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)
AMD 203.76-1.1%Nov 21 9:30 AM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: peter_luc who wrote (35648)4/14/2001 7:56:55 AM
From: ted burtonRead Replies (5) of 275872
 
Peter_luc says "See this great article...".

Hmmm...
The writer makes a huge issue out of the apparent disparity between the 54.7W thermal design power (TDP), and the 73.9W max power. Asserting that many, if not all, power hungry apps will blow away the TDP, & suffer a 50% performance loss! Doesn't anyone wonder why he doesn't offer any concrete examples of this truly nasty behavior?

Answer... There are no nasty examples.

A huge amount of work went into finding a TDP for which the highest power part, running the worst known real app, in the most borderline thermal environment, would suffer an imperceptible performance loss.

When the thermal management kicks in, the author tells you the clock is cut to a horrendous 50% clock rate, and asserts that power hungry apps are going to run 750MHz on a 1.5GHz machine. This is pure hogwash.

If, by some miracle, you find an app that runs at 60W, and due to another stroke of extreme bad luck you happen to have a part at the extreme high power end of the power distribution, and by yet another amazing stroke of bad luck all the manufacturing guardbands AND all the testing guardbands were eaten up by the corresponding theoretical errors they were meant to cover, and by yet another phenomenally bad stroke of luck you really ARE that one person out of a hundred dumb enough to bring his brand new enthusiast machine into the sauna with him, you may actually see the performance drop by between 10 & 20%. The clock would be stopped just often enough to drop the power back to 55W the thermal solution was designed for.

The 50% duty cycle the author/alarmist hyped so heavily is only going to happen on the microsecond scale, & only when the die temperature drifts close to the safe operating limit.

I can't comment on the rest of the article except to state the obvious - this guy's got an axe to grind.

-Ted Burton

P.S. I am an Intel employee, but the views expressed above are my own, and are not necessarily shared by my employer (an official spokesman would not be allowed to resort to name-calling or sarcasm even when it was so well deserved).
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext