where does the crossfire launch leave r520

Nice analysis Walt, and I can't disagree to heavily.

I will however say this: ATI sure scrambled in a hurry to get Crossfire to market after their anti-sli FUD didn't accomplish anything.

Did they spend all that time and R&D funding simply as a marketing expense? (Because sli/crossfire are irrelevant financially according to your theory, so they got involved in this niche solely for the purpose of bragging rights.)

I agree that if my mobo already had multiple PCIx16 slots then I'd be far more likely to play either nVidia's or ATI's little multiple GPU games.

Then again, the reason why I haven't upgraded is because I don't see the need. Now, if there were peripherals other than GPUs that used/required PCIx16 slots (or PCIx8 slots), then I might upgrade and then I might buy a dual PCIx16 motherboard. Because then I might run SLI or Crossfire, or I might run a single card in one slot and a super nacho cheese dispensor in the other.
 
RancidLunchmeat said:
Nice analysis Walt, and I can't disagree to heavily.

I will however say this: ATI sure scrambled in a hurry to get Crossfire to market after their anti-sli FUD didn't accomplish anything.

Did they spend all that time and R&D funding simply as a marketing expense? (Because sli/crossfire are irrelevant financially according to your theory, so they got involved in this niche solely for the purpose of bragging rights.)

I agree that if my mobo already had multiple PCIx16 slots then I'd be far more likely to play either nVidia's or ATI's little multiple GPU games.

Then again, the reason why I haven't upgraded is because I don't see the need. Now, if there were peripherals other than GPUs that used/required PCIx16 slots (or PCIx8 slots), then I might upgrade and then I might buy a dual PCIx16 motherboard. Because then I might run SLI or Crossfire, or I might run a single card in one slot and a super nacho cheese dispensor in the other.

According to Walt's theory, SLI is not a financially good decision. However, that hasn't seemed to stop consumers from purchasing them. SLI is driving Nvidia based motherboard sales as well as video card sales. Their most recent financial reports were very strong and Nvidia claimed that SLI was a big reason for it. I don't see it as being simply a "bragging rights" investment. SLI simply has helped lock consumers into their NForce chipsets. I question ATI's reasoning for allowing other vendors to support SLI with their cards. It evidently has helped Nvidia's chipset sales and I don't see why ATI wouldn't do the same. From the looks of it, these new ATI chipsets look pretty solid. Why not try to take some of VIA's market share and bite into Nvidia's a bit. Then again, what do i know?
 
ondaedg said:
According to Walt's theory, SLI is not a financially good decision. However, that hasn't seemed to stop consumers from purchasing them. SLI is driving Nvidia based motherboard sales as well as video card sales.

I think you're making a bit of a jump there, despite what Nvidia claim in their finance conferences. SLI motherboards are selling because they are currently the best, most featured AMD (and now Intel) motherboards. Only a very tiny proportion of them are going to be kitted out with dual graphics cards, let alone the top performing graphics cards.

SLI isn't selling these motherboards, the quality and features of the motherboard above and beyond the expansion capability of SLI is what's selling them.

Of course Nvidia is going to claim they are selling loads of SLI motherboards, technically they are, but they are not selling dual graphics cards with all of them. Nvidia would look stupid if they told people that their extreme case performance differentiator was just for a tiny minority of people for bragging rights. By embedding SLI within their already successful chipset business, Nvidia can leverage one off the other when it comes to marketing.

That's exactly what they are doing when they tell people how SLI is "selling" loads of SLI motherboards and graphics card, when the vast majority of motherboards arn't using two cards, and the vast majority of card sales are only one single card in a system.
 
RancidLunchmeat said:
Did they spend all that time and R&D funding simply as a marketing expense?

Yup. Because 1) being the manufacturer of the "fastest solution" gives you a "halo effect" which makes all your other products look more enticing (see the car world for countless examples of this), and 2) the thought train "oh, if I buy an SLI mobo and an nVidia card now, I can always add a second one in the future", while it's probably going to be followed up by "hey, I could get another old card, or I could just get a new one with wizzy features... yeah, I'll go with a new one", is fairly common and prompts a lot of purchasing decisions. It's like when you buy a system with certain components because "it'll be much more upgradeable this way" and then junk the whole thing for an entirely fresh build a couple of years down the line.
 
