Intel Files Lawsuit Against Nvidia

[Face]

We have just learned that Intel filed a lawsuit against Nvidia late last night in which it alleges that the four-year old chipset license agreement the two companies signed is not valid for Intel's current and future generation CPUs with integrated memory controllers.

This includes Nehalem - a chip that Nvidia has repeatedly claimed it holds a chipset license for. Intel, as evidenced by this lawsuit, begs to differ.


http://www.bit-tech.net/news/hardware/2009/02/18/intel-files-lawsuit-against-nvidia/1

Update:


NVIDIA responds to Intel court filing regarding chipset license agreement

NVIDIA responded to a Monday court filing in which Intel alleged that the four-year old chipset license agreement the companies signed does not extend to Intel's new integrated memory controllers which are found in the Core i7 (Nehalem) processors. NVIDIA CEO Jen-Hsun Huang says NVIDIA is confident their licenses applies to Nehalem as well, and claims Intel is attempting to stifle innovation to protect its CPU business.

"We are confident that our license, as negotiated, applies," said Jen-Hsun Huang, president and CEO of NVIDIA. "At the heart of this issue is that the CPU has run its course and the soul of the PC is shifting quickly to the GPU. This is clearly an attempt to stifle innovation to protect a decaying CPU business."

http://www.dvhardware.net/article33446.html

http://www.reuters.com/article/marke...0090218?rpc=44

[jasonelmore]

anybody seen the contract? that will tell the truth.. It probably could be interperted both ways.

[Dragy2k]

wow..... uppercut from intel......as much as i disapprove of nvidias marketing ethos......ati need them to be around...

[Face]

NVIDIA responds to Intel court filing regarding chipset license agreement

NVIDIA responded to a Monday court filing in which Intel alleged that the four-year old chipset license agreement the companies signed does not extend to Intel's new integrated memory controllers which are found in the Core i7 (Nehalem) processors. NVIDIA CEO Jen-Hsun Huang says NVIDIA is confident their licenses applies to Nehalem as well, and claims Intel is attempting to stifle innovation to protect its CPU business.

"We are confident that our license, as negotiated, applies," said Jen-Hsun Huang, president and CEO of NVIDIA. "At the heart of this issue is that the CPU has run its course and the soul of the PC is shifting quickly to the GPU. This is clearly an attempt to stifle innovation to protect a decaying CPU business."


http://www.dvhardware.net/article33446.html

http://www.reuters.com/article/marketsNews/idINN1842725320090218?rpc=44

[STaRGaZeR]

"At the heart of this issue is that the CPU has run its course and the soul of the PC is shifting quickly to the GPU[...]

:ROTF:

[Particle]

Wow. Jen-Hsun is reaaaaaally full of himself, isn't he?

[Shintai]

I heard Huang said he would open a can of whoopass mixed with some big bang lawyers. But most likely just rename CPU chipsets to CPU support chips to avoid legal action. Rumours also said that there is a chance he would use AMiga style names. In that case there is no Chipset, Just Angus and Denise :rofl:

[gatecrasherlok]

Intel and Nvidia at it again. They just don't have whats important on the agenda, keeping the consumer happy or keeping their wallets happier. I guess you'll all agree they are both hiding behind the new tech for consumer BS and that the latter involving business monopoly and monneh's is the real reason.

[ghost_recon88]

Why the heck would anyone want an nVidia i7 chipset anyways? With X58 supporting both Crossfire and SLI.

[Vozer]

Wow, Intel want to block Nvidia from joining their Lynnfield and Clarkdale parties too.

[Shintai]

Wow, Intel want to block Nvidia from joining their Lynnfield and Clarkdale parties too.

Not much party there with 5$ southbridges on a "slow" DMI interface from the CPU. Plus its about QPI. Lynnfield and Clarkdale doesnt use QPI like i7. This is only about i7 and Xeons.

[Mk]

Why the heck would anyone want an nVidia i7 chipset anyways? With X58 supporting both Crossfire and SLI.
:up:
I say let them build there chipsets and let's see who buy's them with x58 around:rofl::ROTF:.

[alucasa]

I am not totally sure who's right and wrong here, but personally I hope Intel wins. I just don't like Nvidia anymore.

[Donnie27]

I am not totally sure who's right and wrong here, but personally I hope Intel wins. I just don't like Nvidia anymore.

Me too:up:

The whole NF200 crap caused a lot of bad blood.

[Vozer]

Not much party there with 5$ southbridges on a "slow" DMI interface from the CPU. Plus its about QPI. Lynnfield and Clarkdale doesnt use QPI like i7. This is only about i7 and Xeons.
You're right, Shintai. :) I thought it's about every CPU generation with integrated memory controller (includes Lynnfield and Clarkdale).

[Sparky]

I see this as intel trying to monopolize the market. Get the SLI license from nvidia, then boot them out. Nice :rolleyes:

Even if nvidia's chipsets are not so great, the desire for them (or lack of) would be determined quite quickly by the market. SiS chipsets were pretty lame, but did that warrant forceful removal of them from the market? Not really, they still were around but small in number (do they exist anymore? I haven't checked).

The more chipset makers the merrier IMO. I might not buy their chipset, but someone else might. Depends on features and pricing and all that.

[Chad Boga]

Surely it should be clear cut whether or not the agreement Nvidia and Intel had, allows for them to make a chipset for the i7.

Seems just crazy to me. :shrug:

[highoctane]

At this point I'm thinking this is just Intel being greedy and wanting to monopolize on the i7 platform.

There should at least be one option for a different chipset by a 3rd party.

[Donnie27]

Surely it should be clear cut whether or not the agreement Nvidia and Intel had, allows for them to make a chipset for the i7.

Seems just crazy to me. :shrug:

I'm not sure and don't have the documents in front of me. But from what I've seen and read, Intel and nVidia had a similar agreement to the one AMD and Intel has. Neither of those companies charges the other for the other's tech or IP. If one charges the other, then the deal is off. They agreed to stay out of court otherwise. nVidia wanted Intel's tech free while they wanted Intel to pay them for their tech. So no, I don't blame Intel for not wanting to pay nVidia (directly or indirectly) when nVidia uses their tech free.

There's a 1000 post thread here about the SLI chips nVidia was trying to pretend their cards needed. Hacked drivers proved that was NOT the case. Now sure I could be wrong:D

[Shintai]

People should realise that in the future for both AMD and Intel there aint any room for 3rd party chipsets. Its not about being greedy. Whats next? AMD and Intel monopolize IGPs because they put then on the CPU package/die? Wakeup to reality. nVidia just served its purpose in the chipset business and is no longer needed.

[onewingedangel]

Nvidia have a licence to make chipsets for Intels FSB interface, but haven't licenced the QPI interface.

Nvidia feels that their FSB licence should somehow cover the new QPI interface.

Personally, I see this Intels way. If Nvidia want to develop chipset using Intels QPI interface, they should come to a mutually benefitial agreement to licence the tech.

I think Nvidia's problem was that they don't have all that much leverage in negotiations, as both Intel and AMD move to a 'platforms' approach Nvidia loses out on both chipset sales and GPU sales unless they allow platform independent SLI.

[highoctane]

There should always be an option, how would you like to only be able to buy oem parts for your car.

There is always going to be a need for a motherboard to handle IO functions, why should be have to buy intel chipsets only or amd only regardless of how simple or complex the chipset.

Its simply greedy that these brands want to lock their platforms to themselves, at least intel seems to.

There's plenty of room for third party chipsets as there would be in x86 cpus if intel would license, Intel simply doesn't allow for room for anybody else to exist in its space then who's already there.

How things are not they are simply locking out anybody else, I mean really, what are they afraid of, they're simply squeezing out competition.

[highoctane]

I'm not sure and don't have the documents in front of me. But from what I've seen and read, Intel and nVidia had a similar agreement to the one AMD and Intel has. Neither of those companies charges the other for the other's tech or IP. If one charges the other, then the deal is off. They agreed to stay out of court otherwise. nVidia wanted Intel's tech free while they wanted Intel to pay them for their tech. So no, I don't blame Intel for not wanting to pay nVidia (directly or indirectly) when nVidia uses their tech free.

Well that I had no idea of but I still feel Intel would have resisted licensing qpi to nvidia or any other third party regardless.

[onewingedangel]

Last I heard Intel was open to licencing QPI, but couldn't come to mutually agreeable terms with Nvidia. ie. Nvidia don't want to pay, or give anything up.

As far as DMI equipped processors are concerned, Nvidia may be able to make chipsets for these as DMI is essentially a PCIe x4 connection, but these are essentially southbridges and aren't really able to benefit from Nvidia's IP as much as a QPI enabled chip would (true 3x16 SLI northbridge etc.)

It seems as if Nvidia are trying to play around some (un)specific wording of the previous FSB licence to avoid having to come to terms on licencing QPI.

[Donnie27]

Well that I had no idea of but I still feel Intel would have resisted licensing qpi to nvidia or any other third party regardless.

Maybe, I don't know what they'd do either way if nVidia hadn't pulled what they did. Here some more.

http://www.custompc.co.uk/news/604731/sli-can-run-on-any-x58-board-even-without-nf200-chip.html


The first is obviously nForce 790i for Penryn and prior CPUs, then there’s X58 with an NF200 chip, which will allow four-way and three-way SLI setups, with each slot getting 16 PCI-E 2.0 lanes. Finally, there’s the new licensed native SLI setup on X58 boards, which allows standard two-way SLI via two PCI-E slots with either 16 or eight lanes.

The other catch is that motherboard manufacturers have to send their X58 boards to go through Nvidia’s SLI Certification Labs in order to receive the ‘cookie’ that will then be embedded into the BIOS to enable full SLI support. According to Berraondo, SLI support is only ‘turned on only when our drivers determine that the motherboard has passed a special check that makes sure the key and chipset ID are a match and are in fact, certified to run SLI.’

So Intel would pay nVidia for the "Cookie" (on their boards) while nVidia Paid Intel NOTHING:rofl: Certified my eye LOL!

So yes shintai, folks should wake up! Business isn't carried on like this. Intel-AMD/ATI-IBM and others have always done it the right way. It's goes like; you pay me, I pay you or I get it free if you get it free! nVidia wants; You pay me for mine but I get your tech free.;)

[16floz470ml]

What a turnaround. Intel just released a bios for their own X58 motherboard to support SLI.

[onewingedangel]

What a turnaround. Intel just released a bios for their own X58 motherboard to support SLI.

With a number of x58 oem systems out there, Nvidia needed these systems to support SLI so that they can flog more GPUs - and they charge Intel for the right to licence SLI. Pure profit for Nvidia.

[slapmehard]

Not meaning to ruffal any blue feathers...i hope intel gets spanked in lawer fees and nv gets the judge to toss the lawsuit right out the door.
so when nv buys up ati from amd,red and green will dominate the pc industry,then intel can suck up to them to get the new 4.0 usb specs and the like.
i'll shut up now

[Donnie27]

Not meaning to ruffal any blue feathers...i hope intel gets spanked in lawer fees and nv gets the judge to toss the lawsuit right out the door.
so when nv buys up ati from amd,red and green will dominate the pc industry,then intel can suck up to them to get the new 4.0 usb specs and the like.
i'll shut up now

:rofl::ROTF: Good one!

[Shintai]

There should always be an option, how would you like to only be able to buy oem parts for your car.

There is always going to be a need for a motherboard to handle IO functions, why should be have to buy intel chipsets only or amd only regardless of how simple or complex the chipset.

