Does increase in CPU Heat decrease CPU speed?

[hatimh]

If it does, what kind of figures are we talking?

-----------------------------------
The bigger I am, the harder they fall.

 

[ChipDeath]

Only on pentium 4 chips. They simply go slower and slower until they're no longer at a critical temperature.

Athlons, PIIIs etc, just crash when they get too hot to function properly (and earlier athlons on earlier mobos start smoking and blow up)

THG have a video somewhere of what happens when you remove the heatsink from the CPU while the system is running. the P4 system doesn't even crash - quite impressive stuff.

---
$hit Happens. I just wish it would happen to someone else for a change.

 

[c0d1f1ed]

No.

A sufficiently cooled Pentium 4 (standard cooler is good enough) will constantly run at its highest frequency. The only advantage of extra cooling is that you can potentially overclock your processor. This can gain you 10-30% in performance (or more in extreme cases), but will make it less reliable (increased chance of failure in a few years). If that's acceptable to you and you want to spend the money to keep it cooler, it's worth the try. But personally I find this a marginal performance increase that is hardly really noticable. I myself wait for every next generation of processors and invest in a completely new system...

 

[ChipDeath]

This can gain you 10-30% in performance

depends on the chip. 50% can be attainable from some of them. But on the whole you're right...

Also, just to clarify my earlier post, P4 chips don't start throttling down until they hit a certain critical temperature.. can't remember what it is though, but I think it's somewhat over 75*C

---
$hit Happens. I just wish it would happen to someone else for a change.

 

[jOK]

Mmm...

http://www.xbitlabs.com/articles/cpu/display/p4-temp.html

 

[Guest]

Well also the semiconductors are less effective when you raise the temperature. Im pretty sure a processor running at 10c will run marginally faster than at 50c but even then the difference is minimal and you probably couldnt tell.

 

[slvr_phoenix]

Well also the semiconductors are less effective when you raise the temperature. Im pretty sure a processor running at 10c will run marginally faster than at 50c but even then the difference is minimal and you probably couldnt tell.

Quite true. And as the etching process gets smaller and smaller this may also become more pronounced. (Though this won't really be a problem to worry about for years likely.) The real kicker though is not speed, but longevity. Hot electronics just don't live as long. Keep them cool and they'll last longer. So the <i>real</i> reason to be obsessed about cooling it so maximize the lifespan of your PC.

Admittedly though, what with advancements in components moving along at a fairly regular pace, most people would much prefer to replace their entire PC than to use the same one for ten years. So longevity is not usually nearly the concern in PCs as it is in other electronic devices.

<font color=purple><pre><A HREF="http://www.winamp.com" target="_new">Winamp<b><font color=blue>3</font color=blue></b></A> and freeform skins, the best thing since sliced llama loaf. (Now with more beef.)</pre><p></font color=purple>

 

[JP5]

You are not going to have thermal throttling on your P4 unless there is some kind of problem with your cooling.

 

[Guest]

Tottaly agree with slvr

 

[slvr_phoenix]

I think the funniest thing is that longevity matters a lot in devices like TVs, VCRs, DVD Players, etc. Of those kinds of devices is the new trend of setp-top boxes which are basically stripped-down midget-sized PCs. Yet many of these systems are in fact so HOT because of their effective processing power put into a fanless (or quiet low-air flow enclosure) that they really aren't likely to survive for nearly as long as even a fifty-buck VCR would.

I've even seen a horrible case where the <i>only</i> fan on the system was internally in a completely enclosed box. The box itself was the 'heatsink', except for the major flaw that the majority of the enclosure was padded with foam insulation to dampen the sound of the one fan and the hard drive in the device. The thing was basically a thermal battery because the heat just <i>couldn't</i> be transferred away from the air inside of the enclosure.

Slowly these really bad designs are getting better and the bright sparks that design these things are realizing that at least putting carefully-placed air vents (if not a low-rpm exhaust fan) into these devices is acceptable to consumers and greatly improves the longevity of the device. Still, there are a great many of these bastardized PCs that are used as replacements for other electronics devices that just aren't designed to last for years like the devices that they replace are.

'Tis ironic, no?

