- Status
Not really im not much of a OCer iv used the motherboard oc tool gigabyte provides and gotten it to a stable 2.67 with stock cooling but it gets to hot on 4+ hr gameing so i usally keep it at around 2.4
I was refering to Infineon Micron and Hynix vs rambus go round.
semicondoctor
Distinguished
- Jan 7, 2006
- 4
- 0
- 18,510
You guys have it wrong.
Rambus is about to be completely exonerated by the courts - turns out they got f#cked big time by the memory makers the last 5 years.
Just think about Cell for a moment. Here you have IBM, Sony & Toshiba investing half a billion US$ and years of development in a new generation of CPU. Not that they lack engineers (IBM alone has several thousand), but they nonetheless chose Rambus FlexIO interface & XDR memory technology. Nobody forced them.
Alot of trash has been written about Rambus the last few years, but check out their new inventions and solutions. Almost every week they're granted new patents. Those guys are gonna be the dominant force in memory & high speed interconnects for some time to come.
Rambus is about to be completely exonerated by the courts - turns out they got f#cked big time by the memory makers the last 5 years.
Just think about Cell for a moment. Here you have IBM, Sony & Toshiba investing half a billion US$ and years of development in a new generation of CPU. Not that they lack engineers (IBM alone has several thousand), but they nonetheless chose Rambus FlexIO interface & XDR memory technology. Nobody forced them.
Alot of trash has been written about Rambus the last few years, but check out their new inventions and solutions. Almost every week they're granted new patents. Those guys are gonna be the dominant force in memory & high speed interconnects for some time to come.
worstnightmare
Distinguished
- Jan 7, 2006
- 5
- 0
- 18,510
I want to fill you guys and girls in on some misinformation I read on this thread...
First
I'm sure if someone stole your IP, you would like to be compensated for it as well. RMBS is not different.
RMBS was invited to JEDEC in order to clarify the non disclosure agreements the memory manufactures signed with them. RMBS was ambushed, violated and had their IP cherry picked.
RMBS has NOT LOST ONE LEGAL CASE but won everyone hands down! Recently they were found to not have committed any wrongdoing in California. This may be what prompted AMD to License RMBS. As yo know AMD was on the list to pay for using RMBS's IP, they even were fortunate to be forgiven for past infringement. This is a win win for both companies, look at the stock prices from the past week, all because of the deal with RMBS! This is very positive for AMD and is just the beginning for RMBS. Keep in mind EVERY digital device uses RMBS IP, EVERYONE!!!!!!!!!!
SDRAM, DDR, DDR1, DDR2, DDR3, GDDR, XDR and we are not talking about controller or graphic interfaces.
You may want to consider getting in on the fun, RMBS will appreciate quickly when they conclude with the legal wrangling, which should happen rather quickly. They also have the thieves in check, a few more moves will have Hynix, MU and Sammy in mate. They are looking at 3 times damages, they all have been fined and all had management folks go to jail. Believe me none of the 3 want their dirty laundry aired, can not afford it, too many stockholders, customers etc and way too much to loose. They all admitted to sticking it to RMBS in an effort to put them out of business.
I know you guys n girls do not want to hear the truth, too baddddddd. Do some DD and you may be enlightened, as after reading some post on this thread some have no clue. Perhaps some need to get one... Go to this site and feast your eyes and feed your minds. http://rambus.org/
First
I'm sure if someone stole your IP, you would like to be compensated for it as well. RMBS is not different.
RMBS was invited to JEDEC in order to clarify the non disclosure agreements the memory manufactures signed with them. RMBS was ambushed, violated and had their IP cherry picked.
RMBS has NOT LOST ONE LEGAL CASE but won everyone hands down! Recently they were found to not have committed any wrongdoing in California. This may be what prompted AMD to License RMBS. As yo know AMD was on the list to pay for using RMBS's IP, they even were fortunate to be forgiven for past infringement. This is a win win for both companies, look at the stock prices from the past week, all because of the deal with RMBS! This is very positive for AMD and is just the beginning for RMBS. Keep in mind EVERY digital device uses RMBS IP, EVERYONE!!!!!!!!!!
SDRAM, DDR, DDR1, DDR2, DDR3, GDDR, XDR and we are not talking about controller or graphic interfaces.
You may want to consider getting in on the fun, RMBS will appreciate quickly when they conclude with the legal wrangling, which should happen rather quickly. They also have the thieves in check, a few more moves will have Hynix, MU and Sammy in mate. They are looking at 3 times damages, they all have been fined and all had management folks go to jail. Believe me none of the 3 want their dirty laundry aired, can not afford it, too many stockholders, customers etc and way too much to loose. They all admitted to sticking it to RMBS in an effort to put them out of business.
I know you guys n girls do not want to hear the truth, too baddddddd. Do some DD and you may be enlightened, as after reading some post on this thread some have no clue. Perhaps some need to get one... Go to this site and feast your eyes and feed your minds. http://rambus.org/
thebutcher
Distinguished
- Jan 7, 2006
- 2
- 0
- 18,510
You are looking at it from the big picture. Rambus didn't patent RDRAM - they patented over 20 inventions, many of which were part of RDRAM. (They initially did file it as one patent, but the PTO told them there were numerous inventions in it and they needed to refile for the others.) Some of these inventions were also part of DDR.
If someone patented 20 inventions for a fuel powered car (wheels, transmission, engine, brakes, etc) and then someone comes around with an electric car that uses many of those patents...the fact that they operate 'fundamentally different' is irrelevant to whether many of the 20 patents are included.
from the United States Court of Appeals for the Federal Circuit (Rambus v. Infineon):
and for the record, the written specification for those 4 technologies were identical to the application filed by Rambus in April of 1990. So JEDEC was discussing things 4 years after Rambus filed the original patent on them. and just for kicks..."JEDEC incorporated four technologies into its SDRAM standard that are relevant to this case: programmable CAS latency, programmable burst length, externally supplied reference voltage, and two-bank designs. "
JEDEC IS NOT OPEN SOURCE. They have open meetings, but that does not mean they are developing open standards. They are a standard setting organization, and it would be a HUGE anti-trust violation for them to say 'we only use non-royalty, open-source technologies'.
and finally, for those of you complaining about the pricing of RDRAM, you should really be pissed at Micron, Infineon, Hynix, and Samsung (who have now pleaded guilty to one of the biggest price-fixing cases in history). They lied to Intel about how much RDRAM they would be producing, and kept RDRAM supply artificially low - which led to much higher prices than it should have been.
