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: MileHigh who wrote (14095)	1/23/1999 9:32:00 AM
From: MileHigh	Respond to of 93625

IBM obtained a license, but you have to wonder if they will use it at all.... MileHigh

To: MileHigh who wrote (14095)	1/23/1999 9:38:00 AM
From: capt rocky	Read Replies (2) \| Respond to of 93625

IT'S THE MILEHIGH SHOW!---- starring the new bear of the rambus thread. seriously , most of what you posted makes me want to buy more rmbs. and i will(as soon as cien announces they've bought cisco.((just kidding)) ).since unc. went to cal. i'll do his first post.----------BUY RAMBUS!!!!!!!!!!!!!! capt rocky

To: MileHigh who wrote (14095)	1/25/1999 10:09:00 AM
From: Shibumi	Read Replies (2) \| Respond to of 93625

Re: IBM, Reliance lead march toward Direct Rambus alternative... Thanks for posting this. It's interesting you got called a bear for doing so -- actually, this article increased my belief in Rambus. The section that I thought was the most telling is quoted below: >>"Building a viable 133-MHz spec for system memory is tough," said Pete MacWilliams, an Intel Fellow who heads the Santa Clara company's Memory Enabling group. "Maintaining backward compatibility to PC/100 requires a 3-ns hold time. This makes it very difficult to tighten the access time. The result is that the system timings need to be improved. Assuming the three-quarter DIMM configuration is still the design point, we believe this will require buffers, probably on the motherboard and module. That adds cost and latency," MacWilliams said. Pulling acknowledged it is more difficult to meet the hold time with faster parts, but said Reliance managed to avoid using buffers on the motherboard. The faster SDRAM spec will need them on the module, but Reliance said that using buffered registered DIMMs is already commonplace for high-end systems.<< I'm sure that this might seem very esoteric to many people on this thread. What's basically being said comes down to the following: o Intel: It's too bloody hard to interface to these 133MHz DRAM's. You're going to need to add electronics to handle it -- which adds costs and complexity. o Reliant: You're right, it's hard -- but we can do it without motherboard impact by adding all of the costs to the DIMM's. Please note that what's more important than the disagreement is the implicit agreement by both that 133MHz SDRAM's are hard to handle. What does this say for increased performance SDRAM's -- either with regard to DDR techniques or even worse with regard to frequency increases? It says that you're seeing the nearing of the end of the line for traditional DRAM's with respect to old-style interface technology. The argument in my mind has always been one of timing -- when will Rambus's approach win in the marketplace? There are two major factors working against Rambus: cost of implementation (fab costs are higher since Rambus is new technology and hasn't been optimized -- the issue of being low on the technology S-curve) and supply. It was interesting to note the admission by the SDRAM proponents that some degree of cost was going to have to be absorbed by someone due to the additional buffering. Now, this may be a small cost handled in the DRAM controller or a larger cost handled discretely, but it shows the increasing costs associated with the evolutionary SDRAM-based approach. So the question comes down to supply. It's obvious to me that Rambus is going to have its earnings hurt (as they've announced) by spending a great deal of money supporting the ramp. It's also obvious that Intel is spending a great deal of money (e.g., Micron, Samsung) to help DRAM vendors with this supply. Thus, the only question in my mind is whether Rambus will fail to meet its objectives and Intel will have to support the 133MHz SDRAM due to supply issues. This would be the best buying opportunity of Rambus shares. However, even this very negative eventuality doesn't seem to impact the long-term intrinsic value of the company's technology. Again -- thanks tremendously for posting the article -- I learned a lot from reading it.