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 : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?


To: whortso who wrote (66326)12/27/2001 4:44:13 PM
From: pgerassiRead Replies (2) | Respond to of 275872
 
Dear Whortso:

Most speed distributions are Gaussian in pattern. That is the percentage in each bin (goes as faster that the nominal speed but, slower than the next nominal speed) makes a curve that looks like a Gaussian distribution curve. The two parameters of such a curve is the peak point on the X axis and the standard deviation (the first determines where 50% of the CPUs would bin in a single bin at that speed and the second shows how quickly the 33% of the CPU bin at a speed greater than the peak speed). These are the only two numbers of relevance and barring any unusual attributes or process quirks, this has been noted in almost all speed bin distributions since day 1. Given a peak of 1.5GHz and a standard deviation of 100MHz, the number of 2GHz CPUs are found by taking the 1 minus Gaussian cummulative function of (2GHz - 1.5GHZ)/100MHz or 5. That yields a number of about 1% IIRC. A linear curve like yours means that no speed bins above 2.01GHz and none below 1.5GHz which does not occur in nature. Intel would fire the process manufacturing team as they are doing something obviously wrong (so would AMD for that matter). There are so many defects that could cause a P4 to run slowly but, correctly.

If you drop your pants, the cops and the DA would definitely want to talk with you about indecent exposure. Don't do it!

Pete