overclock damage?

[dmens]

Originally posted by: SX2012
overclocking increases the rate of electromigration along all of the circuits of a CPU. While an overclock may seem stable at first, usually they eventually become unstable sometimes even if you return to stock speeds. Whether its 6 days 6 months or 6 years its never predictable. Also Overvoltage can damage motherboard compontents as well.

besides the fact that there are many other issues besides EM these days, age degradation is predictable, the results are proprietary and not released to the public. just keep the voltage as close to VID as possible and there should be no long term damage, my personal opinion of course. and considering how cpu's are tested to far more rigorous worst case scenarios than even the regular stock cooler, there really is no reason not to take advantage of the extra speed.

 

[dmens]

dup

 

[themisfit610]

Well posted. If you're methodical and aware of what you're doing, you can get significant performance gains from overclocking without much risk to long term system stability.

Take the Q6600 for example. Just about every example of this chip can do 3 GHz (333x9) at, below, or slightly above stock voltage. This ONLY overclocks the CPU (assuming you have at least DDR2-667 RAM), and if you can keep temperatures under control, you're good to go!

Mine does 3 GHz at 1.20v, and loads at 62c after hours of video encoding at 100%. This is a bit warm, but my system is very quiet. I'm not worried at all.

I would agree that most people shouldn't overclock, but if you know what you're doing and you're careful, it can mean a lot of extra performance for close to no extra cost.

~MiSfit

 

[Modelworks]

I needed a pc that was cheap but capable.
I was able to get a gigabyte p35-d3SL + E4500 cpu+2GB ram for 245.00.
That system runs at 3300Mhz perfectly stable.

I would do it again in a second.

Overclocking is safe if you know what you are doing.
Like anything it takes experience.
I've been overclocking since the 486/33 mhz pcs.
Even doing extreme things like drilling out the pins on chips of celerons to enable smp/overclocking.

The first thing I always do on an overclocked pc is install the os and the basic drivers, then I torture test it for hours and hours.
If there are no problems, then I install for everyday use.

Overclocking on a pc setup your not sure of, and with your current running os is asking for trouble.

 

[themisfit610]

EXACTLY.

I think a lot of people don't understand that if you're going to OC, you need to get it all dialed in BEFORE you spend days getting your system all up and running with all your applications and settings. Just load Windows, drivers, and some stress testing / temp monitor software. Get your overclock 24hr stable, and then add applications.

If you're into it, make an image at that point. You should be good to go.

~Misfit

 

[Amaroque]

I don't overvolt my CPU's anymore, due to the fact that I've had too many CPU's "degrade" after only a couple years (one after 6mo) of 24/7 100% load operation. Now I go as high as the CPU will do with the stock Vcore only.

Is it worth getting a couple hundred MHz with a 10-20% increase in Vcore?



This is speculation on my part:

CPU's are MUCH more complicated, and built on MUCH smaller proccess then they were back in the day, and would seem to be subject to such damages such as EM migration a lot easier then they used to be.

 

[SX2012]

calculating overclocking power usage:

http://forums.guru3d.com/showthread.php?t=94726
"So this 36% overclock and 18% core voltage increase has resulted in a 88% wattage increase."

power leakage:

http://www.anandtech.com/showdoc.aspx?i=2343&p=3

The smaller the process the more power that leaks, the more you overclock, the wattage goes up and more of that power leaks

Non-Overclocked Power Consumption of Current CPU's:

http://www.techspot.com/review...reme-qx9650/page6.html

They Overclock to 4GHz but conveniently fail to post their findings! DAMN THEM

Anyways, i seriously doubt overclocking a 2.4Ghz CPU to 3.8 something(which means your playing with the vcore) GHz results in a 30 percent increase over stock draw. Maybe the device you are using to measure power consumption is not accurate.

 

[SX2012]

okay im a BIG exaggerator! but here as you can see going from 3.0Ghz to 4.0Ghz DOUBLES the power requirements

i was wrong to say 250 watts


http://www.tomshardware.com/20...ir_cooling/page14.html

 

[djspl]

Originally posted by: Modelworks
I needed a pc that was cheap but capable.
I was able to get a gigabyte p35-d3SL + E4500 cpu+2GB ram for 245.00.
That system runs at 3300Mhz perfectly stable.

I would do it again in a second.

Overclocking is safe if you know what you are doing.
Like anything it takes experience.
I've been overclocking since the 486/33 mhz pcs.
Even doing extreme things like drilling out the pins on chips of celerons to enable smp/overclocking.

