• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

When you can't sell your hardware on its own merits...

It's staggering that PowerVR had all this amazing technology, yet they didn't tell anyone about it.

Oh how different things could have been.
 
It's staggering that PowerVR had all this amazing technology, yet they didn't tell anyone about it.

Oh how different things could have been.
What are you on about? They did tell people about it and was all over the news and web. That’s how PowerVR make their money. They license and sold technology to companies like STmicro, Sega and more recently even Apple and Intel licence Power graphics technology. PowerVR are one of the largest GPU designers in the world and one of the oldest. If large company's like STmicro and Sega know about the amazing technology and licensed it how can you say they didn't tell anyone about it? More importantly if they licensed it before NVidias patents surly I am correct and what you said is wrong.
 
Last edited:
What are you on about? They did tell people about it and was all over the news and web. That’s how PowerVR make their money. They license and sold technology to companies like STmicro, Sega and more recently even Apple and Intel licence Power graphics technology. PowerVR are one of the largest GPU designers in the world and one of the oldest. If large company's like STmicro and Sega know about the amazing technology and licensed it how can you say they didn't tell anyone about it? More importantly if they licensed it before NVidias patents surly I am correct and what you said is wrong.
I might be wrong, but it just looks to me that Nvidia is jelly of PowerVR's success on the mobile platform occupying such a huge cashcow and they couldn't do anything about it, and trying to compete with them directly has gotten them nowhere. So they just resort to using a different mean to try to getting them out of the picture, as if Nvidia CPU/GPU end-up taking existing mobile space occupied by Qualcomm/PowerVR, it could be very lucrative for them. So whether or not their claim being legit or not, may be it is a worthwhile gamble for them in business sense.

And regarding how odd of that Nvidia is targetting Samsung...I wonder if the reason behind it was that they want to convince/force Samsung to drop the use of Qualcomm/PowerVR and adapt Tegra instead...I mean afterall, Samsung is the largest distrubing power/has the highest turnover on the Android platform :p
 
Last edited:
I might be wrong, but it just looks to me that Nvidia is jelly of PowerVR's success on the mobile platform occupying such a huge cashcow and they couldn't do anything about it, and trying to compete with them directly has gotten them nowhere. So they just resort to using a different mean to try to getting them out of the picture, as if Nvidia CPU/GPU end-up taking existing mobile space occupied by Qualcomm/PowerVR, it could be very lucrative for them. So whether or not their claim being legit or not, may be it is a worthwhile gamble for them in business sense.

And regarding how odd of that Nvidia is targetting Samsung...I wonder if the reason behind it was that they want to convince/force Samsung to drop the use of Qualcomm/PowerVR and adapt Tegra instead...I mean afterall, Samsung is the largest distrubing power/has the highest turnover on the Android platform :p
Right now Arm is getting money from both GPU and CPU from Samsung. NVidia forcing Tegra onto Samsung and going after ARM’s GPU is a very dangerous game. If NVidia hurt Arm to much ARM might withdraw their CPU licence to NVidia. That in turn means NVidia lose Tegra which would be a massive hurt. By being too forceful NVidia could end up an even worse situation then they are already in.

Another problem is the mobile world is in the process of moving to ray tracing (still years out from being widespread). Everyone could just go sod you and abandon today’s style mobile GPU altogether in favour of pure ray tracing chips leaving NVidia out. In fact it wouldn’t surprise me if ray tracing is one of the things that have got NVidia worried and taking aggressive action while they can. If I am right the future of mobile is ray tracing which makes NVidia patents and currently technology obsolete. (in the mobile world that is)
 
Last edited:
Pottsey as I have suggested, maybe if you posted up some evidence rather than just posting up unverified info, people would believe you.
I have searched for PowerVR patents on transform and lighting and I cannot find any, not that it should be our job to find the evidence that shows us all to be wrong. ;)

Just a side note and not saying it is relevant at all.

I assume you know that Gigapixel was one of the companies who also invented some IP for tile based rendering that was bought out by 3DFX who in turn was bought out by Nvidia.
 
If I was Samsung, I would have said the same to Nvidia "Why are you talking to me? You are talking to the wrong guy! If you have graphic patent infringement related issue, go talk to Qualcomm/PowerVR instead, not us!" and ignore them.

