• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

What's actual performance hit with running a single videocard on x8 PCIE mode?

Okay.

I've come to this point where I want to put a waterblock on DFi Ultra-D chipset but videocard is in the way. I was thinking of putting motherboard in SLI mode which will activate the second PCIE (bottom) but it will be at x8.

I wasn't sure if it would be too much of a trade-off having a good system overclock but with x8 PCIE, about 5GB/s bandwidth instead, if I'm right.
 
Originally posted by: myocardia
No, not unless you're talking about a G80 or an R600. For today's cards, there'd be no difference.
If those cards were running in directx 10, there would probably be less of a difference than there is now for cards running directx9.

There really isn't much bandwidth traffic between the graphics card and cpu relative to available bandwidth on a 8x slot. The exception is with the lower end cards that use TurboChache and hypermemory and things like that. Then there is a lot of traffic.

 
Originally posted by: otispunkmeyer
none really


as far as im concerned all PCI-E has given us is SLI and Crossfire.

That and a bunch of little black slots to space out our PCI add-in cards.
 
From 16x to 4x, you lose about 5-10 frames in some FPS's but from 16x to 8x the difference is hardly even noticable, probably less than 5 frames in a FPS.
 
Originally posted by: Cheex
From 16x to 4x, you lose about 5-10 frames in some FPS's but from 16x to 8x the difference is hardly even noticable, probably less than 5 frames in a FPS.

I thought its a 2-3 frame difference from 8x to 4x..? but that was back in the agp era..

So they are finally fully using that 4x bendwidth now?
 
Originally posted by: PhoenixOrion
Okay.

I've come to this point where I want to put a waterblock on DFi Ultra-D chipset but videocard is in the way. I was thinking of putting motherboard in SLI mode which will activate the second PCIE (bottom) but it will be at x8.

I wasn't sure if it would be too much of a trade-off having a good system overclock but with x8 PCIE, about 5GB/s bandwidth instead, if I'm right.

You could probably test this out before putting the waterblock on by just enabling SLI and leaving the card in the first slot. I would think that you could just run 3DMark06 with and without SLI enabled as a good test of this. 3DMark06 isn't so much of a good test for real world performance when comparing two different architectures, but your test would be apples to apples except for the bandwidth available to the card. It might also be a good idea to test if a single card will actually work in the bottom slot without a top card before as well.
 
Originally posted by: beggerking
Originally posted by: Cheex
From 16x to 4x, you lose about 5-10 frames in some FPS's but from 16x to 8x the difference is hardly even noticable, probably less than 5 frames in a FPS.

I thought its a 2-3 frame difference from 8x to 4x..? but that was back in the agp era..

So they are finally fully using that 4x bendwidth now?

AGP era cards couldnt even saturate the agp4x bandwidth, so there was no performance hit going from 8x -> 4x 🙂
 
Don't bother adding a waterblock to your northbridge, it doesn't help anything and only adds more heat to your loop. Get that jing-ting cooler or something.
 
I saw the jingting which was copper.

I saw another heatpipe northbridge cooler that was in all aluminum - I would prefer this one but I already have waterblock on hand from previous build.

I read this text and will definitely do x8/x8 and use bottom pcie slot. I know that x16/x2 config will definitely put a curb on graphics performance with having bus at x2.
 
Back
Top