Its simply greedy that these brands want to lock their platforms to themselves, at least intel seems to.

There's plenty of room for third party chipsets as there would be in x86 cpus if intel would license, Intel simply doesn't allow for room for anybody else to exist in its space then who's already there.

How things are not they are simply locking out anybody else, I mean really, what are they afraid of, they're simply squeezing out competition.

That like saying you want options for the IMC etc too....your option will be to buy AMD or Intel.

[16floz470ml]

I see that a number of people are saying "Why would you buy an Nvidia chipset when there is the X58." On the 775 lga the Nvidia 790i Ultra was/is the best 775 chipset. If they made a chipset that worked with Nehalem and was better I would buy it.

[Arkangyl]

I'll agree that nVidia has pulled some :banana::banana::banana::banana: lately (mostly just that 8xxx, 9xxx and 2xx lines are all based on the same die concept)

HOWEVER, there ought be no reason why we don't want another chipset maker, competition = better products & lower prices (I love my x58, but hell, there's def a place in the i7 market for a $100-150 motherboard nVidia could fill, not to mention a x58 competitor would always be good to have.

[SNiiPE_DoGG]

On the 775 lga the Nvidia 790i Ultra was/is the best 775 chipset.

:rofl::rofl::rofl::rofl::rofl::rofl::rofl: :ROTF::ROTF::ROTF::ROTF::ROTF::ROTF: :rofl::rofl::rofl::rofl::rofl::rofl: :ROTF::ROTF::ROTF::ROTF::ROTF: :rofl::rofl::rofl::rofl: :ROTF::ROTF::ROTF::ROTF::ROTF: :rofl::rofl::rofl::rofl:

wait you were serious?:shrug:

[onewingedangel]

I'll agree that nVidia has pulled some :banana::banana::banana::banana: lately (mostly just that 8xxx, 9xxx and 2xx lines are all based on the same die concept)

HOWEVER, there ought be no reason why we don't want another chipset maker, competition = better products & lower prices (I love my x58, but hell, there's def a place in the i7 market for a $100-150 motherboard nVidia could fill, not to mention a x58 competitor would always be good to have.

It's not as though Nvidia couldn't licence the QPI interface, they just don't want to pay for it.

[Donnie27]

I'll agree that nVidia has pulled some :banana::banana::banana::banana: lately (mostly just that 8xxx, 9xxx and 2xx lines are all based on the same die concept)

HOWEVER, there ought be no reason why we don't want another chipset maker, competition = better products & lower prices (I love my x58, but hell, there's def a place in the i7 market for a $100-150 motherboard nVidia could fill, not to mention a x58 competitor would always be good to have.

I'd like another chipset maker as well. I wish Intel would allow an AMD based Nehalem boards, I liked the last one I saw with a X4-9850 on it:up: I'd like to see nVidia stop acting like something thats halfway between Sony and Donald Trump. I saw Intel being called greedy while greedy nVidia is given a free pass. I'd call it Greedy nVidia giving Greedy Intel an even bigger excuse to be Greedier.

IMHO (and that only) If there are no NF200 or Magic Cookie, there is no dispute. With a new driver patch nVidia will be able to disable SLI on even those boards with Hacked BIOS'.

[16floz470ml]

:rofl::rofl::rofl::rofl::rofl::rofl::rofl: :ROTF::ROTF::ROTF::ROTF::ROTF::ROTF: :rofl::rofl::rofl::rofl::rofl::rofl: :ROTF::ROTF::ROTF::ROTF::ROTF: :rofl::rofl::rofl::rofl: :ROTF::ROTF::ROTF::ROTF::ROTF: :rofl::rofl::rofl::rofl:

wait you were serious?:shrug:

It is not my fault if you could not work your board.

[alucasa]

It's not as though Nvidia couldn't licence the QPI interface, they just don't want to pay for it.

Yep, that's what it looks like since they are claiming that license for FSB extends for QPI.

I guess that, in Nvidia's viewpoint, they are thinking that Intel simply renamed FSB to QPI, you know, like what they do to their products. :p

[SNiiPE_DoGG]

It is not my fault if you could not work your board.

I never had a 790i but I had a 680 and let me tell you, that was PLENTY of nvidia crap for me.

seriously though, in what ways was the 790i the best board for 775? I can think of none.

[16floz470ml]

Well right off the bat it has 3 pcie x 16 slots. Your motherboard for example only has two.

[Vozer]

http://www.dailytech.com/Intel+Sues+NVIDIA+Over+Chipset+Manufacturing/article14322.htm


Intel has filed suit against Nvidia seeking a declaratory judgment over rights associated with two agreements between the companies. The suit seeks to have the court declare that Nvidia is not licensed to produce chipsets that are compatible with any Intel processor that has integrated memory controller functionality, such as Intel's Nehalem microprocessors and that Nvidia has breached the agreement with Intel by falsely claiming that it is licensed. Intel has been in discussions with Nvidia for more than a year attempting to resolve the matter but unfortunately we were unsuccessful. As a result Intel is asking the court to resolve this dispute.

Nvidia believes that our bus license with Intel clearly enables us to build chipsets for Intel CPUs with integrated memory controllers. We are aggressively developing new products for Intel's current front side bus (MCP79 and MCP89) and for Intel’s future bus, DMI.
Somehow I still think in addition to Bloomfield, this time Intel is also trying to prevent Nvidia from coming near their future Lynnfield and Clarkdale CPUs. :p:

[16floz470ml]

Maybe they need to hire new translators or something. WTF?

[D-Cyph3r]

Well right off the bat it has 3 pcie x 16 slots. Your motherboard for example only has two.

Yeah... because Tri-SLI is far more useful than... lets say, proper memory support or a northbridge that doesn't need to be watercooled to overclock to any decent levels.


X48 was/is the best chipset for S775, with P45, X38, P35 and P965 taking the rest of the top 5.

[jasonelmore]

Maybe, I don't know what they'd do either way if nVidia hadn't pulled what they did. Here some more.

http://www.custompc.co.uk/news/604731/sli-can-run-on-any-x58-board-even-without-nf200-chip.html



So Intel would pay nVidia for the "Cookie" (on their boards) while nVidia Paid Intel NOTHING:rofl: Certified my eye LOL!

So yes shintai, folks should wake up! Business isn't carried on like this. Intel-AMD/ATI-IBM and others have always done it the right way. It's goes like; you pay me, I pay you or I get it free if you get it free! nVidia wants; You pay me for mine but I get your tech free.;)


Intel is'nt the one paying for the SLI Capability. Individual Motherboard manufacturers are the one's that are being faced with the choice of putting SLI support on there board or leave it off. Reportately it's costing companies like asus and gigabyte, $5 per manufactured board to carry the SLI code.

[GAR]

I think the P45 was the best chipset for S775. X48 was a let down. 790i was a good chipset as well, id say top 5 for sure. It was the S775 gaming king because of SLI and 3 way SLI.

[16floz470ml]

Yeah... because Tri-SLI is far more useful than... lets say, proper memory support or a northbridge that doesn't need to be watercooled to overclock to any decent levels.


X48 was/is the best chipset for S775, with P45, X38, P35 and P965 taking the rest of the top 5.

Right now, not what my friend told me or what I read on a forum, I have a q9450 and XFX 790i ultra. On all air it is 1:1 at 3.8 ghz with 8gb of memory and is prime 95 stable. All I had to do was set the memory voltage and cpu voltage. All other setting on auto. Too easy. I have not had any problems with the board and I got it in April 08.

The only regret I have with this board is using OCZ memory at first. OCZ memory and their support was a complete waste of time. Once I changed to Crucial that solved any issues.

[Rammsteiner]

Meh, I dont know what's in the contract, but as long as there's nothing stated about memory controllers integrated or external regarding CPU's, I dont think Intel has much to stand on.

Although Ive read several complaints about nVidia chipsets for the Intel platform, they're damn good on AMD platform. I dont know where the issue is on the Intel platform, but if they somehow get that right (for example no need for a memory controller), they would have to offer something good over X58 since that does offer SLI and CFX now. So for example good ref. clocking, improved mGPU performance, hybrid SLI solutions etc.

Hard thing to make nForce worth it over X58, but if they get it working they've quite a huge market. At least eVGA, XFX etc wouldnt think twice to release pure nForce boards, maybe even drop X58 boards. However, if they dont get it working, which is likely since all those things are simply hard to to accomplish, I wouldnt even bother. Or maybe if they would be able to release the same features for 100$ less:rolleyes:

But regarding this whole thing... whatever. There's always something to crap about between them. nVidia made :banana::banana::banana::banana: ups, but so did Intel. Im still convinced they would have sold less X58's if there was no SLI support since I sense a stronger nVidia preferance over any CPU preferance, could be me though.

[twilyth]

How did nvidia get a license to begin with? I assume it was voluntary on Intel's part rather than the long legal battle that resulted in AMD's license.

So if it were voluntary, then there would have to be provisions in there regarding future developments. Either that or a blanket license to Intel tech with no specifics.

Personally, I don't know why nvidia would have agreed to something they they knew would have to be renegotiated. So either intel forced this on them or they thought they might have some advantage over intel when the time came.

I don't know, but there's got to be a lot going on behind the scenes.

[ToTTenTranz]

This would render the Ion platform unable to work with future Atom CPUs with integrated memory controller, like Pineview.

If Intel is right, Ion on Atom had it's days counted before it was born. Atom N270/80 and N330 would be the only Ion-compatible CPUs.

I guess nVidia will run back to VIA with its ass between the legs, asking for forgiveness. And if VIA is already working on a Chrome 500-based IGP, it's a no-go for nVidia.

[v_rr]

Looks like that AMD and Nvidia have different intrepertations of what the contract is.

Nvidia has to focus on GPU and mobile phone only. Focus on Chipset is a no go for AMD and Intel. In Q4 2008 they droped 51% of earnings on chip7 part.

[Donnie27]

Intel is'nt the one paying for the SLI Capability. Individual Motherboard manufacturers are the one's that are being faced with the choice of putting SLI support on there board or leave it off. Reportately it's costing companies like asus and gigabyte, $5 per manufactured board to carry the SLI code.

Intel is paying when we're talking Intel Branded motherboards. I'm not talking about 3rd party boards.

[Firestrider]

Intel is paying when we're talking Intel Branded motherboards. I'm not talking about 3rd party boards.

I thought Intel-branded motherboards were actually manufactured by Foxconn..?

[highoctane]

I'm no fan of Nvidia based chipsets for intel cpus personally but I think Nvidia should be allowed to fail miserably, if thats the case, trying to produce a product rather than being locked out all together.

I really don't know the dealings behind the scenes as far as licensing agreements but I'm sure we all have an idea of how Intel would like things to be.

If Intel simply refuses to license Nvidia to make chipsets for i7 they have simply locked them out, there may not even be an issue with charging a fee at all.

Anyways I guess we'll have to wait and see what a judge decides on the matter or things settle out of court.

[Donnie27]

I thought Intel-branded motherboards were actually manufactured by Foxconn..?

Foxconn used to make them, I'm not sure who makes them now. Yet those costs are passed to Intel no matter who their subcontractor might be. Again Mutual Agreements are based on if License fees are used or not. AMD and Intel and IBM have that arrangement. Once any of them start charging, the other will respond in kind. This has saved both AMD and Intel Millions in Court Costs.

A OWA pointed out, nothings stopping nVidia from paying for a license.