WaltC said:
Personally, I think folks are sadly mistaken who are of the opinion that things like SM3.0 support and SLI/X-Fire support are the things that differentiate one company's products from another. IMO, these are small-time and short-term in greater scheme of things and the battlefield economically will still very much remain in the domain of the single-gpu product line.
SM3.0 isn't in the domain of the single-gpu product line?
 
Bouncing Zabaglione Bros. said:
ondaedg said:
According to Walt's theory, SLI is not a financially good decision. However, that hasn't seemed to stop consumers from purchasing them. SLI is driving Nvidia based motherboard sales as well as video card sales.

I think you're making a bit of a jump there, despite what Nvidia claim in their finance conferences. SLI motherboards are selling because they are currently the best, most featured AMD (and now Intel) motherboards. Only a very tiny proportion of them are going to be kitted out with dual graphics cards, let alone the top performing graphics cards.

SLI isn't selling these motherboards, the quality and features of the motherboard above and beyond the expansion capability of SLI is what's selling them.

Of course Nvidia is going to claim they are selling loads of SLI motherboards, technically they are, but they are not selling dual graphics cards with all of them. Nvidia would look stupid if they told people that their extreme case performance differentiator was just for a tiny minority of people for bragging rights. By embedding SLI within their already successful chipset business, Nvidia can leverage one off the other when it comes to marketing.

That's exactly what they are doing when they tell people how SLI is "selling" loads of SLI motherboards and graphics card, when the vast majority of motherboards arn't using two cards, and the vast majority of card sales are only one single card in a system.

I disagree. There are very few features in the SLI editions of the boards that aren't available in the Ultra boards already. It is the option of adding that second card later that is locking consumers into purchasing the SLI boards. You may be right that the majority will not see a second board. However, it is the future option of adding it that is the convincing factor.
 
Ona,

From what I've read the "crossfire ready" ATI motherboards certainly do have serious issues with their southbridge. Not to mention that nVidia chipsets have always been the preferred AMD solution even for people using ATI GPUs. ATI has to allow other vendors to support crossfire because ATI doesn't have the market presense of nVidia in the motherboard arena. So they can't leverage that to sell more motherboards and vid cards. Essentially, they can't push the motherboard market because their goal is really to just sell vid cards (at this moment).

Charmaka,

I find that to be an incredible expenditure of resources. It also doesn't make sense at this time in the product life cycle. To be honest, I see one of two things happening: either the R520 doesn't launch any time soon, or these crossfire cards never actually make it to market in any meaningful quantities.

ATI might start shipping the "Crossfire ready" motherboards and try to gain market presense in the motherboard arena by promising the potential of Crossfire... but "the masses" (funny way to describe a very small niche of a very small niche..) won't have "master cards" available to them.

And seeing as how the 850 doesn't even work with the 800, I highly doubt any previous generation card will work with the R520.

Seems to me the benefits promised/hoped for from ATI's crossfire over sli don't exist. In terms of "a more elegant solution", in terms of cross-gpu capability, etc..

They did, however, appear to do a great job providing more options for the use of doubling the power of your GPUs rather than just focusing on frame rate.

But one of these products is never going to exist. There's no way ATI can manufacture "master cards" in any meaningful quantity and still launch the R520 within 6 months of the same time. Anybody investing in a new motherboard and another $500 vid card knows enough to read the internet and will see the R520 on the horizon. (or, it won't be on the horizon).
 
Xmas said:
SM3.0 isn't in the domain of the single-gpu product line?

Yes, just as is 8xFSAA. What kind of a sales driver is 8xFSAA, do you think?...;)

To me, the current shipping state of SM3.0 is much like finding a needle in a haystack and then using the magnifying lens of marketing to enlarge its perspective far out of proportion. IE, it is often said to be vastly more important in the scheme of things than it is...;)

If SM3.0 actually was the godsend for 3d that its proponents say it is then I might be persuaded to buy it. But since it isn't even fractionally that important, I'm not persuaded at all...:D In fact, it's difficult to suppress a yawn when talking about it these days. I seem to be missing nothing at all by not possessing SM3.0 hardware, remarkable as that may seem to some.
 