An email by Jeff Mailloux, a senior Micron executive, to Farhad Tabrizi, his counterpart at Hyundai (now Hynix):
why would a micron exec worry about going to jail for talking to a reporter?
shamrokkevin
Distinguished
- Jan 7, 2006
- 1
- 0
- 18,510
Does anyone here have a clue ? I don't even know where to start. First the guy who started this thread didn't even know that rdram transfers data on the rising and
falling edges of a clock cycle.Well Rambus was the first company to use that idea. Jedec only copied the idea and incorporated it into ddr. He also didn't even know that xdr is octo data rate. And He thinks He knows what he's talking about.
Then He states "Infineon paid up because they were being sued. In the end it was cheaper for infineon to pay royalties than to fight in court."Yeah right pal.
Since He also thinks AMD bought a license just to avoid being sued themselves lets see what AMD has to say about it.
Here are couple of quotes taken from the FTC trial. : http://www.ftc.gov/os/adjpro/d9302/040223initialdecision.pdf Maybe you should read the entire ID you might learn something.
1500. Steve Polzin of AMD testifed that he had discussions with DRAM manufacturers in 2000 about alternatives for programmable CAS latency, programmable burst length, and dualedge clocking. (Polzin, Tr. 3988, 3996, 4044). At the time, the DDR2 standard was still winding its way through JEDEC. (polzin, Tr. 4044-45). Polzin understood at the time of these discussions that Rambus patents cover these technologies. (Polzin, Tr. 4047-48). The DDR2 standard, however, still specifies programmable CAS latency, programmable burst length, and dual-edge clocking. (polzin, Tr. 4046-48).
1509. JEDEC ultimately opted to use Rambus s programmable CAS latency technology
in DDR2. (Steve polzin, AMD Tr. 4046; RX 1854 at 12- 14).
1513. JEDEC adopted Rambus s programmable burst length technology in DDR2 despite
complete awareness of Rambus’s issued patents and demands for royalties. (Steve Polzin, AMD Tr. 4046-47).
1514. JEDEC was looking at alternative clocking schemes to avoid Rambus patents.
(Kashinsky, Tr. 2828). JEDEC failed to find an acceptable alternative and adopted Rambus dual-edge clocking technology. (Steve Polzin AMD, Tr. 4047).
1518. Despite the consensus to use SDR (single data rate) in place of dual-edge clocking "provided we can make it work " JEDEC incorporated dual-edge clocking into DDR2. (Steve Polzin AMD, Tr. 4047).
Here are some more "words of wisdom" from mpjesse ."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."? "
Again he has no clue.What crap are you talking about. Do you mean that Rambus actually had the gall to expect to be paid for their IP. Rambus never claimed to invent DDR. But they did invent many of the technologies that jedec incorporated into DDR
In fact it seems to me that the opposite of what you're saying is true. Read this: http://www.highbeam.com/library/docfree.asp?DOCID=1G1:127350805&ctrlInfo=Round18%3AMode18c%3ADocG%3AResult&ao=
Now do you still believe what you wrote?
Now as far as your belief that Rambus was the "bad guy" in this jedec mess and that they joined jedec in odtain "submarime patents on basicaly all memory developed since sdram. Maybe you should read a memo written by a really nice guy named Willi meyers of Infineon back in 92: http://rambus.org/legal/menace.pdf
It sure seems to me that the mm's were the ones playing the games here not Rambus.
falling edges of a clock cycle.Well Rambus was the first company to use that idea. Jedec only copied the idea and incorporated it into ddr. He also didn't even know that xdr is octo data rate. And He thinks He knows what he's talking about.
Then He states "Infineon paid up because they were being sued. In the end it was cheaper for infineon to pay royalties than to fight in court."Yeah right pal.
Since He also thinks AMD bought a license just to avoid being sued themselves lets see what AMD has to say about it.
Here are couple of quotes taken from the FTC trial. : http://www.ftc.gov/os/adjpro/d9302/040223initialdecision.pdf Maybe you should read the entire ID you might learn something.
1500. Steve Polzin of AMD testifed that he had discussions with DRAM manufacturers in 2000 about alternatives for programmable CAS latency, programmable burst length, and dualedge clocking. (Polzin, Tr. 3988, 3996, 4044). At the time, the DDR2 standard was still winding its way through JEDEC. (polzin, Tr. 4044-45). Polzin understood at the time of these discussions that Rambus patents cover these technologies. (Polzin, Tr. 4047-48). The DDR2 standard, however, still specifies programmable CAS latency, programmable burst length, and dual-edge clocking. (polzin, Tr. 4046-48).
1509. JEDEC ultimately opted to use Rambus s programmable CAS latency technology
in DDR2. (Steve polzin, AMD Tr. 4046; RX 1854 at 12- 14).
1513. JEDEC adopted Rambus s programmable burst length technology in DDR2 despite
complete awareness of Rambus’s issued patents and demands for royalties. (Steve Polzin, AMD Tr. 4046-47).
1514. JEDEC was looking at alternative clocking schemes to avoid Rambus patents.
(Kashinsky, Tr. 2828). JEDEC failed to find an acceptable alternative and adopted Rambus dual-edge clocking technology. (Steve Polzin AMD, Tr. 4047).
1518. Despite the consensus to use SDR (single data rate) in place of dual-edge clocking "provided we can make it work " JEDEC incorporated dual-edge clocking into DDR2. (Steve Polzin AMD, Tr. 4047).
Here are some more "words of wisdom" from mpjesse ."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."? "
Again he has no clue.What crap are you talking about. Do you mean that Rambus actually had the gall to expect to be paid for their IP. Rambus never claimed to invent DDR. But they did invent many of the technologies that jedec incorporated into DDR
In fact it seems to me that the opposite of what you're saying is true. Read this: http://www.highbeam.com/library/docfree.asp?DOCID=1G1:127350805&ctrlInfo=Round18%3AMode18c%3ADocG%3AResult&ao=
Now do you still believe what you wrote?
