Sign in with
Sign up | Sign in
Your question
Closed

AMD blows $75 million on Rambus

Tags:
  • CPUs
  • DDR2
  • AMD
  • License
Last response: in CPUs
Share
January 3, 2006 9:12:14 PM

http://www.channelregister.co.uk/2006/01/03/amd_license...

I guess all of Rambus's suing has paid off. AMD would rather license their so called "intellectual property" in the DDR2 standard than risk getting sued. That's basically what this deal boils down to IMO.

It's a sad day when a major company like AMD validates Rambus's claims.

-mpjesse

More about : amd blows million rambus

Anonymous
a b à CPUs
January 3, 2006 9:33:29 PM

Buying better tech is blowing money?

I don't get it.

Maybe you think that since most of the world for some reason hates RAMBUS they should give away their technology?

I don't get it.

AMD never had to validate anything since it has been common knowledge for years that the RAMBUS tech is/was superior.

I don't get it.
January 3, 2006 9:36:49 PM

WOW. (-_-) Rambus actions lately>>> I counter sue you infinity no take backs or turnseys.
Related resources
January 3, 2006 9:47:38 PM

You think by chance AMD is actually licensing tech as so they do not have to reinvent the wheel??
Maybee this way there will be a performance gain with switching to DDR2 instead of for example Intels parallel shift into DDR2.
January 3, 2006 10:20:08 PM

If you read some of the businiss articles reporting on this you will discover that the DDR2 tech is not the highlight. Forbes is speculating this a move toward XDR ( god I hope not RDRAM had/has terrible pricing ).

Also this is not the first company to license this tech. Infineon has also paid up.
January 4, 2006 2:52:37 AM

Infineon paid up because they were being sued. In the end it was cheaper for infineon to pay royalties than to fight in court.

Did everyone forget all the crap Rambus put the entire memory industry through? Did everyone forget that they sued anyone who make DDR when it first came out? Did everyone forget that they claim they invented DDR and DDR2? Did everyone forget that the courts said "uhhhhhh sorry, you don't own any tech that has to do w/ DDR. go fly a kite and good luck with that XDR crap."?

Maybe I missed something here, but RDRAM doesn't even remotely resemble DDR in terms of how it works. Rambus just got pissed off that no one wanted to pay their frigin 5% royalty fees and the entire industry shat on them. And then Rambus really pissed off everyone when they went after JEDEC and DDR. Up until Rambus came along the memory industry and JEDEC openly and freely shared memory technology. Even Intel gave up on Rambus.

And now I see AMD making the same mistake that Intel made 10 years ago. I sincerely hope that Rambus has seen the err of their ways and has changed. Instead of focusing on hunting down every last memory manufacturer for royalties, I hope they focus on technology from now on.

As for XDR, I seriously doubt that AMD will move to it. (and if they do, straight back to intel I go for my next CPU). Until they prove to me otherwise, Rambus is still a crap company. They're no different than SCO in my opinion. This of course is entirely my opinion.

What I don't get is this: if AMD is planning to move to XDR in the future, then why the hell does Rambus want AMD to license the tech??? Wouldn't they make more money off royalties for XDR RIMMs? Maybe Rambus truly has learned their lesson; they can't force royalties down memory manufacturer's throats.

Who knows.

-mpjesse
January 4, 2006 3:03:41 AM

This is what I remember. It may not be perfect, but it is what I remember.
Amd and samsung were colaberating on a new way of using memory. They brought the concept to JEDEC, as an open source standard.
RD Ram made some suggestions on how to improve it. Seemed like a good idea so it became part of the standard.
RD informed everyone de facto, that the concepts they had proposed were patented, so everyone who used DDR would have to pay them royalties.
Does that make RD a bunch of sneaky slimy snakes? All I know for sure is my opinion, and I'm not saying. You be the judge.
January 4, 2006 3:25:36 AM

Quote:
This is what I remember. It may not be perfect, but it is what I remember.
Amd and samsung were colaberating on a new way of using memory. They brought the concept to JEDEC, as an open source standard.
RD Ram made some suggestions on how to improve it. Seemed like a good idea so it became part of the standard.
RD informed everyone de facto, that the concepts they had proposed were patented, so everyone who used DDR would have to pay them royalties.
Does that make RD a bunch of sneaky slimy snakes? All I know for sure is my opinion, and I'm not saying. You be the judge.


