NVIDIA Files Complaints Against Samsung and Qualcomm for Infringing Its GPU Patents

OTOH, they're a public company and under US law - AFAIK - they have the obligation to pursue revenue opportunities or they could be sued by their stockholders. Crazy legislation.
I don't think US law has anything to do with it. It's just that the stock holders own the company, not the CEO. Unless the CEO owns the majority of the stock...
 
Lawsuits, while expensive, typically don't cost US$1.5Bn, though (if you win, I mean...) So yeah, there can be quantifiable benefits. Even if you lose, a court may decide the suing entity isn't entitled to quite as much money as they would like. It's a roll of the dice; you might lose, and if you rather chicken out and pay the ransom then you surely won't win either of course...

Also the winning side usually gets their court costs and attorneys' fees paid for by the losing side.

Nvidia has asked the ITC to block sale or importation of any products found to be infringing, and in the Delaware case it has asked for compensation of up to three times the actual amount of damages found, plus court costs and attorneys' fees.

http://www.theregister.co.uk/2014/09/04/nvidia_sues_qualcomm_and_samsung/
 
One assumes Samsungs mobile's defence on the Qualcomm chips is that it is Qualcomm's problem. They don't have that defense with their Samsung semi's own Exynos line.

The biggest two elephants in the room are Apple and Intel.

Apple is the majority user of IMG PowerVr, but they aren't cited, neither is Intel.

Yes, the PR does say about being in discussions with others, but I'd be surprised if Intel, or in particular Apple just rolled over to get their belly tickled.

I assume an alleged patent infringement is by a physical implementation, hence why chips are being cited and not IP companies. However,amongst other interesting things that this complaint brings up, is that indirectly, Nvidia are getting embroiled with ARM, their CPU IP provider that is central to their Tegra platform.

Correct me if im wrong, but they use ARM Mali Txxx on the Exynos no ? And Adreno obviously on the Snapdragon based chip.

If you can attack someone by include thoses gpu's architecture in your SOC, so can be Apple and many other.
 
Okay, so looking at the patents in question...

6198488: On modern mobile GPUs, where is the transform module? Where is the lighting module? The specific details of this patent only share some very superficial similarities with other designs.. at best.
6992667: On modern mobile GPUs, where is the skinning module? The patent doesn't do a great job explaining masking and swizzling at all, but if they mean the ability to predicate and reorder vertex inputs there are some obvious prior art examples like the PS2 VUs. Any "innovation" that boils down to putting different well understood and utilized algorithms on the same die is a farce and should be overturned, as this is both obvious and expected in the industry.
7038685: Might have something here, although you can find barrel processors with the same basic innovation that predate the patent file by many years.
7015913: Seems redundant vs 7038685.
6697063: Prior use here - the modernish rasterization process with depth buffering and perspective correct interpolation is predated by SGI GPUs (like in the N64) and the screen space tiling is predated by PowerVR PCX1.
7209140: Programmable vertex processing done by anything with a specialized vector coprocessor for this purpose, again looking at SGI machines for example.
6690372: Who is doing this in a hardware unit?
 
Exophase ... with your logical approach it's a shame Intel didn't have you instead of being represented by multiple patent law firms that eventually decided it was better to "licensed Nvidia's GPU patents for $1.5 billion over six years".
 
Exophase ... with your logical approach it's a shame Intel didn't have you instead of being represented by multiple patent law firms that eventually decided it was better to "licensed Nvidia's GPU patents for $1.5 billion over six years".

That cash settlement was due to anti-competitive actions regarding chipset licenses, it was not a licensing fee for GPU patents. The additional broad cross-licensing of patents that accompanied this is a typical way to resolve a settlement and put the two companies on better terms. nVidia got access to completely unrelated Intel patents in addition to the cash payout. Receiving access to these patents doesn't mean that they were deemed to be infringing on them.

That said, defending weak patents does often result in settlements or even court victories. The patent office is full of patents that are very broad, vague, copies of prior products, obvious, or described by open standards. The office simply doesn't have the technical acumen required to properly audit the claims. It doesn't help that the patents are often written to be deliberately confusing and vague, filled with redundant and irrelevant information, legalese, and poor definition of the scope of the invention vs mere examples of its use.

What really gets me is when companies knowingly watch their competitors sell their products for years, only attacking with a lawsuit once they've established a huge ecosystem where they're utterly depending on the technology. This does not strike me as being in the spirit of the patent system.
 
That cash settlement was due to anti-competitive actions regarding chipset licenses, it was not a licensing fee for GPU patents.

That is not true. The cash settlement was about the renewal of the 2004 cross-license.

The Intel situation looks to be much more straightforward. As we mentioned previously, NVIDIA and Intel originally cross-licensed in 2004 so that Intel could build IGPs using NVIDIA patented technologies and methods. That agreement was set to expire this year, which would have been a massive problem for a company whose CPUs almost always include a GPU. Today’s agreement with NVIDIA renews and extends that original agreement: Intel continues to cross-license with NVIDIA, allowing them to produce IGPs that use/infringe on NVIDIA patents. To be clear we believe this is a continuation of existing practices

http://www.anandtech.com/show/4122/intel-settles-with-nvidia-more-money-fewer-problems-no-x86/2
 
Maybe as part of a settlement, NV and Samsung would announce a deal to ship Denver on some Samsung devices.
 
That is not true. The cash settlement was about the renewal of the 2004 cross-license.

