Sign in with
Sign up | Sign in
Your question

NF200 x16x16 vs Intels x8 x8 P67

Last response: in Graphics & Displays
Share
a b U Graphics card
June 17, 2011 12:14:14 AM

Not sure if this has been posted
http://www.tweaktown.com/articles/4147/nvidia_nf200_x16...

It appears the NF 200 only adds a little latency
Maybe next gen well need more than x8
June 17, 2011 7:59:58 AM

You mean next gen graphics cards when refering to next gen needing more than x8?
I hope not or that will make all our moderatley priced sandy bridge mobos (and previous gen mobos) redundant in sli. Although with both companies coming out with new platforms I suppose it wouldn't be a stretch for the card companies to bump it up to x16 only. I dont think they will, I think it will use x8. But the 990 FXA mobos (least the one I looked at) do have 2 x16 slots (they also have x8 for third and somtimes fourth slots)

Was anyone even confused thinking 2 cards at x8 being slower than 2 at x16. The ONLY reason to go the NF 200 is for more than 2 graphics cards isn't it?
m
0
l
June 17, 2011 8:17:48 AM

I know right, thats exactly what I thought lol, is there even any reason to get a NF200 I mean who needs 3 graphics cards? at least for gaming
m
0
l
Related resources
June 17, 2011 8:19:00 AM

Quote:
x8 bandwidth will not limit any card by todays standards (not real world)

http://www.youtube.com/watch?v=rSfifE2Domo


They said the same thing about AGP 8x when it was released. Then they said that about PCIe 8x when we used to use AGP 8x. In another year or two who's know what the demand on the video cards will end up being.
m
0
l
a c 106 U Graphics card
June 17, 2011 5:17:13 PM

Now that Sandy Bridge has a pretty efficient PCI-E controller on-die, there isn't much of a point in the NF200 except for enabling 3-way and 4-way SLI. Remember, PCIE 2.0 8x supplies the same bandwidth as ol PCI-E 1.0 16x. The only way the NF200 would make sense for 2-way SLI is if the two cards in SLI were saturating the the 8x links on Sandy Bridge and the benefit of having the additional 8x lanes to each other outweighed the latency penalty of having the NF 200 between the CPU and the GPUs.
m
0
l
!