why does a cpu get hot?

[zabomb4163]

Only P4's have a thermal shutdown at high temps. P2s and P3s will burn themselves out. All northwoods run at the same voltage no matter what the frequency is - just like AXPs.

NO


http://processorfinder.intel.com/sc...Fam=1026&PkgType=ALL&SysBusSpd=ALL&CorSpd=ALL
-Core Voltage 1.575

http://processorfinder.intel.com/sc...cFam=483&PkgType=ALL&SysBusSpd=ALL&CorSpd=ALL
-Core voltage 1.25V-1.400V

Not all p4s have the same voltage.

just like AXPs.

http://www.amd.com/gb-uk/assets/content_type/DownloadableAssets/Processor_Recognition_Rev05_ENG.pdf

-there are many different voltages for amd xps. i linked you to the chart that will teach you to understand amd cpu codes.

Only P4's have a thermal shutdown at high temps.

-so what exactly does this mean? (from amd64 tech docs)

http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/24659.PDF

3.7 THERMTRIP_L
The AMD Athlon 64 processor provides a hardware-enforced thermal protection mechanism. When
the processor’s die temperature exceeds a specified temperature, the processor is designed to stop its
internal clocks and assert the THERMTRIP_L output.

If the processor’s die temperature still exceeds the thermal trip point when RESET_L is deasserted,
THERMTRIP_L will immediately be reasserted and the processor’s internal clocks will be stopped.

 

[zabomb4163]

This is because of the heatsink attachment, not the chip itself. P4 heatsinks have a higher clamping pressure on the chip. Increased pressure makes the thermal interface layer thinner, and provides less resistance for heat transfer in P4s. If you look at the heatsink ratings identical heatsinks for socket 478 will always have lower c/w values than for socket A.

lets put numbers to it
----------------------------------------------------------------------------------
Intel

http://www.intel.com/design/iio/applnots/27398001.pdf
(intel spec for heatsinks and TIM)
"The heatsink clip provides enough pressure for the TIM to achieve a thermal conductivity of
0.2×C cm2/W"


http://www.honeywell.com/sites/docs/doc1c82bed-f9f96bd175-e0df9bfada07602278603c6cb43673fb.pdf
(spec sheet for company that makes intel heatsinks and TIM)

psi needed for 0.2×C cm2/W - 35psi

35psi*0.336 square inches = 12 pounds of downward force


----------------------------------------------------------------------------------
AMD

http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/21085.pdf
"clips provide approximately 10–15 pounds
of downward pressure ."

-----------------------------------------------------------------------------------
both companies have about the same pressure on the processors. no speculation about it

 

[rudnik68]

lets put numbers to it
----------------------------------------------------------------------------------
Intel

http://www.intel.com/design/iio/applnots/27398001.pdf
(intel spec for heatsinks and TIM)
"The heatsink clip provides enough pressure for the TIM to achieve a thermal conductivity of
0.2×C cm2/W"


http://www.honeywell.com/sites/docs/doc1c82bed-f9f96bd175-e0df9bfada07602278603c6cb43673fb.pdf
(spec sheet for company that makes intel heatsinks and TIM)

psi needed for 0.2×C cm2/W - 35psi

35psi*0.336 square inches = 12 pounds of downward force


----------------------------------------------------------------------------------
AMD

http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/21085.pdf
"clips provide approximately 10–15 pounds
of downward pressure ."

-----------------------------------------------------------------------------------
both companies have about the same pressure on the processors. no speculation about it

Your math is right (thats the first thing I checked). Except you took specs form an Intel 80331 BGA I/O processor that was first manufactured in January 2004.

This is the specs that you should used,. which gives the force of the heatsink retaining mechanism as 60lbs minimum and 75lbs nominal for P4 chassis.

All AXPs don't have the same voltage - you are right. Of the Intel chips, p2, P3 and P4 - only P4's have thermal shutdown. That was the context of the response, although not clear. A64s have a thermal shutdown.

All P4's don't use the same voltage. Northwoods either use 1.50V or 1.525V - I incorrectly believed that all Northwoods were 1.525V. One of your links to intel's site goes to the 3.4EE (not a northwood) the other goes to an odd 3.2GHz northwood with a listed core voltage of 1.25-1.4V. Another sheet links to the same processor and does not give core voltage. I believe the stock voltage for a 3.2GHz northwood is 1.525V. Has anyone seen a northwood with that stock core voltage?

I was misinformed and unclear. Any other misconceptions that I should clear up?

 

[zabomb4163]

One of your links to intel's site goes to the 3.4EE (not a northwood)

the EE is a pentium4 with a larger cache, otherwise they are identical.

 

[rudnik68]

the EE is a pentium4 with a larger cache, otherwise they are identical.

But not identical. Similar doesn't mean that the voltage is the identical. And technically, the EE is not a northwood - its a converted Gallatin Xeon processor. The stepping (M0) from the Intel page that you quoted should tell you that.

Larger issue - why doesn't AMD increase the clamping pressure of their heatsinks to better cool their chips? Does anyone know what the force applied by a A64 heatsink is?

 

[dippy_skoodlez]

But not identical. Similar doesn't mean that the voltage is the identical. And technically, the EE is not a northwood - its a converted Gallatin Xeon processor. The stepping (M0) from the Intel page that you quoted should tell you that.

Larger issue - why doesn't AMD increase the clamping pressure of their heatsinks to better cool their chips? Does anyone know what the force applied by a A64 heatsink is?

They probably would if it was a signifticant enough of a difference.... Maybe we need some tests?

 

[dumbfish]

and what temp to apple chips clock in at?