What is TDP or Thermal Design Power as Fast As Possible

[Student]

I did not find a right place to put it, so,...

 

i just watched his new video: https://www.youtube.com/watch?v=yDWO177BjZY

and thought at: 0:23.

that if you have 2x 250 Watts you need a >500 Watts psu.

i dont the see the error?

Because if a component has a tdp of x it will draw atleased x amount of power, because of the Conservation of energy.

I would evey say it will mostlikely draw exactly x amound of power because where should the energy go? the parts in my computer, don't glow, move, vibrate, spin, or else wise tranform energy (exept for the fans).

for me this video was more confusing than helpfull.

[Askew]

It becomes heat, as all energy ends up eventually.

[CowsGoRoar]

That's what resistors do too. No machine can be 100% efficient. There is always a little bit of energy that is going to be translated into another form of energy, and resistors take the extra energy and heat themselves to prevent explosions.

[steffen_anywhere]

The idea with TDP is that it is the average full load heat dissipation of a processor. It is a specification aimed at being helpful when looking for a suitable cooling solution. Of course @Student all input energy will end up as heat. But components often even exceed their TDP in terms of real-life power consumption, because TDP is really only a guideline to how much a component will use/dissipate (as those two really are the same). So the video explains what TDP is useful for. As I said for choosing cooling and estimating power usage to choose a suitable power supply. It also says what TDP is not: A value that can be added up for every part that will give you an accurate wattage of your PC in total. Hope this helps. Feel free to ask more questions as this answer came out to be somewhat complicated to understand imo. Also, please don;t open new threads for videos, Linus already makes one for each in the Linus' Videos, Reviews and Ramblings section.

[Student]

It becomes heat, as all energy ends up eventually.

Yes and that why i am confused because for processors, all their energy the consume endup in heat so:

tdp >= power draw from psu (if you dont overclock)

[Student]

The idea with TDP is that it is the average full load heat dissipation of a processor. It is a specification aimed at being helpful when looking for a suitable cooling solution. Of course @Student all input energy will end up as heat. But components often even exceed their TDP in terms of real-life power consumption, because TDP is really only a guideline to how much a component will use/dissipate (as those two really are the same). So the video explains what TDP is useful for. As I said for choosing cooling and estimating power usage to choose a suitable power supply. It also says what TDP is not: A value that can be added up for every part that will give you an accurate wattage of your PC in total. Hope this helps. Feel free to ask more questions as this answer came out to be somewhat complicated to understand imo. Also, please don;t open new threads for videos, Linus already makes one for each in the Linus' Videos, Reviews and Ramblings section.

 

But can't you still add up all tdp's and add some watts to choose a psu?

 

and i did not find the post

[Moopey]

But can't you still add up all tdp's and add some watts to choose a psu?

 

and i did not find the post.

 

 

or you could let a calculator do it! (http://extreme.outervision.com/psucalculatorlite.jsp)

 

You didn't find the post because he did the episode on Techquickie and not on his channel.

[LukaP]

I did not find a right place to put it, so,...

 

i just watched his new video: https://www.youtube.com/watch?v=yDWO177BjZY

and thought at: 0:23.

that if you have 2x 250 Watts you need a >500 Watts psu.

i dont the see the error?

Because if a component has a tdp of x it will draw atleased x amount of power, because of the Conservation of energy.

I would evey say it will mostlikely draw exactly x amound of power because where should the energy go? the parts in my computer, don't glow, move, vibrate, spin, or else wise tranform energy (exept for the fans).

for me this video was more confusing than helpfull.

you have work that transistors do, and the energy put into it does not translate into heat but rather into calculations. thats what the arch efficiency means. more watts go into work than heat  

 

if a 250W card only sucked those 250W. it would do nothing just transform electricity into heat (ergo be a resistor), so it takes abit more than that. and how much it does depends on the load

[Student]

you have work that transistors do, and the energy put into it does not translate into heat but rather into calculations. thats what the arch efficiency means. more watts go into work than heat

 

if a 250W card only sucked those 250W. it would do nothing just transform electricity into heat (ergo be a resistor), so it takes abit more than that. and how much it does depends on the load

sorry, but no, there is NO enery cosumed to do calculation, because you don't perform work.

http://en.wikipedia.org/wiki/Forms_of_energy

if 250W go in, in form of electric energy, 250W go out in form of heat.

That is basic conservation of energy.

Thats why i am saying tdp and average energy consumed at full load is equal!

[LukaP]

sorry, but no, there is NO enery cosumed to do calculation, because you don't perform work.

http://en.wikipedia.org/wiki/Forms_of_energy

if 250W go in, in form of electric energy, 250W go out in form of heat.

That is basic conservation of energy.

Thats why i am saying tdp and average energy consumed at full load is equal!

 

have you heard how a transistor works? its a conductivity shift, and to do that, you perform work, and the amout of energy you need for the electrons to tunnel is the difference between energy inputed and TDP. (i study physics so i know this stuff)

[steffen_anywhere]

you have work that transistors do, and the energy put into it does not translate into heat but rather into calculations. thats what the arch efficiency means. more watts go into work than heat

 

if a 250W card only sucked those 250W. it would do nothing just transform electricity into heat (ergo be a resistor), so it takes abit more than that. and how much it does depends on the load

 

Nope, sorry this is completely wrong. All input energy will end up as heat, i.e. degraded energy. The calculations can be seen as working against entropy by creating order. This is comparable to tidying up a room. You spend energy to do it, but after doing so, the room does not have more energy, it is just organized.

 

EDIT: If you studied it in more detail, ok, you have more knowledge on the topic than I do, but in the end, what you mentioned will still end up as heat... Resulting digital data has no potential to do work...