According to Nvidia's statement that's what Samsung have been doing for years hence the lawsuit.
 
What I said was based directly on that statement. To me I think Nvidia is barking up the wrong tree, and that's what Samsung believe as well.

+1, Samsung don't make the chips. Qualcomm is who they need to be barking at.

Perhaps Nvidia see Samsung as an easy target as they gave in to Apple when Apple really didn't have a case.
 
According to Nvidia's statement that's what Samsung have been doing for years hence the lawsuit.
Many of the mobile GPU’s date back 13 years. Some of the PowerVR desktop GPU date back 18 years or 20 if you count development time. Which makes me wonder why now?



Pottsey as I have suggested, maybe if you posted up some evidence rather than just posting up unverified info, people would believe you.
I have searched for PowerVR patents on transform and lighting and I cannot find any, not that it should be our job to find the evidence that shows us all to be wrong. ;)

I thought I posted evidence. I gave you the names of the technology which proves it’s real. Not sure why you cannot find anything a quick search gives me http://www.segatech.com/arcade/naomi2/index.html
“VideoLogic custom transformation & lighting (T&L) chip (Code named: ELAN)”
PowerVR created the technology, patented it then licenced it to Sega.
Ok not a patent but it proves the technology is real and the patents would be older then that.



Just a side note and not saying it is relevant at all.

I assume you know that Gigapixel was one of the companies who also invented some IP for tile based rendering that was bought out by 3DFX who in turn was bought out by Nvidia .
It might be relevant. I am aware but PowerVR predate Gigapixle so it is very unlikely there patents cannot be used against them. But they could be used against ARM and the others.
 
Those companies’ then typically spend 2 to 3 years developing a product like the Dreamcast. The cycle is about 5 years long give or take 1 year. For Sega to have a product out in shops in 2000 they must have at the very least licensed/brought the technology from PoweVR in 1998. That’s a 2 year development cycle for Sega which would be oddly short.

Except we are not talking about a product like the Dreamcast, we are talking about the controller board for an arcade machine which was basically a beefed up version of it's predecessor. The development time is much much shorter, all they had to do was take an existing design and add three chips, If NAOMI 2 took SEGA more than 12 months to design and get to market it would represent a major failure on their part.


That makes no sense PowerVR sold a finale patented finished IP product pre 1999. Gefroce was 1999 yet I am still wrong?

You're saying that the SEGA board that came out in 2000 must have been in development for years, whereas the Nvidia card that came out in 1999 had no development time, can you really not see the flaw in that.


Please don’t make things up that is completely incorrect.

I didn't make anything up, if it's incorrect that's down to wiki (I can't be expected to remember everything, I binned my last PowerVR card when I got a GeForce 3) but according to recorded history the first PowerVR based GPU to have HW T&L would have been STM's 2002ish STG5000 chip, based upon the PowerVR4, but it never came to commercial fruition, it and the KYRO 3 were shelved due to STMicro closing its graphics division.


No PowerVR card was shelved or division closed at least not from the company behind PowerVR.

As above, those ones were when Kyro cards died off and PowerVR exited the desktop GPU market.


The Kyro card did not become uncompetitive due to T&L in fact they had enhanced T&L which made them more competitive.

Sorry but that is complete and utter fanboy ********, I owned one of the last of the Kyros, a Hercules 3D Prophet and it couldn't wipe it's own backside against the Nvidia/ATi cards of the day, I ended up changing it after I went to a Lanparty and people laughed at my Quake 3 performance. Enhanced T&L was a joke, because it was a cool way of saying "fake HW T&L", all it did was trick a program that required HW into thinking it was there so it would launch, then the T&L was passed to the CPU as usual.


I agree with 1999 but what you don’t seem to realise is the PowerVR T&L technology had been in development and sold and licensed out to other company’s before 1999. This is an important factor for the case. How can NVidia say they own T&L when they patented it in 1999 and had a product in 1999 when PowerVR can show they patented it and sold the technology pre 1999?

The issue there is though can they prove that? I haven't seen any evidence of patents or hardware of theirs for T&L that pre-date Nvidias.


Here’s an interesting question what if PowerVR fight back and say NVidia are breaking there FSAA patents and universal shaders and all the other stuff they copied.