Now as far as your belief that Rambus was the "bad guy" in this jedec mess and that they joined jedec in odtain "submarime patents on basicaly all memory developed since sdram. Maybe you should read a memo written by a really nice guy named Willi meyers of Infineon back in 92: http://rambus.org/legal/menace.pdf
It sure seems to me that the mm's were the ones playing the games here not Rambus.
worstnightmare
Distinguished
- Jan 7, 2006
- 5
- 0
- 18,510
I just gotta laugh, master historian, large ego, hero, ha!
You have apsolutely no clue about what you talk about and everyone should know it. You post BS my boy!!! You have NO clue, ie: your clueless to events as well as to the truth concerning RMBS and MU, IFX, HYNIX and SAMMY, the 4 together are thieves, been tried, found guilty, jailed and NOW THEY ARE GOING TO PAY... Why because they have diarria of the mouth like you. The truth is out there in writting, court decisions etc. Go educate yourself before you post anymore of your fantisies. lol... Everyone is going to pay RMBS!!!! Yes Tom wrote alot of crap about RMBS but least you forget who was paying him to do so...lol Doesn't happen any more though, ever wonder why? Had the shit scared out of him but that is not written, some of us are in the know or at least know someone who is... lmao at people like you, your MR. CLUELESS, not a historian, you may have an ego, so does everyone, hero, how can you be when you do not even know the truth when it slaps you.....HA
You have apsolutely no clue about what you talk about and everyone should know it. You post BS my boy!!! You have NO clue, ie: your clueless to events as well as to the truth concerning RMBS and MU, IFX, HYNIX and SAMMY, the 4 together are thieves, been tried, found guilty, jailed and NOW THEY ARE GOING TO PAY... Why because they have diarria of the mouth like you. The truth is out there in writting, court decisions etc. Go educate yourself before you post anymore of your fantisies. lol... Everyone is going to pay RMBS!!!! Yes Tom wrote alot of crap about RMBS but least you forget who was paying him to do so...lol Doesn't happen any more though, ever wonder why? Had the shit scared out of him but that is not written, some of us are in the know or at least know someone who is... lmao at people like you, your MR. CLUELESS, not a historian, you may have an ego, so does everyone, hero, how can you be when you do not even know the truth when it slaps you.....HA
thebutcher
Distinguished
- Jan 7, 2006
- 2
- 0
- 18,510
All patents at issue have the same written specification - which is a description of the invention(s). This has not changed since 1990. What Rambus changed and added was the claims - which is how the invention will be used. It is completely legal under patent law to add claims to an existing invention (no one expects an inventor to predict every single way someone will use their invention at the time they file their patent.)
On its surface, you are right...I don't agree that a company should amend its patent claims to cover a standard that is being worked on. HOWEVER, JEDEC knew about Rambus's patents. Rambus taught many JEDEC companies about their inventions under Non-Disclosure Agreements. The companies studied the patents. Evidence shows that engineers at several companies expressed concerns that Rambus could/would extend their claims.
In the face of that and Willie Meyer's memo in 1992...
"One day all computers will have to be built like this, but hopefully without royalties going to Rambus"
"MAKE IS PUBLIC DOMAIN!"
it is clear what JEDEC companies were trying to do by INVITING Rambus into JEDEC.
Enhance your calm, everyone:
AMD is licensing Rambus technology - thats a fact. We've jumped to the conclusion that AMD wants to introduce RDRAM to the Athlon because that is our consumer experience (and it wasn't a good one.)
A few other simple facts can lead to a different outlook:
1. The people at Rambus aren't st00pid. No matter how much you hate them there are some smart people working at Rambus, and Rambus has some cool technology.
2. The computer industry is not interested in repeating the RDRAM fiasco, and you can bet that Rambus isn't interested in repeating the RDRAM fiasco.
Give AMD some credit - all of their development has been pretty savvy. People made fun of AMD when they licensed/purchased NexGen (anyone here remember the NX586?) and pre-Athlon the industry magazines were writing more about Cyrix. A lot of industry writers doubted the value of 64-bit computing (at least, as opposed to raising the Athlons clock-cycle), and even more people (myself included) questioned the wisdom of an integrated memory controller. In spite all of this, AMD has made steady progress growing their CPUs, and I doubt they plan on sticking their tongue in the same RDRAM light socket as Intel.
3. AMD has big uses for memory technology that has nothing to do memory DIMMs. Remember, the Athlon X2 has an integrated bus for the two processors' cache; by contrast, for Intel to share information from one CPU cache to the next, it has to send the information off the die, through the memory controller, and back to the die. Imagine how much worse this gets on server systems with 2 dual-core CPUs.
The L1 and L2 cache on an Athlon X2 make up the most significant chunk of the processor's silicon, and I can see a LOT of uses for Rambus technology inside the AMD CPU - ESPECIALLY if AMD develops a quad-chip solution. Remember, Rambus technology also includes the ability to 'power-down' portions of memory that are not in use, to intelligently transfer information between physical buffers, to 'glob' information for better fetching, increase the transfer of data on a smaller bus, and increase or decrease memory bus speed as needed to facilitate information transfer.
Forget the DIMMs, in these days when Anandtech and Tom's Hardware are all benchmarking processors against the wattage they use, I can see AMD, with its highly-integrated memory system, looking for ways to make performance gains along the current strengths of their chip. Intel's advantage is not in CPU design - it's in manufacturing, and you can see this in their current move to 65nm and taping out of 45nm before AMD has even finished the migration to 90nm. Intel might be slowly losing market share to AMD, but they're making it up in manufacturing costs. In contrast, for AMD to compete they have to work on their own strengths, and their strength is the capability of their processor and it's integrated memory system.
Currently, dual-core processors are interesting, but they aren't vital. Windows XP does a terrible job of using a dual-core processor, and this isn't likely to change until Vista. If I were to build a computer today I would base it on the Athlon 4000+, not the X2 4800+.
However, it will change. And when the OS, the games, the media, and the applications are all finally optimized for multi-threads, I will bet that you're going to see just how important cross-CPU information is going to be. If AMD can manufacture a quad-core Opteron that is screamingly efficient at handling on-die memory transactions, they will kick much buttage.
The people at AMD aren't dummies.