The first thing I always do on an overclocked pc is install the os and the basic drivers, then I torture test it for hours and hours.
If there are no problems, then I install for everyday use.

Overclocking on a pc setup your not sure of, and with your current running os is asking for trouble.

My exact "new rig" method. Set it up, tweak it, find limits, test it and then re-install the OS after all basic hardware/software/drivers and settings have proven they can play nice with each other.

 

[Acanthus]

Originally posted by: SX2012
calculating overclocking power usage:

http://forums.guru3d.com/showthread.php?t=94726
"So this 36% overclock and 18% core voltage increase has resulted in a 88% wattage increase."

power leakage:

http://www.anandtech.com/showdoc.aspx?i=2343&p=3

The smaller the process the more power that leaks, the more you overclock, the wattage goes up and more of that power leaks

Non-Overclocked Power Consumption of Current CPU's:

http://www.techspot.com/review...reme-qx9650/page6.html

They Overclock to 4GHz but conveniently fail to post their findings! DAMN THEM

Anyways, i seriously doubt overclocking a 2.4Ghz CPU to 3.8 something(which means your playing with the vcore) GHz results in a 30 percent increase over stock draw. Maybe the device you are using to measure power consumption is not accurate.

It was measured from the wall.

Which means total system power consumption.

 

[docmilo]

Have you noticed any spikes in the 3.3v or 5v on PC Probe? I didn't see the specs on your psu. Before the prices of the x2's came down nearly two years ago I bought a 805 Pentium D to play with dual core. I over clocked it from 2.66 to 3.4 without a voltage increase. I still run the chip on that on my home server. After about 6 months the first psu started to crap out.

 

[dmens]

Originally posted by: SX2012
calculating overclocking power usage:

http://forums.guru3d.com/showthread.php?t=94726
"So this 36% overclock and 18% core voltage increase has resulted in a 88% wattage increase."

thats what he gets for raising vcore so ridiculously much. suicidal?

power leakage:

http://www.anandtech.com/showdoc.aspx?i=2343&p=3

The smaller the process the more power that leaks, the more you overclock, the wattage goes up and more of that power leaks

that is *almost* complete crap.

 

[SX2012]

Originally posted by: Acanthus

Originally posted by: SX2012
calculating overclocking power usage:

http://forums.guru3d.com/showthread.php?t=94726
"So this 36% overclock and 18% core voltage increase has resulted in a 88% wattage increase."

power leakage:

http://www.anandtech.com/showdoc.aspx?i=2343&p=3

The smaller the process the more power that leaks, the more you overclock, the wattage goes up and more of that power leaks

Non-Overclocked Power Consumption of Current CPU's:

http://www.techspot.com/review...reme-qx9650/page6.html

They Overclock to 4GHz but conveniently fail to post their findings! DAMN THEM

Anyways, i seriously doubt overclocking a 2.4Ghz CPU to 3.8 something(which means your playing with the vcore) GHz results in a 30 percent increase over stock draw. Maybe the device you are using to measure power consumption is not accurate.

It was measured from the wall.

Which means total system power consumption.

Did you ever find out how much more percent power your CPU uses? Just the CPU

 

[jaga]

Can't say I saw any spikes on the 3.3v or 5v channels. The PSU is an OCZ StealthXStream 600w. Should be strong enough?

 

[Acanthus]

Originally posted by: SX2012

Originally posted by: Acanthus

Originally posted by: SX2012
calculating overclocking power usage:

http://forums.guru3d.com/showthread.php?t=94726
"So this 36% overclock and 18% core voltage increase has resulted in a 88% wattage increase."

power leakage:

http://www.anandtech.com/showdoc.aspx?i=2343&p=3

The smaller the process the more power that leaks, the more you overclock, the wattage goes up and more of that power leaks

Non-Overclocked Power Consumption of Current CPU's:

http://www.techspot.com/review...reme-qx9650/page6.html

They Overclock to 4GHz but conveniently fail to post their findings! DAMN THEM

Anyways, i seriously doubt overclocking a 2.4Ghz CPU to 3.8 something(which means your playing with the vcore) GHz results in a 30 percent increase over stock draw. Maybe the device you are using to measure power consumption is not accurate.

It was measured from the wall.

Which means total system power consumption.

Did you ever find out how much more percent power your CPU uses? Just the CPU

Err for your argument about cost... the power from the wall is what matters.

