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: Zeev Hed who wrote (48321)	7/30/2000 10:09:12 AM
From: Bilow	Read Replies (1) \| Respond to of 93625

Hi Zeev Hed; You wrote: "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 this is true, how come the maximum bandwidth of the FSB of the latest Intel Pentium III chips is only 1.06GB/sec? (I.e. 133MHz x 8 bytes = 1.06GB/sec) This is exactly equal to the maximum bandwidth of PC133 memory. (Funny coincidence.) It seems to me that Intel stuck the bottleneck right on their own chip. No matter how fast they make what goes inside the chip, it still has to communicate with the outside world on that narrow port. Even the P4 has a bandwidth of only 3.2GB/sec, and that is easily attainable by dual DDR channels, as shipped with Nvidia chips from 6 months ago. Current rumor is that Nvidia will use dual DDR channels in the X-Box, and also in chipsets for the Pentium and Athlon in early 2001. Even Intel 4 processor systems still have a bandwidth limit of that 1.06GB/sec. If bandwidth really were a big issue, wouldn't they be using the AMD technique of putting a separate port with each processor? Here's a link to an article about their 16 processor systems. If memory bandwidth is such a big deal, how come they are only claiming "multiple gigabytes/second."? eet.com You've been saying that Intel has supported Rambus because of this alleged bandwidth limit problem for years, but it is nothing more than a reasonable explanation for people who really don't understand the slightest detail of the actual technology. Give me some links. Your theory is at odds with the facts of engineering. My theory is that Intel got stuck to Rambus back when it looked like there wasn't going to be a solution to the pin limitation in packages. That problem got solved, but Intel is a big ship and takes a long time to turn around. For those of us who have been reading the trade press, it is clear that Intel is in the process of turning around. RDRAM got stranded. -- Carl