[jt1]

Intel's large market share of CPU's sold reminds me of the kind of monopoly that the Windows OS holds. For example, Windows can package a browser (Internet Explorer) with its OS to get market share in the browser sector. Similarly, Intel can do this with chipsets. Just change FSB/QPI enough to be considered "different", and Intel can hold exclusive rights on the next chipset.

[DMH]

Intel's large market share of CPU's sold reminds me of the kind of monopoly that the Windows OS holds. For example, Windows can package a browser (Internet Explorer) with its OS to get market share in the browser sector. Similarly, Intel can do this with chipsets. Just change FSB/QPI enough to be considered "different", and Intel can hold exclusive rights on the next chipset.
The point is: QPI is different from FSB,you have license for FSb if FSB is different from QPI then your license doesn't extend to the QPI technology

[Vozer]

A OWA pointed out, nothings stopping nVidia from paying for a license.
But haven't them discussed this matter for more than a year without any success? It's not that easy. I think by insisting on "integrated memory controller functionality", Intel is planning to kick Nvidia out of their chipset market forever.

[XSAlliN]

It's nVidia's fault for annoying Intel, I've seen some topics around here where they attacked Intel's image. Guess Intel wasn't amused with their statements.

[Donnie27]

But haven't them discussed this matter for more than a year without any success? It's not that easy. I think by insisting on "integrated memory controller functionality", Intel is planning to kick Nvidia out of their chipset market forever.

Yes. I DON'T KNOW why the talks failed and all I have to go on are nVidia's past actions. Intel had agreed with nVidia and Intel wasn't the one who added the "Cookie" or the "NF200". This was after Intel have given them the go-ahead for nVidia based Intel motherboards FREE. It is kind of funny here that some folks have forgotten that nVidia got not only FSB but all of the standards FREE!

This has the largest affect on our DIYer market until you start talking about Atom. I don't think Intel wants Fanboys who love nVidia Cards to go with AMD Chipsets exclusively. This fight will be good for AMD. More nVidia Fanboys will lean toward an AMD platform. Most Intel Fanboys will dump nVidia and go with AMD video cards:up: Key, I said most-more, not all. Either way, AMD comes out smelling like a Rose.

Instead of just i5, I'm thinking it might have more to do with Atom. nVidia proposes taking a cheap slow CPU to run the system and having the Video Card provide the muscle. Intel and even AMD will loose money if nVidia could pull that off. Intel can't block x86 that nVidia has a license for but with no BUS, that becomes moot.;)

[Ozzfest05]

ahhh the economy goes down the drain and out comes a fresh lawsuit

[[XC] Oj101]

With the AMD\ATi merger you'd think that Intel and nVidia would be closer together, not trading blows :shrug:

[Donnie27]

With the AMD\ATi merger you'd think that Intel and nVidia would be closer together, not trading blows :shrug:

Yepp!:yepp:

[jt1]

The point is: QPI is different from FSB,you have license for FSb if FSB is different from QPI then your license doesn't extend to the QPI technology
You missed my point. I'm concerned about Intel using its CPU marketshare to dominate the chipset marketshare. When monopolistic behavior goes unchecked, we all lose.

[FischOderAal]

Somehow I still think in addition to Bloomfield, this time Intel is also trying to prevent Nvidia from coming near their future Lynnfield and Clarkdale CPUs. :p:

There's not much to earn with southbridges. More and more will go On-die. So what's left? For instance PATA, SATA and USB...


I guess nVidia will run back to VIA with its ass between the legs, asking for forgiveness. And if VIA is already working on a Chrome 500-based IGP, it's a no-go for nVidia.

Well, not the first time that a company blocks it's own ways with past decisions. NVIDIA needs a x86 chip for the future or all what's left will be graphic cards for enthusiasts...

[Vozer]

Has anyone seen Larrabee involved in all of this? :D IMHO, Intel's ultimate goal behind these actions is to weaken Nvidia and lessen their influence, making it easier to dethrone them in the GPU market later.

[BeepBeep2]

Hey, nomatter who wins...Nvidia could choose to stop Intel's SLi licence...
Intel - SLi = Poo.
Yes, X58(?) still has ATI XFire but what intel enthusiast wants parts that have AMD silkscreened to the GPU PCB for their lovely Intel rig?

Intel is kinda retarded for doing this.
Nvidia's sales will now decline and Intel's will too. We will soon see nvidia halting chipset's to become GPU only, and SLi will go where? Maybe AMD!

[Donnie27]

Has anyone seen Larrabee involved in all of this? :D IMHO, Intel's ultimate goal behind these actions is to weaken Nvidia and lessen their influence, making it easier to dethrone them in the GPU market later.

That wouldn't work if nVidia's not charging for NF200 and the "Cookie". Plus other moves that created bad blood that was mentioned here:D

[Ket]

I hope intel win. QPI is not FSB. nVidia are at their age old tricks again. Who cares if NV go down? I sure as hell don't. Not with intel planning to make serious GPUs to rival what ATi and NV do. However I don't agree with intel having such a monopoly with CPUs/chipsets AND (eventually) GPUs. With a big player like that, it almost makes you hope M$ step up and start making their own GPUs...

[FischOderAal]

Hey, nomatter who wins...Nvidia could choose to stop Intel's SLi licence...
Intel - SLi = Poo.
Yes, X58(?) still has ATI XFire but what intel enthusiast wants parts that have AMD silkscreened to the GPU PCB for their lovely Intel rig?

Intel is kinda retarded for doing this.
Nvidia's sales will now decline and Intel's will too. We will soon see nvidia halting chipset's to become GPU only, and SLi will go where? Maybe AMD!

I don't think Intel forgot to make a contract with NVIDIA ;) Of course there could be some exit-clauses in it but I don't think Intel is that stupid to risk losing SLI.


However I don't agree with intel having such a monopoly with CPUs/chipsets AND (eventually) GPUs. With a big player like that, it almost makes you hope M$ step up and start making their own GPUs...

Nah. Someone with bigger pockets. Samsung anyone :D

[SNiiPE_DoGG]

If anyone is going to make 3rd party i# chipsets I would wan't VIA back in the game :D

@ post above mine: you think intel is risking losing SLI? :rofl: nvidia NEEDS intel to have SLI, otherwise they are gonna lose more sales than even their deep pockets can handle.

[highoctane]

SLI or Crossfire is still a tiny minority of the market really, single card dual gpu solutions offer the same functionality without needling a fancy mb.

So in that regard its not like nvidia would be loosing allot of sales considering single gpu's are the vast majority anyways.

EDIT: But nvidia shoots itself in the foot for not making SLI more accessible like Crossfire. Still not many will go dual gpu but just having the option there for more people means theres a good chance more people might try it out just because the ability to do it is already in place.

[SNiiPE_DoGG]

without SLI capabilities on a platform nvidia couldn't claim the performance crown as easily. That said SO many people buy a card saying "oh, well I'll add another later" even if they do or not its still a selling point for their products that once again they would lose.

[BababooeyHTJ]

Nvidia needs Intel more than Intel needs Nvidia, not that Intel needs Nvidia at all. Why should Intel give Nvidia their tech for free while Nvidia wants to charge Intel? At the end of the day to your average consumer that has problems with their pc due to a crappy Nvidia chipset Intel is at fault. Screw Nvidia and their buisness tactics. I got burned on a 780i and Im glad no one will have to go through what I did again.

[Chruschef]

... seems like intel is pushing nvidia away, as they get closer to releasing their own GPU solution(larabee, integrated video processor on their 32nm westmere)

[dartai7]

If Intel wins, does that mean that Nvidia will be forced to allow SLI on all intel chipsets ?

[Jamesrt2004]

If Intel wins, does that mean that Nvidia will be forced to allow SLI on all intel chipsets ?

nope, means nVidia cant make chipsets for i7

[YukonTrooper]

Jen-Hsun is desperate.

[Zucker2k]

Meh, I dont know what's in the contract, but as long as there's nothing stated about memory controllers integrated or external regarding CPU's, I dont think Intel has much to stand on.

Although Ive read several complaints about nVidia chipsets for the Intel platform, they're damn good on AMD platform. I dont know where the issue is on the Intel platform, but if they somehow get that right (for example no need for a memory controller), they would have to offer something good over X58 since that does offer SLI and CFX now. So for example good ref. clocking, improved mGPU performance, hybrid SLI solutions etc.

Hard thing to make nForce worth it over X58, but if they get it working they've quite a huge market. At least eVGA, XFX etc wouldnt think twice to release pure nForce boards, maybe even drop X58 boards. However, if they dont get it working, which is likely since all those things are simply hard to to accomplish, I wouldnt even bother. Or maybe if they would be able to release the same features for 100$ less:rolleyes:

But regarding this whole thing... whatever. There's always something to crap about between them. nVidia made :banana::banana::banana::banana: ups, but so did Intel. Im still convinced they would have sold less X58's if there was no SLI support since I sense a stronger nVidia preferance over any CPU preferance, could be me though.It is definitely you; what percentage of the market cares for SLI? 0.0001?

Edit: Most are missing the point here; this is no longer about SLI; that is old tech as far as Intel is concerned. The future is netbooks and mainstream desktop offerings with imc. More than anything else, Intel's atom is set to dominanate the netbook market and Nvidia knows this and have become desperate. Of course Intel could reap all the loot or "invite" Nvidia to have a share, your guess is as good as mine what path Intel is choosing. For Intel the way out is that they're arguing a platform or bundle instead of selling the cpus as a unit. What this means is that each atom cpu is sold bundled with a chipset. This effectively locks everybody to the platform since none have a license from Intel to pair the cpus with another chipset. In effect; Intel sells cpu, chipset, igp, etc. A whole load of cash. It's a brilliant move. Ironically, Nvidia is right, the cpu is dead, it's replaced by the platform.

[ToTTenTranz]

With the AMD\ATi merger you'd think that Intel and nVidia would be closer together, not trading blows :shrug:

And they would.. if Intel wasn't planning on fighting nVidia head-on in the future, with Larrabee.

Intel plans things with a lot of anticipation. Remember when they sold the ARM-based XScale CPUs to Marvell back in 2006? That's because they planned that 3 years later they would have to fight ARM CPUs with their own low-power x86.


For Intel, nVidia is no longer a pleasant symbiot, it's a dreadful parasite.

[GAR]

Intel fears nvidia, thats all. After all, a GPU is way more advanced piece of hardware than a cpu is.

[CmB]

Intel fears nvidia, thats all. After all, a GPU is way more advanced piece of hardware than a cpu is.

Interesting point of view. Would you mind elaborating? After all, the prevailing opinion seems to be that nvidia should fear Intel. No sarcasm in this post, just a question.

[Shintai]

Intel fears nvidia, thats all. After all, a GPU is way more advanced piece of hardware than a cpu is.

No..the GPU is multifold more simple than a CPU. Its like comparing a toy car with a real car. GPUs are simple in construction and contains alot of "copy/paste" logic. It takes about 4-5 years to develop a CPU with 1000s of people. A GPU is usually made in 1-1½ years with 100s of people. The record for a GPU is about 3 years I think. With completely new architecture. G80 costed about 475mio$ to develop according to nVidia. K10, K8, Core 2, i7 etc is multi billion$ research pieces.

[ToTTenTranz]

Intel fears nvidia, thats all. After all, a GPU is way more advanced piece of hardware than a cpu is.
:rolleyes:

Honestly, if you're not ever going to have an opinion on your own, you may as well just quote your god Jen-Hsun Huang with bold characters and be done with it.