RancidLunchmeat said:
But one of these products is never going to exist. There's no way ATI can manufacture "master cards" in any meaningful quantity and still launch the R520 within 6 months of the same time.
Quoted for future mirth. :LOL:
 
digitalwanderer said:
RancidLunchmeat said:
But one of these products is never going to exist. There's no way ATI can manufacture "master cards" in any meaningful quantity and still launch the R520 within 6 months of the same time.
Quoted for future mirth. :LOL:
well one card is 130nm or 110nm and the 520 is 90nm. Why would ATI who doesnt make the chip, TMSC does, Have issue producing pcb master cards with differnt chips? You will have to order them from ATI for the first month i bet, and thenn they maybe in Compusa ..... If crossfire is a good sell for ATI , i would think that they will make only the mastercards and not the standard 5xx. So no issues for availbility. It just depends on TSMC, and earthquakes.
 
...and marketing and business timing. ;)

I don't think they're going to have any trouble supplying either Crossfire or R520s, when they release them.
 
RancidLunchmeat said:
Charmaka,

I find that to be an incredible expenditure of resources. It also doesn't make sense at this time in the product life cycle. To be honest, I see one of two things happening: either the R520 doesn't launch any time soon, or these crossfire cards never actually make it to market in any meaningful quantities.

ATI might start shipping the "Crossfire ready" motherboards and try to gain market presense in the motherboard arena by promising the potential of Crossfire... but "the masses" (funny way to describe a very small niche of a very small niche..) won't have "master cards" available to them.

And seeing as how the 850 doesn't even work with the 800, I highly doubt any previous generation card will work with the R520.

Seems to me the benefits promised/hoped for from ATI's crossfire over sli don't exist. In terms of "a more elegant solution", in terms of cross-gpu capability, etc..

They did, however, appear to do a great job providing more options for the use of doubling the power of your GPUs rather than just focusing on frame rate.

But one of these products is never going to exist. There's no way ATI can manufacture "master cards" in any meaningful quantity and still launch the R520 within 6 months of the same time. Anybody investing in a new motherboard and another $500 vid card knows enough to read the internet and will see the R520 on the horizon. (or, it won't be on the
horizon).

Yes, it is a huge cost, but I think the benefits gained by having "fastest product on the market" will probably make up for it, as will those gained from people intending to buy a second card later but then realising it's not worth it (witness sales of SLI mobos, most of which are never going to see a second card). As I said, compare it to the car industry. You think the GT40 (or any similar "halo model") makes business sense on its own? Not in the slightest. But it makes Ford cars more cool, and that shifts more Focuses, which are the bread-and-butter.

And no, I don't think that X8x0 master cards will be out in any volume before R520, but I don't think this matters. For one thing, they'll still sell (as much as they're ever going to) long after the 520's out because they'll enable people with an existing X8x0 card to get a speed boost - most of the sales of these cards will probably be at the end of their lifecycle, when they'll be a cheap upgrade. And for another, for the presumably relatively low cost of slapping the compositing engine they've developed for the 520 onto the 420/480, they get to demo Crossfire early without waiting for the 520, catch up with nVidia in mindshare terms and get people used to it, and have a proper media hoo-ha about just this tech without any extra stuff. If it had launched with the R520, it'd be even further behind SLI, people wouldn't be sufficiently aware of it for it to have a huge effect on the perception of the 520 and it'd be relegated to a small part of the coverage. This way, when they launch the R520 they can just say "and of course it supports Crossfire" and everyone will go "ah, Crossfire, yeah, cool, I know what that is".
 
ondaedg said:
According to Walt's theory, SLI is not a financially good decision.

I don't recall advancing a theory of any type...;) What I did was to simply reiterate the fact that the vast bulk of income and profits for all the companies involved comes out of the single-gpu markets. Intel for instance does more graphics-chip business than either ATi or nVidia, and none of it is derived from any form of "SLI." Indeed, the vast bulk of graphics-derived income for both ATi and nV is generated from single-gpu sales and solutions. There's nothing theoretical at all about these facts.