IIRC it was 3DFX that did FSAA first not PowerVR, so the 3DFX patent would have passed to Nvidia with the companies IP.


That is incorrect. I strongly recommend you read up on the company behind PowerVR as you don’t seem to understand what happened, how the company works. The company behind PowerVR do not produce 3dcards for the most part. They are an IP (Intellectual Property) company. They design technology and sell/licence that technology to other companies. Those other company then create consumer products and none consumer products.

Semantics aside I am fairly sure you understood what I meant :P


Pre 1999 PowerVR sold/licenced the enhanced T&L series 3 chips to STMicro and the hardware T&L chips to Sega. Then in 2000 STMicro had the Kyro cards out in shops which had enhanced T&L and Sega had the hardware T&L arcade board out. Considering the typical development cycle even if we shorten it to the bare minimum of 2 years that puts us at 1998 when PowerVR sold the finished patented technology to both companies. Long before NVidia patent. If you read up on development cycles less than 2 years is unrealistic and if STMicro or Sega brought the T&L technology in 2000 it would have been impossible to have a product out in 2000.

Those STMicro products are irrelevant as they had no HW T&L, the SEGA one (NAOMI 2) could easily have been developed in under a year as it was not a new system just an upgraded version of the previous one. It required no more work than a Playstation motherboard revision (and the PS2 had 14+ motherboard revisions in it's lifetime).
 
Many of the mobile GPU’s date back 13 years. Some of the PowerVR desktop GPU date back 18 years or 20 if you count development time. Which makes me wonder why now?





I thought I posted evidence. I gave you the names of the technology which proves it’s real. Not sure why you cannot find anything a quick search gives me http://www.segatech.com/arcade/naomi2/index.html
“VideoLogic custom transformation & lighting (T&L) chip (Code named: ELAN)”
PowerVR created the technology, patented it then licenced it to Sega.
Ok not a patent but it proves the technology is real and the patents would be older then that.




It might be relevant. I am aware but PowerVR predate Gigapixle so it is very unlikely there patents cannot be used against them. But they could be used against ARM and the others.

Interesting link that.
It shows that the transform and lighting in the Dreamcast was in fact done on a separate geometry co processor, maybe this why Nvidia can patent GPU based T&L.
 
“You're saying that the SEGA board that came out in 2000 must have been in development for years, whereas the Nvidia card that came out in 1999 had no development time, can you really not see the flaw in that.”
I don’t sees the flew as its different company structures. PowerVR complete development of technology then put it up for license sale. Sometime it takes years before anyone takes a license, sometimes no one does. Either way once a license is taken it then typically takes over a year to make its way into the other company's end product. What matters is when PowerVR developed and patented the technology not when a consumer brought an end product. PowerVR are not about end products they are an IP company. 80% of the company is R&D only. There focuses is on deigning and licencing technology not production.

NVidia are not an IP focuses development company. They work in a completely different way.



“I didn't make anything up, if it's incorrect that's down to wiki”
Almost all the T&L comments and history of the companies you posted are wrong and do not match the wiki. Perhaps it was a mistake but it’s still incorrect.



“As above, those ones were when Kyro cards died off and PowerVR exited the desktop GPU market.”
First PowerVR never exited the desktop card market they still have desktop cards today. 2nd the Kyro was not a direct PowerVR product. It was an STmicro product that used technology licensed from PowerVR. As for benchmarks and speeds of Kyro I am not interested in arguing how good or bad a 10+ year old 3dcard performance was. Got better things to do then dig though old benchmarks.



“IIRC it was 3DFX that did FSAA first not PowerVR, so the 3DFX patent would have passed to Nvidia with the companies IP.”
That’s a comment misconception. The PowerVR series 2 chips had FSAA long before anyone else.



“but according to recorded history the first PowerVR based GPU to have HW T&L would have been STM's 2002ish STG5000 chip, based upon the PowerVR4, but it never came to commercial fruition, it and the KYRO 3 were shelved due to STMicro closing its graphics division.”
Well the wiks are wrong then as series 2 had an option for hardware T&L. Series 2 predates 1999. Also remember the wiki are about end consumer products. Not what PowerVR have open to license.
 
Interesting link that.
It shows that the transform and lighting in the Dreamcast was in fact done on a separate geometry co processor, maybe this why Nvidia can patent GPU based T&L.
As I understand it NVidia patent mentions T&L done on the same platform not on the same chip so a co-processor counts.

