Underclocking = Unsafe, depending on voltage

[lcaley]

Thanks lol, I thought so too when I heard it.

 

[iluvgillgill]

evilonigiri how many people does it to take to to agree with me that running stock voltage but underclock will NOT cause any harm to any part of the computer system?its been 2 thread from me and you and you still not agree with not only me.but every single one in here.

say to yourself isnt what i told you is the same as what the other guys is saying?huh?

or do you want to ring up Intel and every motherboard maker to comfirm this?

 

[Shadow703793]

evilonigiri how many people does it to take to to agree with me that running stock voltage but underclock will NOT cause any harm to any part of the computer system?its been 2 thread from me and you and you still not agree with not only me.but every single one in here.

say to yourself isnt what i told you is the same as what the other guys is saying?huh?

or do you want to ring up Intel and every motherboard maker to comfirm this?

Hmm... now that's an idea. 😀.

 

[redzoneos]

lol I'm gonna have to jump on the "disagree with evil" band wagon as well...

The overall logic as explained thus far is flawed...

Overclocking and Overvolting require an addition of some sort, so once you exceed that threshold bad things happen. I'll use a similar analogy as previously mentioned, but one I just randomly created... hopefully it will make sense!

--- Think of a 12 ounce coffee mug... You can pour 12 ounces of blazing hot coffee into this mug... adding extra (overclocking/overvolting) will result in spillage which wastes coffee, and you can potentially get burned (damage to cpu and components)

Underclocking and undervolting require that you subtract from the base requirements.

--- Same 12 ounce coffee mug... You can poor 12 ounces of hot coffee... however since it's 8PM and you don't want to stay up all night, you decide to only poor yourself 6 ounces... there's no spills, no personal injuries... the only result is that you "wasted" the additional 6 ounces of space in that coffee mug...

anyway, the latter explanation demonstrates the following: if you underclock and keep the default voltages... you are wasting energy, nothing more. If you underclock AND undervolt, you are being conservative, and saving energy as well... but at the cost of wasting "space" since the system is clearly capable of more.

 

[Evilonigiri]

I was expecting that no one would be agreeing with me, after all I'm basing this off my knowledge and nothing else.

Anyways, I'll attack this from a different point of view: Underclocking loads is out of Intel's spec, just like Ocing is. I would assume out of spec settings are "dangerous" and thus voids warranty, so underclocking is not entirely safe.

We should ring up Intel and see, really. Or we can debate this forever. After all, I have to maintain my "Master-de-bater" title.

 

[redzoneos]

no it still doesn't imply "dangerous". Intel's undervoltage specs no doubt reflect the minimum requirement for the CPU to operate, at all. If you go below their minimum, the cpu simply will not function.

This is no different then saying that a 60W lightbulb installed in a lamp that provides only 30W will be damaged... not true... it will simply be less bright, or not turn on at all.

 

[Evilonigiri]

This is no different then saying that a 60W lightbulb installed in a lamp that provides only 30W will be damaged... not true... it will simply be less bright, or not turn on at all.

Actually, that's dangerous. There's a risk of catching fire doing that. There is a reason why the manufacturers of a lamp company places a sticker that reads "Warning: To reduce the risk of fire, use 60W or smaller,120v, type A bulb" even though it can do 100W. So putting a 60W bulb on a lamp that provides 30W isn't a good idea and probably quite dangerous.

Anyways, I'll ask this one final question: Is a default Q6600 being harmed more at 1MHz @1.35V or at it's stock specs (2.4GHz @1.35V)? I don't care if the "damage" is infinitesimal or not.

 

[iluvgillgill]

you are talking extreme now.no motherboard will allow you to run any cpu at 1mhz.the lowest is 100mhz X6.and even run at 1.35v it wouldnt be harmful because an SO MANY people has pointed out its still running on manufacture default voltage which is a safe voltage.and you will just end up waste energy and produce unneccessory heat.

 

[4745454b]

1.35v = 1.35v. Again, AFAIK, the frequency doesn't matter. What you need to be worried about with voltage is extra heat, and speeding up electron migration. If you go to low on the voltage, the transistors won't have enough power to turn on or off. (someone posted something about .85v being the lowest "approved" voltage for modern Intel CPUs.)

 

[Evilonigiri]

you are talking extreme now.no motherboard will allow you to run any cpu at 1mhz.the lowest is 100mhz X6.and even run at 1.35v it wouldnt be harmful because an SO MANY people has pointed out its still running on manufacture default voltage which is a safe voltage.and you will just end up waste energy and produce unneccessory heat.

Just because everyone thinks that way doesn't mean it's right. I'm looking for the right answers, not what everybody thinks. Of course there's the possibility that everybody might be right, so I'll stand neutral on this from now on until I find more proof on this matter.

End of debate. For now.

 

[4745454b]

You still haven't told us why you think this way. If you could put forth a rational reason as to why this might be the case, we would have a better understanding of where your coming from. I've asked this before, but you still haven't answered the question.

