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.
Politics : Formerly About Advanced Micro Devices
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: Yousef who wrote (62113)6/17/1999 10:05:00 PM
From: Jeff Fox  Read Replies (2) of 1571691
 
Yousef, re: 3.3v I/O support on .18 processes

The previous 8 messages in this thread were an interesting review of a serious problem . . . for chipsets. For now Intel will use .18 micron process for processors only. Processors do not need to talk to standard busses such as 5v PCI or 3.3v SDRAM, at least for now. You can figure that today's Intel .18 micron process has a single poly gate layer and a single gate oxide thickness.

Now the chipset problem - and also the problem of one-chip processors that talk to memory or I/O. These chips require dual voltage I/O.

You (Yousef) are correct - well, correct up to now anyway. Existing Intel chipset stack two transistors and manage their gate voltages to avoid hot electron stress. BUT - this technique only works well where the higher voltage is less that about 1.7 of the primary supply. (i.e. the technique allows 2x, but certain dynamics on switching cut that down). Something else is needed to support I/O at greater than 2x the primary supply voltage. Well not to worry; Intel circuit designers have good solutions in hand. Chipsets and other variants at Intel are typically a year or more behind the first chip introduction on a new process. This works out well as the process ramps capacity and thus can accept more product volume on its line. The dual voltage ability will be proven and implemented with these products.

RE: Dual poly is a gross "brute force" and expensive solution. It not only adds many process steps, but cuts overall yield. Perhaps this is the only solution for the foundry guys, but this isn't in the cards for an Intel logic process.

re: AMD on chip regulator

Man this sounds fishy. Think for a moment, if the processor is running 1.6v internally on all transistors and is doing this supplied by a 2.5 -> 1.6v on chip regulator then that regulator is burning many watts of power! All that extra heat and it is all total waste. This is a dumb thing to do unless your forced to be socket compatible with a 2.5v motherboard. Perhaps this is an interim thing?

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

HomeMailHotSubjectMarksPeopleMarksKeepersSettings
Terms Of UseContact UsCopyright/IP PolicyPrivacy PolicyAbout UsFAQAdvertise on SI
© 2025 Knight Sac Media.  Data provided by Twelve Data, Alpha Vantage, and CityFALCON News