Widescreen Gaming Forum

BloodTotal, perhaps your fury blinded you but if you hadn't noticed, behardware.com is a sister site to hardware.fr which ranks 3140/153 in traffic on that site you linked.

As for "SLI is required for 3D surround." Knowing Nvidia, they meant "SLI is required for 3D >and< surround."

Their history indicates a very unlikely probability that they'd allow 3 screens off 1 card unless you have one of those sli-on-a-stick cards.

thanks for pointing that out

Of course SLI is required for surround support. The requirement is 3 ports out and nVidia only does 2 ports out per card. That means if nothing else you need a second card for display #3. SLI is used to communicate the data between the 2 cards. What still boggles the mind is why this isn't called SLI Mosaic mode and why it wasn't released years ago.

IMHO, it shows an inherent lack of vision on nVidia's part. It's not like this place hasn't existed or the need for driver level support for surround gaming wasn't needed. It's taken nVidia a better part of a decade to merely acknowledge there is a market for this and only grudgingly after ATI rubbed their noses in it. After watching the CES nVidia keynote stream last night it's become apparent that nVidia is playing catch up on this front and simply doesn't get the audience. They're doing it to feature match ATI, not to actually appease us users after all of this time.

IMHO, nVidia simply doesn't understand the surround gaming market yet. In many ways ATI still doesn't either though. ATI, nVidia and Matrox all still have a learning curve to achieve the needs and understanding of the surround gaming community. While I see Matrox and ATi reaching out in an attempt to better find out the needs of everyone nVidia is just now coming out with 3DVS to the consumer and we still don't yet have a public community rep that is active around here. I wish nVidia rep would be less of a forum lurker and more of a forum poster...

:hide

Well, there is the HDMI port on the new cards, but only nVidia know at the minute whether that has it's own timing/controller and is thus independent of the DVI ports...

Well, there is the HDMI port on the new cards, but only nVidia know at the minute whether that has it's own timing/controller and is thus independent of the DVI ports...

With the compatability with 200-series I can almost bet that GF100 cards are also just 2 port out enabled. I couldn't see them shifting to a 3 port out model and still maintaining compatability with GT200 cards. You'll get no complaints out of me though about multi-card requirement. I'd just as soon be able to shove as many cards in a computer as we have slots and have it scale in a linear manner. I can dream I guess. :wink:

[quote]Well, there is the HDMI port on the new cards, but only nVidia know at the minute whether that has it's own timing/controller and is thus independent of the DVI ports...

With the compatability with 200-series I can almost bet that GF100 cards are also just 2 port out enabled. I couldn't see them shifting to a 3 port out model and still maintaining compatability with GT200 cards. You'll get no complaints out of me though about multi-card requirement. I'd just as soon be able to shove as many cards in a computer as we have slots and have it scale in a linear manner. I can dream I guess. :wink:

I also like the thought of being able to just add cards. If you had three cards, each would be driving its own display. I'm sure that's way easier on the system than one card trying to do all the work. I'll probably buy a third if this is the case. What a great week! I can't wait to hear more.

I'm sure that's way easier on the system than one card trying to do all the work. I'll probably buy a third if this is the case. What a great week! I can't wait to hear more.

Depends upon the mechanics of how it's processed. If it has to push a lot of pixels through the PCIe bus to pull it off, the more displays + the more cards used might = faster degredation of hardware horsepower. Everything is all massive speculation at the moment. I don't think many people know about the inner working of it yet. It looks to be so recent a development I think the feature set is still being figured out. My bet is what we actually saw demoed was SLI Mosaic mode enabled on new hardware and they don't have much to show yet for the actual new consumer version.

So 1 card is going to process 2 monitors, and the second card is going to process 1 monitor?

So 1 card is going to process 2 monitors, and the second card is going to process 1 monitor?

That does seem to be the case. However since it would be running in SLI. I would think the load would be split between the two cards or three to power the 3 displays if it was treated as one resolution/display.

So 1 card is going to process 2 monitors, and the second card is going to process 1 monitor?

Merely being plugged into a card doesn't always mean the GPU on that particular card is doing all the processing. It's too early to say how load balancing is handled. I think we'd all be jumping to unsubstantiated conclusions if we merely guessed at how the mechanics worked and took that as fact. We'll know in short order one way or the other though.

[quote]So 1 card is going to process 2 monitors, and the second card is going to process 1 monitor?

Merely being plugged into a card doesn't always mean the GPU on that particular card is doing all the processing. It's too early to say how load balancing is handled. I think we'd all be jumping to unsubstantiated conclusions if we merely guessed at how the mechanics worked and took that as fact. We'll know in short order one way or the other though.
Agreed.

Most likely course of events would be SLI cards do all the calculations, then it gets split in three and the relevant third piped to each screen. With the SLI bridge and the PCI-E 16x bus, there is lots of bandwidth available for that. Where the DVI port is probably won't matter too much - you simply can't cram three DVI connectors on the back of a single card.

