Which P67 mobo to keep?

A

Armpit

Gawd
Joined
Sep 1, 2002
Messages
557
Asus is sending me a new Maximus IV Extreme after I returned a busted one to them. In the meantime I picked up P8P67 WS Revolution. The board has been running well and I have everything set up the way I want it. Which board should I keep then? I'll sell the one I don't need. Both have pluses and minuses, but for me the SLI PCI-E lane config is most important. The WS Revolution is running dual x16 mode via the NF200 while the M4E runs x8 natively via the CPU. I didn't notice any performance differences personally, and frankly I'm tired of rebuilding this system so I'm leaning on keeping the WS Revolution.
 
If there are compelling reasons to keep the M4E I don't mind swapping it in.
 
Perhaps I should clarify, the PCI-E configuration is not that important but rather the performance of the video cards in SLI.
 
me personally after looking at both boards id go with the maximus[but thats just me] but the revolution is also a great board...i would not buy a board with the nf200 chip on board....x8 x8 would be all you need for gaming..difference between x8 and x16 is almost nothing so why have all the extra heat and besides its an add on chip to get the x8 four times...kinda a gimmick...
 
I think I'll keep the M4E. I did some testing on the P8P67 WS Revolution and at higher resolutions there's a definite stutter in some situations when the cards are set up thru the NF200 chip, this is not there when they're set-up natively at x8 PCI-E. The problem with the latter config on the Revo is that it puts both graphics cards right next to each other and 1-2mm from the CPU heatsink. The stutter is very obvious in 3D Mark 11 at Extreme settings.
 
What resolution are you running? Also, besides 3DMark, was there obvious stuttering in any games?
 
I tried a bunch of games and other tests and after much additional testing I found no difference. The stutter's there no matter how the cards are configured and even when SLI is disabled, so I guess it's the nature of the tests that's to blame. That puts me back to square one though, I think I'll keep the Revolution though because I don't want to take this system apart. Incidentally, with dual x16 mode in Metro 2033 there was an average 4 fps (about 20% boost) increase versus dual x8 mode at 2560x1440 with max everything, less so at lower settings.
 
Oh ok because my WS Revolution + 2 6970 are going to be here 2mro and you had me slightly worried :eek:
 
FYI, here're some test results on the P8P67 WS Revolution and GTX 580 in SLI:

x8/x8:
F1 2010 (Av/Min): Run 1 - 76/71, Run 2 - 76/72
Vantage Extreme (CPU/GPU/Comb): 76051/24841/25707
Heaven 2.5 (max all settings): Run 1 - 923, Run 2 - 931
Metro 2033 (average) - 21

x16/x16:
F1 2010 (Av/Min): Run 1 - 81/73, Run 2 - 82/76
Vantage Extreme (CPU/GPU/Combo): 78875/25266/26155
Heaven 2.5 (max all settings): Run 1 - 946, Run 2 - 950
Metro 2033 (average) - 24

Interestingly, running in dual x16 mode even with the NF200 chip does seem to slightly increase performance. Perhaps this is due to the cards communicating between each other at x16 speeds?
 
Armpit said:
I think I'll keep the M4E. I did some testing on the P8P67 WS Revolution and at higher resolutions there's a definite stutter in some situations when the cards are set up thru the NF200 chip, this is not there when they're set-up natively at x8 PCI-E. The problem with the latter config on the Revo is that it puts both graphics cards right next to each other and 1-2mm from the CPU heatsink. The stutter is very obvious in 3D Mark 11 at Extreme settings.
Click to expand...

Like I said in your other thread, on the WS Revolution ALL full length PCIe slots go through the NF200. On this board you have to put both cards in the blue slots to get x16/x16 which puts them 4 slots apart. The slot layout is actually the best for dual SLI you can get on a P67 board. If you are not going tri-SLI, why don't you get a board without NF200, they are much cheaper.
 
Last edited:
I think the real question then is how much latency does the NF200 introduce at dual x16 mode versus bypassing it in dual x8 mode natively via SB? Cannot seem to find this info anywhere.
 
I doubt that that is something you can measure without protocol analyzers. It is probably in the order of microseconds.
 
What I would like to see is comparison in SLI performance between say a P8P67 Deluxe and the WS Revolution. That at least would tell me what the impact of the NF200 is performance-wise.
 
This thread is a bit confusing. You do realize the NF200 exists on the Maximus 4 Extreme as well, right? If you throw something (anything, really) in the first PCI-E slot and the 2 vid cards in the following ones, it actually will make use of the NF200 chip, rather than just 8x/8x if you use just the 1st and 2nd slots alone for the vid cards.

Personally I'd say get rid of whichever sells first. I've swapped P67 mobos in and out about 5-6 times myself recently and it's annoying - just did it again this morning to get my NH-D14 mounted, backplate on the FT02 and the UD7-B3 don't line up :/.

You do sound like you'd rather keep the M4E overall though.
 
With the M4E the NF200 is only used if you use more than 2 cards or put them in the bottom slots. For most "normal" SLI configurations the NF200 is disabled. I've had the M4E before and only decided to get the Revo while I sent the M4E for RMA, but now I don't feel like swapping them out unless the M4E offers me better SLI performance.
 
Armpit said:
With the M4E the NF200 is only used if you use more than 2 cards or put them in the bottom slots.
Click to expand...

Hey, that's what I said! Although in a somewhat convoluted manner. :)

I recall that I came across a M4E thread that compared it to a board with 8x/8x but I can't for the life of me find it anymore. I think you already did a bit of testing on your own from a separate thread anyways. I think the performance differences are truly negligible for gaming, it's more along the lines of if you want to go through swapping the boards or not.
 
Funny, Tom's just last week did the exact comparison I've been looking for. They compared various configurations, but importantly looked at SLI performance using x8/x8 on a P8P67 Pro (native) versus the P8P67 WS Revolution in x16/x16 via the NF200 chip. Overall they found the Revo board offered up to about 5% performance boost over the Pro at highest resolutions and settings. Interesting, link to the article. I guess that settles my problem and I'll just keep the WS Revo, might as well since I already put the M4E up for sale.
 
Armpit said:
Funny, Tom's just last week did the exact comparison I've been looking for. They compared various configurations, but importantly looked at SLI performance using x8/x8 on a P8P67 Pro (native) versus the P8P67 WS Revolution in x16/x16 via the NF200 chip. Overall they found the Revo board offered up to about 5% performance boost over the Pro at highest resolutions and settings. Interesting, http://www.tomshardware.com/reviews/p67-gaming-3-way-sli-three-card-crossfire,2910.html I guess that settles my problem and I'll just keep the WS Revo, might as well since I already put the M4E up for sale.
Click to expand...

Isn't the main downside to the NF200 16x/16x the stuttering though? For a few % performance hit, I would rather not have the stuttering.
 
I didn't notice any stuttering in my testing in games, I originally thought I did but that's only in benchmarks and not there while gaming. Care to elaborate a bit more on the stuttering issue?
 
You must log in or register to reply here.
Back
Top