Nor is it a theory that far more single-slot PCIe boards (not to forget AGP, of course) are being sold at the moment by all mboard makers than the number of dual-slot PCIe mboards the same companies sell. It's not rocket science to understand that both SLI and X-Fire are incompatible with single-slot PCIe mboards, and so the factual statistics speak for themselves. Based simply on the number of single-slot mboards sold compared with the number of dual-slot mboards sold it is easy to see that both SLI and X-Fire are literally small potatoes in terms of the absolute percentage they might even theoretically hold versus the deployment of single-slot mboards into the markets.

However, that hasn't seemed to stop consumers from purchasing them.

I don't recall stating that no one had ever bought SLI. Some consumers certainly have--just as some consumers buy Apple, and a few years ago some consumers actually bought Microsoft Bob...:D But the fact that some consumers actually will buy almost anything produced at one time or another is no comment on the marketshare those products might garner, and no guarantor of the ultimate success of those products, either.

SLI is driving Nvidia based motherboard sales as well as video card sales.

When nV publishes statistics which state that their total SLI sales exceed those from single-gpu solutions then you'll have a point. But nV hasn't ever said anything like that, have they? And for the reasons stated above they won't ever be stating that, will they? Don't confuse what is "driving" the hype from nV's PR machine with what is driving nV's sales--seems to me that's exactly what you are doing.

Their most recent financial reports were very strong and Nvidia claimed that SLI was a big reason for it. I don't see it as being simply a "bragging rights" investment. SLI simply has helped lock consumers into their NForce chipsets. I question ATI's reasoning for allowing other vendors to support SLI with their cards. It evidently has helped Nvidia's chipset sales and I don't see why ATI wouldn't do the same. From the looks of it, these new ATI chipsets look pretty solid. Why not try to take some of VIA's market share and bite into Nvidia's a bit. Then again, what do i know?

I've been using nV core-logic at home for the last several years--and none of it has anything to do with SLI...;) My comments had to do with the fact that the percentage of the 3d and graphics markets devoted to single-slot/single-gpu solutions is far, far larger than the percentage devoted to dual-slot, SLI/X-Fire solutions. nV is no more "adversely" affected by that than is ATi--it's just the way things are. The overwhelming bulk of the market will always address single-gpu solutions.

The only way to broaden the percentage of dual-slot consumer markets in relation to single-slot markets is to *open up* the standards so that many more companies can take part at a cost they deem supportable. The announcement of X-Fire from ATi simply means that nV's "window of opportunity" with its proprietary SLI motherboard solutions is now drawing to a close. But even so, I believe the overwhelming bulk of nV's and ATi's and certainly Intel's graphics-related earnings henceforth will remain in the single-gpu market arenas. I would hardly classify that observation as a theory...;)
 
WaltC said:
Yes, just as is 8xFSAA. What kind of a sales driver is 8xFSAA, do you think?...;)

To me, the current shipping state of SM3.0 is much like finding a needle in a haystack and then using the magnifying lens of marketing to enlarge its perspective far out of proportion. IE, it is often said to be vastly more important in the scheme of things than it is...;)

If SM3.0 actually was the godsend for 3d that its proponents say it is then I might be persuaded to buy it. But since it isn't even fractionally that important, I'm not persuaded at all...:D In fact, it's difficult to suppress a yawn when talking about it these days. I seem to be missing nothing at all by not possessing SM3.0 hardware, remarkable as that may seem to some.
Well, to me SM3.0 and 8xS would be important points to consider (just as 6x MSAA on ATI's side). They would certainly influence my buying decision as I could use them right now. I do accept your stance on these features, and there's probably a lot of people that share your view, though. But, IYO what are "the things that differentiate one company's products from another" then?
 
Xmas said:
Well, to me SM3.0 and 8xS would be important points to consider (just as 6x MSAA on ATI's side). They would certainly influence my buying decision as I could use them right now. I do accept your stance on these features, and there's probably a lot of people that share your view, though. But, IYO what are "the things that differentiate one company's products from another" then?

