You’re saying that heat and temperature has no effect on the lifespan of a CPU? So if you were to remove the heatsink and fan from your CPU, that would have no adverse effect on the life of your CPU?
No offense, but you really have no clue about what you’re saying. I don’t think your processor would last more than 5 minutes without the heatsink and fan working to dissipate the tremendous heat away from it.
So you’re saying that Ohm’s Law does not apply? There are going to be a lot of very disappointed EE’s out there when this gets out.
Physics is nothing like religion. If it was, we’d have an easier time trying to raise money! /.
I have to call BS on the CPU life not being affected by overclocking unless you change the voltage! I have an engineering sample P4 2.4a northwood, C1 stepping. My CPU is starting to die, it was OK @ 175x17 for a long time with my old ram. Then I got some new ram, and I got better performance using 2-2-2-5 timings @ 165 MHZ bus and the multiplier @ 18. Then a few weeks ago I started getting blue screens, and things started acting funny. My first thought is the ram or the mobo is puking, because I have them pushed close to their limits. I thought for sure it was memory errors, so I ran memtest, nothing came up. Then I ran prime 95, which used to be 24+ hour stable on this beast, and it puked in 15 minutes!!! So I backed the multiplier down to 17 and tried it again, and it passed for an hour until I stopped it. Then I decided if it was going to be that way, I might as well pump up the bus speed to get what I can out of it. Maybe it will die, and I can build a dual core A64 machine. So I’ve got it stable again @ 2.89 ghz, 170x17, which SUX BIG HAIRY DONKEY BALLS!!! I’ve got the worst luck with overclocking, I get hosed on every thing. I only tried a little extra core voltage when I was trying to find the limit of this CPU, never went over 1.65v. I have great cooling, a swiftech MCX4000 heatsink with a temp sensitive fan on it. It never saw over 50c the whole time I was tinkering, and as soon as I figured out it didn’t like voltage I was done. The core voltage has been stock ever since, and that was 2+ years ago. I did have it clocked less than 5 MHZ from it’s absolute max clock speed it would do prime 95 stable for it’s entire life. Now it’s puking, and I really don’t want it to die!!! :Z 
:Z :a
Who’s talking about a processor without heatsink/fan?
No offense, but you really have no clue about what I’m talking about.
As long as the CPU is running way BELOW it’s thermal limit, temperature has no effect.
If it does, then the manufacturer needs to lower the thermal limit to the appropriate number.
Running a processor at 55C vs another at 45C shouldn’t have a noticable difference in lifespan.
By your logic, then that means Prescott CPU’s have 1/6th-1/8th the lifespan of Northwood P4’s.
So you’re trying to tell me people who’s processors at over -100C using liquid Nitrogen/prometia/vaporchill setup lives 16,384 times longer than the same CPU running at 40C?
Assuming a CPU’s lifespan running at 40C is 20 years, then thats as good as saying their processor will live 327,680 years? I doubt it.
Heh…Voltage is part of Ohms law.
And since when did temperature become part of ohm’s law?
That’s your experience.
My PII 266MHz CPU that came with Compaq in 1998 is overclocked to 333MHz and still running fine today @ stock voltage.
An engineering sample is just that, a sample.
Watts is the result of voltage times current. Watts is expressed as a function of BTU’s, or heat. Temperature is a representation of heat. If I went too fast for you let me know.
Strawman fallacy: Attacking a misrepresentation of your opponents stated position
The failure modes of solid state components are complex, other than those which result in immediate damage (eg. excessive overvoltage, overtemperature).
Electromigration is one effect that increases with both applied power and temperature.
Software coolers are another controversial aspect - for a lightly loaded CPU that takes well to HLT-Cooling (like an old Cyrix I had), software cooling can reduce the idling power of the CPU to a very low value indeed.
It does NOT help a CPU that’s at full load, unless it actually reduces the performance (eg. thermal throttling).
I’ve been playing with RMCLOCK on a P4 600 series
http://cpu.rightmark.org/products/rmclock.shtml
NB. This is NOT a HLT-cooler, though it can apply HLT threads to resolve some other issues, NT-class operating systems should use HLT or the appropriate ACPI state as needed.
What this allows , is tighter control of the power states than Windows offers (Windows should be set to “Always On” power profile, if you manage the power this way)
Athlon 64s incorporate Cool and Quiet which is an effective means of throttling the cpu voltage and frequency based on load. i’ve tried it and it works well, i don’t use it because i overclock and run Folding@Home all the time which negates it anyway.
