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 : RMBS: Rambus, Inc. -- Ignore unavailable to you. Want to Upgrade?

To: Jack Hartmann who wrote (6)	12/14/2000 10:51:21 PM
From: Jack Hartmann	Respond to of 79

Good fool post 1. Intel needed Rambus as the article notes, in order to maintain its speed advantage. 2. SDRAM is quite adequate for PCs up to 1 GHz, and so at the time of introduction of RDRAM, SDRAM still had at least two years of productive life, and much lower costs. 3. Intel flubbed with its development of the i820, with the result that RDRAM with a P3 on the desktop provided only minimum performance advantage. 4. VIA and other Taiwanese fielded chipsets that enabled use of SDRAM and the P3 that performed as well as Intel's chipset(s). 5. RDRAM suffered as the article notes because of problems Intel incurred with the 820 and something called the MTH which was supposed to allow use of SDRAM with the 820. 6. RDRAM also suffered because the Intel 820 offered only a single memory channel, unlike its more successful cousin the 840 chipset. This severely limited RDRAM and PC system performance. 7. Asian manufacturers --- now are building large numbers of PCs without Intel chips. This is true today, but only in the sense that many MOBO manufacturers are using VIA chipsets rather than Intel chipsets. All of them are still using Intel CPUs in about 80% of all systems. This will not likely remain true next year. Intel has pointedly not licensed VIA for the P4, which is due to replace the P3 over the next year or so. The chipset and board manufacturers who stuck with Intel are getting licenses to build RDRAM based P4 PCs, and later next year they will likely offer SDDRAM based entry level P4 systems. I expect Intel to return to a market share in chipsets of at least 65% by the end of next year and 75% the following year. 8. DDR is a likely loser. Intel's enemies have been trying desperately to sell DDR as the replacement for RDRAM. It won't wash. a. The MMs can't get DDR to work in PCs at speed. b. DDR doesn't offer the speed that the P4, with its 400 MHZ front side bus, needs. At best DDR is half as fast. It doesn't scale to the higher speed Intel (and AMD) will be introducing over the next year, much less beyond that. c. DDR has noise problems that increase with the speed. 200 MHz works, but offers no measurable performance advantage over SDRAM, while costing substantially more today. d. 266 MHz DDR doesn't work at all today, because of noise and systems integration problems. Despite the fact that Micron announced sytems in October, none are likely to be delivered until 2nd half 2001, if then. And they will perform substantilly slower than a P4 with RDRAM, while likely costing at least as much, because of yield and test problems. f. At the system level DDR will carry a price penalty because of its noise problems. 9. Intel's i820 chip set never achieved great popularity because of the failures asociated with its launch, and because, constrained to only one RDRAM channel, it offered little real performance improvement while requiring use of RDRAM, which was then priced as much as 8X higher than RDRAM. 10. Rambus charges up to 5% in royalties for use of its technology. This has not been a real issue. First the average royalty of SDRAM is only 1%, for RDRAM 2% and for DDR 3%. Only those who sue pay higher royalties. The 5% figure applies only to memory controller circuits. There is only a single memory controller to a PC. Bert makes a big issue of this, but it has not been a limiting factor. 11. The big factors slowing RDRAM success have been the weakness in the 820, the high price associated with launch of a new technology, and the slower that desired ramp, which kept the prices higher than they would have been. All of these factors are dissappearing with the successful launch of the dual channel P4, the PS2, and the many new products now reaching the market using RDRAM. 12. RDRAM is now a new standard. DDR is the one suffering and struggling to get out the door and find a home. No one is likely to use it other than AMD, and even they are in doubt. All the major OEMs have signed on to the RDRAM P4 program. ONly Micron and NEC have signed on to AMD's DDR/Athalon program, and neither of them can ship it at speed. 13. Rambus and RDRAM will benefit from all of these factors to become the dominant form factor in the memory business. Author: DougK99 Jack

To: Jack Hartmann who wrote (6)	12/15/2000 1:50:17 PM
From: Don Green	Read Replies (1) \| Respond to of 79

DRAM makers sit on stocks By: Drew Cullen Posted: 15/12/2000 at 16:49 GMT Spot market DRAM prices rose for the week of November 20, but mainly because manufacturers are sitting on stocks, ICIS-LOR reports. The DRAM makers don't like today's current price slump -and they expect prices to swing back upwards in line with anticipated increased demand next spring. But this means holding onto stocks for three to four months. But can all the makers last this long? We doubt it very much. ICIS-LOR's 30 rolling average of 128Mb DRAM (PC133, 16Mx8) prices for large volume buyers in the week to November 24, 2000 was US$13.82 in North America, US$9.92 in Europe, and US$13.56 in Asia. Prices were stable in North America, down 7.07 per cent in Europe and 6.76 per cent down in Asia, from the previous week. AsiaBiztech reports. The spot market for memory modules saw 128MB DIMMicrosoft (PC133) down 7.14 percent from the previous week to US$62.58 in North America, 6.88 per cent down to US$60.89 in Europe, and 4.86 per cent down to US$56.32 in Asia. ®

To: Jack Hartmann who wrote (6)	12/18/2000 9:39:15 AM
From: tinkershaw	Read Replies (1) \| Respond to of 79

The reason I like RMBS is (1) Jack, this is no Yahoo! post, that is a response I gave to a poster on the Avanex SI board (since obviously we had run out of cool stuff tot talk about regarding AVNX) just off the cuff. Not tough to explain why one likes RMBS;) Anyway, interesting to see how information works its way around. But lets give credit where credit is do. That, like most quality posts on the web is an SI post. Tinker