In my opinion, yes. Rambus knew that JEDEC has always been an open standard forum. Did they really think that everyone would say "Gee thanks Rambus, here's your cut! Oh wait, JEDEC is supposed to be open standard. Go to hell." Maybe they did, maybe they didn't. The point is Rambus should never have gone to JEDEC if they always had the intention of licensing RDRAM to members of JEDEC. That's a little bit like SCO joining the Open Source Initiative. They stand for everything that is not JEDEC. And they know it. Here's a brief summary of the lawsuits take from www.wikipedia.org:

Quote:
In the early 1990s, Rambus was invited to join the Joint Electron Device Engineering Council. Rambus had been trying to interest memory manufacturers in licensing their proprietary memory interface, and all companies had signed non-disclosure agreements to view Rambus' technical data. During the later Infineon v. Rambus trial, Infineon memos from a meeting with representatives of other manufacturers surfaced, including the line "[O]ne day all computers will be built this way, but hopefully without the royalties going to Rambus," and continuing with a strategy discussion including options for buying and burying Rambus' technology. As Rambus continued its participation in JEDEC, it became apparent that their idea for a standard was not going to be allowed to be presented for a vote, and Rambus withdrew from the organization in 1995. Memos from Rambus at that time showed they were tailoring new patent applications to cover features of SDRAM being discussed, which were public knowledge (JEDEC meetings were not considered secret) and perfectly legal for patent owners who have patented underlying innovations, but were seen as evidence of bad faith by the jury in the first Infineon v. Rambus trial. The Federal Court of Appeals rejected this theory of bad faith in its decision overturning the fraud conviction Infineon achieved in the first trial (see below).

In 2000, Rambus began filing lawsuits against the largest memory manufacturers, claiming that they owned SDRAM and DDR technology. Seven manufacturers, including Samsung, quickly settled with Rambus and agreed to pay royalties on SDRAM and DDR memory. When Rambus sued Infineon Technologies, however, Micron and Hynix joined forces with Infineon to fight the lawsuit, countersuing with claims of fraud. This trio of memory manufacturers became known as "The Three Amigos". In May 2001, Rambus was found guilty of fraud for having claimed that they owned SDRAM and DDR technology, and all infringement claims against memory manufacturers were dismissed. In January 2003, the Federal Court of Appeals overturned the fraud verdict of the jury trial in Virginia under Judge Payne, issued a new claims construction, and remanded the case back to Virginia for re-trial on infringement. In October 2003, the US Supreme Court refused to a hear the case. Thus, the case returned to Virginia per the Federal Court of Appeals ruling.

In January 2005, Rambus filed four more lawsuits against memory chip makers Hynix Semiconductor, Nanya Technology, Inotera Memories and Infineon Technology claiming that DDR 2, GDDR 2 and GDDR 3 chips contain Rambus technology. In March 2005, Rambus had its claim for patent infringements against Infineon dismissed. Rambus was accused of shredding key documents prior to court hearings, the judge agreed and dismissed Rambus's case against Infineon. This sent Rambus to the settlement table with Infineon. Infineon has agreed to pay Rambus quarterly license fees of $5.9m and in return, both companies ceased all litigation against each other. The agreement runs from November 2005 to November 2007. After this date, if Rambus has enough other agreements in place, Infineon may make extra payments up to $100m. Currently, cases involving Micron and Hynix remain in court. In June 2005, Rambus also sued one of its strongest proponents, Samsung, the world's largest memory manufacturer, and terminated the license. Samsung had promoted Rambus's RDRAM and currently remains a licensee of Rambus's XDR memory.

In May 2002, the United Stated Federal Trade Commission filed charges against Rambus for antitrust violations. Specifically, the FTC Complaint asserted that through the use of patent continuations and divisionals, Rambus pursued a strategy of expanding the scope of its patent claims to encompass the emerging SDRAM standard. The FTC's antitrust allegations against Rambus went to trial in the summer of 2003 after the organisation formally accused Rambus of anti-competitive behaviour the previous June, itself the result of an investigation launched in May 2002 at the behest of the memory manufacturers. The FTC's chief administrative law judge dismissed the antitrust claims against Rambus in 2004, saying that the memory industry had no reasonable alternatives to Rambus technology and was aware of the potential scope of Rambus patent rights, according to the company. Soon after, FTC investigators filed a brief to appeal against that ruling.