AMD is licensing Rambus technology - thats a fact. We've jumped to the conclusion that AMD wants to introduce RDRAM to the Athlon because that is our consumer experience (and it wasn't a good one.)
A few other simple facts can lead to a different outlook:
1. The people at Rambus aren't st00pid. No matter how much you hate them there are some smart people working at Rambus, and Rambus has some cool technology.
2. The computer industry is not interested in repeating the RDRAM fiasco, and you can bet that Rambus isn't interested in repeating the RDRAM fiasco.
Give AMD some credit - all of their development has been pretty savvy. People made fun of AMD when they licensed/purchased NexGen (anyone here remember the NX586?) and pre-Athlon the industry magazines were writing more about Cyrix. A lot of industry writers doubted the value of 64-bit computing (at least, as opposed to raising the Athlons clock-cycle), and even more people (myself included) questioned the wisdom of an integrated memory controller. In spite all of this, AMD has made steady progress growing their CPUs, and I doubt they plan on sticking their tongue in the same RDRAM light socket as Intel.
3. AMD has big uses for memory technology that has nothing to do memory DIMMs. Remember, the Athlon X2 has an integrated bus for the two processors' cache; by contrast, for Intel to share information from one CPU cache to the next, it has to send the information off the die, through the memory controller, and back to the die. Imagine how much worse this gets on server systems with 2 dual-core CPUs.
The L1 and L2 cache on an Athlon X2 make up the most significant chunk of the processor's silicon, and I can see a LOT of uses for Rambus technology inside the AMD CPU - ESPECIALLY if AMD develops a quad-chip solution. Remember, Rambus technology also includes the ability to 'power-down' portions of memory that are not in use, to intelligently transfer information between physical buffers, to 'glob' information for better fetching, increase the transfer of data on a smaller bus, and increase or decrease memory bus speed as needed to facilitate information transfer.
Forget the DIMMs, in these days when Anandtech and Tom's Hardware are all benchmarking processors against the wattage they use, I can see AMD, with its highly-integrated memory system, looking for ways to make performance gains along the current strengths of their chip. Intel's advantage is not in CPU design - it's in manufacturing, and you can see this in their current move to 65nm and taping out of 45nm before AMD has even finished the migration to 90nm. Intel might be slowly losing market share to AMD, but they're making it up in manufacturing costs. In contrast, for AMD to compete they have to work on their own strengths, and their strength is the capability of their processor and it's integrated memory system.
Currently, dual-core processors are interesting, but they aren't vital. Windows XP does a terrible job of using a dual-core processor, and this isn't likely to change until Vista. If I were to build a computer today I would base it on the Athlon 4000+, not the X2 4800+.
However, it will change. And when the OS, the games, the media, and the applications are all finally optimized for multi-threads, I will bet that you're going to see just how important cross-CPU information is going to be. If AMD can manufacture a quad-core Opteron that is screamingly efficient at handling on-die memory transactions, they will kick much buttage.
The people at AMD aren't dummies.
- Dec 31, 2007
- 48,582
- 92
- 66,190
I see the problem here, the problem is, you're confrontational. Someday that's going to hurt you my boy. So why don't you just prove me wrong. No, I'm not speaking about some mispoken statement, I'm speaking of a whole concept. Prove the concept wrong.
You can't. Because in the end, you're using "projection", that is, you're accusing me of having traits you see in yourself. Hmmph.
You can't. Because in the end, you're using "projection", that is, you're accusing me of having traits you see in yourself. Hmmph.
- Dec 31, 2007
- 48,582
- 92
- 66,190
Hey, I never claimed that AMD would be using RDRAM. I always assumed they licensed a technology that would enhance one of their products, but I never assumed which technology that would be. My talk of RDRAM was just in response to other people's comments.
read this
http://www.hardwarecentral.com/hardwarecentral/reports/1686/1/
then this
http://www.hardwarecentral.com/hardwarecentral/reviews/1787/1/
this is real reason for crappy proformance
http://www.geek.com/news/geeknews/2000nov/chi20001129003081.htm
- Dec 31, 2007
- 48,582
- 92
- 66,190
RDRAM operated very well in P4 systems. It just couldn't offer tangible benefits to PIII systems, because PIII's used a slow bus.
SiS even developed a quad-channel RAMBUS chipset for the P4, but it failed in the market place. Basically you could put two 32-bit RIMMS (dual channel on one RIMM) in dual-channel mode (Quad since each RIMM was already dual channel) for enough bandwidth to satisfy even the hugriest P4.
SiS even developed a quad-channel RAMBUS chipset for the P4, but it failed in the market place. Basically you could put two 32-bit RIMMS (dual channel on one RIMM) in dual-channel mode (Quad since each RIMM was already dual channel) for enough bandwidth to satisfy even the hugriest P4.
Shamrokevin and Worstnightmare your info and perspectives on relevant subjects being discussed are appreciated but whats with the personal attacks? MPJ admitted his mistakes, and Crash merely relayed his opinions. Not everyone has time to stock the RAMBUS events and procedings over the last 10 years or whatever. Besides some of these folks were probably not even off the boob when all of this mess got started.
Shoot, your arguments seem fairly respectable except for that distracting namecalling.
Shoot, your arguments seem fairly respectable except for that distracting namecalling.
worstnightmare
Distinguished
- Jan 7, 2006
- 5
- 0
- 18,510
hmmm, lets see... you surprise me with your in-site. Yes your are partially correct on all accounts. With your in site obviously you do know the truth. I can only ask why then do your statements not reflect the truth?
I intentionally was confrontational about what you consider a misspoken statement. I felt in needed to be brought out that in fact your statements are categorically both wrong and improper. There has been enough untrue and misleading information published in the past 6 years concerning RMBS and the IP they own. It is now finally coming out who was wrong, who committed crimes, who mislead the industry and profoundly who owns certain IP. Look at the Summery Judgements HYNIX has already lost on, they infringe, Samsung infringes as well as Micron and BTW there are many many others that need to be brought to the table to sign. AMD is just the beginning. Obviously the IP is the best otherwise no one would be using it and no one has designed around it. Read the FTC report, everyone had the opportunity to design around but chose to roll the dice to see if they could get away with the IP for FREE. Actually I believe they thought RMBS would go away or out of business. Well they lost. The PC and many other applications may not exist today at the speeds they do if it was not for RMBS!
