NVIDIA shows signs ... [2008 - 2017]

[aaronspink]

The problem comes in when they join a board like JEDEC in which you are supposed to disclose any relevant patents that you hold with regards to any standards that are being proposed.

actually this wasn't really true until after the whole rambus/jedec thing happened. From the history and documents available, everyone was playing fast and loose with the patents. See, most of the members had outside x-licenses already, etc.

And that license fee's from those patents will be "reasonable and non-discriminatory." That way memory makers making memory to JEDEC standards wouldn't have to pay unreasonable royalties when making JEDEC standardized memory.

define reasonable.

Slimy yes, but not exactly a patent troll. IE - they don't submit patents for the sole purpose of litigation. They actually do some good design work. But the way they use those patents to make their own work more attractive is very very slimy.

by that definition, all patents are slimy. Patents have the express purpose of protecting your work. It appears that rambus simply used them as defined.

 

[aaronspink]

Not really, No. But thats personal opinion. I dont feel anything rambus has done anything that couldn't have been done by someone else in the industry. Or that anything they have done is really revolutionary.

that's interesting. They've continuously had the best signaling technology in the industry for ~10 years now and have been able to demonstrate that technology in low cost consumer products. You may not understand the technology enough to understand how far ahead they really are.

Is rambus reputation deserved? I don't know. What I do know is you hear more about rambus sueing everybody than about anything they have ever made in the last 7 years.

Then you have selective hearing.

 

[aaronspink]

Which in itself (as SB points out) was in violation of agreements. Rambus gambled that they had more to gain than to loose by their conduct, leveraging and arguably manipulating the patent legal system.
The reason many of us would like to see Rambus dead, is that if this would prove to be correct, it would be a disaster for all sorts of cooperative bodies trying to promote common standards. Even as it is, their actions has already to some degree damaged trust (and by consequence, efficiency) of such groups.
They need to be ostracised.

It might do you good to actually understand the enviroment and the issues and contracts involved before spouting out nonsense.

 

[aaronspink]

No, however in the specific case of Rambus with regards to JEDEC and SDRAM/DDR manufacturers, the court agreed that Rambus' actions were in violation and thus their litigation was thrown out when it was finally brought to court.

except IIRC, that was overturned.

If Rambus had managed to inflate the price of DDR at the crucial time when the open DDR standard was in competition with the proprietary Rambus memory standard, Rambus stood to gain billions as Intel at the time was pushing hard for it. However the cost to produce Rambus memory at the time was almost twice as high as DDR making it unattractive in comparison.

you mean like the DRAM companies have done time and again with both standard and non-standard drams?

 

[Silent_Buddha]

define reasonable.

That's not up to me. That's up to JEDEC the organization that set the standards along with the memory manufacturer's that are a member of that organization.

by that definition, all patents are slimy. Patents have the express purpose of protecting your work. It appears that rambus simply used them as defined.

Not really as they had to sign agreements and contracts with regards to patents held that are relevant to the memory standards that were being conceived while a member of JEDEC. As well not using the information received as to what technology will be used for the memory standard to tailor and file patents while a member of JEDEC with regards to standards under developement.

Those were the main falling points for Rambus in the specific case of SDRAM and more importantly DDR.

They've been perfectly fine with regards to protecting their patents in other cases. Although they may have over reached in some cases. For example the recent case against Nvidia.

And this has nothing to do with whether their technology is good or not. I personally find some of their solutions interesting. However, I still don't like the company itself for what it tried to pull with regards to SDRAM and DDR.

Regards,
SB

 

[aaronspink]

That's not up to me. That's up to JEDEC the organization that set the standards along with the memory manufacturer's that are a member of that organization.

Not really as they had to sign agreements and contracts with regards to patents held that are relevant to the memory standards that were being conceived while a member of JEDEC. As well not using the information received as to what technology will be used for the memory standard to tailor and file patents while a member of JEDEC with regards to standards under developement.

Oh, you mean the contracts that were found in court to be void and unenforceable? You mean the contracts that were willfully, deliberately, and continuously violated by members of JEDEC without recourse?

Or do you mean the contracts that the memory companies tried to get Rambus to sign which would of given all their IP to the memory companies/JEDEC that rambus explicitly refused to sign?

If anything, the bigger villians in the whole case are the memory companies and by extension JEDEC itself.

 

[ChrisRay]

that's interesting. They've continuously had the best signaling technology in the industry for ~10 years now and have been able to demonstrate that technology in low cost consumer products. You may not understand the technology enough to understand how far ahead they really are.

Maybe not. Never claimed to be an expert on rambus technology. If you were actually not so busy trying to troll my post. You'd see that my original post was speculation as to why people ((and myself)) feel that rambus isnt' paticularly useful to the industry. In the end. I still don't feel they are. And it'd take alot to change my mind.

Then you have selective hearing.

It has nothing do with selective hearing. It has to do with the news thats reported on rambus. Google up rambus sometime.

The first 6 links are rambus dedicated links. The next 50-60 are about litigations, lawsuits, antitrust. and the sort. Like I said. All you hear about is rambus lawsuits. If you want to blame someone for not getting their product information out the door and marketed. Blame rambus. As it is. The only thing news websites have to report on is all the lawsuits.

http://www.google.com/search?q=ramb...rg.mozilla:en-US:official&hs=jEg&start=0&sa=N

 

[rpg.314]

Think about it. Why is Rambus a classic patent troll? If they sue X, X can countersue saying that that the IP they license out is infringing and demand injunction. Can't they?

 

[Entropy]

It might do you good to actually understand the enviroment and the issues and contracts involved before spouting out nonsense.

It's a bit odd that you say that I'm mistaken in that their actions have repercussions, when you yourself write:

actually this wasn't really true until after the whole rambus/jedec thing happened.

You're right that I'm not an electronics guy by profession. I do have some intimate familiarity with patents though, and through my professonal position with collaborative groups where trust is a necessary ingredient.

Furthermore, the Rambus case has been widely publisized, and I was curious about it since the very beginning, since I followed discussions between proponents of serialized memory interfaces from very early on due to my interest in high performance computing.

My opinion is clearly my own, but I don't hide it - their actions need to be punished, the legal system doesn't help, indeed it's the tool Rambus meant to exploit! So their peers need to ostracise them.

 

[aaronspink]

Furthermore, the Rambus case has been widely publisized, and I was curious about it since the very beginning, since I followed discussions between proponents of serialized memory interfaces from very early on due to my interest in high performance computing.

My opinion is clearly my own, but I don't hide it - their actions need to be punished, the legal system doesn't help, indeed it's the tool Rambus meant to exploit! So their peers need to ostracise them.

The only reason they joined JEDEC in the first place is because they were pushed to join by the various dram companies that make up JEDEC AFTER rambus has approached them about licensing the rambus IP for enhanced/future dram products.

No one is clean here, but rambus is not worse than any of the various DRAM companies who had a strategy to try to get ahold of rambus's IP via nefarious means.

 

[aaronspink]

Maybe not. Never claimed to be an expert on rambus technology. If you were actually not so busy trying to troll my post. You'd see that my original post was speculation as to why people ((and myself)) feel that rambus isnt' paticularly useful to the industry. In the end. I still don't feel they are. And it'd take alot to change my mind.

then in fact you are the one that is trolling the thread. Rambus has in fact done a large amount of research and enhancing the state of the art in signaling technologies.

The first 6 links are rambus dedicated links. The next 50-60 are about litigations, lawsuits, antitrust. and the sort. Like I said. All you hear about is rambus lawsuits. If you want to blame someone for not getting their product information out the door and marketed. Blame rambus. As it is. The only thing news websites have to report on is all the lawsuits.

Rambus does a fairly good job of getting their product information out to though that use it, which are not end users. Their product marketing isn't aimed at the general consumer.

 

[ChrisRay]

then in fact you are the one that is trolling the thread. Rambus has in fact done a large amount of research and enhancing the state of the art in signaling technologies.