 

sorry, but no, there is NO enery cosumed to do calculation, because you don't perform work.

http://en.wikipedia.org/wiki/Forms_of_energy

if 250W go in, in form of electric energy, 250W go out in form of heat.

That is basic conservation of energy.

Thats why i am saying tdp and average energy consumed at full load is equal!

 

Yes exactly right. But the thing is that TDP is usually used in a very loose way. An entry level i5 is rated at 84Watts. So is a 4770K. At stock speeds even, the 4770K will still consume a bit more as it is simply clocked higher (I know it is probably higher binned or something, but it will still consume more).

[LukaP]

Nope, sorry this is completely wrong. All input energy will end up as heat, i.e. degraded energy. The calculations can be seen as working against entropy by creating order. This is comparable to tidying up a room. You spend energy to do it, but after doing so, the room does not have more energy, it is just organized.

 

EDIT: If you studied it in more detail, ok, you have more knowledge on the topic than I do, but in the end, what you mentioned will still end up as heat... Resulting digital data has no potential to do work...

you are kinda right , but not completly. the energy that gets locked in the electrons will first end up as kinetic energy, and then the electrons will bind into other mass. and due to mass defect and some other effects, energy is expelled in a way of heat, but a part of that energy is also taken as electric energy through the diplay cable, into the cpu (energy is lost to heat on the entire way everywhere) and when it comes to the display scaler same happens. then the energy that is left (there has to be energy/mass to transfer information, its a basic law) is amplified with more power to then drive the panel. there all the remaining energy is output as light  

 

so yes in essence, all energy does convert to a basic form od heat/light, but not all of it does so in the GPU itself, since it needs to transfer infromation along too  

[Student]

have you heard how a transistor works? its a conductivity shift, and to do that, you perform work, and the amout of energy you need for the electrons to tunnel is the difference between energy inputed and TDP. (i study physics so i know this stuff)

 

you are kinda right , but not completly. the energy that gets locked in the electrons will first end up as kinetic energy, and then the electrons will bind into other mass. and due to mass defect and some other effects, energy is expelled in a way of heat, but a part of that energy is also taken as electric energy through the diplay cable, into the cpu (energy is lost to heat on the entire way everywhere) and when it comes to the display scaler same happens. then the energy that is left (there has to be energy/mass to transfer information, its a basic law) is amplified with more power to then drive the panel. there all the remaining energy is output as light

 

so yes in essence, all energy does convert to a basic form od heat/light, but not all of it does so in the GPU itself, since it needs to transfer infromation along too

the energy that goes out the connectors in neglectable, because that are very smal numbers compared to the chip itself. signals are transferd with very low electric current.

 

Yes exactly right. But the thing is that TDP is usually used in a very loose way. An entry level i5 is rated at 84Watts. So is a 4770K. At stock speeds even, the 4770K will still consume a bit more as it is simply clocked higher (I know it is probably higher binned or something, but it will still consume more).

That makes sense, but it is not what tdp is defined as.

TDP is defined as the maximum head dissipation. There is nothing wrong with an i5 and i7 at the same TDP because it probibly is the TDP of the i7 and the i5 will never consume more than that. Which is what is happenig:

 

Since TDP is defined for families, not individual processors, MHz/TDP W are not useful for comparing processors using the same internal structure.

http://en.wikipedia.org/wiki/List_of_CPU_power_dissipation#Microprocessors

 

The TDP is typically not the largest amount of heat the CPU could ever generate (peak power), such as by running a power virus, but rather the maximum amount of heat that it would generate when running "real applications." This ensures the computer will be able to handle essentially all applications without exceeding its thermal envelope, or requiring a cooling system for the maximum theoretical power (which would cost more but in favor of extra headroom for processing power).^[2 

 

Some sources state that the peak power for a microprocessor is usually 1.5 times the TDP rating.^[3] However, the TDP is a conventional figure while its measurement methodology has been the subject of controversy. In particular, until around 2006 AMD used to report the maximum power draw of its processors as TDP, but Intelchanged this practice with the introduction of its Conroe family of processors.^[4]

http://en.wikipedia.org/wiki/Thermal_design_power

 

 

same for PSUs:

the BeQuiet DarkPowerPro 550w has continues Power of 550w and peak power of 610w.

 

TDP is still the only reference point for coosing our PSU and if we have same extra wattage, everything will be fine, even if we overclock and run some extreme tests.

but for the average user they will and are fine with the next greater psu, if they add up all TDPs.

 

if not please show some source that has scientivic tests on powerconsumption vs tdp.

 

 

[steffen_anywhere]

 

the energy that goes out the connectors in neglectable, because that are very smal numbers compared to the chip itself. signals are transferd with very low electric current.

 

That makes sense, but it is not what tdp is defined as.

TDP is defined as the maximum head dissipation. There is nothing wrong with an i5 and i7 at the same TDP because it probibly is the TDP of the i7 and the i5 will never consume more than that. Which is what is happenig:

 

 

same for PSUs:

the BeQuiet DarkPowerPro 550w has continues Power of 550w and peak power of 610w.

 

TDP is still the only reference point for coosing our PSU and if we have same extra wattage, everything will be fine, even if we overclock and run some extreme tests.

but for the average user they will and are fine with the next greater psu, if they add up all TDPs.

 

if not please show some source that has scientivic tests on powerconsumption vs tdp.

 

 

 

Yes this is what TDP is. You are basically right. However, most manufacturers do list both power consumption and recommended power supply wattage. TDP, being "Thermal Design Power" is still really aimed at cooling solutions, but yes you are right that those are obviously closely related.

 

EDIT: Also, I'm saying Hi to a fellow German LOL