SI
SI
discoversearch

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
RMBS 89.73+2.3%2:46 PM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: Dave B who wrote (18278)4/6/1999 5:47:00 PM
From: Tenchusatsu  Read Replies (3) of 93625
 
Dave and Rambus investors,

Dave asked me to post on the Rambus thread and answer two questions for him (as I understood it):

1) Is DDR SDRAM more unstable than RDRAM?

In a nutshell, yes. For any flavor of SDRAM, the memory interface has four different signal groups (address, data, control, and clock), each with different electrical characteristics. Increasing the speed of the interface means that all the signal groups will change non-uniformly. DDR SDRAM tries to solve this problem by only increasing the speed of the data signal group. But even this method isn't as clean as the RDRAM method, which uses one uniform group of "transmission lines" for address, control, and data. One reason is that there are no "stubs" on the RDRAM lines, while SDRAM requires a "stub" for each memory module. More stubs increases the electrical loading on the line, which makes higher speeds very difficult to achieve.

2) What are some of the difficulties that could be causing RDRAM delays?

I really don't know all the details, and even if I did, I don't think I could reveal any. There are always headaches associated with validating a radically different architecture, and RDRAM is no exception. Because of its "transmission line" interface, hardware testing equipment has to be redesigned in order to inspect this interface at full speed. The lines themselves also have to adhere to strict compliance rules, such as making sure that the impedance along the line doesn't change in the middle. And the RDRAM interface protocol is rather complicated because of the need to pack the address, control, and data information into "packets" in order to transmit them across the uniform transmission lines.

My guess is that the difficulties that RDRAM developers are currently experiencing may just be due to Murphy's Law, i.e. anything that can go wrong will go wrong.

And that's my two cents. By the way, I currently have a small position in RMBS, but I plan on adding to it soon, considering that even with all the headaches, RMBS will be the future of DRAM.

Tenchusatsu
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext