Jump to content

Linus, on the topic of TDP

@LinusTech , on the livestream you said there is "somewhat of a correlation" between the power a chip consumes and the heat it puts out. This is quite an understatement. In silicon chips, the power a chip consumes equals the heat it puts out.

If you've had physics in school, you should know that you can not create or destroy energy, energy can only be transformed. This means that all the electrical energy (or power, if you look at it on timely basis) a chip consumes, must go somewhere, be transformed into some sort of a different energy. Seeing that there are no moving parts inside a silicon chip, the only way a chip can consume electrical energy is by wasting it as heat. So in a chip, electrical energy is transformed into heat. This means that the electricity a chip consumes is equal to the heat it puts out.

As simple as that. It doesn't matter to what you're referring with TDP (heat or consumed electricity), both of the ratings are exactly the same.

On the topic of the power consumption curve related to overclocking: it is only when you start upping the voltage that the curve becomes quadratic. As long as you only change the frequency and don't touch voltage delivered to the chip, its power draw (and therefor heat output) is linear.

Hope this clears up some of the confusion.

Link to comment
Share on other sites

Link to post
Share on other sites

This is all well and true, but the TDP doesn't necessarily refer to the potential power draw from a chip, but rather how much heat the cooling system should be designed to handle in normal operations. So while running normally (under normal stress) the power draw should average at about TDP, but it will probably rise above TDP in shorter bursts. Stressing the CPU in prime95 will most likely bring it above TDP

I bet most of the confusion about TDP lies within what it actually is supposed to be. Max power draw/output is something else

Link to comment
Share on other sites

Link to post
Share on other sites

Stressing the CPU in prime95 will most likely bring it above TDP

That's true, but I do not see what that has anything to do with what I've wrote. TDP is indeed a term that says something about the normal heat output of a chip. A chip can indeed go beyond its TDP, even at stock clock and voltage. The only thing I wanted to clarify is that while TDP is saying something about heat, it is  equally saying something about the power draw. I wanted to clarify that there is no difference between consumed power and wasted heat.

Link to comment
Share on other sites

Link to post
Share on other sites

That's true, but I do not see what that has anything to do with what I've wrote. TDP is indeed a term that says something about the normal heat output of a chip. A chip can indeed go beyond its TDP, even at stock clock and voltage. The only thing I wanted to clarify is that while TDP is saying something about heat, it is  equally saying something about the power draw. I wanted to clarify that there is no difference between consumed power and wasted heat.

What it had to do with what you wrote was that you were trying to clear up some confusion about a subject, and I supplied some extra information about a closely related subject where it most likely is even more confusion. Of course, if you intended the thread to be a monologue where you talk about the most basic of thermodynamics, I can see why what I wrote would bother you :-p

Link to comment
Share on other sites

Link to post
Share on other sites

What it had to do with what you wrote was that you were trying to clear up some confusion about a subject, and I supplied some extra information about a closely related subject where it most likely is even more confusion. Of course, if you intended the thread to be a monologue where you talk about the most basic of thermodynamics, I can see why what I wrote would bother you :-p

It doesn't bother me at all ;) I'm on this forum primarily to learn. Sorry if I seemed bothered by your post, I'm not even a little bit.

Link to comment
Share on other sites

Link to post
Share on other sites

It doesn't bother me at all ;) I'm on this forum primarily to learn. Sorry if I seemed bothered by your post, I'm not even a little bit.

It's all cool! ;-)

Link to comment
Share on other sites

Link to post
Share on other sites

I've heard Linus discuss essentially this exact point in one of his videos, though I can't remember which, so I don't think it's something he does not understand.

 

Semantics are very important in his videos since, as you've just indicated, people pick up on incorrect statements very quickly. This is why he often words what he says in a less-than-absolute manner. In this case, I think he was trying to be modest in his understanding of the relationship between power consumption and heat output.

 

While I can't speak for Linus, I'm sure he would appreciate the clarification for those who need more of an in-depth understanding :)

Link to comment
Share on other sites

Link to post
Share on other sites

While I can't speak for Linus, I'm sure he would appreciate the clarification for those who need more of an in-depth understanding :)

Well, I hope the explanations above help somebody, be it Linus or someone else :)

Link to comment
Share on other sites

Link to post
Share on other sites

  • 4 months later...

@LinusTech , on the livestream you said there is "somewhat of a correlation" between the power a chip consumes and the heat it puts out. This is quite an understatement. In silicon chips, the power a chip consumes equals the heat it puts out.

 

i was just searching for some clarity on this issue as i remember that live stream when linus said above.  so you are saying no energy is used(wasted) by the CPU for all the processes it performs?  energy going in ALL comes out as heat?

Link to comment
Share on other sites

Link to post
Share on other sites

  • 2 months later...

Correct me if I'm wrong as I want to learn.

 

Lets use a chip with a 100W TDP. Is this basically saying that the cooling system must be able to cool a chip that on average will expel the heat of a 100W load?

PC: Ryzen 5 2600 // 16 GB Corsair 3200 // MSI RX 580 8 GB // 500 GB WD Blue M.2 (sata) // Silverstone Raven RVZ03B // Fedora 33

LAPTOP: i7 5700HQ // 16 GB Kingston 1866 // GTX 970M // 250 GB Samsung 850 EVO M.2 // 1 TB HDD // Windows 10

ETC: (2) Dell U2515H (2560x1440) // Corsair K63 // Logitech G603 // FiiO E10 // Beyerdynamic DT 770 PRO (250 ohm) // Audio Technica ATH-M60x

 

Link to comment
Share on other sites

Link to post
Share on other sites

Correct me if I'm wrong as I want to learn.

 

Lets use a chip with a 100W TDP. Is this basically saying that the cooling system must be able to cool a chip that on average will expel the heat of a 100W load?

A chip that has a TDP of 100W will, at full power and without any tweaking, consume 100W of eletrical power and expel 100W of heat to the cooling system. This is in normal cases. The chip's actual power draw will vary because of tolerances in the manufacturing process. The TDP gives a guideline to system manufacturers.

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×