Optimum temperature for P4 Prescott?

[Bobby]

My new CPU is an Intel P4 (Prescott) running at 3.0GHz. The CPU temperature
is between 62-70C. Is that OK?

It's higher than I'm used to (usually AMD CPUs which run cooler than this).

Cheers.

Bobby

 

[DaveW]

70 C is the upper LIMIT for the safe operation of a Prescott CPU. Yours is
running too hot if those are the average temperatures.

 

[Spajky]

70 C is the upper LIMIT for the safe operation of a Prescott CPU. Yours is
running too hot if those are the average temperatures.

limit temps are ment of a IHS long term full load; he has /looks like
onDie temp reporting, which is approx. at least 10° higher than that
of IHS/, but anyway if those temps are correct reported, he needs
additional cooling. But maybe his MoBo is not reporting him correct
temps ...

 

[Ken]

My new CPU is an Intel P4 (Prescott) running at 3.0GHz.
The CPU temperature is between 62-70C. Is that OK?

No.

 

[YavolGurrick]

My new CPU is an Intel P4 (Prescott) running at 3.0GHz. The CPU
temperature is between 62-70C. Is that OK?

This is not good. What motherboard you using? Which heatsink/fan? Is the
fan actually spinning?

I've a P4 2.4c clocked to 3.2GHz and my maximum temp. at full load is around
the 55 degs. C mark. Idle, I'm looking at 38-42 (and that's with the fan
running
at only 60% of maximum). BTW, I'm using a stock Intel P4 cooler, here.

 

[Glitch]

My new CPU is an Intel P4 (Prescott) running at 3.0GHz. The CPU temperature
is between 62-70C. Is that OK?

No,I believe those temps are too high,and you should consider a new
heatsink/fan.I'd recommend the Zalman 7000Al-Cu if you're using sck.478.

 

[willard.myron]

Path:
news.demon.co.uk!mutlu.news.demon.net!peer-uk.news.demon.net!kibo.news.demon.net!demon!newsfeed.wirehub.nl!news.cambrium.nl!npeer.de.kpn-eurorings.net!fu-berlin.de!uni-berlin.de!not-for-mail
From: "Bobby" <bobby@europe.com>
Newsgroups: alt.comp.hardware,comp.sys.intel
Subject: Optimum temperature for P4 Prescott?
Date: Sat, 6 Nov 2004 23:27:22 -0000
Lines: 10
Message-ID: <2v54urF2gftraU1@uni-berlin.de>
X-Trace: news.uni-berlin.de
tdFkbK2pkzfqtY4MOR/FGwwke3Z71qfhGPjV7SBdbdRX4FOITM
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.2180
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.2180
X-RFC2646: Format=Flowed; Original
Xref: news.demon.co.uk alt.comp.hardware:84653 comp.sys.intel:18822

My new CPU is an Intel P4 (Prescott) running at 3.0GHz. The CPU
temperature
is between 62-70C. Is that OK?

It's higher than I'm used to (usually AMD CPUs which run cooler than
this).

Cheers.

Bobby

Are you using an LGA775 processor?. If so, did it come fitted with a grey
thermal pad?. If it did, you can reduce the CPU temperature by 10C just by
removing the pad and using a heatsink compound instead.

Consider using a different heatsink:
http://www.thermaltake.com/coolers/comboCool/cl-p0015tower102/cl-p0015tower102.htm#logoInfo

Willard.

 

[WebWalker]

My new CPU is an Intel P4 (Prescott) running at 3.0GHz. The CPU temperature
is between 62-70C. Is that OK?

It's higher than I'm used to (usually AMD CPUs which run cooler than this).

What motherboard are you using?
Try to flash the BIOS for the latest update and see whether its helps
(the update help to correct wrong temperature reading).
If no, then you need a better after market heatsink plus fan.

 

[Bobby]

I bought the Intel Pentium Prescott kit which included the CPU (3GHz),
cooler and fan.

The cooler had a solid paste base which sat on the CPU. But it wasn't very
big or very thick. When I saw it I thought: "That's not much paste to cool
the CPU". And I was right. It wasn't. The CPU runs very hot.

I have now replaced my Pentium pack with an Athlon 64 bit pack which runs at
37-44C - that 24C cooler than the Prescott.

Prescott heating seems to be a known issue:

http://207.44.152.242/index.html

Bobby

 

[JK]

The power consumption statistics for different processors are very interesting.
Notice that the Pentium 4 550 (3.4 ghz) consumed around 40% more power idle and around 60% more power under load than the Athlon 64 3500+ 90nm system.

http://www.tech-report.com/reviews/2004q4/p4-overclock/index.x?pg=10

 

[Ronald]

My new CPU is an Intel P4 (Prescott) running at 3.0GHz. The CPU
temperature is between 62-70C. Is that OK?