[FischOderAal]

Maybe he forgot to make [irony]-tags? :shrug:

[zanzabar]

With the AMD\ATi merger you'd think that Intel and nVidia would be closer together, not trading blows :shrug:

thats what NV wanted but NV screwd intel with the crap that they put out and it was an embarrassment for intel. and any1 saying that NV makes good chipsets on amd, they always made the lowest tier, u had via/s3 and amd/ati above them in every generation of socketA, 754, and 939 then in am2 NV payed amd to not relies a chipset affter the ati merger and look what that gave them.

and for the 16x 3 way, why would u even want that using a routing chip and only having 36 usable lanes like the x38/48 and just wiring them electrically 16x wastes space on the pcb and makes things do more than they need for a little more synthetic bandwidth but more latency, and if u want to go like that u can get an asus board with x48 and 64 lanes with duel plx chips, with routing u can make as amny pipes as u want but u cant use more bandwidth than the IO so its pointless

[ToTTenTranz]

thats what NV wanted but NV screwd intel with the crap that they put out and it was an embarrassment for intel.

Big decisions in big companies like Intel, nVidia and AMD are never made with revenge in mind.


We all joke and comment about the "can of whoop ass" episode, but don't think even for a second that Intel's decisions have ever had anything to do with that. There's too much at stake, so people who make the decisions are all "grown-ups". Revenge, pride, justice and other terms only exist in the PR department.

When Jen-Hsun Huang made that statement, the only thing he accomplished was to reduce himself to an idiot among neutral observers (and to a "cool guy" among his fanboys).

[perkam]

Nvidia is quickly becoming the underdog in this industry with AMD and Intel both wanting a common platform for their CPU and Motherboard product lines.

ATI was right when it told Nvidia that the writing is on the wall back when it was acquired by AMD....................BUT........................ ..

Nvidia is smarter than most companies and even in times of desperation, it can decide to sell off non-essential components of its business to stay profitable, if it comes to that.

Perkam

[ownage]

No..the GPU is multifold more simple than a CPU. Its like comparing a toy car with a real car. GPUs are simple in construction and contains alot of "copy/paste" logic. It takes about 4-5 years to develop a CPU with 1000s of people. A GPU is usually made in 1-1½ years with 100s of people. The record for a GPU is about 3 years I think. With completely new architecture. G80 costed about 475mio$ to develop according to nVidia. K10, K8, Core 2, i7 etc is multi billion$ research pieces.

I think its more then 3 years. G80 started back in 2001 (but with some breaks).
RV770 was started in 2005. R600 in 2002 if my memory doesn't play tricks with me.
But indeed with lot less people to develop.

[zanzabar]

intel wants to hold their market share (thats the point of the biz model for a duopoly) and keep oems happy with low rma costs, if they let some1 make an inferior product that courses problems then they wont give them a new license. so if some1 lowers your cpu sales then u cut them off so u can protect your core market

[ToTTenTranz]

Nvidia is smarter than most companies and even in times of desperation, it can decide to sell off non-essential components of its business to stay profitable, if it comes to that.

Perkam

And what would they sell and to whom?

Although they do have the best platform for low-power x86 IMO, their chipset business is about to die.

All the GPU computing stuff is directly related to their GPU hardware, so they cannot sell it without selling their GPUs business (which is what? 85% of current nVidia personnel?).

The only thing they can sell is Tegra and Goforce, but that's really their last way out to remain profitable when they loose all the low-end GPU business to AMD and Intel IGPs.

[rozzyroz]

nvidia is in a bad spot right now... unles they offer a product that is lightyears ahead of either intel or amd's graphics solutions, they will fade away from the industry.

nvidia is crying foul or whatever, but the truth of the matter is that when amd bought ati, the future of the industry was taken to a new direction. the platform direction that amd/ati were talking about when the joned together will be how buisiness will be done in the future.

amd already has their platform thing going on, but its a matter of time before larabee comes to market and intel is in the same boat. if nvidia had any insight, they would consider joining intel, or face fading away from the industry.

[crash5s]

Big decisions in big companies like Intel, nVidia and AMD are never made with revenge in mind.


We all joke and comment about the "can of whoop ass" episode, but don't think even for a second that Intel's decisions have ever had anything to do with that. There's too much at stake, so people who make the decisions are all "grown-ups". Revenge, pride, justice and other terms only exist in the PR department.

When Jen-Hsun Huang made that statement, the only thing he accomplished was to reduce himself to an idiot among neutral observers (and to a "cool guy" among his fanboys).

I don't think that intel is out for revenge, but Huang certainly did make a jackass out of himself and has been goading intel with a stick.

Really though though only reason for an nvidia chipset was SLI. I'd feel better for nvidia if they could make a working chipset, but I'm not going to hold my breath.

[ToTTenTranz]

if nvidia had any insight, they would consider joining intel, or face fading away from the industry.

Or merging with VIA.
With the emergence of netbooks and increased importance of low-power CPUs, I do see VIA as a potential major player in this future CPU+platform market.
The Nano CPU is excellent for netbooks and, paired with Ion, it would blow away every Atom platform we've seen in Intel roadmaps so far.

And I still don't really get why nVidia and VIA broke up their "marriage" a while ago. Maybe VIA got confident of their S3 GPUs, or nVidia thought that pairing with Atom would get them a lot more OEM design wins.

[crash5s]

Or merging with VIA.
With the emergence of netbooks and increased importance of low-power CPUs, I do see VIA as a potential major player in this future CPU+platform market.
The Nano CPU is excellent for netbooks and, paired with Ion, it would blow away every Atom platform we've seen in Intel roadmaps so far.

And I still don't really get why nVidia and VIA broke up their "marriage" a while ago. Maybe VIA got confident of their S3 GPUs, or nVidia thought that pairing with Atom would get them a lot more OEM design wins.

Do we know that chipset isn't bugged though? I mean after 680i, 790i, g92's failing, laptop problems, ect, do you really think that Ion won't be plagued with problems?

nvidia seems to be imploding lately. The can of whoop-ass moment just made it more funny then tragic.

[ToTTenTranz]

Do we know that chipset isn't bugged though? I mean after 680i, 790i, g92's failing, laptop problems, ect, do you really think that Ion won't be plagued with problems?

nvidia seems to be imploding lately. The can of whoop-ass moment just made it more funny then tragic.

We know the chipset isn't bugged because it's being used in millions of Macbooks and Macbook Pros, with no known major problems. Their low-cost Geforce 8200 chipset didn't have any problems either AFAIK.

[ownage]

Good or bad chipsets, either way it doesn't say much nVidia sells allot (or not) or has a healthy future in the 'chipset' market.

[zanzabar]

We know the chipset isn't bugged because it's being used in millions of Macbooks and Macbook Pros, with no known major problems. Their low-cost Geforce 8200 chipset didn't have any problems either AFAIK.

u apparently havnt a laptop fail on u for no apparent reason and rma 6 times with HP from an NV IGP

and apple is replacing the mac book with amd gpu and intel chipsets in q3, when the NV contract expires. they are only using it now becouse they have to

and ive already seen 3 dead mac books with the new NV chipset so go figure

[zsamz_]

u apparently havnt one fail on u for no apparent reason and rma 6 times with HP from an NV IGP

and apple is replacing the mac book with amd gpu and intel chipsets in q3, when the NV contract expires. they are only using it now becouse they have to

and ive already seen 3 dead mac books with the new NV chipset so go figure

i had a fever today n i almost bought a 790i :rofl:
came back to earth after i remembered my 680i board :rofl: it died n took my sticks with it

[alpha0ne]

Jen-Hsun is desperate.

Desperate to keep his job more likely :yepp: :rofl:

Nv only need a new CEO/director with a little more maturity to solve most of their current problems though in the long term Nv deciding to enter the chipset business could be their undoing

[Donnie27]

SLI or Crossfire is still a tiny minority of the market really, single card dual gpu solutions offer the same functionality without needling a fancy mb.

So in that regard its not like nvidia would be loosing allot of sales considering single gpu's are the vast majority anyways.

EDIT: But nvidia shoots itself in the foot for not making SLI more accessible like Crossfire. Still not many will go dual gpu but just having the option there for more people means theres a good chance more people might try it out just because the ability to do it is already in place.

Tiny yes, but you're still talking about a few million boards. Not exactly Chump Change when DIY market are expected to be small vs the overall market. nVidia's problem is they are full of yes men and no one has the ballz to tell the Boss when to step back and not be so aggressive.

Yet folks talk about Steve Jobs having a big ego:rofl:

Yes guys I've been following the ION and Atom arguments.

http://www.xbitlabs.com/news/mobile/display/20081217162305_Nvidia_Launches_GeForce_9400M_Chips et_Ion_Platform_for_Intel_Atom.html


According to media reports, Intel sells its Atom processors bundled with supporting chipsets, hence, the market of supporting third-party core-logic is absent these days. Additionally, it is unclear whether Nvidia has legal rights to produce and sell Intel Atom-compatible chipsets.

Much larger market than some stupid SLI. Greed on SLI blinded nVidia and they couldn't see where the real profits would be.




The Nvidia Ion platform combines the GeForce 9400 mGPU with the Intel Atom CPU. That combination would be only possible if Intel started selling Atom processors without the Intel chipset (945GSE and 945GC). However, Intel replied to hardware makers that this wouldn't happen, at least anytime near soon. DigiTimes got an Intel comment that "...We have no plans to validate the Nvidia MCP79 chipset on Atom-based nettop or netbook platforms. Intel also has no plans to form a partnership with Nvidia to support nettop or netbook platforms based on the Intel Atom CPU,...". Nvidia is expected to respond soon, even there are many people who believe that such platform would be seen at the market...Only time will tell if Nvidia will partner with Intel.

[Greg83]

i had a fever today n i almost bought a 790i :rofl:
came back to earth after i remembered my 680i board :rofl: it died n took my sticks with it

:rofl: , if nothing else. I hope this results in never having to worry about being tricked by the another 680i.

If I were intel and had purchased a 680i board, to have it kill my ram(s) and then die itself to have to rma and repeat 10 times.
I too would be sueing nvidia for any future attempts at making a motherboard that uses my cpu in it.

[bahlgren342]

I have an nForce 4 board at the moment. Can't wait to burn it after I upgrade. :)

[hollo]

no reason why NV couldn't get motherboards made that have 128/256mb GDDR2/3 soldered to the board along with a GPU (eg 9400) that connects to the existing PCI-e connections when activated in the BIOS, and passes the PCI-e lanes to a PCI-e x16 connector when disabled, basically like the ATI rage chips with their 4 or 8mb vram that you get on server boards

personally i'd buy a board like that over a motherboard with a GPU that accessed system memory in a second. i hope they lose this case just so they'll go ahead and do it!

[saaya]

Why the heck would anyone want an nVidia i7 chipset anyways? With X58 supporting both Crossfire and SLI.
hmmm lets see, lower qpi multipliers for higher bclocks, possibly higher than 4ghz/8GTs qpi clocks, much much better pciE overclocking...

im not saying nvidias chipsets are great, but if intel managed to lock them out then intel owns the whole platform, and they can do whatever they want.

apart from beeing able to force some crazy drm onto us, pushing prices to ridiculous levels, limiting or locking overclocking on at least some segments, the most important thing is: innovation and diversification!

if your the only chipset provider for 90% of the markets cpus, then there is almost no pressure ot diversify or innovate. it doesnt mean intel wont diversify their products and wont innovate to meet all segments demands, but it makes it very likely that sooner or later innovation and diversification will get worse, worse and worse.

