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 96.40+5.4%Dec 19 9:30 AM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: Bilow who wrote (34770)11/17/1999 11:02:00 AM
From: Ian Anderson  Read Replies (2) of 93625
 
Since I don't believe the channel length is an issue, I don't think that it's relevant. In any case we are talking maybe 1" longer.

Rambus is a set of transmission lines. Provided they are properly terminated and laid out you can run the channel down the hall, and the only effects would be 1)increased latency 2)some signal attenuation. If I wanted a high speed bus to run down the hall, and was designing my own controller, I might well consider using Rambus.

Using the 820, the round trip delay it can handle is the limiting factor, since the 820 only allows 5 time domains (if they are all working ???) which is 6.25ns round trip delay to the furthest RDRAM.

To make the maths easy assume thats 3ns each way. A reasonable assumption about the propagation speed on a loaded bus is that it is 0.5 times the speed of light, which is 0.5ns per foot. The distributed capacitance is what makes it slower. That makes the maximum Rambus channel length the 820 can support about 18". Sounds like there is plenty of margin for two RIMS (4" each) plus 3x1" interconnects.

Of course if the reason Intel delayed is that not all the time domains work, it might be another story.

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