I haven't trolled anything. I stated an opinion about what and what I thought that others thought rambus as a company ((including myself)). If I was wrong about rambuss contributions no matter how meager I consider them to be then so be it. Considering there was discussion about why rambus is not well thought of company by many enthusiasts. I was hardly out of line. This is not an opinion that I just made up. Simply based on my experience with people's commentary on the company and how they view them. There's nothing "trollish" about that. I didn't read anywhere in the rules that I couldn't post said opinion or my thoughts and experience. If you have something meaningful to say. Then please say it. If we don't understand it. Then please explain it.

Otherwise, Just let it go.

Rambus does a fairly good job of getting their product information out to though that use it, which are not end users. Their product marketing isn't aimed at the general consumer.

Maybe. But that doesn't change what I said. Nor your comment about me "having selective hearing". Which was wrong and you know it.

Public information has painted rambus in a bad light in the enthusiast and hardware community. If all the good things they supposedly do for the industry are kept secret. Nobody is going to know about it. And all everyone hears about is the legal material. I still haven't heard anything that makes rambus company that unreplacable in this industry. Or hell. I haven't heard anything that would make me miss rambus if they were to dissapear.

Which comes back to my original question. Does rambus deserve its reputation? Let me remind you that I said. That I don't know. Only what I have seen and heard about them.

 

[aaronspink]

Which comes back to my original question. Does rambus deserve its reputation? Let me remind you that I said. That I don't know. Only what I have seen and heard about them.

what do you suggest they do, pay people to post positive things about them all over the internet as some companies have done in the past and probably still do?

 

[ChrisRay]

I honestly have no expectations of rambus, I thought that was clear. But you can't blame people for having a bad opinion of them either. Because as I said. All you hear is negative press about them.

 

[John Reynolds]

what do you suggest they do, pay people to post positive things about them all over the internet as some companies have done in the past and probably still do?

Ouch! Unnecessary roughness, 5 yard penalty Mr. Spink.

 

[rjc]

Two articles from Ms Barak at the Inquirer on Nvidia Analysts Conference:

Nvidia lays out Plans for Analysts:

Nvidia says it is expecting to see Tegra account for about 50 per cent of its overall revenue longer term. This depends on the market and whether the company can maintain average selling prices and margins to justify the 500-600 people it has working in its Tegra group and the $100 million per quarter in operating expenses.

Nvidia a Company in Transition:

Nvidia told analysts it was putting a lot of faith into "co-processing" and CEO Jen-Hsun Huang emphasized his firm's determination to diversify by aiming for a future product mix of Tegra at 50% of sales and GeForce 25% to 30%, with Quadro and other professional kit making up the balance.

Big call on Tegra, market is huge but competition is fierce, good luck guys! . Hmmm, looks like the previous question about consumer GPUs returning to 40%+ margin has been answered: No

It kind of looks like the future will be a massive increase in the number of gpu units sold, but most at pretty much commodity prices.

 

[JoshMST]

Threw this one up last night. I honestly am confused from a business perspective why NVIDIA is making the decisions they are in the chipset arena.

http://www.pcper.com/article.php?aid=734

Hopefully NV will respond to this one.

 

[neliz]

Threw this one up last night. I honestly am confused from a business perspective why NVIDIA is making the decisions they are in the chipset arena.

http://www.pcper.com/article.php?aid=734

Hopefully NV will respond to this one.

Did they ever respond to Charlie's "nv chipsets are doomed!" articles?

Nice read though

 

[JoshMST]

Does NVIDIA even respond to Charlie anymore?

 

[Arun]

I really shouldn't be taking too much time to reply (or even read this thread but heh), however... Josh, I think you're missing three things strategically:
1) NV's chipset business has basically ALWAYS lost money operationally before MCP61's very successful intro in August 2006. There is little to no money to be made in the discrete chipset business.
2) Intel's S775 business isn't going away quite as fast as people think. I can't find the graphs anymore, but ir will remain a large percentage of mix for a long time. Any OEM that wants to use Ion2 with a Wolfdale(-3M) in 2010/2011 will be able to, no matter what Intel thinks. Laptops are slightly more complicated.
3) Seems to me Jen-Hsun loves Rayfield (who doesn't?) and so I suspect he's willing to bet quite a bit on Atom eventually becoming irrelevant relative to Tegra. Can't blame him, to be honest...