But Brad's right - we'll know more when they're 'in the wild'. As nice as any info would be, anything we get from nVidia prior to launch (like any other company) will have been through ten layers of marketing first. ;)

you simply can't cram three DVI connectors on the back of a single card.

Not directly no, but there are ways to do so.

3 hdmi connectors with either short hdmi-dvi patch cables, adapters or bundle 3 hdmi to dvi cables.
2 dual-link dvi connectors with one with a double single-link adapter.

True either of these wouldn't allow more than single link resolutions but except for 30" few would notice.

Heck, make the card with an extra pcb and the third connector below the other two, it's been done before.

If you have R&D money to burn, make a model with 3 displayport outputs with 3 displayport to dual-link dvi active adapters bundled.

In the end, there's very little backing to the 'need 2 cards for more than 2 outputs' both ATI and Nvidia have been burning us with so far. Even their own workstation cards prove that, many of which have 3+ outputs from single slot cards.

It depends really if Fermi will have a 1 card solution or not for 3 monitors. According to Dave, ATI didn't just slap on some display ports to make up for having only two timing signal generators. They added also a lot of display pipelines in the 5000 series. Fermi might need to use a secondary card if they don't have this. Too early to tell, since the card isn't released yet though. I hope they get a single card solution too, but its great that they have the backwards compatibility, so people with SLI gets it by default. That increases the userbase much more and is good incentive for developers to add multi-monitor support! :rockout

Sorry, I couldn't resist, but:

Nvidia doesn't comment on unreleased products (or so they say), so we might not know before its released. But, they hold a lot of press conferences and such 6 months before its released, so info might slip out before that.... :nudgenudge :lol:

I wouldn't consider just 1 previous generation of cards to be any real amount of backwards compatibility though. I mean, they haven't fixed issues in 8800 cards for what? 17x.x to 19x.x drivers now?

And it's likely to only be for some of the 2xx cards as well, not even all of em.

Anyone know if it will be possible to mix monitors? Would it be possible to have 3 monitors "connected" together in 3D and have another monitor in 2D?

This stuff kind of scares me a little bit, because I can see myself dropping a large quantity of cash only to not have it work as I intended.

Anyone know if it will be possible to mix monitors? Would it be possible to have 3 monitors "connected" together in 3D and have another monitor in 2D?

This stuff kind of scares me a little bit, because I can see myself dropping a large quantity of cash only to not have it work as I intended.

It's way too early to definitively say anything at this point. Expect more to unfold the closer Fermi comes to release. With how demanding people are around here for specs, we'll find out though.

something to ponder ..


VESA details its plans for DisplayPort v1.2
1/10/2010 by: Anshel Sag - Get more from this author


The standard that was recently passed right before CES is to be known as DisplayPort 1.2. This revision of the DisplayPort connector by VESA is designed to increase the functionality of DisplayPort as well as further solidify it's place as a replacement for almost all types of video signal transmission.

For those that might not know, VESA is a collection of companies who are responsible for helping create the ecosystem that DisplayPort exists within. Meaning, they are the companies that manufacture the connectors themselves, the companies that install them on their computers, the companies who make the cables, and finally... the company who manufactures the display itself. Although, some of the developments mentioned were actually previewed during CES 2009. This is serving as a validation of those previews as well as a detailed insight into exactly what the improvements are.

Some notable improvements go along the lines of performance and usability. And to be quite honest, the increase in performance is what really enables the DisplayPort to become such a broad functioning and useful connector. Because of the performance increases, the DisplayPort devices can simply do more with the same connector. This performance increase is actually double that of v1.1a which is what is currently being used by consumers and businesses. It will support multiple stream support over a single connector as well as support for a higher speed auxiliary channel. They also integrated MiniDP which was originally developed by Apple and included that into the DP 1.2 standard. All of this is accomplished meanwhile still maintaining full backwards compatibility.

Since performance has been doubled to 5.4Gbps per lane, DP is capable of delivering beyond full hd stereo 3D at true 120Hz . There is also improvement in the resolution and color range that DisplayPort is capable of delivering as a result of these performance increases. It can support full 4K resolution among many other lower resolutions. DP 1.2 will also support dual 30” monitors at 2560x1600 off of one DP 1.2 source or four 1920x1200 monitors. So imagine being able to take your laptop to work, and with the right adapters, you can drive four or five 24” monitors off of your laptop, GPU willing of course.

http://www.brightsideofnews.com/news/2010/1/10/vesa-details-its-plans-for-displayport-v12.aspx

I'm Tri-SLI GTX260. Of the people who used TH2G, many are nVidia based and multi-GPU, so the adoption rate at least around here will be a lot easier (if the drivers mature to a decent level).

What are your minimum framrates like that with GTX260 tri-SLI setup? I'm interested in getting a second GTX260 for this, but I've heard minimum framrates aren't helped out much by dual and tri-card solutions...