-mpjesse
January 4, 2006 4:41:18 AM

Ya know I was never aware of all this crap before, but after reading that info boy are they the Biggest $#@(*&^ jerks on the planet. Thanks for the info mpjesse, it astounds me how someone can join a group that is dealing purely on good faith and the betterment of an industry and screw them over so royally. Unbelievable, makes me just shake my head in disgust.
January 4, 2006 4:47:15 AM

Here's the best part. All the ram makers have paid royalties to RD.
This is a front end cost. You ended up paying for it, plus carrying charges, plus holding fees. In the end, you have paid a good chunk of $ because of it.
They think they can continue doing it for ever.
a b à CPUs
January 4, 2006 6:49:43 AM

Ah, but RDRAM was DDR, PC1066 ran at 533MHz clock rate, 1066MHz data rate...of course RDRAM was serial

Hey, isn't the new FB-DIMM standard also serial?
January 4, 2006 3:25:53 PM

Quote:
Ah, but RDRAM was DDR, PC1066 ran at 533MHz clock rate, 1066MHz data rate...of course RDRAM was serial

Hey, isn't the new FB-DIMM standard also serial?


Yeah but the primary difference in technology between DDR and SDRAM/RDRAM is DDR's capability of processing 2 bits per cycle (hence the double date rate). RDRAM doesn't do this. Correct me if I'm wrong, but RDRAM operates on multiple 16bit channels. They can get the clock speed so high because the bus is only 16bits wide. The theoretical bandwidth relies on parallelism where DDR SDRAM doesn't. And as far as I know RDRAM only processes 1 bit per cycle.

Again, I could be wrong on that portion. So someone correct me if I am. Here's an excerpt on RDRAM from Wikipedia.org:

Quote:
The Direct RDRAM, features an architecture and a protocol that were designed to achieve high bandwidth. The Rambus channel architecture has a single-device upgrade granularity, offering engineers the ability to balance performance requirements against system capacity and component count. The narrow, high-performance channel also offers performance and capacity scalability through the use of multiple channels in parallel. In addition, the validation program created by Intel and Rambus promotes system stability by ensuring that devices and modules conform to published specifications. Although RDRAMs have a low pin count, a single device is capable of providing up to 1.6 GB/s bandwidth. Memory systems that use RIMMs (Rambus inline memory modules), also known as RDRAM modules, employ a narrow, uniform-impedance transmission line, the Rambus Channel, to connect the memory controller to a set of RIMMs. Low pin count and uniform interconnection topology allow easy routing and reduction of pin count on the memory controller. While a single channel is capable of supplying 1.6 GB/s of bandwidth, multiple channels can be used in parallel to increase this number. Systems that use, for example, the Intel 840 chipset have two parallel Rambus channels, and are able to handle up to 3.2 GB/s.


The similarities between RDRAM and DDR end at the "channel" level. Initially, DDR didn't support "channels" or "dual channels." Of course, it does now... but again that's where the similarities end. So I guess Rambus sorta has a claim to fame regarding "dual channel."

-mpjesse
Anonymous
a b à CPUs
January 4, 2006 4:10:09 PM

As crashman pointed out RDRAM is also dual data rate in the same sense that what is now commonly called "DDR" ( actually should be called DDR SDRAM) is dual data rate. RDRAM kicked some ass and it is criminal that RAMBUS got so greedy to turn of the entire PC world. Even richie rich Intel couldn't stomach the cost of the memory even for a processor architecture designed to use it which didn't do as well without it.
January 4, 2006 4:54:05 PM

Although XDR seems unlikely, it would kick major ass. 256MB of the 512 in the PS3 is XDR 3200Mhz. Retarded fast
January 4, 2006 6:34:59 PM

I think people should read the news and take an unbiased view before trashing Rambus and making them the bad-boys of the memory industry.

Wasn't it Samsung et al who recently accepted guilt in price fixing of DRAM and recieved the second largest anti-trust fine? Most of the other memory makers such as Hynix, Infineon and Micron had also been involved price fixing DRAM too. Although Micron cooperated with prosecutors so they didn't face charged.

Open you eyes and see what really happened, the memory makers have been trying to trash Rambus for years to hide the lies, and they should soon get they comeupance when all the court cases go Rambus's way. The memory makers lowered the price of DRAM so that Intel would adopt non-rambus technology so they wouldn't have to pay royalties to Rambus. And then when Intel obliged, DRAM price quadrupled over night.

