r9 360, does not need extra power from psu... Could this mean efficiency

[GoldSrc]

Wow that was painful to read. With 80% efficiency PSUs, 20 % of the power is indeed being turned into heat, inside the psu, the remaining 80% is being turned into heat in the components that it powers. You do realise that those graphs show system power consumption, motherboards, hard drives, gpus, etc. don't run on pixie dust, they use power. The tdp of the processor, is for the processor alone, and the VRMs have an efficiency rating too, like power supplies.

If you understand that PSUs don't turn 99% of the energy they use into heat, why did you mention it in the first place?

Also, the energy that is lost to heat from other components is not 99%, we're not talking about incandescent light bulbs here.

 

But if you have some information of some current electronic device that turns 99% of the energy into heat, I would really like to see it.

And I'm pretty sure that incandescent light bulbs are the only thing I know of that turns more than 95% of the energy it uses into heat.

[GoldSrc]

"Total system power consumption" includes CPU, chip set(Z77), ram, SSD/HDD, etc. You can't say i7 3770K uses ~50W when idling, because you don't know the power consumption of other components.

I know, but 450W for the entire PC is a good average value for your average PC with a single GPU.

[Cheddle]

the reason TDP gets interchanged is because they might as well be the same thing, heat and power draw... almost all the electricity used in a solid state IC's IS leaked as heat...

 

If a video card is emitting 250w of heat, I can assure you its pulling almost bang on 250w of power... the only reason its NOT pulling exactly the same amount there is approximately 6 watts per GPU fan @ 100% fan speed (12v @ 0.5amps) and an almost negligible amount lost over the monitor cables (when in use only)

 

Yes TDP means a specific thing, but TDP may as well be draw power... it doesn't matter in the context of computer components at all. - even manufacturers interchange them horridly

 

This aside back to the original topic, if the TDP of a card is 75 watts or less it can run over the PCIE slot alone... any higher and its going to need additional power, if the card uses the same GPU then there is no way you can reduce its power consumption unless you reduce its performance. Very very rarely a manufacturer will change the transistor type (I've never heard of a size change) on a GPU and keep the same design.

 

off topic again - To be totally honest there is no reason why a single 6 pin cant provide 500+watts of power. its really dumb... you see the power cable running into your pc? that sucker can carry 2,400watts... its got 2 active pins and one ground pin... I assume there is some really bad old PCIe standard that limits supply per pin...

[Megahurt]

If you understand that PSUs don't turn 99% of the energy they use into heat, why did you mention it in the first place?

Also, the energy that is lost to heat from other components is not 99%, we're not talking about incandescent light bulbs here.

 

But if you have some information of some current electronic device that turns 99% of the energy into heat, I would really like to see it.

And I'm pretty sure that incandescent light bulbs are the only thing I know of that turns more than 95% of the energy it uses into heat.

A power supply "uses" 20% of the energy, and more than 99% of that gets converted to heat, and  more than 99 % of the remaining 80% will get turned into heat by the rest of the system's components.

 

http://superuser.com/questions/148070/where-does-power-consumption-go-in-a-computer

 

http://electronics.stackexchange.com/questions/79166/where-does-all-the-power-consumed-by-a-cpu-go

 

http://www.silentpcreview.com/forums/viewtopic.php?t=44418

 

http://www.instructables.com/answers/efficiency-not-efficacy-of-CPUscomputers-is-all-th/

 

Edit : found an actual test :

 

https://www.pugetsystems.com/labs/articles/Gaming-PC-vs-Space-Heater-Efficiency-511/

[rjfaber91]

Yes TDP means a specific thing, but TDP may as well be draw power... it doesn't matter in the context of computer components at all. - even manufacturers interchange them horridly

 

Would you mind stepping into the AMD headquarters and telling them that? My Phenom II X6 is rated at 125W TDP, and on idle it uses 129W, but as soon as there's even a minute load, it goes up to 147W. Now, I know that CPUs are horribly inefficient compared to GPUs, but the difference between TDP and actual power consumption really isn't something to be ignored...

[Glenwing]

Would you mind stepping into the AMD headquarters and telling them that? My Phenom II X6 is rated at 125W TDP, and on idle it uses 129W, but as soon as there's even a minute load, it goes up to 147W. Now, I know that CPUs are horribly inefficient compared to GPUs, but the difference between TDP and actual power consumption really isn't something to be ignored...

Is it overclocked?

Also, how are you measuring your power consumption?

[Megahurt]

Would you mind stepping into the AMD headquarters and telling them that? My Phenom II X6 is rated at 125W TDP, and on idle it uses 129W, but as soon as there's even a minute load, it goes up to 147W. Now, I know that CPUs are horribly inefficient compared to GPUs, but the difference between TDP and actual power consumption really isn't something to be ignored...

You mean your pc uses 129W at idle, which means all the components + efficiency factor of the PSU.

[rjfaber91]

You mean your pc uses 129W at idle, which means all the components + efficiency factor of the PSU.

 

Hah, I wish. No, it's only the CPU. It's also not overclocked, incidentally...

[Megahurt]

Hah, I wish. No, it's only the CPU. It's also not overclocked, incidentally...

Don't know where you get those numbers from, but you're simply wrong. AMD's stock coolers can barely handle the 125W tdp, if your numbers were true, AMD would have had a lot  of problems with them, yet they didn't.

[typographie]

Hah, I wish. No, it's only the CPU. It's also not overclocked, incidentally...