<font color=purple><pre><A HREF="http://www.winamp.com" target="_new">Winamp<b><font color=blue>3</font color=blue></b></A> and freeform skins, the best thing since sliced llama loaf. (Now with more beef.)</pre><p></font color=purple>

 

[hatimh]

Thanks for answering my question so extensively all of you.

It's really hot here and my chip is idling at 49 deg. I have a Globalwin CAK2-38 meant to be a decent cooler. The only problem is the damn thing is TOO noisy. So I did the "7V trick" on the fan so it doesn't make so much noise. But it is really hot weather here at the mo but it'll cool down later.

So far I haven't crashed at load it maxes at 54 deg. I'm running an ASUS A7N8X rev 1.2 but 198 fsb 11x so 2180 MHz 1.725 VCore.

The only reason I'm not going higher at the moment is because my PSU's 5v rail drops too low (in a game) when I increase the mult and VCore. BSOD

-----------------------------------
The bigger I am, the harder they fall.

 

[slvr_phoenix]

Maxing out at 54deg doesn't sound bad at all if that's <i>after</i> the 7V fan mod. Hell, as long as it's stable (and not a P4 being thermally throttled, which yours isn't) temps really don't matter.

That is, unless you want it to run for fifteen years, in which case you wouldn't be OCing in the first place, which brings us back to temps just not mattering so long as it's stable.

So enjoy. Get something quieter if you want, but it's possible that airflow through the case itself could help more than the CPU heatsink. What kind of case setup do you have?

<font color=purple><pre><A HREF="http://www.winamp.com" target="_new">Winamp<b><font color=blue>3</font color=blue></b></A> and freeform skins, the best thing since sliced llama loaf. (Now with more beef.)</pre><p></font color=purple>

 

[hatimh]

I certainly learnt something very important today. You know....with what you mentioned about the heat. Up till now I've placed so much emphasis on keeping cpu temps down thinking that's the main reason instability is caused.

I mean I KNOW for sure that well within 5 years time I'll need a new machine, chances are...in a couple years I'll ahve upgraded again. That's with most people I guess.

OK, about my case well I have 1 80mm lower case fan intake and and a HARDCORE (mains powered) 120mm upper case fan outtake. I think I'm gonna maybe have to cut up the front of my case coz I have another 120mm case fan and maybe I'll use that for the intake instead.

-----------------------------------
The bigger I am, the harder they fall.

 

[slvr_phoenix]

Your airflow sounds like it should be just fine. Aren't 120mm fans great? I can't wait for my second 120mm to be delivered.

But yeah, temperature <i>is</i> important, though mostly just because in the past AMD's CPUs <i>really</i> didn't take to overheating very well. At least now there's shut-off protection and on-die thermal diodes to monitor the temps, which helps a lot. (Though still not as elegant as Intel's P4 solution.)

But I've personally found that instaiblity itself is quite often just an insufficient power supply and/or a funky stick of RAM. It's rarely heat and if it is heat even then it's most often overheating of the video card or the RAM, not the CPU. Good airflow through the case does wonders for correcting any of these kinds of problems.

So as long as your processor isn't frying, your case has good airflow, and your power supply meets the demands of your hardware, then you're rocking.

Of course nothing will save you from bad drivers though. If you buy hardware from a company that hires monkeys to write their firmware then you're pretty screwed no matter what you do. Luckily those kinds of issues only come out whenever significant OS changes occur. (Such as when Microsoft introduced Windows 2000 and the hardware vendors were all going nuts trying to write drivers for it because the old Win9x ones weren't usable for Win2K.) Since WinXP is based on the same technology as Win2K, we've seen a pretty nice stretch of time now where only really bad companies have bad drivers for their hardware.

But anywho it sounds to me like you've found the line between stability and BSODs for your system and you're good to go. You could try to lower your temps, but it probably won't do much for you.

<font color=purple><pre><A HREF="http://www.winamp.com" target="_new">Winamp<b><font color=blue>3</font color=blue></b></A> and freeform skins, the best thing since sliced llama loaf. (Now with more beef.)</pre><p></font color=purple>

 

[Flinx]

Use the side of the Case for another intake.
Go have a look at an Antec case.


The loving are the daring!