Will ATi jump on the SLI bandwagon?

[Karma Police]

Evans & Sutherland had a dual R200 card. they also have dual R300 and quad R300 cards.

SGI has highend visualization systems (Onyx 4 Ultimate Vision) with upto 32 ATI VPUs.

It should be no problem for ATI to introduce cards with 2, or perhaps even as many as 4 VPUs on board.

whether it's SLI or some other implementation (like what E&S and SGI have) doesn't really matter. as long as we can get a boost in performance, both geometry-wise and pixelfill-wise, it should be great 8)

I hear you, but it just seems much more consumer friendly to have it in SLI form. Where you can buy one now & another later when u get more money. If they put 2 chips in a MAXX form, the price will just be too high, I think.

 

[Megadrive1988]

^I don't disagree, as long as we get something, be it SLI or whatever implementation is best.

 

[quest55720]

Evans & Sutherland had a dual R200 card. they also have dual R300 and quad R300 cards.

SGI has highend visualization systems (Onyx 4 Ultimate Vision) with upto 32 ATI VPUs.

It should be no problem for ATI to introduce cards with 2, or perhaps even as many as 4 VPUs on board.

whether it's SLI or some other implementation (like what E&S and SGI have) doesn't really matter. as long as we can get a boost in performance, both geometry-wise and pixelfill-wise, it should be great 8)

I hear you, but it just seems much more consumer friendly to have it in SLI form. Where you can buy one now & another later when u get more money. If they put 2 chips in a MAXX form, the price will just be too high, I think.

I think right now for me atleast a dual on chip x800xt would be great. I do not want to buy a new MB, proccessor and ram for SLI. When I upgrade to PCI-X MB then yes SLI will be nice. That will be atleast 1 more generation of PCI-X motherboards because I don't want to be a beta tester. Heck for the price of the new MB,ram,proccessor and videocard I could get a dual x800xt and be happy for atleast 18 months.

 

[suryad]

Agree with quest tho I think upgrading within 18 months is kind of a bit extreme. I am thinking that anything based on the x800 and 6800 platform will be running every game at playable fps even till the year 2008! I am speculating of course but it seems to me that if tech demos done by Epic games and Artifical studios can run on existing hardware and run quiet well then the current high end should carry us through till 2008. My 2 cents. Of course that does not mean the hardware industry will be sitting still...heck I am already saving for a top of the line desktop that I will purchase in 2007-8 timeframe!!

 

[Dave Baumann]

if ATI should truly have a multi-board config in it's plans after all and/or releases a R4xx based refresh late enough, it could only indicate that R520 shouldn't be expected as early as many believe or would like to believe.

If there were to be something introduced I don't think it would represent any ramifications on the lon or shortterm roadmap, merely an anti-marketing tool.

 

[jvd]

I def think the r520 can be a 3 x jump over the r420 and nv40 series .

Yes bandwidth wont increase to 3x times what it is now (next year maybe 800mhz ram imho ) But certianly the shader performance can increase , leading to more shader intense games , which the newer boards will have a much bigger lead over the current boards , perhaps even up to 3x the performance of the current boards .

But even if they can only make a leap of twice the performance . 2x the x800xt pe performance that will be faster than 2 6800ultras in sli mode.

I highly doubt they will get 100% performance from an sli set up but even so.

You have 2 6800ultra , each has 2 slot cooling , each needs 2 power connectors , and you need a special mobo for it .

Or a single nv50 or r520 that perhaps (From ati's past trends at least ) uses single slot cooling , needs no special mobos and most likely (again with ati's past trends ) only needs a single molex controler .


The majorty of users are going to go with the newer card over 2 of the older ones .

 

[DSN2K]

hope ATI dont waste time/money/effort on anything like this. whose going to waste all that money when they can get a better single slot solution 6-12 months down the line.

 

[Bjorn]

hope ATI dont waste time/money/effort on anything like this. whose going to waste all that money when they can get a better single slot solution 6-12 months down the line.

The 9800 wasn't twice as fast as the 9700. And it was a bit more then 6-12 months between the 9700 and X800.

 

[trinibwoy]

I think neither R520 nor NV50 will be near 3 times X800XTPE performance. Memory bandwidth just isn't keeping up with that.

It should be no problem for ATI to introduce PC gaming-grade cards with 2, or perhaps even as many as 4 VPUs, on board.

But, hear is my question: which would be better, in your opinion, a single R520 with 24 pipes, or 2 6800's in SLI? Or is that a rhetorical question until we get more info?

These last three posts got me thinking. In a situation where ATI puts multiple gpu's on one card where is it going to get the bandwidth to feed all of them? At least in SLI although you don't have twice the mem capacity you do have twice the bandwidth.

 

[Fodder]

You have 2 6800ultra , each has 2 slot cooling , each needs 2 power connectors , and you need a special mobo for it .

... the PCIE 6800U only has one connector, albeit the new six-pinner, though I think the PCIE X800XT has it too?

These last three posts got me thinking. In a situation where ATI puts multiple gpu's on one card where is it going to get the bandwidth to feed all of them?

Semi-related, with ATI's multi-GPU tech does each chip use dedicated memory on a separate bus, like the XGI cards?