[zanzabar]

i would like to see via, or ati. ati wont happen but the r600 was great and they actually unlinked the memory from the FSB without using variable multis something that NV could never do. and via would be great but i doubt that they could pay for a license

[drizzt5]

I never had a 790i but I had a 680 and let me tell you, that was PLENTY of nvidia crap for me.

seriously though, in what ways was the 790i the best board for 775? I can think of none.



Well right off the bat it has 3 pcie x 16 slots. Your motherboard for example only has two.

Your joking....? I can't tell.

[ToTTenTranz]

and apple is replacing the mac book with amd gpu and intel chipsets in q3, when the NV contract expires. they are only using it now becouse they have to


Source?

Also, vanilla macbooks with discrete GPUs? I find that hard to believe..

[bing]

"We are confident that our license, as negotiated, applies," said Jen-Hsun Huang, president and CEO of NVIDIA. "At the heart of this issue is that the CPU has run its course and the soul of the PC is shifting quickly to the GPU. This is clearly an attempt to stifle innovation to protect a decaying CPU business."

Just wondering what is the word "GPU" stands for in Huang's mind ? General Processing Unit ?

How bout the "CPU" ? Prolly for him is Crappy Processing Unit ??

Just can't believe this dude, what a joke !

[eXa]

no reason why NV couldn't get motherboards made that have 128/256mb GDDR2/3 soldered to the board along with a GPU (eg 9400) that connects to the existing PCI-e connections when activated in the BIOS, and passes the PCI-e lanes to a PCI-e x16 connector when disabled, basically like the ATI rage chips with their 4 or 8mb vram that you get on server boards


That sentence is freaking impossible... i get a headache halfway through it.

[zanzabar]

Source?

Also, vanilla macbooks with discrete GPUs? I find that hard to believe..

sorry cant give one, but NV is being dropped. if u look back im like 99% on calling things like this, but sometimes a month or so off

but its going to go like this, apple is switching the newer intel platform with ddr3 sodimm then in the same time frame or just affter they will ofter the 46xx mobile

[Sr7]

sorry cant give one, but NV is being dropped. if u look back im like 99% on calling things like this, but sometimes a month or so off

but its going to go like this, apple is switching the newer intel platform with ddr3 sodimm then in the same time frame or just affter they will ofter the 46xx mobile

sigh... if only you knew.. :shrug:

[Sr7]

No..the GPU is multifold more simple than a CPU. Its like comparing a toy car with a real car. GPUs are simple in construction and contains alot of "copy/paste" logic. It takes about 4-5 years to develop a CPU with 1000s of people. A GPU is usually made in 1-1½ years with 100s of people. The record for a GPU is about 3 years I think. With completely new architecture. G80 costed about 475mio$ to develop according to nVidia. K10, K8, Core 2, i7 etc is multi billion$ research pieces.

This statement more than any other makes it completely clear that you make things up when you don't know the answer.

The GPU is just as complex if not moreso to design. It does *not* take 1-1/12 years for a new architecture. It takes 3-4 years. It also cost $1 billion for developing G80, not 500 million.

Additionally, 70% or so of the CPU die area is cache, not fundamental logic. This is contrasted with the GPUs relatively high % of logic vs. cache.

If you want to talk copy-paste logic you can probably look at multi-core on-die in a CPU.

Since it's so cheap, so fast, and so easy to design a GPU according to you, why hasn't Intel just made a directly competitive traditional GPU ASAP to put ATI and NV out of business? Why are they taking the larrabee/x86 route and risking a lot on something that doesn't have an ecosystem to support it? Do you think it will be easy to thread applications 16-64 ways?

They're making an awfully big bet on an unproven technology. Even if it works well, winning over developers to move to a GPU which has no installed base is a whole other struggle on top of that.

[ToTTenTranz]

This statement more than any other makes it completely clear that you make things up when you don't know the answer.

The GPU is just as complex if not moreso to design. It does *not* take 1-1/12 years for a new architecture. It takes 3-4 years. It also cost $1 billion for developing G80, not 500 million.

Additionally, 70% or so of the CPU die area is cache, not fundamental logic. This is contrasted with the GPUs relatively high % of logic vs. cache.


If you want to talk copy-paste logic you can probably look at multi-core on-die in a CPU.

So according to you, every engineer who works in Intel and AMD must be pretty stupid, if they only come up with a new design every 5 or 6 years, they're more in number and spend more money on R&D.

Tell me, what's the difference between each of GT200's 240 SPs? And each of RV770's 40 TMUs? There's no copy/paste in there?





Since it's so cheap, so fast, and so easy to design a GPU according to you, why hasn't Intel just made a directly competitive traditional GPU ASAP to put ATI and NV out of business?

Because they lack:

1 - Engineering experience for 3D-oriented hardware;
2 - Engineering experience for driver development;
3 - Patents for for efficient 3D architectures. nVidia and ATI have been registering those for more than 10 years now.





They're making an awfully big bet on an unproven technology. Even if it works well, winning over developers to move to a GPU which has no installed base is a whole other struggle on top of that.

Thank god for those guys who don't "fear the unproven". If everyone had that, we would still be using the fingers in our hands to make calculations.

Some people just don't get the point of science at all...

[dinos22]

i'm pretty sure they had little support from Intel to develop their chipsets and still managed to do a cracker job....not perfect but considering their improvement over older gen and issues with support nvidia did pretty good with their chipset. It's a real shame that 790i had some issues from previous gen otherwise as it shown amazing improvement to previous gen particularly overclocking and clock for clock performance wise against Intel chipsets. i never thought i'd see the day nvidia beat intel boards in SuperPi efficiency for example. That takes some serious skill on their part.

i really hope intel lose this lawsuit for everyone's sake because once we start losing that competition we will LOSE......WE WILL LOSE......that is correct. Remember this, companies dont innovate as much when they have a big advantage, they get lazy and :banana::banana::banana::banana:y and just work their old tech to death until someones flies past them in a huge way....all manufacturers have been in this position including the three companies being discussed here.

Why do you guys think ATI sells their cards and CPUs at these prices now.......do you remember the days of 9800 cards and K8.......FX CPUs etc......how the time changed hey ;)

well kiss all that goodbuy if you want Nvidia to die. That's an incredibly foolish thing to say, incredibly foolish. I dont think some people are stopping to think before they post here.

[ownage]

How little support did they have?
I can remember how Intel screwed DFI in the past by giving them 2 chipsets to develop there boards while ASUS received 100.
If nVidia also get screwed like that, then I can't blame nVidia for the quality.
As as customer that doesn't change much, I won't buy anything bad. That's how things work.

[Shintai]

The GPU is just as complex if not moreso to design. It does *not* take 1-1/12 years for a new architecture. It takes 3-4 years. It also cost $1 billion for developing G80, not 500 million.


Nope, 475million.

http://www.anandtech.com/video/showdoc.aspx?i=2870&p=5

And a CPU is multitude times more complex in instruction handing, logic complexity, validation etc. And your cache arguement doesnt hold water. Your multicore is like shooting yourself in the foot. Each TMU, shader etc is a pure copy of the other, and there tend to be alot of that in GPUs.

Just because you make a GPU with x amount of transistors doesnt make it complex in any way besides manufactoring.

[Mr Roboto]

Maybe I'm missing something here but what's the big deal? This obviously isn't over the Atom\Ion platform right? Does this mean Nvidia are getting ready to release an i7 based chipset?

[Linux_Box]

Competition is normally a good thing for consumers. The move by Intel doesn't appear to be a good thing in the long run for consumers. Hopefully Intel will prove me wrong on this!! :)

[ToTTenTranz]

Maybe I'm missing something here but what's the big deal? This obviously isn't over the Atom\Ion platform right? Does this mean Nvidia are getting ready to release an i7 based chipset?

What you're missing is that the lawsuit isn't because of i7 only. It's about every Intel CPU with a memory controller, and i7 just happens to be the first.
The next CPU with IMC in schedule is Atom Pineview, hence the Ion discussion.

[Donnie27]

What you're missing is that the lawsuit isn't because of i7 only. It's about every Intel CPU with a memory controller, and i7 just happens to be the first.
The next CPU with IMC in schedule is Atom Pineview, hence the Ion discussion.

Thank you!



Competition is normally a good thing for consumers. The move by Intel doesn't appear to be a good thing in the long run for consumers. Hopefully Intel will prove me wrong on this!!


No, that's not true at all. No one's stopping nVidia from paying for license. If Intel was shutting nVidia out by keeping them from PAYING like any one else, then all the talk of squashing competition might be valid.

If I'd invested Billions in MY IP, I'd want to collect license fees to recoup some of my expenditures as well. No matter what Socialist and Communist say, that's what truly drives innovation. To continue to grow, Intel has to compete with itself.

[Mr Roboto]

What you're missing is that the lawsuit isn't because of i7 only. It's about every Intel CPU with a memory controller, and i7 just happens to be the first.
The next CPU with IMC in schedule is Atom Pineview, hence the Ion discussion.

OK I see thanks. I didn't think Nvidia was releasing anymore Intel based chipsets.

Just saw this at VR-Zone


NVIDIA has responded to a Monday court filing in which Intel alleged that the 4-year-old chipset license agreement the companies signed does not extend to Intel’s future generation CPUs with “integrated” memory controllers, such as Nehalem. Nvidia's CEO Jen-Hsun Huang reiterated that they are confident that their license, as negotiated, applies and that this is clearly an attempt to stifle innovation to protect a decaying CPU business.

Update #1 (19th Feb) : http://forums.vr-zone.com/news-around-web/394487/intel-court-filing-update.html

Update #2 (19th Feb) : The News has managed to scan a copy of the court filing.

Update #3 (19th Feb) : It looks like a reenact of the saga between Intel and VIA where Intel tried to block VIA on having the Pentium 4 FSB license back in 2001. As a result, many board makers were reluctant to make motherboards based on VIA chipsets, fearing of getting involved in the lawsuit. Therefore, we asked the same question if the motherboard makers will make boards based on Nvidia's Nehalem chipset. Most makers replied that they are waiting for further instructions from both sides and one of the makers is reluctant citing poor support from Nvidia as one of the reason.

Poor support from Nvidia you say, eh?

[SNiiPE_DoGG]

Poor support from Nvidia you say, eh?

that would be foxconn IIRC, they scrapped the dreadnought citing that the 790 chipset was so awful they couldn't work with it any more.

[tam2]

Dumb question here, does a company that owns a X license for A part allowed to ban other company that make supporting tech for that A part?
The way i see it, if it does allowed, then its possible in the future we will have an all intel parts from gfx to usb stick. Banning other chipset other than theirs is just a beginning.

-tam2-

[Mr Roboto]

that would be foxconn IIRC, they scrapped the dreadnought citing that the 790 chipset was so awful they couldn't work with it any more.
Yeah once my 680i was up and running it was great but when there were problems it was a nightmare. These chipsets are SO quirky and just have really weird issues unlike anything I've ever experienced. I imagine the 790i is similar since it's just a derivative of these early chipsets.

I remember hearing about Foxconn ditching the 790i. Gigabyte pulled their in house designed 680i because of these quirks. Most of the other big motherboard companies made sure NOT to follow Nvidia's reference design. Asus seems to be the only ones who created a successful in house version.

Too bad because with some work they could have been really good, although take SLI out of the equation and Intel based chipsets are clearly better.

[Donnie27]

Yeah once my 680i was up and running it was great but when there were problems it was a nightmare. These chipsets are SO quirky and just have really weird issues unlike anything I've ever experienced. I imagine the 790i is similar since it's just a derivative of these early chipsets.

I remember hearing about Foxconn ditching the 790i. Gigabyte pulled their in house designed 680i because of these quirks. Most of the other big motherboard companies made sure NOT to follow Nvidia's reference design. Asus seems to be the only ones who created a successful in house version.

Too bad because with some work they could have been really good, although take SLI out of the equation and Intel based chipsets are clearly better.

VIA lost that court battle though. They didn't have to pay Intel because they'd changed it (tech) enough that it almost didn't infringe anymore.

[Face]

If anyone feels like doing a lil' reading:

Intel court filing Vs. Nvidia:

http://www.itexaminer.com/images/Intel%20REDACTED%20Complaint.pdf

[SNiiPE_DoGG]

well I've just read through it and everything useful to use in determining anything about this court battle has been redacted. :\ (IE the actual CLA and contested portions)

[dinos22]

that would be foxconn IIRC, they scrapped the dreadnought citing that the 790 chipset was so awful they couldn't work with it any more.

nah i think i have a pretty good idea who they're referring to ;) and i dont think it's them

[saaya]

So Intel would pay nVidia for the "Cookie" (on their boards) while nVidia Paid Intel NOTHING:rofl: Certified my eye LOL!

So yes shintai, folks should wake up! Business isn't carried on like this. Intel-AMD/ATI-IBM and others have always done it the right way. It's goes like; you pay me, I pay you or I get it free if you get it free! nVidia wants; You pay me for mine but I get your tech free.;)
i think you shouldnt look at everything from intels perspective only...
imo both intel and nvidia are trying to misuse their market dominance to push their will onto people, but in this particular case intel is wrong.

but why should nvidia pay intel? for what? what technology does nvidia get from intel? essentially they are licensing their sli driver, a piece of software...
does any software vendor pay intel to run on their cpu/chipset?
i dont see why nvidia should be paying intel for beeing allowed to run their driver on an intel based platform. if nvidia really HAD to pay intel for this, then why doesnt intel just lock sli on their chipset?

first of all intel would get bad letter bombs from customers all around the world, then theyd get sued for anticompetitive behaviour and misusing their chipset monopoly, and then theyd be abandoned in the high end since nvidia sli would be amd only and shift the platform performance towards amd in a very notable fashion ;)


Maybe I'm missing something here but what's the big deal? This obviously isn't over the Atom\Ion platform right? Does this mean Nvidia are getting ready to release an i7 based chipset?
laptops
entry level igp
highend desktop

but most of all its a long term strategic thing... the platform wars have begun and amd intel and nvidia all want to sell the whole pc platform, basically an entire pc, all made by intel, nvidia or amd. thats their future plan...

nvidia will probably have to team up with via while amd and intel have all they need. nvidia knows that via isnt healthy and will limit them so of course jensen is upset that he cant use intel cpus and license their bus to build a nvidia platform with intel cpu...


that would be foxconn IIRC, they scrapped the dreadnought citing that the 790 chipset was so awful they couldn't work with it any more.thats an overstatement... but if you do a bit of research you might stumble over several cancelled 790i boards that nobody ever talked about though.

asus rog pinot noir 790i - prototype was shown at cebit last year
asus 790i reworked - never saw a prototype of this
dfi 790i - never saw a prototype
giga 790i - never seen a prototype of this either

[Donnie27]

i think you shouldnt look at everything from intels perspective only...
imo both intel and nvidia are trying to misuse their market dominance to push their will onto people, but in this particular case intel is wrong.

but why should nvidia pay intel? for what? what technology does nvidia get from intel? essentially they are licensing their sli driver, a piece of software...
does any software vendor pay intel to run on their cpu/chipset?
i dont see why nvidia should be paying intel for beeing allowed to run their driver on an intel based platform. if nvidia really HAD to pay intel for this, then why doesnt intel just lock sli on their chipset?



I know there is a prevailing sentiment that many here think IP and Patents are wrong. This leads folks to think License Fees are repugnant and or evil. The business world wouldn't be worth a flip if that were the case. There are thousands R&D companies making our lives better every hour. These folks can't perform their magic for Free:rolleyes:

I'm NOT on just Intel's side and just as I've argued for AMD and thought they were rightly awarded their case vs. Intel on the whole X86 fight from the early 90's. I was for Intergraph (http://www.internetnews.com/infra/article.php/1480431) long before they won vs Intel. I said AMD would have to pay them too.

My view is that everyone should either;
A. Pay each other for their respective tech/IP.
Or
B. Agree to mutually use each others' IP free of Charge as AMD and Intel currently does.
Not
C. Everyone pay nVidia for their IP while nVidia gets everyone's else's IP free.

Sorry Saaya but you and nVidia are wrong here. So of course I agree with Intel on this simply because they made the agreement with nVidia and nVidia found it hard to break their old Habits. nVidia is aggressively greedy and stengey, a bad combo:rofl: nVidia doesn't pull the crap they pulled I'm about 85% certain we're not talking about this right now.

[saaya]

My view is that everyone should either;
A. Pay each other for their respective tech/IP.
Or
B. Agree to mutually use each others' IP free of Charge as AMD and Intel currently does. totally agree! actually companies should be FORCED to license their IP to others and the fees should be controlled by an industry body to make sure innovation isnt slowed down and nobody can bully anybody with their market share or money. as soon as ip is not available to everybody innovation slows down and political and mafia like relationships emerge, which in the end bend over the end user...

we came up with the idea to use 775 mounting holes on 1366, asus copied it, we hold a patent on it, but... whats the point? innovation drives the market and suing each other for using and not using each others ip is ridiculous and only gets you down in dirty mudfights hurting everybody, especially the customers cause in the end its THEM who pay for all this legal bs the big companies are pulling off. shamino came up with an external panel to access the clockgen and voltages while the board was running, without using the boards resources... again asus copied it, and foxconn could sue them, but again whats the point?

i just dont see how intel is beeing treated bad here, since atm its perfectly normal to use the pc infrastucture without having to pay any licenses and its not sneaky or greedy from nvidia to do this. IF they would have paid Intel to have sli on x58, then it would be intel abusing their market position, forcing nvidia to pay or else they get kicked out. which they are now trying to do with the chipset business. THATS wrong. intel is the one screwing nvidia and the market here by not wanting anybody else building chipsets! just let nvidia build their chipsets and license them and make extra money on it, whats the big deal?

i totally agree that actually there should be a constant flow of licensing between all parties that share a platform, but right now thats not the case, and as such nvidia is doing what everybody else is doing and its not wrong...
my 2cents...

[mikeyakame]

Hmmm I might have this wrong as I read the complaint PDF this morning.

I think this court hearing is more about NVIDIA making wild false claims towards OEs, vendors and even customers which allegedly do not correlate to the chipset licensing agreement they have with Intel. Intel wants NVIDIA to make a public statement about the false claims and NVIDIA refuses to do this because they believe they are in the right, even after being given the chance and not responding to Intel's formal letters. Intel is claiming that these false statements are both hurting them financially now and will even more so in the future. The rest of the pieces of the puzzle fall into their place beyond this.

[DMH]

totally agree! actually companies should be FORCED to license their IP to others and the fees should be controlled by an industry body to make sure innovation isnt slowed down and nobody can bully anybody with their market share or money. as soon as ip is not available to everybody innovation slows down and political and mafia like relationships emerge, which in the end bend over the end user...

we came up with the idea to use 775 mounting holes on 1366, asus copied it, we hold a patent on it, but... whats the point? innovation drives the market and suing each other for using and not using each others ip is ridiculous and only gets you down in dirty mudfights hurting everybody, especially the customers cause in the end its THEM who pay for all this legal bs the big companies are pulling off. shamino came up with an external panel to access the clockgen and voltages while the board was running, without using the boards resources... again asus copied it, and foxconn could sue them, but again whats the point?

i just dont see how intel is beeing treated bad here, since atm its perfectly normal to use the pc infrastucture without having to pay any licenses and its not sneaky or greedy from nvidia to do this. IF they would have paid Intel to have sli on x58, then it would be intel abusing their market position, forcing nvidia to pay or else they get kicked out. which they are now trying to do with the chipset business. THATS wrong. intel is the one screwing nvidia and the market here by not wanting anybody else building chipsets! just let nvidia build their chipsets and license them and make extra money on it, whats the big deal?

i totally agree that actually there should be a constant flow of licensing between all parties that share a platform, but right now thats not the case, and as such nvidia is doing what everybody else is doing and its not wrong...
my 2cents...
I believe that everyone that want's to use Intel CPU's and being able to develop a chipset to them need to pay them

[SNiiPE_DoGG]

Nvidia does not have an attitude of sharing their technologies as the standard, and intel should behave the same or else it is intel's loss

[highoctane]

If anyone feels like doing a lil' reading:

Intel court filing Vs. Nvidia:

http://www.itexaminer.com/images/Intel%20REDACTED%20Complaint.pdf

Anybody have a copy of the CLA?

[hollo]

laptops
entry level igpNV can stick a small GPU+vram on the mobo alongside a intel chipset.
highend desktopi'm not really sure what NV have to offer there, except maintaining their own manufacturing e-peen.

suppose you want i7 CPU power but don't game - how about a x58 mobo with a 9400 (with 128 or 256mb ddr3 sideport) connected to the chipset with a PCI-e2 x4 link (a single PCI-e2 lane may be enough)?
a board like that with a working hybrid-power implementation that'll power up the GTX295 in the PCI-e x16 slot when you open a game, and switch back to the 9400 when you quit to desktop, would be a sure-sell. but NV don't need to design the entire chipset to do that.

the atom platform may be different - if intel will only sell atom + a chipset with a IGP then nvidia won't be able to easily add their own GPU since the end customer will have to pay for two GPUs, and that'll kill the profit margin. even if NV did want to add a non-chipset integrated GPU to the atom platform it would need it's own memory controller, which would be acceptable on a 15W mobile CPU, but may be a TDP killer on a 1-2W atom. atom is the kind of platform where it really makes sense for the GPU to borrow the CPU's memory controller, for all other platforms it seems pretty optional to me.

interesting post about foxconn IP btw, thx

[Donnie27]

totally agree! actually companies should be FORCED to license their IP to others and the fees should be controlled by an industry body to make sure innovation isnt slowed down and nobody can bully anybody with their market share or money. as soon as ip is not available to everybody innovation slows down and political and mafia like relationships emerge, which in the end bend over the end user...

we came up with the idea to use 775 mounting holes on 1366, asus copied it, we hold a patent on it, but... whats the point? innovation drives the market and suing each other for using and not using each others ip is ridiculous and only gets you down in dirty mudfights hurting everybody, especially the customers cause in the end its THEM who pay for all this legal bs the big companies are pulling off. shamino came up with an external panel to access the clockgen and voltages while the board was running, without using the boards resources... again asus copied it, and foxconn could sue them, but again whats the point?

i just dont see how intel is beeing treated bad here, since atm its perfectly normal to use the pc infrastucture without having to pay any licenses and its not sneaky or greedy from nvidia to do this. IF they would have paid Intel to have sli on x58, then it would be intel abusing their market position, forcing nvidia to pay or else they get kicked out. which they are now trying to do with the chipset business. THATS wrong. intel is the one screwing nvidia and the market here by not wanting anybody else building chipsets! just let nvidia build their chipsets and license them and make extra money on it, whats the big deal?

i totally agree that actually there should be a constant flow of licensing between all parties that share a platform, but right now thats not the case, and as such nvidia is doing what everybody else is doing and its not wrong...
my 2cents...

Each company should and does have the choice to charge or not for tech they invested in.

