NVidia's Dispute over Nehalem Licensing - SLI not involved

As posted by Bit-tech

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.
Might as well link the Daily Tech version of the story while I'm here (even though it references Bit-tech):

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

Which is useful because it includes this snippet I missed before:

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.
Implying that the battle is over DMI. I think DMI has been suggested before as the point of conflict.

Jawed
 
From the sounds of it...

Nvidia believes the current agreement allows them to build chipsets for Intel CPUs regardless of what type of "bus" it uses.

Intel believes the current agreement only applies to ones currently available. And future busses need to be licensed seperately.

Nvidia is going ahead with developement on chipsets supporting newer Intel busses and Intel is seeking a court order to make them stop unless they license the ability to do so.

At least that's the gist of what I get from this.

Regards,
SB
 
why do they need a license

So they can say that their chipset is certified for those processors. otherwise intel would just do something funky during the processor startup and leave the chipsets dead.
Why do you need a license for SLi?
 
you shouldnt do if the board has 2 pci-e slots sli should work they are standard slots nothing has to be done to them or added to make sli work (what next hard drives wont raid on non licensed sata chipsets)
its only becasue nv has put some mechanism in its drivers to disable it on chipsets they dont like, what next a single gfx card not working unless the board maker pays nvidia
its a form of blackmail if you dont pay us we will disable sli on your boards...
 
ACtually the review of the X58 boards seemed to indicate there was more to it than that as they sucked at sli until the most recent bios release that actually fixed all the issues.

And while I believe Sli should be available on all boards I do understand that Nvidia would not want it to work like crap and dilute mindshare.
 
Thought you'd like this :

As Intel took its case against Nvidia to court, Nvidia CEO Jen-Hsun Huang could not resist making the oft-repeated assertion that the GPU is in, and CPU is out--a thinly veiled reference to the graphics chip maker's credo that PC processor mind share is shifting from Intel to Nvidia.


Here is the statement that Huang inserted into the Thursday Nvidia release about the Intel court filing. "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."
 
what is it with intel, can they not design a decent gpu
maybe they should of bought ati
or is it that if they did and everyone had decent graphics importance would shift more toward the gpu and away from the cpu so they keep making sucky ones ;)
 
For an internal only, confidential, document I'm fairly sure you can quote whoever you like, and Fudzillas sources are breaking their confidentiality agreement.
Yup, you're exactly right AFAICT. Although was it really internal only? Might this not be shown to potential customers? I'll admit to have no idea as to the legal status there if the quotes were truly out of context.

This part is hilarious though:
urthermore, Intel says that Nvidia's Ion is expected to consume more power and add unnecessary cost to system Bill of Material (BOM)
Errr, right. The only reason it'll increase the BOM is because you are artificially forcing customers to pay the full CPU+Chipset price anyway. If you looked at the real manufacturing costs of both solutions, it's pretty clear that's pure BS. I'd be very interested in truly reliable and fair power tests in non-3D workloads, though.
 
Yup, you're exactly right AFAICT. Although was it really internal only? Might this not be shown to potential customers? I'll admit to have no idea as to the legal status there if the quotes were truly out of context.

This part is hilarious though:
Errr, right. The only reason it'll increase the BOM is because you are artificially forcing customers to pay the full CPU+Chipset price anyway. If you looked at the real manufacturing costs of both solutions, it's pretty clear that's pure BS.

Competitive Position Guides are an internal briefing document, you use it to answer questions perhaps from press or customers as to Intels position.
To be honest I think they are probably right on most the points they make too - HD videos relevence to the market is in my mind argueable.
3D gaming performance is poor even with Ion as the CPU can't keep pace even with an IGP
Without CPU+chipset bundling the BOM would probably go up as I can't MCP79 isn't a small chip and its manufactured on a far newer process than 945GC so either the margins are going to be LOLsmall or it'll push the price up.

As for quotes and context, they'll include links to the articles so people can see for themselves anyway, as it is this is a non-story, its just Fuad trying to pretend people give a damn what he writes (badly).

I'd be very interested in truly reliable and fair power tests in non-3D workloads, though.

For a Nettop platform I think Anandtech can obligue? http://www.anandtech.com/video/showdoc.aspx?i=3509&p=6

5W extra idle and load, so 33% idle, 25% load.

I know we are all hardware snobs here but for the average Atom buyer I would say 945GC is sufficient, these are cheap internet and home office systems and while Ion may give them a few extra capabilities it just means people are going to be disappointed in other areas as they'll expect increased performance across the board and it can't deliver.
 
To be honest I think they are probably right on most the points they make too - HD videos relevence to the market is in my mind argueable.
3D gaming performance is poor even with Ion as the CPU can't keep pace even with an IGP
Everything is useless. Doesn't mean useless stuff doesn't sell. How many people buy much more expensive CPUs than they'll ever need for the entire lifetime of the system?
FWIW, I agree gaming is a gimmick, although personally I still think HD video is useful.

Without CPU+chipset bundling the BOM would probably go up as I can't MCP79 isn't a small chip and its manufactured on a far newer process than 945GC so either the margins are going to be LOLsmall or it'll push the price up.
You're only partially right... According to this, the 945's die size is ~101.76mm²: http://www.intel.com/assets/pdf/designguide/307504.pdf
And the ICH9's die size is ~64.78mm²: http://www.intel.com/Assets/PDF/designguide/316974.pdf

I know Atom uses ICH7, but the die size of that chip isn't public. Presumably it isn't smaller than 50mm² though, which means the total die size is... ~150mm². What is MCP79's die size? ~150mm². The big difference, of course, is that MCP79 is on 65nm and isn't manufactured on old fully amortized fabs; but there are a few other factors to consider such as:
- 1 package vs 2 packages.
- ICH7 doesn't even integrate PS/2, so they need a SuperIO chip in just about every setup - probably including most netbooks... Similarly, ICH7 doesn't integrate an Ethernet MAC.
- I have yet to see an Atom PCB that is quite as small as Ion, so that also implies higher integration for things most people don't usually think of.

In the end it's likely that even if Intel was 100% fair, MCP79 would still increase the BoM, but it'd probably be a small increase and maybe even within the margin of error TBH...

As for quotes and context, they'll include links to the articles so people can see for themselves anyway, as it is this is a non-story, its just Fuad trying to pretend people give a damn what he writes (badly).
Fair enough, I didn't see the presentation so I wasn't sure they included links to the articles.

For a Nettop platform I think Anandtech can obligue? http://www.anandtech.com/video/showdoc.aspx?i=3509&p=6

5W extra idle and load, so 33% idle, 25% load.[/QUOTE]Intriguing, although that is Atom 230 vs 330 and that's probably MCP79U, not MCP79MX etc... Still not a good sign, obviously.

I know we are all hardware snobs here but for the average Atom buyer I would say 945GC is sufficient, these are cheap internet and home office systems and while Ion may give them a few extra capabilities it just means people are going to be disappointed in other areas as they'll expect increased performance across the board and it can't deliver.
I pretty much agree, Atom is still too slow. I'd much rather wait for a Cortex-A9 netbook than one based on Pineview, thank you very much! ;)

(oh hey, this did turn out into a pretty big post - and it ended into a mention of the Cortex-A9! Maybe I'm being a tad too predicable...)
 
You're only partially right... According to this, the 945's die size is ~101.76mm²: http://www.intel.com/assets/pdf/designguide/307504.pdf
And the ICH9's die size is ~64.78mm²: http://www.intel.com/Assets/PDF/designguide/316974.pdf

I know Atom uses ICH7, but the die size of that chip isn't public. Presumably it isn't smaller than 50mm² though, which means the total die size is... ~150mm². What is MCP79's die size? ~150mm². The big difference, of course, is that MCP79 is on 65nm and isn't manufactured on old fully amortized fabs; but there are a few other factors to consider such as:
- 1 package vs 2 packages.
- ICH7 doesn't even integrate PS/2, so they need a SuperIO chip in just about every setup - probably including most netbooks... Similarly, ICH7 doesn't integrate an Ethernet MAC.
- I have yet to see an Atom PCB that is quite as small as Ion, so that also implies higher integration for things most people don't usually think of.

I don't think any Nettops or Netbooks are using PS2, but other than that I'd probably agree.
The ION reference platform is using 2 PCBs however and is presumably optimised for size rather than cost (how many layers is it compared to a typical Atom board for example?).

Intriguing, although that is Atom 230 vs 330 and that's probably MCP79U, not MCP79MX etc... Still not a good sign, obviously.

D'oh my mistake there.

I pretty much agree, Atom is still too slow. I'd much rather wait for a Cortex-A9 netbook than one based on Pineview, thank you very much! ;)

(oh hey, this did turn out into a pretty big post - and it ended into a mention of the Cortex-A9! Maybe I'm being a tad too predicable...)

Cortex-A9 with ZRAM surely? ;)
 
Enemas for leaky Intel staff

http://www.theinquirer.net/inquirer/opinion/145/1051145/intel-broadsides-nvidia-powerpoint-assault

We feel a bit bad for our friends at Bit-Tech, since, following our INQuiry into the origin and purpose of the presentation, an Intel spinner forced the site to take down slides it had posted.

He said: "Talked to Tim [Smalley] - he will take that foil down from the story. It is an Intel internal foil marked Intel confidential. Make sure you don't post it either, or the legal beagles will give me an enema".
:LOL:

Jawed
 
For a Nettop platform I think Anandtech can obligue? http://www.anandtech.com/video/showdoc.aspx?i=3509&p=6
The first page of that article here has under the cost of Ion:
There are a few dozen Ion reference platforms out in the hands of OEMs and decision makers in the industry. NVIDIA expects Ion to add between $50 and $100 to the cost of a typical Atom machine.

That seems quite a bit, say adds 33% to BOM maybe more?

Anyway looks like Ion is trying to repeat the business model nvidia had in the regular mobile(notebook/laptop) market where they would sell add in gpus into a large percentage of the mid and high end models and leave alone the low end models. Which is great as long as there are actually sales of mid/high end models.
 
NVidia didn't take long

http://www.bit-tech.net/news/hardware/2009/02/25/nvidia-responds-to-intel-document/1

Nvidia also dismissed Intel’s claims about power consumption and battery life, saying that "peak power is a very poor measurement" that doesn’t represent real-world use of a netbook. To make its point, Nvidia published results from MobileMark 2007, which tests battery life using standard apps and video playback. The results showed that an Intel 945GSE-based system lasted for two hours, 40 minutes, while an Nvidia Ion system lasted for two hours, 31 minutes.
Jawed
 
Back
Top