci-e 2.0 slot that runs at x4. I dont care if the 560 will be bogged at x4 as long as it will allow my primary to run at x16.

Someone confirm for me?

More about : asus z77 motherboard video card setup

a b Ĉ ASUS
September 9, 2012 8:12:54 PM

It'll be fine (even in the second PCIe 3.0 slot because, remember, PCIe 3.0 x8 = PCIe 2.0 x16), but it's not even necessary. The 680 is more than powerful enough on its own to handle PhysX and there aren't many PhysX games anyway. I'd just sell the 560, honestly.
m
0
l
September 9, 2012 8:38:18 PM

Im aware it would still work fine, and that the 560 as a PhysX card is overkill and probably not nessecary. I'm just curious to know if I can get the 680 (a PCI-E 3.0 card) to run at x16 plugged into the primary PCI-E 3.0 lane, with the 560 plugged into the seemingly independant PCI-E 2.0 x4 lane.
m
0
l
Related resources
a b Ĉ ASUS
September 9, 2012 8:49:52 PM

kewlguy239 said:
I'm just curious to know if I can get the 680 (a PCI-E 3.0 card) to run at x16 plugged into the primary PCI-E 3.0 lane, with the 560 plugged into the seemingly independant PCI-E 2.0 x4 lane.


That part, I honestly don't know, and I'm curious about it myself. A little research (Google) may be in order.
m
0
l
a b Ĉ ASUS
September 9, 2012 9:05:40 PM

Took me a while, but I found it. Not specifically about using a second card in the 2.0 x4 slot, but it has a bit that explains the 2.0 x4 slot. Read the part in bold.

Quote:
You cannot 3-way SLI on Sabertooth Z77.
SLI requires at least x8 per lane.
If I am not mistaken, Sabertooth gets 16 lanes from CPU and puts them towards slot 1 and 2 in x8 and x8 configuration.
It also gets x4 for the last slot from motherboards PCH.
The lanes that come from CPU are either PCI-E 3.0 or PCI-E 2.0 depending on CPU.
The lanes from PCH are PCI-E 2.0 on Intel series 6 and 7 motherboards.
Now, SLI is impossible due to PCI-E 3.0 x8/ PCI-E 3.0 x8/ PCI-E 2.0 x4 on Sabertooth Z77 (assuming that Ivy Bridge is installed for PCI-E 3.0); however you can crossfire with that configuration.
With x4 2.0 there will be a quite a bit impact on the performance. Some boards like my, use x8 3.0, x4 3.0, x4 3.0 which is still equal to x16 2.0, x8 2.0, x8 2.0 which is still OK for Crossfire without impacting almost any performance at all (There was an article about this, I will add a link if I find it).
The only way to provide X16 3.0 lanes across all lanes on Z77 is to use PLX chip which is quite expensive.


That would seem to suggest that it is completely independent of the 3.0 slots.

That answers that, but I still say that it wouldn't even matter. x8 3.0 (if that was the case and it won't be) has more than enough bandwidth for a 680 and using the 560 for PhysX is still totally pointless. It's not even worth worrying about.
m
0
l
September 10, 2012 2:54:50 PM

yeah, that about solves that. it looks like in theory i could run the 680 at x16 with another card at x4 on the PCI-E 2.0 lane. I do realize how pointless it is and how minimal of a performance increase it would be, but the fun part of being an enthusiast is not doing whats practical, but trying whats possible! Thanks very much for the research. I was going nuts with the wording in the forums I had read and am grateful someone finally confirmed my suspicions and could back up the theory.
m
0
l