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 -- Ignore unavailable to you. Want to Upgrade?


To: Rich1 who wrote (79938)11/23/2001 10:49:08 AM
From: blake_paterson  Read Replies (1) | Respond to of 93625
 
Quad Band 'a pipe dream', DDR barely works

By Mike Magee, 23/11/2001 08:03:28 BST

213.219.40.69

ENGINEERS CLOSE TO THE SIGNAL action have told the INQUIRER that Kentron and maybe Via's cunning plans to implement quad band memory (QBM) are almost certainly doomed to fail.
And, worse than that, DDR (double data rate) memory, even in its current recension is, as Intel engineers have indicated during this year, far from being the panacea some propose, with great sensitivity to device loading.

According to one senior memory engineer, the introduction of FET switches in the data path greatly complicate the design and introduce another source of noise to DDR. "Few designers even know that FET switches can be the largest source of crosstalk in a system," he suggested.

And packaging for the controller and the memory are the largest problems that will need to be dealt with, with the QBM document we pointed to yesterday not even discussing the matter.

"Via has an extremely poor record of dealing with these problems in their chipsets. It usually takes Via one or two revisions of their chipsets and packages to 'get it right'," he alleged. And, at these data rates, it gets harder and harder to perfect the techniques, a point illustrated by the problems Rambus experienced.

The Kentron simulations are too simplistic and, the engineer said, will not stand up to rigorous characterisation.

"They most certainly did not characterise the crosstalk of the FETs and the packages in either their simulations or their testing," he added. µ


AHHH, THE TRUTH EMERGES... Hell, I should have bolded the entire damn story line <g>

BP