 

[Martimus]

Electromigration happens on *ALL* CPUs, if you stay with the CPU manufacturers stated "safe range" (which is in their whitepapers) you will not significantly reduce the life of the CPU as long as you are also in the safe range for thermals.

I used to test components to determine their life, and I can tell you that increased current will quicken electromigration. Usually by a large margin. Increasing voltage will increase the current (ohms law), so your chip is more likely to fail the higher you push the voltages.

 

[SX2012]

Its not the only power from the wall that matters! (Pick whatever numbers you want)If your whole system uses 300 watts(hypothetically) under load with the CPU overclocked, and 210 when not overclocking the CPU. That means that the CPU is increasing your TOTAL draw by 30 percent like you said yours was.

However, the CPU is using probably using double the power it should be and only giving back half of that in performance because of all the leakage.

See this link here?

http://www.tomshardware.com/20...ir_cooling/page14.html

To get that extra 1 GHz a lot of power is lost because it leaks from the CPU. The smaller the process size the less insulation on the transistors

Going from 3GHz to 4GHz is a 33 percent boost in performance but it takes twice as much power to do it.

33% more speed 100% more power

I must admit though, the CPU scales power consumption nicely to 3.66GHz. 45nm is pretty impressive.

 

[myocardia]

Originally posted by: Martimus

Electromigration happens on *ALL* CPUs, if you stay with the CPU manufacturers stated "safe range" (which is in their whitepapers) you will not significantly reduce the life of the CPU as long as you are also in the safe range for thermals.

I used to test components to determine their life, and I can tell you that increased current will quicken electromigration. Usually by a large margin. Increasing voltage will increase the current (ohms law), so your chip is more likely to fail the higher you push the voltages.

And yet, as Acanthus said in the post you quoted, as long as it's within the manufacturers stated safe range, you'll long have gotten rid of the processor, before you start having problems. The real problem people have is trying to run bleeding-edge 24/7. If your processor/system can do 3.54 Ghz, you should not be trying to run it @ 3.54 Ghz 24/7.

 

[Amaroque]

Originally posted by: myocardia
The real problem people have is trying to run bleeding-edge 24/7. If your processor/system can do 3.54 Ghz, you should not be trying to run it @ 3.54 Ghz 24/7.

QFT That's why I only go as high as the CPU will with stock Vcore. Running my systems 24/7, and overvolting don't mix well.

If you turn it on for a few hours a day to browse, and game, then no problems (unless you have a crappy CPU).

 

[Acanthus]

Originally posted by: Martimus

Electromigration happens on *ALL* CPUs, if you stay with the CPU manufacturers stated "safe range" (which is in their whitepapers) you will not significantly reduce the life of the CPU as long as you are also in the safe range for thermals.

I used to test components to determine their life, and I can tell you that increased current will quicken electromigration. Usually by a large margin. Increasing voltage will increase the current (ohms law), so your chip is more likely to fail the higher you push the voltages.

How does what you say contradict my post?

 

[Acanthus]

Originally posted by: SX2012
Its not the only power from the wall that matters! (Pick whatever numbers you want)If your whole system uses 300 watts(hypothetically) under load with the CPU overclocked, and 210 when not overclocking the CPU. That means that the CPU is increasing your TOTAL draw by 30 percent like you said yours was.

However, the CPU is using probably using double the power it should be and only giving back half of that in performance because of all the leakage.

See this link here?

http://www.tomshardware.com/20...ir_cooling/page14.html

To get that extra 1 GHz a lot of power is lost because it leaks from the CPU. The smaller the process size the less insulation on the transistors

Going from 3GHz to 4GHz is a 33 percent boost in performance but it takes twice as much power to do it.

33% more speed 100% more power

I must admit though, the CPU scales power consumption nicely to 3.66GHz. 45nm is pretty impressive.

You were arguing the cost of overclocking, as in upgraded cooling and the additional COST of power from overclocking vs just buying a faster chip.

You implied that power consumption of the CPU is doubled, which you can do if you leave the safe range of the processor, but if you leave the safe range, its called "suicide benching" for a reason.

Electromigration is not a significant factor either if you stay in the safe range with safe temps and voltages.

You seem to be arguing for the sake of arguing and not presenting anything new to the thread. Shifting the subject to CPU power consumption seems pretty moot when the argument to begin with was about power cost.

Once again, im in for $535 for the cooling system and the CPU in the PC in my sig. With my whopping extra $3 a month on my power bill. Where can i pick that other 3.8ghz stock speed quad core for $600?