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 (48940)	8/4/2000 3:07:32 AM
From: Bilow	Read Replies (3) \| Respond to of 93625

Hi NightOwl; Re embedded and RDRAM (and ESDRAM, for that matter). The basic reason that you have to have fancy DRAM interfaces (like SDRAM, ESDRAM, RDRAM, DDR, etc.) is that you are connecting two (or more) chips, each made of Silicon and capable of incredibly high densities and speeds, with Printed Circuit Board material capable of much, much, much less density or speed. When you go to embedded, that conversion, from Si to PCB, disappears, and so does the need for patented, or JEDEC standard memory interfaces. Embedded bandwidths are huge. Incredibly high. Far beyond what is needed for a long time to come. And when higher bandwidths are needed, it is clear that embedded DRAM will be able to provide it. Where embedded fails is in the size of memory available. But as integration continues, you can expect to see more and more chip area dedicated to various memories. With CPUs, we have already seen the embedding of cache memories. That used to be a nice market for SRAM makers, now it is gone. Eventually, the DRAM market goes down the same hole. The rise of embedded is identical to the decline and elimination of the need for standard DRAM interfaces. Given that these stocks are trading at very high P/E multiples, I would say that they are not really good ideas for the long term buy and hold type. That doesn't mean that you can't make a lot of money on them over the short term, just that the market for these products is one that is not going to keep growing at the speed of the chip industry in general. -- Carl

To: NightOwl who wrote (48940)	8/4/2000 2:05:12 PM
From: pheilman_	Read Replies (2) \| Respond to of 93625

Nightowl, Carl covered most of the points of embedded DRAM. But, having worked on one of these chips, I thought I would add a bit more. The bandwidth is great. The bus width is as wide as you want it, 128-256 bits at a time, no problem. The structure of the DRAM can be modified to match your application. Knowing your access needs, the DRAM can be built with more or fewer banks. There can be significant power savings, as signals don't have to drive out onto PCB traces. There is some problem with the speed of the logic that can be fabricated on the die with the DRAM, due to optimization of memory. Denser than SRAM, not nearly as fast. It seemed like a great idea in 95-96 when DRAM prices were very high. Conversely the DRAM fabs were uninterested in converting capacity from DRAM to embedded DRAM designs. The fabs were more interested after the price of DRAM collapsed to the trend line in 98. And it is very unlikely that IP from Rambus is involved in the embedded designs. Most of the claims from Rambus patents start off with "A multiplicity of chips..." Paul