Fact is I paid more for my memory because of the memory makers, not Rambus, so they are the ones who should go to hell.
Anonymous
a b à CPUs
January 5, 2006 12:21:18 AM

You know it seems to me there would just have to be a market for a chipset which bolted some of that XDR stuff onto a modern P4. I'm thinking Intels P4 glory days are largely being hidden by the peashooter tech they offer. I wonder how much $$$ some rich consumers would cough up to have that. Some smart alec would make up some PC's if there was a chipset which worked. Any one have some cash? I have an idea!
January 5, 2006 2:35:01 AM

Again, unless I'm misinformed, RDRAM is not double data rate. RDRAM does not process two bits per cycle in the same way DDR SDRAM does.

In DDR, the entire DIMM is single channel; dual channel is achieved at the memory controller level- not the RAM. High bandwidth is achieved by processing 2 bits per cycle (hence double data rate).

In RDRAM, the RIMM is split into multiple channels and high bandwidth is achieved by the northbridge/processor using all those channels in parallel. But each channel still only processes 1 bit per cycle.

This is a sticking point with me because they are fundamentally different in how they operate. Do they provide the same result in the end? Yes. But the point is Rambus thought that they owned rights to DDR. How can they own rights to something so fundamentally different?

Obviously this AMD deal is extremely significant for both sides. Rambus gets some much needed revenue and AMD gets something new and possibly dangerous. Instead of Rambus going to the hungry wolves (Samsing, Infineon, Micron, etc) and trying to royalty fees, they went to the always opportunistic AMD and got some $$$. But mark my words, if Rambus tries to force XDR/RDRAM license fees down all the memory manufacturers' throats again (i.e., AMD puts an XDR memory controller on chip) it'll be another Rambus/Intel fiasco all over again. And companies like nVidia, ATI, and VIA WILL find a way around forcing users to buy and use XDR RAM. Just like VIA did w/ RDRAM.

EDIT: XDR is more like DDR/DDR2 than RDRAM. XDR DRAM can process 8 bits per clock cycle.

-mpjesse
a b à CPUs
January 5, 2006 4:01:18 AM

I'm serious, PC800 runs 400MHz and moves ~3200MB/s across a 32-bit bus using the i850 series chipset, using two 16-bit pathways. So where does the 800 in PC800 come from? The data rate, really! 800x32/8=3200. RDRAM uses a DDR bus!
Anonymous
a b à CPUs
January 5, 2006 3:23:21 PM

I haven't worked with DDRSDRAM yet (it's coming) and predict I may never work with RDRAM, but my brains memory and apparently Crashmans memory is saying RDRAM is DDR. I am sticking with that. I suppose it wouldn't be too tough to find proof but I'm lazy.

RDRAM was used in an eight bit wide bus in the 820, 840 days then Intel moved to 16 bit busses in the 850 days. With the 840 and 850 chipsets you could configure these modules to be used in dual channel mode. The difference between RDRAM and DDRSDRAM is that it RDRAM uses a serial packet based protocol on each bit not the single clock parallel reading and writing like in DDRSDRAM.

I am agreeing that it is lame to have to pay for good stuff but that is what patents are all about. The first guy to come up with something gets to reap the financial rewards if they properly tie up all of the legal loose ends. In this case if RAMBUS has patents saying they invented DDR then the people using this tech darn well better be paying them to use it.

There is enough money and legal advice available to the JEDEC committee members that they ought to have known. I mean duh, when they are all sitting around saying to themselves "hey yeah DDR that rocks, I'm sure glad RAMBUS told us about it" the lights should be coming on that it is a new invention and someone will likely want money for it.

Please don't misunderstand me. I'm not saying it didn't suck paying the premium for RDRAM. What I am saying is that: while it sucked it blew away the alternative, ie the VIA DDRSDRAM approach, which even in dual channel mode all these years later is still trying to hold a candle to RDRAM's performance while coupled to the P4. The inventors should be paid for their work, at least by my interpretation of the events and laws.
Anonymous
a b à CPUs
January 5, 2006 3:26:23 PM

Quote:
folks like these need to go broke and sell out to better qualified minds


Huh? They came up with a technology that allowed PC's to step into the modern ages and they should go broke because you want it and think you should get it for free.

You kids these days, you worry me.
January 5, 2006 3:58:59 PM

Quote:
I haven't worked with DDRSDRAM yet (it's coming) and predict I may never work with RDRAM, but my brains memory and apparently Crashmans memory is saying RDRAM is DDR. I am sticking with that. I suppose it wouldn't be too tough to find proof but I'm lazy.


I'll do it for you. But you're not going to like the "proof."

