• 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.

nVidia SLI - any info?

I wasn't aware Nvidia had Scan Line Interleaving technology. Although they probaably own the patents due to taking over 3Dfx.

Oh you must be talking about the Alienware rig that has 2 PCI-e 16x slots for dual video cards. Not sure how thats done. But I am pretty sure you wont be able to easily find a mobo with 2 of thos slots.
 
Yep ... talking about the Alienware system.

I remember "back in the day" that 3DFx had some two video card solutions.
Is this similar?
 
Originally posted by: keysplayr2003
Oh you must be talking about the Alienware rig that has 2 PCI-e 16x slots for dual video cards. Not sure how thats done. But I am pretty sure you wont be able to easily find a mobo with 2 of thos slots.
Click to expand...
Many motherboards will offer at least dual PCI-e slots because it speeds up certain other I/O cards (1000mbps Ethernet, for one). PCI-e didn't come about explicitly for video card use.

This is not similar to SLI. SLI basically allowed two PCI cards to split the work of rendering a scene, theoretically giving better rendering performance, which it did, but the side effect was that it dropped IQ quite a bit. It will be interesting to see how NVidia and ATI take advantage of this though, we may well see a dual monitor gaming solutions in the future -- be it by driver, or by a proprietary internal connector like Matrox has on many of the Millennium line. For now you could use it to use two PCI-e capable graphics cards for your desktop if you really wanted to, but then, most can do that with AGP + PCI at the moment.
 
Many motherboards will offer at least dual PCI-e slots because it speeds up certain other I/O cards (1000mbps Ethernet, for one). PCI-e didn't come about explicitly for video card use.
Click to expand...

it really depends on the number of PCIe lanes avaialable

http://www.tomshardware.com/motherboard/20040619/socket_775-28.html

A motherboard will need a lot of lanes to accompdate 2 PCIe 16x slots as well as the various 1x, 2x, and 4x that will be present as well.
 
Yes, unfortunately we will need northbridges with 32 lanes on them, I don't think any of the early northbridges will be having 32 lanes.

PCI-e implementations are kind of odd in that they're generally putting 16 lanes on the northbridge for the graphics, then some lanes on the southbridge for the add-in cards. The graphics and periperhal PCI-e slots are being treated completely seperately as they are now in AGP/PCI implementations. I would have expected that to change somewhat as PCI-e is PCI-e, and you'd think it would be cheaper/easier to pu all the lanes on either the NB or SB chips.

Course the nForce3 solutions will probably have them all on one chip...because there is only one chip =)
 
I am curious then, how does Alienware make their Dual VGA system?
From a press release/technology report on X-bit Labs I read recently (but was written much ealier in the year) Alien wear "claims" to be using an Intel chipset solution that will be released "later in the year" and some of their own "proprietry" solutions.

I wonder ... how many x16 PCI Express connectors are on some of these upcoming motherboards? (maybe recalled? Read about the 915/925 ICH6 problems).

What happens if a mobo maker uses an x16 physical connector on the mobo (915/925) for that additional connection? How does that affect performance? I imagine that the PCI Express video cards can "slow down" to work with the slower PCI Express bus speeds butusing the that x16 "physcial" connector.

This is all so new and interresting ... I am just very confused/interrested.

Thank you all for your assistance/information
 
Originally posted by: MDE
Nvidia uses a PCI "merger" card to link the cards together.
Click to expand...

What is this "merger" card you speak of?
How can two video cards reside on the same motherboard? (I don't mean 1 AGP and one PCI)
 
Originally posted by: Kimo96734
Originally posted by: MDE
Nvidia uses a PCI "merger" card to link the cards together.
Click to expand...

What is this "merger" card you speak of?
How can two video cards reside on the same motherboard? (I don't mean 1 AGP and one PCI)
Click to expand...

Its the same idea as having 1AGP and 1PCI, only its 2 16X Pci-Express slots instead.
 
If this is anything like what 3dfx did...it's going to do quite a # on anything that isn't using the same tech.
 
Yes, 2x PCI-express motherboards are set to come out for this. I think the nforce4 will have it. Anyways, very nice looking. Especially if you have the money to buy it.
 
does anyone else think that this was a big waste of money for nvidia? Not only are the 2 video cards going to cost you around $1000, but you are talking hundreds more $$$ to buy into a Xeon system with 2 PCIe slots versus a vanilla P4. So few people are going to buy into this, that no game developer will take advantage of the features.
 
If this merely doubles fill-rate, then most fill-rate limited games will get a large boost. Doom 3 is one example. You'll be able to use very high resolutions. There's nothing developers need to do.
 
NO MORE DOUBTS ABOUT THIS
NVIDIA SLI

Geforce 6800 family enables this tech that comes AGAIN.

Question for ppl that know about sli:

is it posible to plug 3 displays using connectors of both cards?
this article specifies that one of the cards acts lik a merger, so the information rendered by second card is sent to main card then merged and sent tru the connectors of this card,...
i wanna surround gaming =)
 
I want surround gaming, but I want it on one semicircle screen. I cant get into gaming with multiple monitors when the monitors frame splits them up.
 
Leave it to Nvidia's marketing department to call a solution SLI which clearly is not SLI. :roll:

Obviously they intend to invoke the memories of 3dfx's SLI. But since it is similar to Alienware's method I doubt it will be nearly as good. The video of Alienware's demo left a lot to be desired. I was apparent that the graphics on the top and bottom half of the screen were not always in sync. If you think trilinear filtering issues are bad, wait until you have the upper half and the lower half of your screen showing different frames.
 
Originally posted by: SilentRunning
Leave it to Nvidia's marketing department to call a solution SLI which clearly is not SLI. :roll:

Obviously they intend to invoke the memories of 3dfx's SLI. But since it is similar to Alienware's method I doubt it will be nearly as good. The video of Alienware's demo left a lot to be desired. I was apparent that the graphics on the top and bottom half of the screen were not always in sync. If you think trilinear filtering issues are bad, wait until you have the upper half and the lower half of your screen showing different frames.
Click to expand...

NVs version doesnt render like AW's
NVs uses load balancing for render eveyrthing out. Besides Ill trust NVs version over AWs version
 

Boy, you guys sure can't read.


The 'New' SLI stands for 'Scalable Link Interface'

And it comes down to using algorithims to split the workload among two (or more) cards.

It should be approx 90% faster than one card alone, due to overhead.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top