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: Tenchusatsu who wrote (57651)10/14/2000 5:08:55 AM
From: Bilow  Read Replies (1) | Respond to of 93625
 
Hi Tenchusatsu; You wrote: "DDR was demonstrated on open-case systems. I was told that those were 266 MHz DDR DIMMs, and the motherboards were high-volume 4-layer motherboards."

Don't be taken in by the twisted liars of DDR!!! Those weren't really working boards. It was all faked, and how could you have counted the layers in those motherboards anyway? Did you peel one apart? Bet you didn't did you, huh? Huh? See! It's all just FUD. Every real engineer knows that DDR is impossible with less than 12 layers, and that's not FUD, Intel's Gelsinger himself said so:

VIA to Produce Core Logic for DDR-SDRAM July 2000
"We will support DDR-SDRAM operating at 100MHz for the server segment," said Patrick Gelsinger, vice president of Intel's desktop product group. "The electrical noise problems associated with DDR can be solved using 12-layer boards which are OK for servers but not for desktops. For desktops you need a narrow, fast memory flow, not a wide, slow memory transfer like DDR offers. Intel will not offer DDR SDRAM for desktops for at least the next two years."
nikkeibp.asiabiztech.com

I know that VIA says it's possible to make DDR motherboards with only 4 layers, but even they say that putting 4 DIMMs on the board is going to complicate things to 6 layers. Since everything they say is, at best, a half truth, that means that to get 4 DIMMs actually requires 6/0.5 = 12 layers, just like Intel says. Here's the VIA link:

Pro266 SDR and DDR Motherboard Design
via.com.tw

The worst problem for DDR is that since the design wasn't done by a single company, every time anyone makes the slightest change to anything, you have to go through the full testing to see if it still works. That means that they can't make small changes to the design easily, and that means that RDRAM will advance against DDR, cause RDRAM is so robust. Here's a quote showing exactly how unstable DDR really is:

The other key point is just because one module passed, if you make any tweaks at all to the artwork - which we don't recommend, by the way - but if they do change something, change something in their fab, switch to a new PCB vendor, any change at all, they need to go back through the validation process to make sure that that design with that particular vendor is still solid and still meeting the specifications.
intel.broadcast.com

-- Carl

P.S. What is the feeling around Intel about how long the i820 has to live? It seems like the i815 is replacing it.



To: Tenchusatsu who wrote (57651)10/14/2000 7:32:22 AM
From: Scumbria  Respond to of 93625
 
Ten,

AMD's 760 chipset isn't going to appear in systems until Q1 2001 at the earliest. I don't know where Tom's Hardware gets their inside info. I met the AMD spokesman personally, and he himself said that systems won't appear until Q1 2001 at the earliest.

When the AMD guy told you this, were you wearing your MPF badge with the word "Intel" on it? ;^)

Scumbria



To: Tenchusatsu who wrote (57651)10/14/2000 4:01:40 PM
From: Dan3  Read Replies (1) | Respond to of 93625
 
Re: I met the AMD spokesman personally, and he himself said that systems won't appear until Q1 2001 at the earliest.

You weren't wearing a badge that identified you as an Intel employee, were you?

If you were, I'm sure you received nothing but optimistically skewed full disclosure of all plans and schedules.

:-)

Dan