This is the definition of DDR as stated by Wikipedia.org:
http://en.wikipedia.org/wiki/Double_data_rate

Quote:
In computing, a computer bus operating with double data rate transfers data on both the rising and falling edges of the clock signal, effectively nearly doubling the data transmission rate without having to deal with the additional problems of timing skew that increasing the number of data lines would introduce. This is also known as double pumped, dual-pumped, and double transition.


Read that very carefully. RDRAM does not transfer data on both the rising and falling edges of the clock cycle. This is the fundemental difference I've stated like 28 times in this thread. Notice it says "without having to deal with additional problems of timing skew that increasing the number of data lines would introduce."

That's exactly what RDRAM does. RDRAM adds the number of data lines or "channels" to increase bandwidth.

This is the definition of RDRAM as stated by wikipedia.org:
http://en.wikipedia.org/wiki/RDRAM

Quote:
The Direct RDRAM, features an architecture and a protocol that were designed to achieve high bandwidth. The Rambus channel architecture has a single-device upgrade granularity, offering engineers the ability to balance performance requirements against system capacity and component count. The narrow, high-performance channel also offers performance and capacity scalability through the use of multiple channels in parallel. Although RDRAMs have a low pin count, a single device is capable of providing up to 1.6 GB/s bandwidth. Memory systems that use RIMMs (Rambus inline memory modules), also known as RDRAM modules, employ a narrow, uniform-impedance transmission line, the Rambus Channel, to connect the memory controller to a set of RIMMs. Low pin count and uniform interconnection topology allow easy routing and reduction of pin count on the memory controller. While a single channel is capable of supplying 1.6 GB/s of bandwidth, multiple channels can be used in parallel to increase this number.


Do you see anything in that definition that says anything about RDRAM transmitting a bit on both the rise and fall of the clock cycle?

No.

RDRAM is NOT DDR.

Again, do they provide the same result in the end? Yes. But so does a Pentium 4 and an Athlon 64. Yet they are fundamentally different!

Now.. if you want to get into XDR then yes, XDR DRAM is DDR by nature. It transmits 4 bits per clock cycle on the rise and fall. Technically XDR is Quad Data Rate, but that's a whole 'nother issue.

-mpjesse
Anonymous
a b à CPUs
January 5, 2006 4:43:09 PM

Whether you are right or wrong I would hardly call what you have posted some proof. Again keep in mind I'm just saying I don't really remember what it is and don't have any emotional or other prejudice either way and could really care less. It will be the judges responsibility to sort out all of these details. You posted a definition of what DDR is, thanks but not necessary. Then you posted a description of RDRAM which tells nothing of it's bandwidth and where the ratings come from. Besides....wikipedia.org?....I don't necessarily think that some random blog should be considered gods law or something.

PS now that your hassling me on it I went and did a google search. The first link on the list http://www.overclockers.com/articles146/ by a respected reviewer contradicts what you are saying. I stopped there, that was enough for me.
January 5, 2006 5:32:48 PM

Can we just agree to disagree? LOL.

I'm not saying that RDRAM doesn't process 2 bits per cycle- I'm saying that it does it in a very different way from DDR. DDR uses the 2 bits per clock cycle on one big channel while RDRAM uses smaller parallel channels to achieve 2 bits per clock cycle.

The article you posted is somewhat ambiguous on the topic (see below for which excerpt I'm refering to).

Quote:
RDRAM is a narrow, high speed serial connection.
SDRAM/DDR is a wider, lower speed parallel connection. (DDR is simply SDRAM that squeezes out two actions per clock cycle; something RDRAM already does. I'll use the term "SDRAM" when what I'm saying applies to both current SDRAM and DDR, and "DDR" when it applies only to that).


He states that RDRAM does what DDR does... but he doesn't state how; which is the issue of our current topic. :-)

As for Wikipedia.org, I agree it's far from a definitive source. However, from my experience it's always been an accurate one. And their explanation of both technologies are exactly how I remember it.

Here's a couple simple articles that describe how RDRAM works:
http://www.karbosguide.com/hardware/module2e4.htm
http://www.ngcsu.edu/Adminsrv/Infotech/StuHlpDk/hguide.... (scroll down to "Memory" on this one)

-mpjesse
a b à CPUs
January 5, 2006 7:27:07 PM

Here you go It's hard to find this OLD information any longer, but if you wish you can download the reference info for certain RDRAM chips, which shows the same thing in picture form.

Now that you see what's going on, please do update Wikipedia, as I've had problems with a few of their definitions before.

