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: ted burton who wrote (35957)4/16/2001 8:10:38 PM
From: kash johalRespond to of 275872
 
ted,

re: power consumption

In this case you are right.

Both Intel and AMD way overspec their CPU's.

And as the MHZ ramp continues they have to get closer to the real limits.

I am sure Intel is fine for these type of apps.

And the McComas article was a complete red herring.

regards,

Kash



To: ted burton who wrote (35957)4/17/2001 2:21:03 AM
From: pgerassiRead Replies (2) | Respond to of 275872
 
Dear Ted Burton:

I have done these checks on Athlon Tbird, and K6-2 & K6-3 at home. At work, I have done this on P3, P2, and Celeron. Have you seen different? Then prove that you have taken a P4, packed the case with heat generating goodies, left the stock fan on, and then shoved it all into a closet with no airflow. Then look at the Scimark bench scores from at first, and then 72 hours of continuous usage. Wait until the closet air measures 50C (122F) or more before the second run. (If it does not rise above that, you are not running the P4 with all the goodies (burners, high speed hard drives, Nvidea Geforce 2 Ultra, etc.), the closet has good airflow, or you are air conditioning it.)

Oops! You see it not running as high a score the second time over that first run. You just observed trottling. It is easy for a system running in an air conditioned environment, with the case cover removed, and fans blowing air all over to not observe this "feature, not fraud". It is hard when all is out to get you (the definition of worst case).

Also, you do not need to measure wattage, only temperature. As I pointed out, the trottling circuit is spec'd to operate at a fixed temperature. It can be idle, if the temperature is above critical (72C or 162F), trottling will be continuous enabled. Wattage only comes into design of the Thermal Management System. The goal of the system is to be below a critical teperature which until recently was the maximum allowed die temperature. The wattage specs and the design ambient allows the HSF and case fans to be properly sized and positioned (it also has a basis for choosing things like peltier wattage and cooling capacity of gas cooling units).

Thus you can see with even simple motherboard status of things like die temp (from a thermal diode in a P4), core voltage, motherboard temp, and lastly any digital thermometer to measure inside air temp. Using the the units of Centigrade (C), volts (V), and freq (fixed by the CPU), you can perform the simple checks yourself. Take the air temperature after the system has been off for a while (all the temps will be the same). Now turn the system on and measure the values just after boot up (most bioses have a screen to see this). Let it wait some hours simply idling and then check the reading again. Run Scimark and record the values shown over time. Rerun Scimark until the values reach their maximum. Now enable Seti, record the values over a period say 24 hours (they should stablize after a while). You can repeat for each load you wish to place on the P4 system. See, how easy it is! If the die ever exceeds 72C, do the scores start to drop in benchmarks?

That is the real beef with the specs. Intel publishes the wrong ones in the datasheet and then has the gall to market these specs as the real ones. There is a big difference between 54.7W and 73W, between max die temp of 90C and a max performance die temp of 72C. Taken together, these make for a HSF half of the real size needed or less. Shame on Intel!

Pete