Bottom line is, where talking about money and technology. The technology is fine that is why it is in use today. However many have lost small fortunes because of the thefts and down right lies that have occurred. Me included. It is and was completely unnecessary. Free enterprise is great! Theft, cherry picking and conspiring is not. Somewhere sometime it needs to change and it is beginning too.
I believe the AMD deal will open many other doors for AMD and their fine products. I also believe many other applications will come from it, some will benefit only AMD and some will benefit the free enterprise concept so others may grow.
My statements although confrontational were meant to spur dialog and I encourage others to explore what is really going on or happened. Some on this thread are under misconceptions based on other peoples statements to include your statements, as opposed to the reality that is finally being exposed.
Well wishes to you sir
I intentionally was confrontational about what you consider a misspoken statement. I felt in needed to be brought out that in fact your statements are categorically both wrong and improper. There has been enough untrue and misleading information published in the past 6 years concerning RMBS and the IP they own. It is now finally coming out who was wrong, who committed crimes, who mislead the industry and profoundly who owns certain IP. Look at the Summery Judgements HYNIX has already lost on, they infringe, Samsung infringes as well as Micron and BTW there are many many others that need to be brought to the table to sign. AMD is just the beginning. Obviously the IP is the best otherwise no one would be using it and no one has designed around it. Read the FTC report, everyone had the opportunity to design around but chose to roll the dice to see if they could get away with the IP for FREE. Actually I believe they thought RMBS would go away or out of business. Well they lost. The PC and many other applications may not exist today at the speeds they do if it was not for RMBS!
Bottom line is, where talking about money and technology. The technology is fine that is why it is in use today. However many have lost small fortunes because of the thefts and down right lies that have occurred. Me included. It is and was completely unnecessary. Free enterprise is great! Theft, cherry picking and conspiring is not. Somewhere sometime it needs to change and it is beginning too.
I believe the AMD deal will open many other doors for AMD and their fine products. I also believe many other applications will come from it, some will benefit only AMD and some will benefit the free enterprise concept so others may grow.
My statements although confrontational were meant to spur dialog and I encourage others to explore what is really going on or happened. Some on this thread are under misconceptions based on other peoples statements to include your statements, as opposed to the reality that is finally being exposed.
Well wishes to you sir
worstnightmare
Distinguished
- Jan 7, 2006
- 5
- 0
- 18,510
I do not believe I used information strictly from the companies website. I used 5 years of information, wether it was from RMBS website, court documents, personal conversations or from a website that has recorded, dated and documented the saga as it unfolded. I am not the most astute ( not an EE) on the saga but I do have an understanding of how and why the perceptions are. That being said there are 2 sides to everything and I have not heard all the opposing arguements
worstnightmare
Distinguished
- Jan 7, 2006
- 5
- 0
- 18,510
I agree with your observations and admit I should have chosen verbage more carefully. LOL, some were still on the BOOB, Thanks for the humor!
- Dec 31, 2007
- 48,582
- 92
- 66,190
I try to keep things simple. And sometimes things get long when I compile a bunch of simple statements.
The matter in question was weather Rambus had the rights to specific technologies listed in broader patents they were suing over. The problem was that those specific portions incorporated other people's ideas, stuff that they took from the JEDEC meetings. Technologies developed by collaboration belonged to all members of JEDEC and wouldn't be a valid portion of a patent by a single member, JEDEC is all about establishing standards for the entire group.
Dr. Tom was part of all this, so he was called in to testify against Rambus. That is to say, Tom himself was an insider, so that leant credibility to THG's take on the trials.
A lot of what I said had nothing to do with the patent squabble as I was simply relaying some of the problems that hurt Rambus' reputation with enthusiasts. It turns out that half the marks against Rambus had nothing to do with them.
The matter in question was weather Rambus had the rights to specific technologies listed in broader patents they were suing over. The problem was that those specific portions incorporated other people's ideas, stuff that they took from the JEDEC meetings. Technologies developed by collaboration belonged to all members of JEDEC and wouldn't be a valid portion of a patent by a single member, JEDEC is all about establishing standards for the entire group.
Dr. Tom was part of all this, so he was called in to testify against Rambus. That is to say, Tom himself was an insider, so that leant credibility to THG's take on the trials.
A lot of what I said had nothing to do with the patent squabble as I was simply relaying some of the problems that hurt Rambus' reputation with enthusiasts. It turns out that half the marks against Rambus had nothing to do with them.
Wow its always interesting to see such a spirited "debate".
I used to use quotes on the internet. Its just hard to know what is rhetoric, fact, opinion or a mix of all of the above.
Does anyone else remember Micron showing up with Proprietary DDR first on the shelves in their machines before it was available mainstream?
I am really looking for confermation here from someone else that was working in the industry and not a handpicked quote.
Either way
I remember hating Rambus and the Evil Empire because of the Memory and proprietary parts and crazy pricing.
Its so easy to manipulate info these days everyone is a thief or a saint depending on which page you read. :>
Old dogs can learn new tricks they just cant preform them.
Sunset1
I used to use quotes on the internet. Its just hard to know what is rhetoric, fact, opinion or a mix of all of the above.
Does anyone else remember Micron showing up with Proprietary DDR first on the shelves in their machines before it was available mainstream?
I am really looking for confermation here from someone else that was working in the industry and not a handpicked quote.
Either way
I remember hating Rambus and the Evil Empire because of the Memory and proprietary parts and crazy pricing.
Its so easy to manipulate info these days everyone is a thief or a saint depending on which page you read. :>
Old dogs can learn new tricks they just cant preform them.
Sunset1
I just wanted to point out the i820 was not a bad chipset and that there was outher major factors in rambus not porforming well. I agree most of the statments just wanted to point out the flaws in some.It also may or may not be true that rambus was designed for p4. It may be it just worked better with the p4 b/c of the higher clock speeds that rambus can handle. Rambus was designed with scale beyond courrent clock speeds and as you can see in the link it proformed really well with p3(outproformed the sdram).
http://www.hardwarecentral.com/hardwarecentral/reviews/1787/17/
http://www.hardwarecentral.com/hardwarecentral/reviews/1787/17/
Justmyopinionnotthatitmat
Distinguished
- Jan 10, 2006
- 2
- 0
- 18,510
I’ve been reading the posts of crashman and mpjesse, and just have to chime in, since these guys are just spouting the Rambus FUD that the electronic rags picked up, and don’t seem to have a clue as to what the real story is..
First off, Rambus doesn’t claim they invented DDR, or DDR2, or even SDRAM. What they invented were technologies that are used in those memory types. In Rambus’ ‘898 patent application, which was filed in 1991 before they ever set foot in JEDEC, they described 11 separate technologies for use in high speed synchronous memory. Remember, this was at a time when EDO and FPM were just coming out as standards.
The 11 inventions were:
a) Highly Packetized/Multiplexed Bus
b) Programmable Burst Length
c) Programmable CAS latency
d) Externally supplied Reference Voltage
e) Two Bank Design
f) Source Synchronous Clocking
g) Low voltage swing signaling
h) Dual clock edge
i) On-chip PLL/DLL
These had never been used before in high speed memory systems. Because there was more than one identifiable invention (remember, this was Rambus’ first ever patent application) the patent office said, pick one invention, and file the others as separate inventions later. Rambus picked a, and wrote claims that described how it is implemented or used. In other words, the description of that patent, and all of the divisional/continuation patents have the same description of the 11 inventions, but what was patented or “claimed” is only one of them. But Rambus still gets the benefit of the 1991 filing date on the later applications, because that is when they told the patent office about the inventions.
Again, before they ever joined JEDEC Rambus met with all of the major dram companies, and taught them about Rambus’ inventions under non disclosure agreements. They were laughed at, told it would never work, but in the FTC and Virginia trials we found out that behind Rambus’ back, the DRAM companies were calling the Rambus designs “revolutionary”, and that “some day all drams will be made like this, hopefully without the royalties going to Rambus”.
Rambus was invited to join JEDEC. When Rambus’ received their first patent in 1993, they disclosed it to JEDEC. And asked on two occasions to present their technology to JEDEC. They were refused twice, the ONLY company that JEDEC has kept from presenting. JEDEC started working on the SDRAM memory design, and decided to incorporate b, c, d, and e of the Rambus inventions.
Mitsubishi Electric reviewed Rambus’ patent that was disclosed to JEDEC, and concluded that Rambus could, and probably would file additional claims that would cover the use of their technologies in SDRAM. Under patent law, it is permissible to file additional claims to cover the use of your technology in other applications that you hadn’t foreseen, or hadn’t intended.
A lot has been said about JEDEC being an open source, open forum, royalty free, etc, and that Rambus was required to disclose that they could file additional claims on their patents. That is just plain a bunch of BS that was spouted by the witnesses in the Virginia trial. Yet at the FTC trial, we learned that just the opposite was true.
JEDEC is not open source, royalty free. In fact, the FTC judge noted that Rambus was asking less than the royalties being paid and/or cross licensed for by the JEDEC member companies. We also learned that JEDEC had no reporting requirements, and only required presenters to disclose their patent intentions. Ye we also learned that there were many cases where companies presented, didn’t disclose patents, and then asked for royalties, and JEDEC made no stink about it. In fact, the FTC judge noted that the complaint counsel witnesses that testified about JEDEC’s disclosure requirements and Rambus conduct had testimony that sharply contrasted the written records and documentation. He darn near flat out said the witnesses were obviously lying.
So what was the real problem behind Rambus? Again in the FTC trial we learned that the DRAM companies felt that if Rambus’ business model took hold, they would be relegated to nothing more than foundries for Intel, with Intel and Rambus doing the design work, and the DRAM companies having nothing to do but make chips. When it was clear to Rambus that JEDEC wasn’t going to let them present, and was going to just rip off Rambus’ designs, they quit JEDEC.
After Rambus left, JEDEC began work on DDR. DDR uses b,c,d,e,f,g,h,i. The JEDEC member companies stated that Rambus’ patents likely covered these technologies, but JEDEC determined that it was still the best way to make faster DRAMs. At the same time the Synclink consortium was working on SLDRAM, which Rambus flat out told them was a blatant rip off of RDRAM. The same companies that were in JEDEC were also in Synclink.
Rambus’ other patents from the original 1991 application were applied for after Rambus left JEDEC, but all of the patents have the same specification as the 1991 application. To say that Rambus stole ideas from JEDEC is just plain wrong and laughable. What Rambus described in 1991 was the use of those technologies in a high speed memory system with a narrow bus. This was listed as the preferred method, but other methods were possible. They incorporated claims describing how the inventions could be used in a wider bus as well in the later applications. This is perfectly legal under US patent law. It was Rambus’ ideas, not JEDECs.
So why did JEDEC include those technologies, if Rambus was going to patnet them? We learned in the FTC trial that the JEDEC members thought that the patents wouldn’t issue due to prior art. We later learned (and that’s what the antitrust trial is about) that some of the companies teamed up to drive RDRAM out of the market, and Rambus out of business.
But Rambus’ patents did issue, and they filed suit against Hitachi, and then Infineon. Immediately Micron and Hynix filed suit against Rambus, and sicced the FTC on Rambus.
As all of this progressed JEDEC started working on DDR2. Rambus informed JEDEC of the patents that Rambus held, yet JEDEC still went ahead and incorporated even more of Rambus inventions into DDR2. DDR2 is RDRAM with a wide bus. Same on die terminations, short stubs, voltage, voltage swing, etc.
So to say that Rambus stole anything is far from the truth. The truth is, Rambus figured out how to make high speed synchronous memory before anyone else did, and they want to be paid for the use of their inventions. The JEDEC companies don’t want to pay, because they don’t want to lose control of the design process. But they continue to use Rambus techniques, because they can’t figure out anyway to do it better.
And that’s the real story in a nutshell.
First off, Rambus doesn’t claim they invented DDR, or DDR2, or even SDRAM. What they invented were technologies that are used in those memory types. In Rambus’ ‘898 patent application, which was filed in 1991 before they ever set foot in JEDEC, they described 11 separate technologies for use in high speed synchronous memory. Remember, this was at a time when EDO and FPM were just coming out as standards.
