So the new standard is clearly the one set by AMD/Micron.
Dan,
Okaaaaay. If you say so.
I am frankly tired to death of talking about the technology. Please review the following very balanced (if you read it all you will see definite personal bias against RDRAM on the behalf of the author) exhaustive benchmarks:
tech-report.com
In addition, if one thinks RDRAM is dead, then one really does not understand the P4 architecture, or the importance of the P4 to Intel. The P4 is Intel's equivalent to the x86 for the next 5 years or so. At every new product introduction Intel has had enormous competition. Going back to before the 8086. This new product transition is no different than the prior transitions. The P4 is it for INTC. If this architecture fails you will have AMD gaining so much marketshare that even the whisper of INTC as a gorilla will not be heard. In fact AMD will become the first monkey in my knowledge to conquer a gorilla at their own game. Given this history and with this much at stake, to think that RDRAM will not be supported going forward is to totally misconstrue exactly what makes the P4 special and absolutely imperative for Intel. The P4 is not backwards looking. It is not made to make Word run faster. We can only type so fast. It is made to make Quake run faster, to enable voice recognition, multi-tasking, SETI alien tracking, number crunching in the calculation of global warming, CAD applications, Toy Story III, home streaming media movies, digital photography manipulation, etc. In a word it was made for very "intense" computing. The type of computing we all expect to become the norm very shortly as applications catch-up, yet once again, to technology.
But hey, maybe I'm wrong. Maybe we have reached the end of where the PC goes. We are satisfied. Excel can't be made to work any faster. Therefore the P4 is a product without a market. We have reached the epitome of desktop computers. Hmmmm, maybe we have. If such is the case Rambus certainly is dead because SDRAM can run your Excel, under 95% of circumstances, just as well as can a P4 equipped RDRAM system.
I will just end with the comment that Intel may be the single best high-tech marketing firm of our age. Intel's motto is differentiate or die. Intel is also intensely aware of the need to provide the best product, that is the best "product," not the best "device." DDR is a device, RDRAM is a device, P4 is a device. The package is what all these devices from Intel do for the customer. It includes standardization, software value-chain, support, distribution, production capacity, product attributes, software compatibility (SSE2), advertisement, Intel brand equity, etc. Benchmarks just describe devices and are only one part of the total product.
As a quick example, when Intel beat Motorola in the transition to the 8086 chip, Intel did not do so because the 8086 was a better device. On the contrary, internally, Intel admitted that Motorola's offering was technologically superior in many ways. What Motorola didn't offer was a better system, with software support, specialized, expert sales force, leading reputation for future integration and a credible road-map thereon, best service in the business, a guarantee that if you commit to Intel your products will work, our service department will make sure of it, and a whole host of other factors which by the time it was over with made Intel into the CPU gorilla.
Intel won this market with a gorilla grip because Intel puts together the best products. If Intel thought that involved DDR it would not have spent billions on pushing RDRAM and instead just took the much easier and cheaper path to DDR. DDR would have been a very simple transition (theoretically as getting DDR 266 to work is proving quite a chore). No need to pay out hundreds of millions to subsidize DRAMs to make the transition. No need to put out the crappy 820, etc. Intel would not have architectured the replacement to the x86 chips (the P4) to be so dependent upon RDRAM for future performance if it did not think it was putting together the best "product" possible. But hey, even Intel can make mistakes. Perhaps Intel will concede its market leadership role to AMD and Samsung will do the same to Micron.
On the note of the best product. If AMD/Micron were setting the standard, with chips producing a great product, then why is it that DELL, who also prides itself on providing the best practical technoloy "products" to their customers (I emphasize "PRACTICAL" as DELL does not sell cutting edge, early adopter products to corporate America)will not use AMD? If AMD is cheaper, performs better the tasks of corporate America, DELL would be crazy to not do it. So why does DELL persist? Because AMD does not provide the complete package to DELL or its customers as INTC does. Selling CPUs is a lot more than just the microprocessor or the DRAM or any of the devices individually. It is the ability to create the best product. At this INTC shines perhaps better than any company in history.
I also tire of these value-chain and marketing arguments. The facts are in front of our faces. Make what you will of them.
Tinker
I am not concerned one iota anymore of the acceptance of Rambus products in the marketplace. Rambus is there, and there to stay. The only thing of concern on my mind is the forthcoming lawsuits. I am at the point where Qualcomm investors were in 1998 or so. Not real concerned about CDMA in the marketplace. Instead I'm just worried about the lawsuits which seek to deny the enabling companies ability to collect royalties on the technology. Please see
eb-mag.com
for a good overview of the antitrust suits. I typed the link in, and did not paste it, so if it does not work give me a holler. The only thing between Rambus and enormous wealth at this juncture is these lawsuits IMHO. |