Did you notice my last post? 1.35v = 1.35v. Do you see the significance of that statement? (hint: there is no frequency listed in the formula.) To little voltage and the chip won't have enough power to flip the transistors. To much power and the heat gets to high and the gates on the transistors will suffer from electron migration. Again, I never once mentioned frequency "explaining" this. Voltage is one thing, frequency is something else.

 

[Evilonigiri]

You still haven't told us why you think this way. If you could put forth a rational reason as to why this might be the case, we would have a better understanding of where your coming from. I've asked this before, but you still haven't answered the question.

I told you before didn't I? I said overvolting is a bad idea. And frequency IS involved with the voltage. That's how my thinking goes. Just like how higher frequency requires higher voltage, lower frequency needs lower voltage.

Did you notice my last post? 1.35v = 1.35v. Do you see the significance of that statement? (hint: there is no frequency listed in the formula.) To little voltage and the chip won't have enough power to flip the transistors. To much power and the heat gets to high and the gates on the transistors will suffer from electron migration. Again, I never once mentioned frequency "explaining" this. Voltage is one thing, frequency is something else.

Yes I did read your post. I also understand what you and all the others against me is thinking. It goes like this: The cpu needs at least (let's say) 1V to function. Any lower and it won't work properly. The cpu is also rated up to 1.5V. Any higher is considered damaging. It doesn't matter what the speed is or isn't. Thus I went from my point of view to being neutral, until I can find more evidence for or against myself.

 

[4745454b]

You said:

My thinking is overvolting something will do more damage than not overvolting. Surely overvolting will do more harm then good?

That has nothing to do with undervolting.

I said overvolting is a bad idea. And frequency IS involved with the voltage. That's how my thinking goes.

Correct, you said overvolting is a bad idea, and I'll doubt you'll find anyone who disagrees. (there might be some that do, shave 5 years off of a 10 year CPU, and who would want to run a 5yo CPU? If you take time into consideration, overvolting might make sense...)

Although you refuse to say it, I think I know where the fault in your thinking is. Many people up the voltage to their CPU as they hit higher and higher frequencies. This can help stabilize a CPU if its right at the edge. This isn't the case of more frequency = more voltage, so less frequency = less voltage. I'm not 100% sure why upping the voltage helps stabilize a CPU, although I'm sure it has something to do with helping the transistors switch on and off fast enough at the increased clock speed. Remember that a transistor on the 65nm process can switch (at a lower voltage to) faster then a transistor on the 90nm process. If you want a 90nm transistor to switch as fast as a 65nm transistor, you'd probably have to give it more juice, though why that works I don't know.

If you lower the speed of the chip, you don't need them to switch as fast. Because of this, you can decrease the voltage of the chip seeing as the transistors don't need the "extra" juice. You don't have to however. Simply because you lowered the voltage of the chip doesn't mean the transistors gates are thinner. They are still as wide/thick as they've always been.

 

[Evilonigiri]

Although you refuse to say it, I think I know where the fault in your thinking is. Many people up the voltage to their CPU as they hit higher and higher frequencies. This can help stabilize a CPU if its right at the edge. This isn't the case of more frequency = more voltage, so less frequency = less voltage. I'm not 100% sure why upping the voltage helps stabilize a CPU, although I'm sure it has something to do with helping the transistors switch on and off fast enough at the increased clock speed. Remember that a transistor on the 65nm process can switch (at a lower voltage to) faster then a transistor on the 90nm process. If you want a 90nm transistor to switch as fast as a 65nm transistor, you'd probably have to give it more juice, though why that works I don't know.

I refused to say it? I thought I expressed myself clearly in the examples I gave before. I guess it wasn't clear enough, my apologies.

That is true. My main argument was more voltage than necessary is more damaging than the same voltage with what it requires, no matter how little this damage is. The cpu gets all the current regardless of whether it needs it or not, and so when giving it more than it needs, it is "damaging", even though it can handle it perfectly fine.

If you lower the speed of the chip, you don't need them to switch as fast. Because of this, you can decrease the voltage of the chip seeing as the transistors don't need the "extra" juice. You don't have to however. Simply because you lowered the voltage of the chip doesn't mean the transistors gates are thinner. They are still as wide/thick as they've always been.

I suppose you are right. And so a 1MHz cpu at 1.3V will have as much "damage" as the same cpu clocked at 4GHz and 1.3V? This I can not comprehend, perhaps due to my lack of knowledge.

Because of my ignorance, I'm withdrawing until I know more. I've been using my common sense (If you can call it that) and experience, along with some "different" thinking and induce what I am arguing about. Can't debate fully without the knowledge you know.

 

[4745454b]

LOL, trust me, the knowledge here isn't that much... (not compared to JJ or some others.)

I and the others COULD be wrong, but I don't see how. Intel said the chip was fine at 1.35v. I'm not going to argue with them. With what I know of CPU arch, running 1.35v is fine, it doesn't matter at what frequency you are talking about.

Try to find more info about this. Perhaps Intel has a nice white paper about frequency and speed. Maybe you can show us the true light on this subject. I can't speak for the others, but I'm always willing to learn something new.