The 11 inventions were:
a) Highly Packetized/Multiplexed Bus
b) Programmable Burst Length
c) Programmable CAS latency
d) Externally supplied Reference Voltage
e) Two Bank Design
f) Source Synchronous Clocking
g) Low voltage swing signaling
h) Dual clock edge
i) On-chip PLL/DLL
These had never been used before in high speed memory systems. Because there was more than one identifiable invention (remember, this was Rambus’ first ever patent application) the patent office said, pick one invention, and file the others as separate inventions later. Rambus picked a, and wrote claims that described how it is implemented or used. In other words, the description of that patent, and all of the divisional/continuation patents have the same description of the 11 inventions, but what was patented or “claimed” is only one of them. But Rambus still gets the benefit of the 1991 filing date on the later applications, because that is when they told the patent office about the inventions.
Again, before they ever joined JEDEC Rambus met with all of the major dram companies, and taught them about Rambus’ inventions under non disclosure agreements. They were laughed at, told it would never work, but in the FTC and Virginia trials we found out that behind Rambus’ back, the DRAM companies were calling the Rambus designs “revolutionary”, and that “some day all drams will be made like this, hopefully without the royalties going to Rambus”.
Rambus was invited to join JEDEC. When Rambus’ received their first patent in 1993, they disclosed it to JEDEC. And asked on two occasions to present their technology to JEDEC. They were refused twice, the ONLY company that JEDEC has kept from presenting. JEDEC started working on the SDRAM memory design, and decided to incorporate b, c, d, and e of the Rambus inventions.
Mitsubishi Electric reviewed Rambus’ patent that was disclosed to JEDEC, and concluded that Rambus could, and probably would file additional claims that would cover the use of their technologies in SDRAM. Under patent law, it is permissible to file additional claims to cover the use of your technology in other applications that you hadn’t foreseen, or hadn’t intended.
A lot has been said about JEDEC being an open source, open forum, royalty free, etc, and that Rambus was required to disclose that they could file additional claims on their patents. That is just plain a bunch of BS that was spouted by the witnesses in the Virginia trial. Yet at the FTC trial, we learned that just the opposite was true.
JEDEC is not open source, royalty free. In fact, the FTC judge noted that Rambus was asking less than the royalties being paid and/or cross licensed for by the JEDEC member companies. We also learned that JEDEC had no reporting requirements, and only required presenters to disclose their patent intentions. Ye we also learned that there were many cases where companies presented, didn’t disclose patents, and then asked for royalties, and JEDEC made no stink about it. In fact, the FTC judge noted that the complaint counsel witnesses that testified about JEDEC’s disclosure requirements and Rambus conduct had testimony that sharply contrasted the written records and documentation. He darn near flat out said the witnesses were obviously lying.
So what was the real problem behind Rambus? Again in the FTC trial we learned that the DRAM companies felt that if Rambus’ business model took hold, they would be relegated to nothing more than foundries for Intel, with Intel and Rambus doing the design work, and the DRAM companies having nothing to do but make chips. When it was clear to Rambus that JEDEC wasn’t going to let them present, and was going to just rip off Rambus’ designs, they quit JEDEC.
After Rambus left, JEDEC began work on DDR. DDR uses b,c,d,e,f,g,h,i. The JEDEC member companies stated that Rambus’ patents likely covered these technologies, but JEDEC determined that it was still the best way to make faster DRAMs. At the same time the Synclink consortium was working on SLDRAM, which Rambus flat out told them was a blatant rip off of RDRAM. The same companies that were in JEDEC were also in Synclink.
Rambus’ other patents from the original 1991 application were applied for after Rambus left JEDEC, but all of the patents have the same specification as the 1991 application. To say that Rambus stole ideas from JEDEC is just plain wrong and laughable. What Rambus described in 1991 was the use of those technologies in a high speed memory system with a narrow bus. This was listed as the preferred method, but other methods were possible. They incorporated claims describing how the inventions could be used in a wider bus as well in the later applications. This is perfectly legal under US patent law. It was Rambus’ ideas, not JEDECs.
So why did JEDEC include those technologies, if Rambus was going to patnet them? We learned in the FTC trial that the JEDEC members thought that the patents wouldn’t issue due to prior art. We later learned (and that’s what the antitrust trial is about) that some of the companies teamed up to drive RDRAM out of the market, and Rambus out of business.
But Rambus’ patents did issue, and they filed suit against Hitachi, and then Infineon. Immediately Micron and Hynix filed suit against Rambus, and sicced the FTC on Rambus.
As all of this progressed JEDEC started working on DDR2. Rambus informed JEDEC of the patents that Rambus held, yet JEDEC still went ahead and incorporated even more of Rambus inventions into DDR2. DDR2 is RDRAM with a wide bus. Same on die terminations, short stubs, voltage, voltage swing, etc.
So to say that Rambus stole anything is far from the truth. The truth is, Rambus figured out how to make high speed synchronous memory before anyone else did, and they want to be paid for the use of their inventions. The JEDEC companies don’t want to pay, because they don’t want to lose control of the design process. But they continue to use Rambus techniques, because they can’t figure out anyway to do it better.
And that’s the real story in a nutshell.
Justmyopinionnotthatitmat
Distinguished
- Jan 10, 2006
- 2
- 0
- 18,510
I figure that next someone will cry out about the EXORBITANT royalties that Rambus was asking for.
On SDRAM, Rambus was asking 0.75% of the chip price (not the completed module). At the peak SDRAM was selling for about $1.5 per MB retail for the completed module, the chips on the module were selling for roughly $10 each. and most systems had 64MB to 128MB. So Rambus' royalty would have been 60 cents per computer. Wow, that would have broke me.
When chip prices fell, computers used 128MB to 256MB, some up to 512MB, but chip prices fell to around $4 per chip. So the Rambus royalty would have been around 60 cents to $2.40 cents.
On RDRAM the rate was higher, 1.5%. Even when RDRAM modules peaked at around $1000 per system, the royalty was in the neighborhood of $12. And of course, RDRAM prices fell pretty rapidly, but we now know RDRAM prices were being manipulated (from the DOJ indictments and settlements against the DRAM companies).
Rambus asked for higher rates for DDR, 3.5%. DDR prices started out at the same as SDRAM. Of course, we now know that the companies were manipulating those prices as well. The Rambus royalties in this case would have been in the neighborhood of $1 to $2.60 per system.
