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.

Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?

To: Tony Viola who wrote (72445)	9/20/1999 1:39:00 AM
From: Elmer	Read Replies (1) \| Respond to of 1573096

Re: "They data lines run at 400 MHz, producing 800 Megabits per second per line because there is data present on both edges of the data lines. " I think it's 800 Mega transfers X16 bits = 1.6 GBytes per sec. The data lines are running at 800MHz. The clock is 400 MHz and the data (2 bytes) is on both edges. EP

To: Tony Viola who wrote (72445)	9/20/1999 10:16:00 AM
From: Ali Chen	Read Replies (3) \| Respond to of 1573096

Tony, <you make it sound like there are 800 MHz signals that are required to be routed. I think we've been around on this before, but there are no 800 MHz signals.> You think half right but mostly wrong. Yes, formally no signal is switching at 800MHz. However, the RAMBUS timing requires about 100-150 picosecond time budget for egde skews, which means that the board designer must handle tracing/material properties up to 10-20GHz range (accounting for edge rise/fall times of 30-50ps). Can you comprehend what 50 picosecond is? Even ambient humidity could be a factor with this requirements! Even fingerprints on board! Could you recollect the "little industry-wide problems" with adoption of 100MHz memory? I can remind you how it was "solved": Intel provided memory makers with DIMMs layout that fits Intel's custom timing on their BX chipset, with traces tweaked down to 1/4 of inch. Please note that it was a move from 1500pS to 1000ps timing scales. Now intel wants to push industry 10 times that far. Could you tell me from your general experience what the chances are? I think that it is technically illiterate managers like yourself who dragged Intel into premature switch into techology that current MB makers simply cannot afford.