The $1.5b settlement was to prevent nVidia from being awarded in court over Intel over a breach of a licensing agreement specifically regarding chipsets.

http://www.dailytech.com/FTC+Reviewing+Intel+NVIDIA+Legal+Spat/article17036.htm

In other words, it was due to damages claimed by nVidia, NOT a cost to renew the cross licensing agreement. It was because Intel claimed that the license did not grant nVidia the right to make Nehalem compatible chipsets because of the IMC, and were able to block nVidia from releasing such chipsets by court order. Intel must have realized that nVidia would be able to overturn this prohibition and thus owe damages.

Now look back at the original 2004 cross-license agreement:

http://www.nvidia.com/object/IO_17070.html

Do you think Intel paid a big cash payment to nVidia to be eligible for that cross license agreement then? No, it was an agreement of mutual benefit, covering an unknown swath of patents (by no means should we assume it was, for instance, limited to these 7 patents nVidia is citing now).

nVidia and Intel renewing the original cross license agreement is not part of the $1.5b in damages they agreed to pay, it was a separate part of the settlement.
 
Maybe as part of a settlement, NV and Samsung would announce a deal to ship Denver on some Samsung devices.

I wouldn't be surprised if Samsung is secretely fiddling around with a custom CPU core based on ARM ISA; I'm just worried that it might stink more than their former own GPU development attempts :LOL:

Another idea would be either Samsung or QCOM to buy NVIDIA and we can all live happily ever after (or something equally cheesy....). Rumors that large amounts of Samsung's personell was willing to commit suicide with the notion of having Jensen as CEO couldn't get verified until now.... :runaway:
 
Maybe as part of a settlement, NV and Samsung would announce a deal to ship Denver on some Samsung devices.

That would not be in the best interest of Nvidia.

Nvidia is interested in a more constant royalty revenue stream so having a (court determined) royalty amount per SOC would be great in todays shipping mobile quantities and much much better in future shipping mobile quantities.
 
I wouldn't be surprised if Samsung is secretely fiddling around with a custom CPU core based on ARM ISA; I'm just worried that it might stink more than their former own GPU development attempts :LOL:

Another idea would be either Samsung or QCOM to buy NVIDIA and we can all live happily ever after (or something equally cheesy....). Rumors that large amounts of Samsung's personell was willing to commit suicide with the notion of having Jensen as CEO couldn't get verified until now.... :runaway:

Lol.. maybe not suicide, but they could go work for Hyundai instead. But what will Nvidia do of Samsung City .. Called it Jensen city or Nvidia City ? And im ask me what will happend with Lehmann Brothers Asia .. do we need to rename it to Jensen and brothers ?

Anyway, i wait the next flies of sue against the Samsung VR gear ..
 
Last edited by a moderator:
Being that nVidia is forcing the issue, having these patents invalidated in court could ultimately be healthy for the industry.

Given the evolution of the research, API formulation and specification, and simultaneous multi-vendor development and implementation of many fundamental graphics approaches, nVidia has their work cut out for them in proving these patents are actionable.
 
Last edited by a moderator:
Assuming, for the sake of argument, that NVIDIA's claims are valid, I could understand why they'd sue Qualcomm. But Samsung seems to be not only a strange target, but also a dangerous one.

Samsung makes laptops and buys GPUs to equip them. If I were them right now, I'd be very reluctant to buy those GPUs from NVIDIA. Other big OEMs (Lenovo, Asus, etc.) sell mobile devices with Qualcomm SOCs and laptops with discrete GPUs. They're not being sued at this time, but it stands to reason that they would be next in line.

It just seems like a very dangerous move.

Edit: after a quick look at Samsung's website, it seems that only one of their products uses discrete graphics, and it's from AMD. Maybe that's why Samsung is being sued and not Apple.
 
No, I'm saying that if I were Samsung, I wouldn't buy products from a company suing me.
Ok.

Either way, if Samsung is anything like the big conglomerate I've worked for (they are), the Samsung Laptop division isn't going to care one bit about what's going on at the Samsung Mobile group. (In fact, if they are like the one I've worked for, the other divisions might be cheering right now. :D )

Samsung Laptops is probably a different legal entity, it wants to make products that people want to buy. Minor stuff like this in a remote other part of the company is no big deal. (Yes: minor stuff. This is going to be about a few $100M maximum per year. Peanuts for a division that makes tens of billions in profits per year.)
 
Assuming, for the sake of argument, that NVIDIA's claims are valid, I could understand why they'd sue Qualcomm. But Samsung seems to be not only a strange target, but also a dangerous one.
Isn't the subtext of the suit that passing the buck on licensing fees between Samsung and Qualcomm isn't a reason to ignore license fees at all (as far as Nvidia is concerned ofc) ?
Involving both companies also opens the door to both possibly having to defend their positions against each other as well as the Nvidia.
No, I'm saying that if I were Samsung, I wouldn't buy products from a company suing me.
They aren't buying products from Nvidia in any case, so what's the difference?

From a purely strategic PoV, what does Nvidia lose by pursuing this?
Worst case scenario: No case to answer, they're out legal fees, and the patents aren't used for further litigation.
Best case? Part or complete judgement in their favour and set legal precedent and a likely never ending revenue stream?
If you were a gambler (and JHH's personality says that he probably is), do you think the possible return of SoC royalties offsets the expenditure of a few million in legal fees?

As for Qualcomm, they seem to love the courtroom - Bandspeed, a looming antitrust case in China, Adaptix, ParkerVision coming back for another chunk of cash, and a possible EU case to answer - co-incidentally (?)initiated by Icera in June 2010.

I'd really like to find out what kind of cross-license and IP license structure ARM has in place. Considering ARM Holdings aren't shy about telling world+dog how many license holders they have, the only real info seems to come from the licensees press releases.
 
Back
Top