I think you go it mixed up. I said "If Intel has to pay nVidia for the "Cookie and NF200" when clearly neither is needed, then nVidia should have to pay Intel as well. What's going on between your company and Asus is a prime example and NOTHING like what's going on between Intel and nVidia.

I disagree 100 about what creates innovation. Innovation starts before the first motherboard is shipped in this case. Everybody with the exception if nVidia and many times VIA pay fees and acts in accordance with established business practices. Getting someone to follow the LAW and Rules is not bullying BTW. If I invest money to come up with better Mouse Trap, Damned right I want to be paid for it. Then someone else spends money to try and out do me and the beat goes on:up: Folks not paying for my or your ideas is what slows down innovation, just the opposite of what your saying. Why? Those creating the innovation will go broke if they don't get paid.

So think Intel shouldn't have paid Intergraph? Guys, you can't have it both ways. We ALL KNOW SLI doesn't need a Cookie or NF200 and nVidia shouldn't be charging anyone for something we don't need to run two of their cards at one time. It is the essence of Sleaze & Greed and we talking about Intel is Wrong?

I remind folks with Idiotic talk of Intel buying off courts and etc..... Intel lost to AMD, Intergraph (twice) and others:rolleyes:

[Sr7]

Each company should and does have the choice to charge or not for tech they invested in.

I think you go it mixed up. I said "If Intel has to pay nVidia for the "Cookie and NF200" when clearly neither is needed, then nVidia should have to pay Intel as well. What's going on between your company and Asus is a prime example and NOTHING like what's going on between Intel and nVidia.

I disagree 100 about what creates innovation. Innovation starts before the first motherboard is shipped in this case. Everybody with the exception if nVidia and many times VIA pay fees and acts in accordance with established business practices. Getting someone to follow the LAW and Rules is not bullying BTW. If I invest money to come up with better Mouse Trap, Damned right I want to be paid for it. Then someone else spends money to try and out do me and the beat goes on:up: Folks not paying for my or your ideas is what slows down innovation, just the opposite of what your saying. Why? Those creating the innovation will go broke if they don't get paid.

So think Intel shouldn't have paid Intergraph? Guys, you can't have it both ways. We ALL KNOW SLI doesn't need a Cookie or NF200 and nVidia shouldn't be charging anyone for something we don't need to run two of their cards at one time. It is the essence of Sleaze & Greed and we talking about Intel is Wrong?

I remind folks with Idiotic talk of Intel buying off courts and etc..... Intel lost to AMD, Intergraph (twice) and others:rolleyes:

It costs a lot of money to maintain SLI and Crossfire as they are essentially glorified driver hacks.. How can you think it's any less worthy of licensing than a chipset that you design once and print out in silicon thousands/millions of times?

[fireice2]

NO to Intel monopolizing the i7 Platform!

[Anemone]

What do you think SLI is? a monopoly! Nvidia just wants to be paid three times for it. They wanted to base the entire existence of their chipsets on it. Only they wanted to charge people 3x over, chipset, card #1 and card #2, and eventually card #3. They negotiated a cost to build chipsets for the Intel platform and were welcomed for the innovation, but they did not have the skill (yet) to get them durable and bug free, especially when pushed beyond their limits. But they could have gained that skill over time.

To offer history in this, folks will have to go back and research what Intel charged Nvidia to build for 478 and 775 platforms. I'd bet it was a fairly low amount, but I'll stand to be corrected if that's not true. Along comes SLI and, in effect, it's like the days of Intel's having a math processor as a separate add on chip. You want more performance, you sell someone a second GPU, which gives that to them. And you make more profit. Everyone thinks that the profit is all in the high end chips, but it's not. It's in the midrange. So having customers buy 2x or 3x midrange cards made Nvidia MORE money than if they had sold one high end card.

On high profit margin cards (Quadro's), SLI has worked everywhere for a very long time. There is no issue with chipsets or whatnot. So the whole "driver" cost is kind moot. It's being developed, and the cost paid for. It's just what market gets to "have" it is the only question.

All Intel asked for is a retraction of comments, and a public admission that Nvidia was not suitably licensed. Why would Nvidia withhold a license on I7? For two reasons, and I'd bet that both are equally weighty. One is obviously the SLI fighting that went on. SLI should not be used as a Nvidia tool for selling chipsets, because no excessive or restrictive measure was put on Nvidia for what CPU's they could develop chipsets for (775, 478). Intel didn't restrict what "market" (server, workstation, mainstream) Nvidia wanted to target, and didn't charge them extra for any given market target. In fact, if we go way back, Intel could simply (like Via) denied Nvidia the right to make chipsets at all and then their wouldn't be any SLI (oh wait, yes there would be but it would work on Intel chipsets go figure). Intel simply asked that Nvidia treat Intel the same way that Nvidia had been treated. "Here is your license. Now we'll compete on features and abilities in the marketplace."

Nvidia made two errors. They withheld a feature that propped up their chipset business, which I'll get more into below, and they made some very lousy chipsets which harmed the Intel end user experience, something I suspect Intel has been fuming about for years (775 socket era).

A lot of folks have said that SLI is Nvidia's baby and Intel should pay for it. Do we pay the principal founders for SATA, USB, pci-e, pci? Yes you do, but you do it via a specific route. The industry generally gets together to hammer out the standard, and then folks make money selling devices that use it. And to the degree to which each company can sell parts that use the feature, or optimize the features speed, latency and other aspects, a given company will make more or less money based on the quality of the parts and optimizing of the standard used with it's parts. These "understood" practices break down from time to time. And Intel doesn't play fair all the time. These moments often result in legal proceedings to work out the details, which is often a kinder wording for "deciding who gets what portion of the profits".

In this case, Nvidia has worked very hard to keep SLI for itself, except on the Quadro series. Moreover, as we saw with Skulltrail, they even decide "when" to enable SLI on even "SLI enabled" chipsets. Intel indicated that there should either be a reasonable fee for enabling SLI on other chipsets (emphasis on reasonable), or Nvidia would feel the same sting by getting charged for all the other industry infrastructure that they were making free use of, in making profits with SLI. Now that time has come. The sockets are changing, SATA 3 and USB 3 are coming, and everyone is running around making sure they line up their agreements and fees (where needed) to be ready for the wave of new products. But not Nvidia. During the 775 era Nvidia degraded the user experience, and didn't play fair regarding what to charge for and what to make proprietary. You don't see a wealth of companies standing up for Nvidia, do you? Nvidia was urged, many times, over the past several years, to reconsider its position. It has the ability to get better at building and eventually make some very good chipsets, with some features not seen before on chipsets. But given how they feel they wish to play, they are going to get cut off.

This is how the next few years will play out:


GPU's will become powerful enough that they will become sideline or "included on main silicon" items. Contrary to what Nvidia proclaims, it's actually their silicon that is going the way of the sound card, not the other way around.
Nvidia's chipset business will go the way of Via's. Cut off (due to their own misbehavior) they will not be allowed on any I7 or future platform. They will continue to make chipsets for AMD, and this will be where they will have to flourish or wither and die.
The next generation of GPU's (G300, ATI 8xx series) will be more than capable of enough power on a single card. At that point, even if the quad resolution displays come about, the power will be there to drive them.
Crossfire works, but becomes relegated to benching. SLI is the same.
AMD/ATI meanwhile will join Intel in eventually offering their "GPU" silicon as an on die feature. Remember the days when math chips were separate? Same thing happens with GPU ability. AMD and Intel get to play, and Nvidia plays in other pastures.
The first step in seeing this come about is to have a direct pci-e link to the GPU on silicon (core I5 series).
The second step is that Nvidia will not be making core I7 or core I5 chipsets, probably ever, but there might be room to negotiate.
Nvidia meanwhile will branch to other areas and abilities, much like Creative.

[hollo]

supercomputers started off as single-core solutions, but now they have thousands of cores, often mixtures of CPUs from different manufacturers, even nvidia GPUs are turning up in them.

the single-socket desktop became parallel decades later - hyperthreaded single cores, duals, hyperthreaded duals, quads, hyperthreaded quads, and soon we'll have hyperthreaded hex-cores.

so i think multi-GPU will stay, and expand. a multi-GPU setup is like a multi-processor server.
going again by the principle that history repeats - in the past if you wanted multiple CPU cores you needed to buy a multi-socket motherboard, now you can get multi-core CPUs. in the past if you wanted multiple GPUs you needed a multi-PCI-e motherboard, but in the future multi-GPU single cards will become more popular, ATI have made multi-GPU their high end card policy already. quad-sli/xfire on a single PCB? it's coming. wanna know a upcoming gaming technology that'll get everybody buying multiple cards? stereoscopic vision. we'll have each card working on a different picture instead of a single picture being split between two cards, so scaling will be perfect. in cases where two GPUs scale well when working on a single picture stereoscopic 3d will make quad GPU setups interesting instead of a micro-stutter-fest.

but nvidia have taken on AMD and Intel instead of just ATI by giving their GPUs a CPU feature - programmability. and it wasn't a half-assed effort. sub-$100 NV cards beat a overclocked core2quad at folding. NV GPU owners have a seriously powerful piece of silicon that's unused when they're not gaming, but in theory it could accelerate any desktop app, and obselete the CPU in many.

i'm not sure how this will pan out for nvidia, they may not have enough money to invest to get CUDA into the market. they need to lower the cost of entry for software companies, ie they need to sponsor CUDA development for the really big companies, but that isn't cheap. intel is going to rain on their parade pretty hard with larrabee which will already be compatible with x86, so today's programmers will find it easier to work with. can you imagine how much NV would like to have the world's biggest software company dedicating most of its time to programming CUDA? now you see NV's problem - microsoft dedicates most of its time to x86. larrabee may well accelerate microsoft windows and any program that runs on windows out of the box. if you're a occasional gamer, or not into the graphically stressful first person shooters, do you buy a $200 NV card, or a $200 Intel card with the gfx power of a $100 NV card + the ability to speed up any multi-threaded windows app, and multitasking in general?

in the longer run i agree with you that GPUs and CPUs will become the same thing. a GPU with CUDA is a great maths processor, and like the maths processor it will end up on-die. there's so much waste in having the GPU and CPU seperate - two memory controllers, two sets of memory, two programming languages...

if i were Jen-Hsun Huang there's one thing i would do today - i'd get a few hundred motherboards specially made, and find elaborate and entertaining ways of giving them away:
GTX285 core (maybe two, so it'd be the most powerful single PCB card in existence) where the CPU should be. GDDR5 where the ram should be. a slot that resembles a PCI-e x16 slot which you can add one of two small add-in cards to, one has a LGA1366 socket + 3 DDR3 slots on it, one has a AM3 socket + 2 or 4 DDR3 slots. the GTX285 should have a tru120e on it, and the CPU should have a thin GPU-style cooler that cools the CPU, chipsets and VRM.
an essential marketing stunt :yepp:

[Donnie27]

It costs a lot of money to maintain SLI and Crossfire as they are essentially glorified driver hacks.. How can you think it's any less worthy of licensing than a chipset that you design once and print out in silicon thousands/millions of times?

Less as in the percentage of Products with CF or SLI installed on them vs. Chip Sets. Simply put, there are way the hell more chip sets than CF or SLI. nVidia sells two cards for folks to USE on Intel boards, paying extra for Cookies and or NF200 is BOGUS IMHO. Those extra costs aren't consumer friendly either:rolleyes:

If CF and SLI demands a Fee, Tax or charge, and Intel pays it, so should nVidia pay for Intel's tech=P (repeated too many times) Neither company deserves any kind of special treatment or a etc.... AMD and Intel avoided this by following that some old IP type of agreement that's been around since World War 2. Why are you guys making this simple case of nVidia's greed into something more than it is?