The patent in question is http://www.google.com/patents/US6198488 to quote it “Transform, lighting and rasterization system embodied on a single semiconductor platform” To me the entire board with co processor is one platform.
 
As I understand it NVidia patent mentions T&L done on the same platform not on the same chip so a co-processor counts.

The patent in question is http://www.google.com/patents/US6198488 to quote it “Transform, lighting and rasterization system embodied on a single semiconductor platform” To me the entire board with co processor is one platform.

Well the definition of a Semiconductor would seem to disagree with you there, a semiconductor is a single integrated chip not a collection of them on a circuit board. Therefore 'a single semiconductor platform' is referring to a single chip.


1. Any of various solid crystalline substances, such as germanium or silicon, having electrical conductivity greater than insulators but less than good conductors, and used especially as a base material for computer chips and other electronic devices.

2. An integrated circuit or other electronic component containing a semiconductor as a base material.

Source http://www.thefreedictionary.com/semiconductor
 
To me "semiconductor platform" describes the new stacking technologies which include off-die RAM stacks on interposers. Platform to me implies non-monolithic...
 
The consequences of nV winning this one would be pretty catastrophic for the whole industry. They could put all GPU makers out of business, surely? Not that I think they'll win.
 
The consequences of nV winning this one would be pretty catastrophic for the whole industry. They could put all GPU makers out of business, surely? Not that I think they'll win.
It might be in the short term but if the mobile worlds moves towards ray tracing it could end up catastrophic for NVidia who will end up with obsolete mobile technology and useless patents for mobile along with having upset every major company in the mobile world. Who would work with them after that? What would they do? Ray tracing is so different from what NV do they cannot stop it .Nv could force the company's to move to ray tracing to get away from NV. So it could really backfire on Nv. Not the mention the possibility of Nv losing the Tegra CPU which will kill Tegra.



Well the definition of a Semiconductor would seem to disagree with you there, a semiconductor is a single integrated chip not a collection of them on a circuit board. Therefore 'a single semiconductor platform' is referring to a single chip.
Source http://www.thefreedictionary.com/semiconductor
I agree a semiconductor is a single integrated chip but I thought when you add the word platform it means the entire group of integrated chips or underlying hardware for a single system. Am I wrong in that? A hardware platform is normally a group of chips isn’t it?

Another interesting question can you argue moving a co processer on board a chip is evolution and not something you can patent? I am by no means anywhere near an expert on the US patents system but I thought ideas and evolution of systems couldn’t be patented over there. You can patent the method but not the idea? Is that right?
 
Interesting link that.
It shows that the transform and lighting in the Dreamcast was in fact done on a separate geometry co processor, maybe this why Nvidia can patent GPU based T&L.

Actually that link is about an arcade board that came out in 2000, as mentioned earlier the Dreamcast never had HW T&L.
 
What matters is when PowerVR developed and patented the technology not when a consumer brought an end product.

And you don't think that if they had developed and patented the technology before Nvidia that it would have made it into an end product sooner? Doesn't that sound likely as HW T&L was in demand at the time? but according to you they managed to design and patent it and it then took their licensees so long to get a product out that both Nvidia and ATi had beaten them too it by years... Oh and despite being an IP company they never took any issue with Nvidia gaining market dominance via the use of one of their patents... right...


As for benchmarks and speeds of Kyro I am not interested in arguing how good or bad a 10+ year old 3dcard performance was. Got better things to do then dig though old benchmarks.

I got my Kyro II SE on launch and it was slightly weaker than my friends GF2 and destroyed by another friends GF3 (I knew it wasn't going to be as much bang for buck when I bought it, I just wanted something that wasn't ATi/Nvidia because I'm odd lol). However as HW T&L became commonplace in games it just got left behind, I think I replaced it with a GF4 Ti4200 if memory serves right.


Well the wiks are wrong then as series 2 had an option for hardware T&L. Series 2 predates 1999. Also remember the wiki are about end consumer products. Not what PowerVR have open to license.

Series 2 didn't have an option for HW T&L, that's the whole reason the NAOMI 2 needed to have an additional T&L co-processor, because its series 2 GPU's couldn't do it!
 
Back
Top Bottom