Nvidia Sues Qualcomm, Samsung Over GPU Patents

[Silverforce11]

http://en.wikipedia.org/wiki/GeForce_256


http://en.wikipedia.org/wiki/Graphics_processing_unit

Nvidia defines the GPU as a graphics chip that also calculate geometry and lighting. They were the first with such a chip, and they created the term "GPU" aka graphics processing unit.

Gee, I could have sworn I had many graphics cards before the Geforce.

What was it called back then?

 

[dacostafilipe]

nVidia invented the GPU.

"GPU" was invented by nVidia just like VPU was invented by ATI. Both are just something the marketing department "invented" to show us how advanced/awesome the stuff is.

So yes, nVidia invented the GPU and ATI the APU ... but they did not invent the video accelerator.

They where the first implementing T&L into hardware, but it was used long before in software. The same goes for shaders.

 

[NTMBK]

They invented the term "GPU - Graphics Processing Unit":
http://www.nvidia.com/object/IO_20020111_5424.html

NV10 was the first chip which integrated a hardware transformation and lightning unit in hardware.

Inventing a freaking marketing term is not the same as having a patentable technology.

I think any sane jury would agree that the Voodoo 1,2,3, the ATi Rage, etc all qualify as GPUs.

If NVidia does, in fact, have a patent on putting T&L onto the graphics card then that is a different matter. The question now is, "is this patent valid"?

 

[tviceman]

Gee, I could have sworn I had many graphics cards before the Geforce.

What was it called back then?

Video accelerators??? I don't remember. Did those video cards have geometry and lighting capabilities? Because you are the aforementioned authority on all things video chip, does that mean Nvidia's own created term is wrong and therefore should be invalidated because you owned something prior that Geforce that put pixels on a screen?

Also, http://forums.anandtech.com/showpost.php?p=36682219&postcount=22

 

[NTMBK]

Okay, so the actual patent is: https://www.google.com/patents/US6198488

But seriously, is putting it all together on a single piece of silicon patentable? This one is utterly laughable too:

https://www.google.com/patents/US7038685 Because freaking multithreading hasn't been around since the 60s or anything... Just saying "on a graphics processor" doesn't make it novel. It's as bad as the trolls with "on the internet!" patents.

 

[sontin]

"GPU" was invented by nVidia just like VPU was invented by ATI. Both are just something the marketing department "invented" to show us how advanced/awesome the stuff is.

So yes, nVidia invented the GPU and ATI the APU ... but they did not invent the video accelerator.

They where the first implementing T&L into hardware, but it was used long before in software. The same goes for shaders.

Problem is they patented it:
https://www.google.com/patents/US6198488

Okay, so the actual patent is: https://www.google.com/patents/US6198488

But seriously, is putting it all together on a single piece of silicon patentable? This one is utterly laughable too:

What? They did it a a time where no other company was able to do it. ImgTech wasnt able to it with Kyro 2...

 

[NTMBK]

Gee, I could have sworn I had many graphics cards before the Geforce.

What was it called back then?

"3D accelerator".

 

[Silverforce11]

"3D accelerator".

I bought Video Cards all these years, back to my first Rendition Verite and later 3dfx Voodoo addon.

Doesn't matter squat what its called, if NV think they can patent programmable shaders or multi-threading on a processor and it holds up in court.. well, let the court case begin and I will enjoy laughing at their demise.

This is nothing more than a bad tactic to get cross-license agreement so they don't have to pay royalties on Tegra. They are barking up the wrong tree, certainly against Samsung who has had extensive experience in patent court cases.

 

[NTMBK]

If anyone wants to peruse the patents in question, El Reg has linked to them all: http://www.theregister.co.uk/2014/09/04/nvidia_sues_qualcomm_and_samsung/

 

[Keysplayr]

If anyone wants to peruse the patents in question, El Reg has linked to them all: http://www.theregister.co.uk/2014/09/04/nvidia_sues_qualcomm_and_samsung/

Yeesh that is a lot of info in each patent.

 

[Silverforce11]

"Rendering pipeline"

 

[Cookie Monster]

There is a reason why Intel decided to go ahead with the 1.5 billion licensing deal for the next 6 years with nVIDIA.

Im sure they've done their homework, and when your a company that outlasted almost all of them in one market/technology segment, theres going to be plentiful of ammunition when it comes to this.

 

[DeathReborn]

If they can't make a GPU/Graphice Core without violating IP they should be banned from sale unless they pay requisite fees to allow them to continue. Going after Samsung might seem bad but remember Exynos is GPU "neutral" in the fact they've used PowerVR & Mali so they make a choice on that side.

Personally I think Nvidia should be paid licensing fees by IMGtec, ARM & Qualcomm (assuming they all breach), maybe Qualcomm will instead cross license Modem/GPU tech instead of fees going both ways. There is an element of "can't compete, sue" in this but it's understandable given the alleged IP "theft".

 

[Mand]

But seriously, is putting it all together on a single piece of silicon patentable?

In a word, yes. Unique implementations of things that hadn't been combined before are patentable. Yes, the components may have existed before, but the combination is novel, and that gets a patent. It's why there's a "prior art" section in patent filings.

 

[sontin]

There hadn't existed a hardware transformation and lightning unit on the same chip before nVidia introduced Geforce SDR (NV10). Before NV10 all these "GPUs" depended on the CPU for the calculations.

 

[rgallant]

samsung lawyers last statement before the jury go's in to debate on the verdict :

should you find my client guilty don't forget to leave your illegal smart phone's in that garbage can before leaving the court room .

 

[Khato]

Qualcomm had more revenue from licensing cellular modem IP in Q2 than Nvidia's total profit in the past two years. Everyone who makes a cellular modem pays money to Qualcomm.

Explain that to me and how it makes more sense than companies paying royalties to Nvidia and/or AMD for GPU technology, unless you think the same STFU mentality should be applied to Qualcomm.

Precisely. And I'd be quite surprised if what NVIDIA's looking for here is a cross-licensing agreement with Qualcomm across CPU, GPU, and modem IP. If you want to complain about abusive patents just go take a look at Qualcomm. None of what Qualcomm has on the modem side is any more innovative or unique than what NVIDIA has for graphics, it's just that no one else was building up a portfolio like Qualcomm was back in the 3G development time frame because it was expected to just be another commodity product. (aka, the time before smartphones.) Now Qualcomm is reaping the benefits rather abusively - every smartphone either pays them their huge royalties or gets them an SoC sale.

 

[Headfoot]

ITT: people unable to discern the difference between copyright and patent decry both systems as broken

 

[Exophase]

There hadn't existed a hardware transformation and lightning unit on the same chip before nVidia introduced Geforce SDR (NV10). Before NV10 all these "GPUs" depended on the CPU for the calculations.

They weren't the first to do it just because they were the first on a PC add-on card, which has zero to do with their claim. Look at SGI workstations, look at N64 which has a vertex processor (which is on the same chip as the rasterizer, not that it should even matter)

 

[monstercameron]

arent quallcomm's modem patents standards essential?

 

[ams23]

They weren't the first to do it just because they were the first on a PC add-on card, which has zero to do with their claim. Look at SGI workstations, look at N64 which has a vertex processor (which is on the same chip as the rasterizer, not that it should even matter)

Speaking of SGI, instead of going through with actual litigation, SGI formed a "strategic alliance" with NVIDIA more than ten years ago where both companies entered into a patent cross-licensing agreement.

 

[Exophase]

Speaking of SGI, instead of going through with actual litigation, SGI formed a "strategic alliance" with NVIDIA more than ten years ago where both companies entered into a patent cross-licensing agreement.

And? SGI is not defending their patents here, nor is nVidia defending patents they acquired from SGI. nVidia is defending patents that for the most part describe technology that existed before the patents were written. It doesn't matter if nVidia made an alliance with those technology creators to avoid litigation.

 

[Keysplayr]

And? SGI is not defending their patents here, nor is nVidia defending patents they acquired from SGI. nVidia is defending patents that for the most part describe technology that existed before the patents were written. It doesn't matter if nVidia made an alliance with those technology creators to avoid litigation.

What is your point though? A patent is a patent is a patent. Wouldn't you agree? If a patent is truly found to be infringed upon, then restitution should be agreed upon and paid. If it is found that no patent was infringed upon, then a new suit should be brought about to pay for the legal fees incurred from initial suit. IMHO.

 

[xpea]

And? SGI is not defending their patents here, nor is nVidia defending patents they acquired from SGI. nVidia is defending patents that for the most part describe technology that existed before the patents were written. It doesn't matter if nVidia made an alliance with those technology creators to avoid litigation.

well going this route, nothing can receive a patent as new tech is always an evolution of prior art...
But between you an me, when Geforce256 aka NV10 came to the market, with hardware T&L, it was revolutionary. Nothing was like that at that time. Same for NV30 with hardware programmable shaders (even if this GPU suffered from many flaws). Each time, competitors came very late to the market with their similar tech.
You may disagree with nvidia claims but nobody will argue against the prominent role they played in the graphic technology evolution. They invest heavily in R&D, they patent their ideas first and they put it in real products...

 

[lefty2]

Nvidia said they started talks with Qualcomm in 2012, but the first Adreno GPU was released in 2008, so why did Nvidia wait 4 years before they decided that they were infringing?