As I remember things, RDRAM went to the JEDEC meetings, took ideas from the open source, incorporated them into their own works, patented the works, then sued members of the project for violating the patented open source work! And that would be why they lost so many lawsuits!
a b à CPUs
January 5, 2006 7:37:40 PM

Hey I got another link for you:

http://www.intel.com/technology/magazine/computing/dt01021.pdf

Quote:
To understand why RDRAM is still the right choice for the high-performance market segment, let’s examine the
potential bandwidth an Intel 850 chipset-based system can provide. A single RDRAM memory channel operates at 400
MHz with a bus width of 2 bytes and, very much like DDR technology, RDRAM can transfer data on the rising and
falling edges of a clock cycle. Therefore, memory bandwidth for a single channel of RDRAM is 400 MHz x 2 data
read/cycle x 2 bytes bus width, or 1,600 MB/s.
January 5, 2006 8:17:14 PM

I hope AMD will never use Rambus memory.

Now we got a new way of finding out AMD fanboys: if someone in the Intel/Rambus days thought Rambus was evil, but after AMD/Rambus deal starts to think "Rambus isn't that bad/evil" then he is surely a fanboy.
January 5, 2006 8:26:45 PM

I posted that article earlier. I read through it and I don't see any new information that contradicts what wikipedia.org says about RDRAM.

This new article obviously goes into further depth about RDRAM, but I don't see anything contradicted.

Did I miss something? Seriously... I don't mean to be rude (As you know I have the utmost repsect for you!), but is there something I don't see?

EDIT: ignore this post and read my next.

-mpjesse
January 5, 2006 8:28:43 PM

LOL. Nevermind my previous post.

I'll be damned... you even got Wikipedia to change their definition. Damn fine job.

I stand corrected. :-)

-mpjesse
Anonymous
a b à CPUs
January 5, 2006 8:31:31 PM

OK go write "I love RAMBUS" 100 times and all will be well. :wink:
January 5, 2006 8:36:33 PM

I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus
I love Rambus

:-)

-mpjesse
Anonymous
a b à CPUs
January 5, 2006 8:56:41 PM

lol
January 5, 2006 9:15:29 PM

Here it says XDR is octo data rate.

anandtech

Could be a good fit.

Although I Loathe Rambus for the submarine patent thingy, they do kind of have the patent on this stuff. So I hate the suits at Rambus but like the techs. Story of my life.
January 5, 2006 9:28:09 PM

Yeah it is octo. My bad again.

-mpjesse
January 6, 2006 2:54:57 AM

Amd and rambus, yup you can call me a fan boy, ------ right after they pay me back the thousands thier royalties have cost me.
If Amd only sells chips with rambus solutions, I will never buy another Amd chip.
a b à CPUs
January 6, 2006 3:25:17 AM

Hmm, not even if Intel cancels the drop of Netburst and your only other choice becomes a dual-core 4GHz 200W P4?
January 6, 2006 3:40:56 AM

My problem isn't so much with netburst, as with the leakage current associated with the prescotts. A scotty, with FDSOI might even be able to do dual 4ghz. Of course there are a few other changes that would also help, like a true HTT, and an onboard memory controller or two, with a true 400 mhz memory bus to each.
January 6, 2006 9:48:51 AM

Is it just me or is every one just jumping on rambus like they do intel. Theres no real reason to be mad at rambus for sueing people over property rights thats why we have courts. If everything was free market there whould be little incentive for companys to innovate or push to make the most cutting edege parts for us consumers. I for one whould love to see amd and rdram you could overclock like crazy and rdram is alredy fast. As far as prices go theres not much diffrence from rdram and high proformance ddr2.
January 6, 2006 10:10:14 AM

Ultimately it was the consumers who suffered due to evil business policy of Rambus. So they have very valid reason to panic if they can see possible comeback of Rambus in the PC market.
a b à CPUs
January 6, 2006 10:10:24 AM

If you missed the point, you're brainless. The point was, RAMBUS went to JEDEC discussions on DDR technology, took those developements and patented them. Now before you go off saying "why not, they own them", no they didn't. JEDEC is an open forum for memory manufacturers.

They let a few years pass. Then when they figured nobody would have the records to prove they didn't actually own DDR technology, they sued RAM companies for using it. That is, they sued the CO-DEVELOPERS from an OPEN DEVELOPEMENT FORUM for using technology owned by the ENTIRE forum.

That would be like Sun patenting Linux. Really. And if that's OK, well, maybe I can get a patent on water.
January 6, 2006 7:08:14 PM

I know exactally what the point was but it seems you missed mine business is cruel companys do this kind of stuff every day and thats why we have courts . Almost every action taken agenst rambus has failed or been overturned even the antitrust case was dismissed. If i had controle of rambus id probley have done the same thing.

Quote:
In the early 1990s, Rambus was invited to join the Joint Electron Device Engineering Council. Rambus had been trying to interest memory manufacturers in licensing their proprietary memory interface, and all companies had signed non-disclosure agreements to view Rambus' technical data. During the later Infineon v. Rambus trial, Infineon memos from a meeting with representatives of other manufacturers surfaced, including the line "[O]ne day all computers will be built this way, but hopefully without the royalties going to Rambus," and continuing with a strategy discussion including options for buying and burying Rambus' technology.

Seriously you want me to belive rambus is the bad guy in this?
a b à CPUs
January 6, 2006 9:25:31 PM

Not all of it, just the parts where they took technology from the community and integrated it into their own technologies, pattented it, then sued companies for using their own ideas.

Tom was part of all this, you can read some of the stuff at the guide.

Memory companies are generally just as bad, looking at Samsung's recently created DDR2 shortage as an example. But here you have an exceptionally deceptive company taking on a whole industry of lesser evils, so to speak.
January 7, 2006 4:17:40 AM

sorry for posting in the middle of the list but might have you considered rambus as all it was toughted for if NOT for the extravagant price? Amd and the rest of the enthusisists at that time hoped but did not have anything in their cubbord till later and hoped that the new ddr? ddr2?
up from pc133 standard was going to help out the situation with the new hypertransport bus improvements.

Sorry If im a little crossed up Ive been out of the game for a tad with illness but getting back in the fray.

Good post tho.
Sunset1
a b à CPUs
January 7, 2006 5:47:06 AM

I actually have a couple answers here: The hoopla against RAMBUS started with Intel replacing the highly regarded BX chipset with the i820 and i840 RDRAM chipsets for the PIII. Performance was horrible, costs was horrible, and VIA started taking away Intel's chipset sales.

Intel intentionally handicapped the i820 to make it slower than the i840, so that the i840 appeared to be a superior solution. But it was also significantly higher priced.

People refused to put RDRAM on the PIII because it was such a horrible deal, so Intel came out with a Memory Translator Hub, that allowed SDRAM to work with the i820 chipset. But there was some kind of noise issue on the bus that caused data loss, and Intel ended up recalling every last one of the i820/SDRAM boards. In fact, they'd give you a choice, $150 or a free i820/RDRAM board with a free 128MB stick of RDRAM.

Meanwhile Intel had their bargain basement i810, which they improved to support 133MHz bus and UDMA100 in the i810E. And with all the sales problems and the recall, they eventually added a couple features to the i810 (most importantly support for an AGP slot) and called the new chipset the i815, which was replaced almost imediately with the i815E.

Now the i815E wasn't as fast as the BX, and only supported a maximum of 512MB and a maximum of 4 banks. Some boards offered 3 DIMMs for use with single-sided RAM, most only had 2 DIMMs so that you couldn't physically instal more than 4 banks.

Hatred for RAMBUS grew throughout Intel's blunders, and the price stayed extremely high. And both companies got each other's enemies because Intel made a signficant corporate investment in RAMBUS.

All became clear when Intel released the P4 and i850 chipset. RAMBUS was never intended for the PIII, but rather Intel tried to force it onto the PIII market in order to increase availability and bring PC800 to maturity. RAMBUS chips that failed at 800 could always be retested at a lower speed, and if they passed, sold at that lower speed.

So part of Intel's plan worked, at P4 launch RDRAM at PC800 was mature and readily available. But then the P4 flopped completely with enthusiast, because the first chip, the P4 1.4, was slower than the PIII 1000EB in many applications. Many people remembered the problems of RDRAM from their PIII experience, and Intel further tarnished RAMBUS' reputation.

You think the general public is clueless about computers? Back then it was at least twice as bad, probably more like 4x, that is to say the general public had 1/4 the PC knowledge the do today. So P4's actually sold fairly well, and when the P4 had established its name with the clueless, Intel released an even slower P4 1.3. The collaboration between Intel and RAMBUS had now brought enthusiasts to FULL FURRY.