For me they are in this order:

(1) Overall gaming IQ
(2) Drivers
(3) Frame-rate performance

I enjoy running both AF and FSAA simultaneously, and I haven't seen anything yet to surpass my x800 on that score visually. The Catalyst program is, I think, without peer in the industry and is certainly better than anything I ever saw from 3dfx or nVidia in the past (I think the monthly updates are great.) I hesitated even listing frame-rate performance here because almost all of the products currently sold will produce high enough frame rates for seamless play--but this again spins back to (1) above because very high frame rates which are dependent on what I would consider to be IQ degradation are just of no interest to me. Frame-rate is only important to me if the IQ is superlative--that's when I look at frame rates. Lower on the scale I would list, finally:

(4) PR maturity and veracity

I don't particularly enjoy it when a company tries to persuade me that bad is good and that black is white...;) It's faintly insulting and leads me to view anything such a company states with a high degree of skepticism. You might just say it puts a bad taste in my mouth for anything said company might attempt to sell me...;) I'll conclude here by saying that veracity is far more important to me than perfection (as nobody's perfect--but some pretend to it nonetheless.)
 
Walt

Intel for instance does more graphics-chip business than either ATi or nVidia, and none of it is derived from any form of "SLI." Indeed, the vast bulk of graphics-derived income for both ATi and nV is generated from single-gpu sales and solutions. There's nothing theoretical at all about these facts.

Why bring Intel into this mix? Discrete graphics chipsets can not be sold to a consumer, nor added as an add-in card. Intel selling integrated graphics chips is not the evidence that you're looking for. Dig deeper.

Based simply on the number of single-slot mboards sold compared with the number of dual-slot mboards sold it is easy to see that both SLI and X-Fire are literally small potatoes in terms of the absolute percentage they might even theoretically hold versus the deployment of single-slot mboards into the markets.

Can you provide factual evidence of this? Linkies please. There is a market for this or else ATI, Nvidia, and now VIA wouldn't be participating in this. There is nothing theoretical about that fact. 8)

I don't recall stating that no one had ever bought SLI. Some consumers certainly have--just as some consumers buy Apple, and a few years ago some consumers actually bought Microsoft Bob...Very Happy But the fact that some consumers actually will buy almost anything produced at one time or another is no comment on the marketshare those products might garner, and no guarantor of the ultimate success of those products, either.

Ultimate success is shown by financial evidence. Nvidia recorded near record profits recently. Much of that was contributed to its chipset division.
Like you said though, don't believe what the PR machine says!

I've been using nV core-logic at home for the last several years--and none of it has anything to do with SLI...Wink

I find that hardly a surprise considering there were no SLI solutions available. Way to point that out. I thought I had read somewhere that you were an attorney. If that is true, have you ever entered a toaster oven into evidence even though it had absolutely nothing to do with the case whatsoever?
 
Onda,

A perfect time to use the Chewbacca defense, but you go with the more original toaster. Nice, but I would have liked the SP reference.

The most salient part of your evidence is VIA, actually. ATI and nVidia both have other reasons to push SLI. What reason does VIA (or actually why did Intel buckle under to nVidia first?) for pushing for SLI/Crossfire capable motherboards?

If there's no market, or it's not significant, why would motherboard chipset manufacturers be either bending over to nVidia (Intel) or trying to leverage ATI's "open" method against nVidia (VIA) to get them to make their multi-gpu method "open" as well?

The only reason I can see is that both Intel and VIA saw their markets shrink because people were purchasing SLI capable motherboards in quantities that caused them concern.

How many people actually put multiple GPUs into those boards doesn't really matter, does it? In fact, I only see it matter to ATI... which is why I think the timing of Crossfire is suspect and rather silly. Crossfire is aimed at the same market as the R520. People who already own 850s and even 800s are the same people who will go out and buy the R520. Maybe they'll buy a "Crossfire capable" mobo as well. I highly doubt they'll buy a "Crossfire capable" mobo, plus a "master card" and then turn around and buy a R520 two months later. (Whether or not they ever buy the second R520, I think is highly unlikely)
 
Back
Top