Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Rambus (RMBS) - Eagle or Penguin -- Ignore unavailable to you. Want to Upgrade?

To: NightOwl who wrote (48315)	7/30/2000 4:54:00 AM
From: Bilow	Respond to of 93625

Hi NightOwl; Re why INTC got in bed with RMBS (and contracted a nasty case of Awkward, Inconvenient, DRAM Syndrome). In the early 90s, Intel (and everybody else) concluded that processor were going to grow to big for their britches. That is, they looked at the trends in processor memory bandwidth usage and concluded that by 1999, in order to get enough data into a CPU chip, they would need more pins than would be available with the packages that were likely to be available by then. In that circumstance, a higher bandwidth per pin memory technology was on everybody's mind. SLDRAM, DDR SDRAM (the earlier variety), RDRAM, all these were thought up then. But the predictions of processors being limited by their memory bandwidths failed to come to pass. There were two reasons for this. First, the processor designers added larger and more efficient on-chip cache memories, and these provided most of the needed bandwidth. Second, the packaging industry found ways of adding huge numbers of cheap pins to packages. Thus there turned out to be two ways around the memory bandwidth problem. RDRAM was always the memory technology that was going to solve the memory bandwidth problems of the near future. The problem is that the near future kept getting farther and farther away. Cheaper technologies keep intervening, and eliminating the need for these packetized memory technologies. At this point in time, it is too late for RDRAM to catch on, and the industry is not likely to adopt another Rambus standard in the future. There were just too many Rambus engineering mistakes in Direct RDRAM for anybody to seriously consider them. In addition, they have alienated a substantial percentage of the industry. So DDR will get replaced by DDR-II, and then, probably, by whatever the ADT consortium cooks up. That should keep us for the next 8 years or so. By then, I expect SOC and embedded to be the dominant solutions. At that time, the memory interface problem will disappear from history, never to reappear again. (Though there will always be high speed interfaces needed to connect chips, and those speeds will continue to increase for long into the future.) -- Carl

To: NightOwl who wrote (48315)	7/30/2000 8:50:36 AM
From: Zeev Hed	Read Replies (1) \| Respond to of 93625

Nightowl, you finally got it, that is exactly where we are headed. You will not lose the random walk of traditional PC (but may not get much of an oomph with RMBS with highly ransomized short packets), but you will enable seamless streaming applications. The proof of the "merging" of PC computing, communication and entertainment is the AOL/Warner wedding. Will it go there? I don't know, but it sure looks as if that the general strategy. Will it succeed? I do not know. When computing and communications merging was "attempted" on a grand scale before (IBM/NCR) it was foiled. Watch how successful the warnert/AOL wedding is to get an inkling as to the success of this trend. However, to allow such a future merging, the underlying hardware (and software) should be unabling and take into account the most demanding applications. The way I see, INTC does not want any part in the computing platform to serve as a bottle neck to ever increasing CPU speed. If they accepted slow evolution of SDRAM into DDR DRAM, they would soon find themselves producing premium (high frequency) CPU's that cannot be exploited at max speed because of slower memory, and thus enable to charge the huge premiums for the fast parts. Profits would disapear (commodity state). That is one reason Grove and Gate are "in bed" with Wintel. Zeev

To: NightOwl who wrote (48315)	7/30/2000 8:51:05 AM
From: Zeev Hed	Read Replies (1) \| Respond to of 93625

Nightowl, you finally got it, that is exactly where we are headed. You will not lose the random walk of traditional PC (but may not get much of an oomph with RMBS with highly randomized short packets), but you will enable seamless streaming applications. The proof of the "merging" of PC computing, communication and entertainment is the AOL/Warner wedding. Will it go there? I don't know, but it sure looks as if that the general strategy. Will it succeed? I do not know. When computing and communications merging was "attempted" on a grand scale before (IBM/NCR) it was foiled. Watch how successful the warnert/AOL wedding is to get an inkling as to the success of this trend. However, to allow such a future merging, the underlying hardware (and software) should be unabling and take into account the most demanding applications. The way I see, INTC does not want any part in the computing platform to serve as a bottle neck to ever increasing CPU speed. If they accepted slow evolution of SDRAM into DDR DRAM, they would soon find themselves producing premium (high frequency) CPU's that cannot be exploited at max speed because of slower memory, and thus enable to charge the huge premiums for the fast parts. Profits would disapear (commodity state). That is one reason Grove and Gate are "in bed" with Wintel. Zeev