It's higher than I'm used to (usually AMD CPUs which run cooler than
this).

Way too high for normal use. Im burning in my Pentium 4E 3.0 Ghz with HT
activated by running two instances of SETI@home, which will use up the full
CPU power, all of the time. Running temp is 62C constantly, at full load for
days at a stretch. System temperature is 26C.
This is a seperately soled, boxed CPU with standard cooler. No overclocking.
Normal running temperatures should be well below 50C.

Regards Ronald

 

[Noozer]

Way too high for normal use. Im burning in my Pentium 4E 3.0 Ghz with HT
activated by running two instances of SETI@home, which will use up the full
CPU power, all of the time. Running temp is 62C constantly, at full load for
days at a stretch. System temperature is 26C.
This is a seperately soled, boxed CPU with standard cooler. No overclocking.
Normal running temperatures should be well below 50C.

Not for a Prescott CPU... Those are crap and I'd avoid them if at all
possible. Looks like AMD is the clear winner these days for most things.
Really ticks me off because I just moved to Intel last year too. : (

I think that Intel is trying to get into the coffe maker business with their
CPU's...

 

[Ronald]

Not for a Prescott CPU... Those are crap and I'd avoid them if at all
possible. Looks like AMD is the clear winner these days for most
things. Really ticks me off because I just moved to Intel last year
too. : (

Mine are actual readings. Pleases substanciate your 'crap' opinion.

 

[kony]

Mine are actual readings. Pleases substanciate your 'crap' opinion.

Crap opinions aside, "normal" readings can easily vary by
20C due to heatsink finish irregularities, differences in
case airflow, ambient room temp, other parts heating up case
air, and even minor differences in motherboard vcore
level... and we haven't even considered any method of
determining a baseline for what "full" or "idle" load is,
since neither is ever at 100% / 0% utilization.

3+ GHz Prescott can be expected to be in the 50-65C range
with anything but a very good cooler and mild external
forces playing against it.

 

[Guus Ellenkamp]

2.8C I assume is NOT a Prescott. Prescott gets hot. I have similar readings
with 50% load on a 2.8 system.

 

[Guus Ellenkamp]

I have similar readings under 50% and up load. I put extra fans and put the
system in aircon room, but this didn't make it better, sometimes worse. My
experience is the CPU heatsink fan adapts itself to the MB and case
temperature. Only option I heard of is buying better heatsink to lower
temperature.

 

[James Boswell]

2.8C I assume is NOT a Prescott. Prescott gets hot. I have similar
readings with 50% load on a 2.8 system.

Aye, a P4C is a Northwood with Hyperthreading and an 800Mhz FSB


-JB

 

[Dorothy Bradbury]

A hot CPU can be caused by 3 things:
o Insufficient CPU cooler
---- although the Intel supplied cooler is adequate
o Insufficient case airflow
---- which is a common problem with the hotter CPUs
o Incorrectly seated heatsink
---- Intel's physical fitment is good - better than early Athlon clip/broken-die
---- however it IS possible to miss-align the heatsink during fitment = hot CPU

First - check the heatsink is properly seated.

Second - if the heatsink has been reseated, the TIM pad may need replacing:
o Thermal Interface Material, TIM, are generally a one-use compound
o Performance is good - TIM have come a long way from the early offerings
o However it can be beaten - good heatsink compound is available, cheaply

If you use a liquid/paste type compound, remember less is better:
o Two metal surfaces only truly contact at 3 points
---- the rest of the interface is full of microscopic voids
o Those microscopic voids are filled with air
---- air is an insulator - it does not conduct heat well
o Thermal compound is simply a better conductor of heat than air
---- it is not a better conductor of heat than metal

So the objective with thermal compound is enough to fill in the microscopic
voids in the metal surfaces - and not spread like butter on a toast, spread thin.

Assuming the heatsink is seated correctly, a CPU temp of 62-70oC is too hot:
o It could be the temperature is being miss-reported
---- however even with a 10% error depreciating the 70oC figure, it is still high
o It is more likely the case temperature is too high
---- I suspect your case temperature is being reported as 40oC

If your case temperature is being reported as 25-30oC, whilst CPU temperature
is being reported as 62-70oC that indicates a reporting or heatsink fitment error.

However, assuming a case temperature of ~40oC...

Case temperature is important with Prescott CPUs:
o Northwood CPUs had a low idle thermal output
---- sat idle they emit ~30-35W, fully loaded they emit 75-89W
o Prescott CPUs have a higher idle thermal output
---- sat idle they emit ~50-55W, fully loaded they emit 89-110W

Case temperature of 40oC is too hot
o CPU-coolers use case-ambient airflow to cool the CPU
o The lower case-ambient, the greater the temperature delta between case & CPU
o A high case-ambient results in the CPU cooler losing efficiency = hotter CPU

Hence Intel refer to an "approved chassis" in relation to Prescott CPUs, and the
idea of BTX is to get case-ambient (as seen by the CPU) closer to room-ambient.

Elevated case temperatures will elevate the CPU temperature:
o CPU coolers typically recirculate 40-70% of their own heated air
---- in doing so their airflow cfm must be increased to compensate
o Case exhaust fans should exhaust air quicker than the CPU cooler heats it
---- the ideal is to get case-temperature as close to room temperature as possible
---- a case temperature of 40oC in a room ambient of 25oC will mean a hot CPU

Airflow through a case can be optimised - from exhaust to intake:
o Exhaust fans must overcome resistance of fingerguards
---- punched case-metal grills can flow just 45-55% air
---- round wire grills flow 89-91% air & make less noise = cooler CPU
o Exhaust fans should be sited correctly
---- fan exhaust ports should be prioritised near to the CPU & at the top
---- exhaust fan behind a CPU reduces CPU cooler airflow recirculation = cooler CPU
o Exhaust fans should not be short-circuited
---- correct airflow path in a PC is exhaust at top-rear, intake at front-bottom
---- an empty rear fan port or missing drive bay cover can short-circuit airflow
---- short-circuited airflow will increase CPU cooler airflow recirculation = hotter CPU
o Intake fans are best at cooling HDs sited by them
---- an intake fan acts serial to an exhaust fan - thus is not additive in airflow cfm
---- however, serial fans are additive in pressure - so do assist an exhaust fan
o Intake fans, indeed the whole case, must breathe through the smallest intake port
---- typically this is viewed as the intake punched-holes in the case
-------- this is partly correct, and often such metal grills flow just 45-55% air
-------- the intake of a fan is also more sensitive to obstruction regarding noise
---- however more often the smallest intake port is in the front plastic fascia
-------- intake cutouts of 3"x1" are less than half the c/sectional area of 1x 80mm fan

So high case temperatures, if present, will produce a high CPU temperature.
o The solution is to address case (system) level cooling first
o This may be a matter of blocking off an empty exhaust fan port - or adding a fan
o One variable at a time should be changed - not throwing money/fans at the solution

If, conversely, case temperatures are relatively low, and CPU temperature is high
o The solution is the CPU cooler is insufficient in airflow for the application
o Since this is an Intel cooler, I would tend to suspect case airflow & case
temperatures

You can have a situation where case temperature is reported as (relatively) low,
due to the sensor area receiving airflow but the CPU cooler is recirculating its air.
An example would be older ATA cables obstructing the CPU cooler intake path.

Prescott CPUs will run warmer than a Northwood - at least in an identical system.
Cooling a Prescott CPU is not particularly difficult, if you build a PC yourself then
you are assumed to be the System Integrator. That gives you flexibility, but it also
requires you to do some simple benchmarking & correction if temps are not suitable.

So...
o Check your case temperatures
o If they are 40oC you have "insufficient case airflow changes per hour"
o The solution is to improve airflow - via another fan or correcting an airflow problem

Since an Intel cooler is in use I'd be inclined to suspect case temps are too high.

 

[Eddie]

A hot CPU can be caused by 3 things:

So...
o Check your case temperatures
o If they are 40oC you have "insufficient case airflow changes per hour"
o The solution is to improve airflow - via another fan or correcting an airflow problem

Since an Intel cooler is in use I'd be inclined to suspect case temps are too high.

Fantastic, clear comment. Ta Dorothy. Just one question re fans. (prescott
P4. 3.0E) I have only the power supply fan and an extra exhaust at the rear
of the case. If I put in an identical fan to that in the rear, at the front
of the case, but reverse its face so it draws air in, will this improve
circulation for the CPU?

Eddie

 

[DevilsPGD]

In message
<41970c2c$0$25776$5a62ac22@per-qv1-newsreader-01.iinet.net.au> "Eddie"

Fantastic, clear comment. Ta Dorothy. Just one question re fans. (prescott
P4. 3.0E) I have only the power supply fan and an extra exhaust at the rear
of the case. If I put in an identical fan to that in the rear, at the front
of the case, but reverse its face so it draws air in, will this improve
circulation for the CPU?

Short answer: Probably.