I agree with your points. Failure mode analysis is an extremely complex subject and something best left to the chip designers. What a lot of people fail to realize about overclocking is that you don’t receive those extra CPU cycles for free, Additional power (current) is required if the stock voltage is retained and that results in additional heat generated, with the end result being premature failure if the additional heat is not effectively removed. I have no doubt that the short term effects are neglegible but feel that the long term effects are still to be determined. There have been a lot of numbers thrown around on this thread and I haven’t seen any basis for their use, other than a quote from a web page proclaiming the beneficial effects of a software program to reduce CPU utilization when not required. One thing is for certain and that is heat is the enemy of electronic components and any steps taken to reduce it will result in a longer service life for your system.
May we strive to be the people our dogs think we are /.
Well, they still have the same core as any other CPU, they just give you unlocked multipliers. Your P2 was overclocked well under what the core was capable of. If you had a p2 @ 600 mhz, it would be a different story. I had this close to it’s max for it’s entire life. So I’d say it can and does have an effect on the life of the CPU regardless of the temperature. It also depends on your luck, some people get away with overclocking forever. So if you didn’t have your cpu die on you, doesn’t mean that it can’t or won’t happen. It’s not like there’s anything special about the core of my chip, I just have the option to change the multiplier. Good for you, your P2 might live forever, but it’s not a P4. Processors can and do die from overclocking, just because it didn’t happen to you, doesn’t mean they don’t. 
Hehe… Guess there’s no need to try the program then. Don’t want reduced performance for only a 5C loss of heat.
I feel sorry for those who buy highend prescott P4 3.6GHz CPU’s or higher and don’t know that more than half the time is on thermal throttling at 3.0/3.2GHz.
I don’t leave my comp on 168hrs/week for it to idle while not being used.
If my comp is on, it means I’m using it. 
(I feel the same way about screensavers as well…I don’t use them)
I totally agree 
Throttling only occurs when a temperature threshold has been reached or is rapidly approached during operation. Adequate cooling solutions will eliminate or reduce the throttling effect. An excellent review of the process for your review.Link
“By the way if anyone here is in advertising or marketing… kill yourself.” /.
Luck? well maybe. Each chip is different and they don’t have the same overclocking potential.
I wouldn’t attribute all the systems I’ve overclocked to luck though.
I’d attribute my experiences to good airflow, decent case, thermal paste, etc…more than luck or anything else.
Good for you, your P2 might live forever, but it’s not a P4. Processors can and do die from overclocking, just because it didn’t happen to you, doesn’t mean they don’t.
See my sig.
My P4 2.4C GHz is overclocked pretty close to it’s limit. 3.5GHz is pretty much the limit on air, and thats only if you have a decent case with good airflow/quality thermal paste.
My Barton 2500+ is pretty much at it’s limit on air. I think you’ll find that 2.2GHz is pretty much the limit that 90-95% of overclockers reach on air for this CPU.
Both systems have been overclocked since Spring 2003 and are running perfectly fine. I have no doubt that either systems can still live for at least 5 more years in their current state.
No short term effects has been noticed, and longterm effects are still to be determined.
Oh and I have no intent of using liquid nitrogen or water cooling setup.
I have to agree that there are requirements to overclock with some level of success and airflow is at the top. I haven’t seen any research that proves the advantage of exotic thermal transfer compounds (Artic Silver etc.) but a quality HSF rather than the stock unit is advised. I see that you use the stock Intel HSF and have been successful with it but my Prescott’s required aftermarket solutions without overclocking.
Physics and politics agree: Hot air rises. /.
Interesting review to say at least.
Too bad Dell dosen’t provide adequate cooling solutions for the average joe running Prescott.
If a 3.2E prescott on full load is running 4C off the throttling value, I can only imagine what a 3.6/3.8E would be doing.
Average joe won’t even know that his comp is throttling.
Stock HSF was good enough on Northwood processors…for mine at least. The same can’t be said for Prescott processors.
The P4 2.4 Northwood all came with the same exact coolers as the 3.0/3.2 Northwood chips.
The overclocking potential was there.
50C on full load for Northwood vs 76C full load prescott running stock HSF at 3.2GHz is a big difference.
People who buy Dell are not concerned or possibly aware at what temperature their computer is running. Hard to tell if your computer is throttling when you’re web surfing, reading e-mail or listening to Brittany Spears.
Oh my GOD!!! There’s a message in my Alpha-bits. It says “OOOO.” /.
My 3.2E and 3.0E both run at around 56 C + - with Thermalright XP-90’s at full load. Much better than the 70 C+ with the stock HSF
Support bacteria! It’s the only culture some people have. /.