Possible Explanation for Varying Socket-Thermistor Compression Ratios

[Mikewarrior2]

One thing I have completely forgotten about.

With any socket-thermistor setup, you get, at best, 1% of the thermistor contacting backside PCB. The other 99% are contacting socket air, which might explain why any small amount of airflow can disrupt the socket-thermistor reading.


Mike

 

[jsbush]

Mike do you know why AMD didn`t include one in there cpu like intel did?

 

[Mikewarrior2]

Jsbush,

I wish i knew... Chances are pretty high that the internal diode is already in the CPU, but the Marketing people may have nixed the ability to read it(ie removal of one pin or not revealing what voltages to apply to the two pins to read it).

HOpefully Palamino will let you read the diode, albeit i'm not very optimistic... I could just be a marketing thing, and they don't want people to see how much warmer their chips are running compared to p3s, but p4s run just as warm as 1gig t-birds, so I don't know for sure.


Mike

 

[jsbush]

Well I don`t really see why the average computer user would care about the heat it makes.

Anyways, thanks for the info.

 

[Mikewarrior2]

true... i'd guess 1% of computer users would care...



Mike

 

[smp]

AMD didn't put one in cause after the billions of transistors, they didn't have the space. P4's run as hot as tbirds? I thought that the P4 review at anandtech said that they still run cooler.. oh well.. not going that route anyways.

 

[Mikewarrior2]

Internal diodes take up virtually no space on a .18uM process. And its not like htey were running out of room on top of the pcb... . Its probably there for them to use when they bin their chips.

As far as the p4, a p4 1.5ghz maxes out at 58W, so around the same as a t-bird at 1.0ghz 1.8V.


Mike