Rambus will probably ask for in the neighborhood of 5% for DDR2, since it uses even more of Rambus' technology.
So what did you get from the DRAM companies eliminating RDRAM from the market place? In a way, I guess you got MORE choices. RDRAM was only made in 4 flavours, and it worked in any motherboard you plugged it into. RDRAm still commandes a good price on eBay, because there is still a demand for rock solid, stable systems.
With DDR, you get 1600, 2100, 2700, 3200, 3500 etc. And you get a flash back to the old days of EDO and early SDRAM, where only certain modules worked with certain motherboards, and you couldn't buy generic memory and be sure it would work.
But once RDRAM was out of the market, the DRAM companies jacked up the DDR prices practically overnight, which is what led DELL to compalin to the Department of Justice, and the subsequent settlements paid to the DOJ by the DRAM companies.
So bottom line, the royalties didn't kill Rambus, the minor cost would have just gone through to the consumer, who wouldn't have even noticed a buck or two of extra cost.
The real problem is that the DRAM companies didn't want the little upstart company challenging the goliaths ideas of what YOU as the customer should get to have.
On SDRAM, Rambus was asking 0.75% of the chip price (not the completed module). At the peak SDRAM was selling for about $1.5 per MB retail for the completed module, the chips on the module were selling for roughly $10 each. and most systems had 64MB to 128MB. So Rambus' royalty would have been 60 cents per computer. Wow, that would have broke me.
When chip prices fell, computers used 128MB to 256MB, some up to 512MB, but chip prices fell to around $4 per chip. So the Rambus royalty would have been around 60 cents to $2.40 cents.
On RDRAM the rate was higher, 1.5%. Even when RDRAM modules peaked at around $1000 per system, the royalty was in the neighborhood of $12. And of course, RDRAM prices fell pretty rapidly, but we now know RDRAM prices were being manipulated (from the DOJ indictments and settlements against the DRAM companies).
Rambus asked for higher rates for DDR, 3.5%. DDR prices started out at the same as SDRAM. Of course, we now know that the companies were manipulating those prices as well. The Rambus royalties in this case would have been in the neighborhood of $1 to $2.60 per system.
Rambus will probably ask for in the neighborhood of 5% for DDR2, since it uses even more of Rambus' technology.
So what did you get from the DRAM companies eliminating RDRAM from the market place? In a way, I guess you got MORE choices. RDRAM was only made in 4 flavours, and it worked in any motherboard you plugged it into. RDRAm still commandes a good price on eBay, because there is still a demand for rock solid, stable systems.
With DDR, you get 1600, 2100, 2700, 3200, 3500 etc. And you get a flash back to the old days of EDO and early SDRAM, where only certain modules worked with certain motherboards, and you couldn't buy generic memory and be sure it would work.
But once RDRAM was out of the market, the DRAM companies jacked up the DDR prices practically overnight, which is what led DELL to compalin to the Department of Justice, and the subsequent settlements paid to the DOJ by the DRAM companies.
So bottom line, the royalties didn't kill Rambus, the minor cost would have just gone through to the consumer, who wouldn't have even noticed a buck or two of extra cost.
The real problem is that the DRAM companies didn't want the little upstart company challenging the goliaths ideas of what YOU as the customer should get to have.
peach
Distinguished
- Jun 28, 2001
- 271
- 0
- 18,780
One thing proud people value more than money is their pride. Rambus challenged the industry on both.
First off I have no affiliation with any parties involved although I might go buy some RMBS so as to at least make some money out of the hour I just put into the reading of this spirited thread & its links - my thanks to all involved.
I really didn't enjoy reading the personal attacks (aoplogies noted), but I might guess that had we experienced what worstnightmare has these last few years, we might have not held back so well. It sounds as though he was personally vested in the situtation. I hope he's in a position to reap future benefits, although some things can't be undone.
I know there were a LOT of links in the thread and this one is at rambus's domain, but if havn't read it you really should. It is pretty short and VERY telling. here - Rambus is a deadly menace to an established Computer industry This was written in 1992 by Will Meyer of Infineon before the awarded patent to Rambus but after the intial one was applied for 1991 with all of the said technologies presented. The fear is written all over it and all Rambus was at the time was a small company with some great ideas. Meyers' thought had to be a global sentiment across the major players (or the established computer industry).
The worst part is - they almost got away with it, and kind of still have. I mean Crash's Historical title is *only* due to an inordinate amount of time spent writing posts, nonetheless I have come to repsect his posts. So I figure it's not often he is caught advocating arguments (imo) apparently orchestrated as a media disinformation plot by HUGE companies. Companies the public known to pricefix. So if someone as 'informed' as crash is clearly uninformed, or should I say misdirected, is it fair to say the general enthusiast is? I was. Prior to the thread if you asked me Rambus - Good or Bad? Right or Wrong? Wronged little guy or greedy & underhanded. I'd have checked against them 3 times, and I (prideful as well) wouldn't have considered myself terribly ignorant. Doesn't the general consumer have an enthusiast that they rely on for opinions like the Rambus fiasco? I worry - has Rambus lost the PR war? Maybe a rising stock and few more court wins will reverse that.
You know what my biggest beef is? So much of the public world is done in fear of "the man" and their retaliation. The forums of the internet are one of the few places where "the man" doesn't rule by sheer financial might. Three cheers for unedited free speech!
*rant over*
turpit
Splendid
- Feb 12, 2006
- 6,373
- 0
- 25,780
First: THIS THREAD IS OVER 2 YEARS OLD. RELEVENCY TO CURRENT SITUATION = NIL
Second: Chrashman is a "someone". He knows his stuff better than any 100 posters on this site combined
Third: Jesse hasnt been here for well over a year.
Fourth: Irrelevent thread necromacy = BAD
- Status
TRENDING THREADS
-
-
Question Which bottleneck/benchmark tools actually give the most reliable results?- Started by HamishZwar
- Replies: 0
-
-
Question Has anyone seen AMD Adrenaline Warning ⚠️- Started by Mrpockets151
- Replies: 4
-
-
Question High performance automotive radiator coolant for PC water loop coolant?
- Started by fcar1999ta
- Replies: 3
-
Question Cloned XP image on ssd will not boot unless sata port configured as IDE
- Started by kkroeker
- Replies: 3
Facebook
Twitter
Instagram