 

[trinibwoy]

The majorty of users are going to go with the newer card over 2 of the older ones .

Agreed. No one said SLI is a mass market option. But it is an option nonetheless.

 

[jvd]

The majorty of users are going to go with the newer card over 2 of the older ones .

Agreed. No one said SLI is a mass market option. But it is an option nonetheless.

and one that causes nvidia long term harm for short term gain.

LIghting my self on fire is an option. Not going to run out and do that though.


All nvidia is doing with sli is hurting its self .

They don't release enough connectors and people are going to be pissed , they release enough and they will eat into future high end sales .

Sli wont eat into ati's sales as much as nvidia's either. An ati fan is going to buy ati regardless . So its really just going to hurt nvidia alot going foward .

 

[trinibwoy]

All nvidia is doing with sli is hurting its self .

I don't think most people (besides investors) particularly care about what nvidia does to itself as long as they produce good hardware. I am actually surprised at your concern for your arch nemesis

They don't release enough connectors and people are going to be pissed , they release enough and they will eat into future high end sales .

You really have a penchant for the negative don't you? You will probably throw a party if Nvidia falls flat on its face on this SLI initiative. While everyone else is looking forward to the new technology you are the single cheerleader hoping that they fail. I wonder why?

Sli wont eat into ati's sales as much as nvidia's either. An ati fan is going to buy ati regardless . So its really just going to hurt nvidia alot going foward .

Haha this was funny. Most people are not 'fans' of any kind jvd. I know that might be hard for you to accept but that is the reality of it. If someone with money to burn wants the absolute fastest rig do you think he will settle for a single R420 when he can do 2xNV40?

 

[trinibwoy]

SLI boards soon? For SLI to be an attractive option for me these boards better not have more than a $50 premium over non-SLI boards and the performance better be at least 175% of a single card.

 

[jvd]

don't think most people (besides investors) particularly care about what nvidia does to itself as long as they produce good hardware. I am actually surprised at your concern for your arch nemesis

I care , I want the company to be around in 2 -5 years , i don't want them to go under because they made stupid choices . Much like 3dfx did .

You really have a penchant for the negative don't you? You will probably throw a party if Nvidia falls flat on its face on this SLI initiative. While everyone else is looking forward to the new technology you are the single cheerleader hoping that they fail. I wonder why?

Its better than just looking at the bright side all the time . What you don't seem to understand is that no matter what happens nvidia has already fallen flat on its face with sli. If it happens today or with the next gen , they fell on thier face.

This isn't back with the voodoo 2 chips. They wont have 1-2 years where the only card worht buying is the 6800ultra card .

But you seem to not want to talk about anything but the good side. So its no use in tlaking to u .

Haha this was funny. Most people are not 'fans' of any kind jvd. I know that might be hard for you to accept but that is the reality of it. If someone with money to burn wants the absolute fastest rig do you think he will settle for a single R420 when he can do 2xNV40?

Some times i feel like i'm talking to a child .


The majority of people don't care about sli, they will use whats ever in thier dell computer , they wont go out and look for a rig that will let them put a second video card in.

A smaller portion knows nothing but has friends that will upgrade for them.

Then there are teh gamers who would know about sli and out of them most of them are fans. If you know what sli is and your watching it closely then your a fan of one of the video card companys .


Yes , i think he will. Considering he wouldn't be able to buy an sli rig right now. But in the future it depends . Does the person want to spend 1k on just his video system ? perhaps , if so then that is good for him. Its going to be a hand full of people.



Anyway as i said its like talking to a brickwall , so i will leave you to your fights of fancy .

 

[digitalwanderer]

SLI boards soon?

Ok, the Inquirerererer says that VIA will be ready to ship SLI boards at the end of next month....but it doesn't say if they are going to or not.

I have a feeling they can't release 'em until nVidia does and that is why it is worded that way, but it's just a hunch right now.

 

[trinibwoy]

Why would I dwell on the negatives ? If SLI doesn't work out I can just go buy a single card - no big deal. What is the purpose of constantly focusing on why SLI would not work unless you have some personal desire to see it fail? Myself and others have commented on both the benefits of SLI and the scenarios in which it does not work. You, however, have entered every single SLI thread with the usual 'points' decrying any scenario in which SLI is useful. Who is the stubborn child here?

 

[trinibwoy]

SLI boards soon?

Ok, the Inquirerererer says that VIA will be ready to ship SLI boards at the end of next month....but it doesn't say if they are going to or not.

I have a feeling they can't release 'em until nVidia does and that is why it is worded that way, but it's just a hunch right now.

Yeah I don't think anyone is clear on that relationship. But either way my plan for SLI would be 2xGT refresh parts. What's the timeline on those?

 

[trinibwoy]

Its better than just looking at the bright side all the time .

But you seem to not want to talk about anything but the good side. So its no use in tlaking to u .

I see a pattern here. I think you should see a therapist. I'm sure there are people in Iraq who have a more positive outlook on life than you do

Anyway as i said its like talking to a brickwall , so i will leave you to your fights of fancy .

Flights not fights

 

[Sxotty]

How expesnive is it to implement SLi like Nv did? It doesn't seem like it would cost much more than $10 per board, I would even think it less like perhaps $1.00 If that is the case a company might as well so they can say they have that feature.