AMD went to a DDR CPU bus with their Athlon back in the PIII days. Enthusiasts got their matching DDR SDRAM a couple years following the Athlon launch. Intel was fully Anti-AMD technology, so when DDR guys asked for a P4 chipset with DDR SDRAM, Intel basically told them to go to hell. Intel later released a chipset for PC133 on the P4 (Horrible), updated the P4 core to a more efficient version, and under market pressure finally released a DDR SDRAM chipset for the P4.

Now that's less than half the storry, there were a lot of other companies involved, including VIA with their stolen P4 bus, and to explain the whole thing would take several hours! But the end result was Intel cutting their ties with RAMBUS, RAMBUS stock dropping like a rock, and RAMBUS seeking other ways to make money.

RAMBUS had an insurance plan: Being a technology development company, they patented a huge number of ideas. Durring their time with JEDEC, as memory companies were hammering out the details of DDR SDRAM, they took a bunch of their notes from the meetings and included those technologies in some of their patents. They were able to sneek this past nearly everyone! The plan was, if the company ever needed quick cash, they could suddenly announce their patents had been violated and demand royalties.

Several RAM manufacturers sold out. They put up the cash, rather than fight. I believe it was Micron and Infineon who joined together to fight RAMBUS assertions, and of course won.

But durring that epic battle of the evil punk and the overbearing giants, DDR SDRAM prices went nuts. And enthusiasts got really pissed off. It was easy to blame RAMBUS because of the lawsuit of course, but meanwhile some of these "innocent" RAM companies were being indicted for PRICE FIXING.

Well of course these companies could claim that the price fixing was needed to make up for lost income over the suite, but really they just kept their mouths shut and paid the fines. RAMBUS was now the enthusiasts worst enemy.
January 7, 2006 6:56:57 AM

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

Quote:
It's important to note that RDRAM uses Double Data Rate in a similar manner that competing DDR SDRAM has. Accordingly, PC800 has an 800MHz data rate using a 400MHz clock rate, PC1066 has a 1066MHz data rate using a 533MHz clock rate, etc. This is best described in Intel’s Developer UPDATE Magazine June 2002 article “Mainstream PC Memory Architectures for the Intel® Pentium® 4 Processor”

This use of Double Data Rate technology was included in RAMBUS’ patent portfolio, after being discussed openly in JEDEC meetings. RAMBUS used the patent to file several lawsuits against DDR SDRAM manufacturers. While a few companies settled, others challenged the suite, and ownership of said intellectual property was eventually denied.


This is the bottom line DDR was patented by RAMBUS and all the other company tried to get buy the royaliest and change the way SDRAM by transmitting a bit on both the rise and falled, that is why uit got ugly
a b à CPUs
January 7, 2006 7:29:12 AM

Yes, I wrote that. I didn't want to make it long. RAMBUS was party to those meetings and anything discussed there was property of the entire group. They later patented the shared technology. Since they didn't have any sole claim to shared technology, they lost their court case.

In other words, the courts ruled the patents were fraudulent.
January 7, 2006 7:49:58 AM

WoW I didnt know that...
a b à CPUs
January 7, 2006 9:56:57 AM

Yes, in fact, RAMBUS's suite turned against them so badly that the courts awarded the defendants (memory companies) a substantial settlement to reimburse their defense fees and so forth. That judgement was later deemed excessive in an appeal, and it went back and forth so many times in the appeals process I really don't know whether they even paid anything.
January 7, 2006 10:30:45 AM

While most you said is true crashman what is missleading is "I believe it was Micron and Infineon who joined together to fight RAMBUS assertions, and of course won." what you failed to mention is the verdict was overturned by a fedral appeals court judge. I have a socket 478 board with rambus with a 2GHz p4 with 850 chipset and it outperformes a almost identical 2.4GHz amd comp I have. So i can olny say iv had good experinces with rambus and cant see why everyone bashing them and not much has been said about samsung one of the bigest chip makers for memory in the world that supplyes memory to Ocz Corsair Mushkin just to name a few for there price inflation on memory. Cutting off supply to raise prices isnt so "innocent" neither is the antitrust case samnsung lost. Like I said before this is business and business can get ugly I guess it is easyer to blame rambus even when outher companys had played a hand in what happend.
a b à CPUs
January 7, 2006 11:56:30 AM

No, I mentioned the awards were ruled excessive. That's the part that was overturned. I lost track on the THIRD go-around.

Or are you speaking of the earlier case that was overturned twice?
      • 1 / 2
      • 2
      • Newest
!