[Shintai]

so i think multi-GPU will stay, and expand. a multi-GPU setup is like a multi-processor server.

Multi GPUs are nothing like multi CPUs. SLi/CF is more like 2 clustered servers if you like. You cant draw parallels between CPUs and GPUs as you do. They are worlds apart.

Personally I think multi GPUs is only here for a limited time until we do it all in the CPU. Sooner or later both the multi and single GPU setups will fade off.

[hollo]

a misunderstanding - i don't mean on a micro-architectural level
i'm referring to the evolution of the platform as it becomes more parallel...

desktop/server - single socket & single core, then multi-socket and single core, then multi-socket and multi-core
SLI/crossfire - single AGP and single GPU, then multi-PCI-e and single GPU, then multi-PCI-e and multi-GPU

...identical, in a sense.

edit: forgot the voodoo - multi-PCI and single GPU....

[Ket]

This is all kind of ironic, really :D nVidia killed the legend that is 3Dfx (look around, that hardware was so advanced for its time, among those devoted to 3Dfx still those GPUs can run the likes of Doom3, and probably others I don't know about)

What goes around comes around. Karma is not without a sence of humor.

[iboomalot]

Multi GPUs are nothing like multi CPUs. SLi/CF is more like 2 clustered servers if you like. You cant draw parallels between CPUs and GPUs as you do. They are worlds apart.

Personally I think multi GPUs is only here for a limited time until we do it all in the CPU. Sooner or later both the multi and single GPU setups will fade off.


limited time atleast for the next 10+ yrs

too much heat involved and voltage to push a monitor for gaming.

a 4780x2 generates alot of heat and put on a CPU chip will run way to hot.

This of course takes into account performance of the same calibur.

will the new westmere even compete against a ati 9700 pro????

guess we will find out

[hollo]

could a TRU120Ex2 with double the fin depth and 12 heatpipes instead of 6 handle a phenom2 MCMed to two 4870s?
i'd say.. easily.

[saaya]

donnie, microsoft doesnt NEED a license key for windows and no application NEEDS a license key or serial number either, does that mean they should pay intel licensing fees for beeing able to use their plattform?

Anemone, how the h3ll is sli a monopoly?
could you please explain this in some more detail?


The next generation of GPU's (G300, ATI 8xx series) will be more than capable of enough power on a single card. At that point, even if the quad resolution displays come about, the power will be there to drive them.
thats taking the trend of the past years and flipping it around 180degrees... could you please explain why the trend should change all the time?

whats driving pc upgrades and innovation? its almost entirely gaming and some internet and multi media. now whats better for this, slow cpu and fast gpu or fast cpu and slow gpu? then think again about which chip is going to be integrated into what... ;)

jensen is right about that part, and why do you think intel is going for gpus? just for fun? they know its more important than cpus and itll be even more important in the future as gpu processoring demand is growing way faster than cpu processing demand.

[saaya]

could a TRU120Ex2 with double the fin depth and 12 heatpipes instead of 6 handle a phenom2 MCMed to two 4870s?
i'd say.. easily.
probably, but why would you do that?
you want to cool several chips as good as you can, so you can clock them high and max them out performance wise. then what uber genius would come up with the idea to put it all on a 40x40mm package? :P

thats what ive been saying for a long time, the whole point of FUSION doesnt add up... it makes sense for the mainstream and igp, but people always thought integrating gpus would mean better performance... well no...

both cpus and gpus are tdp limited, thats the whole point of overclocking, removing the tdp limit with better cooling and then beeing able to clock higher or bump voltages to clock even higher.

so we have two parts that are tdp limited, then how is putting them next to each other or merge them going to improve performance? :D

what benefits do you get from putting them on the same package or same piece of silicon?
more bandwidth...
was there a notable boost from pciE 1.1 to 2.0?
nope... then isnt it obvious that increasing the bandwidth between cpu and gpu isnt a limiting factor?

the cpu memory bandwidth is a joke compared to gpus, so again, thats limiting, not speeding things up! unless you give them seperate memory interfaces which means loads of pins and defeats the purpose of putting them on the same package. or you let the cpu use the gpu mem, but while it has massive bandiwdth the latency is terrible, and the cpu would suffer and not benefit from that. so AGAIN, it doesnt make sense...

the only good point is to save money if you put two mainstream or entry level parts on one package rather than two, or the same silicon die. and to everybody who thinks this is a revolution, amd geode gx cpus have had integrated gpus for a decade, and they are far from the only cpus with integrated gpus... its just uncommon outside of the ce segment, thats all...

[Greg83]

Currently, X58 boards are getting a SLI cookie for $5 per board sold or so.

This means, intel , to compete vs asus, gigabye, evga, etc, has to pay nvidia $5 to get the SLI cookie for there boards too.

Since intel has to then, get there board approved by nV :rofl::ROTF: for it to get the cookie added to the bios.


So for intel, to then charge nvidia back $7.50 per board sold to approve this new SLI cookie input into all x58 boards using the cookie, would only be fair IMHO :up:


and during all this, AMD and Intel can co-operate fine, and crossfire is on X58 boards w/o lawsuits

[Vinas]

I see that a number of people are saying "Why would you buy an Nvidia chipset when there is the X58." On the 775 lga the Nvidia 790i Ultra was/is the best 775 chipset. If they made a chipset that worked with Nehalem and was better I would buy it.
Maybe compared to the P965... Other than that no thanks. The 790i was just as terrible as the 780 IMO with it's super low clocks and pathetic raid support.

[Donnie27]

donnie, microsoft doesnt NEED a license key for windows and no application NEEDS a license key or serial number either, does that mean they should pay intel licensing fees for beeing able to use their plattform?

Anemone, how the h3ll is sli a monopoly?
could you please explain this in some more detail?


thats taking the trend of the past years and flipping it around 180degrees... could you please explain why the trend should change all the time?

whats driving pc upgrades and innovation? its almost entirely gaming and some internet and multi media. now whats better for this, slow cpu and fast gpu or fast cpu and slow gpu? then think again about which chip is going to be integrated into what... ;)

jensen is right about that part, and why do you think intel is going for gpus? just for fun? they know its more important than cpus and itll be even more important in the future as gpu processoring demand is growing way faster than cpu processing demand.

My point exactly! MS and Intel follows that same old agreement I mentioned earlier. They have a Mutual Agreement just like the ones with AMD, IBM, and everyone else not named nVidia and VIA.

Intel and MS also sit on many of the same standards boards and don't need side agreements, nVidia doesn't. As was mentioned, SLI is a monopoly because no 3rd party is allowed to build NF200, BIOS Cookies and etc.......

nVidia uses it (or tries to use it) as SLI a leverage when it should be an Open Standard. Just like some of the Stuff Intel has championed like PCI-E, AGP, USB and yada yada! Intel did give nVidia an open shot and nVidia didn't return the favor. As aways nVidia wants all incoming tech FOC and wants all THEIR out going tech paid for. That ain't going work without repercussions.

[hollo]

probably, but why would you do that? you'd be able to upgrade v-ram! :rofl:
you want to cool several chips as good as you can, so you can clock them high and max them out performance wise. then what uber genius would come up with the idea to put it all on a 40x40mm package? :Pa understandable POV on XS, from someone involved in designing enthusiast motherboards :p:

but would you really care whether you buy one high end heatpipe cooler for your CPU and one for your GPU, or whether you buy one that costs twice as much for your GPU/CPU, if you'd get the same temperatures in the end?
thats what ive been saying for a long time, the whole point of FUSION doesnt add up... it makes sense for the mainstream and igp, but people always thought integrating gpus would mean better performance... well no...

both cpus and gpus are tdp limited, thats the whole point of overclocking, removing the tdp limit with better cooling and then beeing able to clock higher or bump voltages to clock even higher.

so we have two parts that are tdp limited, then how is putting them next to each other or merge them going to improve performance? :DTDP has become outrageous on the latest GPUs, but at the same time the operating temperatures (for CPUs and GPUs) have expanded (up and down) and our heatsinks have become quite powerful. a tru120e can do nearly 200W, and it's not even that big - you could easily make it twice as deep and 50% taller and it'd handle 600W if cooling power scaled perfectly with size, 500W would definitely work. that'd cool a stock nehalem + GTX295!

back when CPUs didn't need heatsinks i'm sure it would have been hard to see today's coolers coming.. so i wouldn't say a future CPU/GPU combo would be impossible, or even inconvenient, due to thermals. maybe AMD will start selling heatpipe coolers again (i really wish intel would stick a small heatpipe in the middle of their stock coolers)
what benefits do you get from putting them on the same package or same piece of silicon?
more bandwidth...
was there a notable boost from pciE 1.1 to 2.0?
nope... then isnt it obvious that increasing the bandwidth between cpu and gpu isnt a limiting factor?true, although i imagine the longer term processing core could be something like 50% x86, 50% GPU/CUDA (not those instruction sets, but... those functions compared to today's cores, and the manufacturer would make chips with varying ratios of CPU or GPU function to cater to different needs in the market,). or, to put it another way: a lot like larrabee ..
x86 with larrabee specific instructions..
the cpu memory bandwidth is a joke compared to gpus, so again, thats limiting, not speeding things up! unless you give them seperate memory interfaces which means loads of pins and defeats the purpose of putting them on the same package. or you let the cpu use the gpu mem, but while it has massive bandiwdth the latency is terrible, and the cpu would suffer and not benefit from that. so AGAIN, it doesnt make sense...definitely technical challenges, much harder than solving the TDP problem i expect.

shared memory controller: it'd have to be very wide.
if the CPU was going to have to suffer slightly due to memory latency it could be compensated for by adding extra cache
the only good point is to save money if you put two mainstream or entry level parts on one package rather than two, or the same silicon die. and to everybody who thinks this is a revolution, amd geode gx cpus have had integrated gpus for a decade, and they are far from the only cpus with integrated gpus... its just uncommon outside of the ce segment, thats all...the way i see it is... it's going to take time but it's not impossible. consider the i5 westmere MCM - here's a MCM a crazed billionaire could get made today, long before Intel or AMD or anyone else design and manufacture 'the ultimate gaming/multimedia core' with CPU/GPGPU/GPU logic all mixed in together:

GTX285 has just under 160GB/s theoretical memory bandwidth. triple channel DDR3-2000 has just under 50GB/s. DDR3 is lower latency than GDDR3, so let's say two triple-channel DDR3 memory controllers could feed a GTX285 pretty well, ~100GB/s bandwidth.

so -
32nm westmere quad w/o memory controller <-QPI-> 45nm x58gtx285 with 384bit MC
MCM on a single package, 6 DIMMs beside the socket
with a TRU120Ex2
delicious :ROTF:

^ of course that's not gonna happen.


phenom 2 with quad-channel DDR3-2000 for ~64GB/s
<-hypertransport 3->
HD4860 (4850 has 64GB/s, 4870 has 115GB/s)

<-hypertransport 3->
790fx & sb750

hehe
[/dream]

[moiraesfate]

Looks to me like an attempt to create a monopoly by intel. It would be bad.

[tam2]

SLI is an optional tech, while CPU with chipset support is a non replaceable-crucial yet universal part in modern computing. And thats already uncompetitive, and now they dont want other company to make product derivative components for their CPU.
Wonder what would happen if someday they able to cramp everything from graphic to usb interface into their cpu....

-tam2-

[hollo]

this