 

http://www.anandtech.com/show/3674/amds-sixcore-phenom-ii-x6-1090t-1055t-reviewed/10

These are numbers for total system consumption, including either a GTX 280 or HD 5870.

 

I have to ask, where are you reading your power consumption figures?

[SkilledRebuilds]

3570k using 100w isn't BS.

I had my 2600k pullimg up to 118w on its own when OC'd.

[Techhog]

If that were the case, power supplies would never have 80+% efficiency.

Electronics are more efficient than they were years ago, an 80% efficient PSU turns 20% into heat not 99%.

 

And you're wrong, the 3570K uses around 50W at idle and +100W at full load.

3570K load

3570K_load.png

3570K idle

3570K_idle.png

 

And since most PCs don't use more than 450-500W on full load, those numbers are consistent.

But hey, you're free to do your own test if you don't trust Anandtech ones.

Protip: Read the text on images before posting them. Specifically, read the line where it mentions SYSTEM power consumption. The only thing you can get from this is that the 3570K probably uses around 40W more under load than it does at idle.

[Techhog]

3570k using 100w isn't BS.

I had my 2600k pullimg up to 118w on its own when OC'd.

Do I even need to add anything to this?

[Megahurt]

3570k using 100w isn't BS.

I had my 2600k pullimg up to 118w on its own when OC'd.

TDP only matters at stock, once you overclock it goes out the window.

[sgloux3470]

In electronics, I'd be willing to bet that  more than 99% of the energy gets converted into heat. There is no physical work being done so the only possible way it gets converted is heat. Also I call bullshit on that 100W with the 3570k, and MAJOR bullshit on the 50W at idle.

http://www.guru3d.com/articles-pages/core-i5-3570k-review,10.html

 

Power consumption measurements are largely a clusterfuck of poor methodology so I'd take it with a grain of salt, but these results do reflect what he stated.

[Cheddle]

Would you mind stepping into the AMD headquarters and telling them that? My Phenom II X6 is rated at 125W TDP, and on idle it uses 129W, but as soon as there's even a minute load, it goes up to 147W. Now, I know that CPUs are horribly inefficient compared to GPUs, but the difference between TDP and actual power consumption really isn't something to be ignored...

 

If your CPU consumes 147watts of electricity - It will emit 147 watts of heat.

 

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

[Megahurt]

http://www.guru3d.com/articles-pages/core-i5-3570k-review,10.html

 

Power consumption measurements are largely a clusterfuck of poor methodology so I'd take it with a grain of salt, but these results do reflect what he stated.

From your own link : "Bear in mind that we measure the ENTIRE PC"

[sgloux3470]

From your own link : "Bear in mind that we measure the ENTIRE PC"

 

True, but with only the CPU/Mobo/Memory and one SSD installed.  It's pretty safe to say that the CPU is 95% of that usage.

[Megahurt]

True, but with only the CPU/Mobo/Memory and one SSD installed.  It's pretty safe to say that the CPU is 95% of that usage.

Lol no, the cpu itself probably consumes around 5W at idle, they are incredibly efficient when doing nothing.

[sgloux3470]

Lol no, the cpu itself probably consumes around 5W at idle, they are incredibly efficient when doing nothing.

 

Well even if so, it's still the bare minimum power you're going to be drawing from the wall when actually using the chip.  

 

I find it hard to believe that the motherboard/memory/SSD is consuming the majority of the wattage though.  

[Megahurt]

Well even if so, it's still the bare minimum power you're going to be drawing from the wall when actually using the chip.  

 

I find it hard to believe that the motherboard/memory/SSD is consuming the majority of the wattage though.  

At idle yes, wouldn't suprise me if the motherboard + memory were in the range of 40-50W, and don't forget that the psu could have an efficiency of 70-90%, so 95% would litterally be impossible.

 

Edit : I don't know how recent it is but :

http://www.buildcomputers.net/power-consumption-of-pc-components.html

[GoldSrc]

A power supply "uses" 20% of the energy, and more than 99% of that gets converted to heat, and  more than 99 % of the remaining 80% will get turned into heat by the rest of the system's components.

 

http://superuser.com/questions/148070/where-does-power-consumption-go-in-a-computer

 

http://electronics.stackexchange.com/questions/79166/where-does-all-the-power-consumed-by-a-cpu-go

 

http://www.silentpcreview.com/forums/viewtopic.php?t=44418

 

http://www.instructables.com/answers/efficiency-not-efficacy-of-CPUscomputers-is-all-th/

 

Edit : found an actual test :

 

https://www.pugetsystems.com/labs/articles/Gaming-PC-vs-Space-Heater-Efficiency-511/

 

Wait what?

How can 99% of that 20% be converted into heat?, that's the most inefficient PC I've ever heard of, that just doesn't sound right.

 

 

 

Protip: Read the text on images before posting them. Specifically, read the line where it mentions SYSTEM power consumption. The only thing you can get from this is that the 3570K probably uses around 40W more under load than it does at idle.

 

Wait a minute, how can you come up with "the 3570K uses around 40W more under load", when you're actually measuring the entire PC? and seeing the difference between idle and load values?

Add to that that you may know the power consumption of the mobo alone and figure out the real idle power consumption of the CPU, and add a few watts just to be sure, but not 40W.

[Megahurt]

Wait what?

How can 99% of that 20% be converted into heat?, that's the most inefficient PC I've ever heard of, that just doesn't sound right.

 

Either you're trolling or you just don't and won't get it. The bottom line is that you can assume that all of the power a system draws will be converted to heat, and you'll be VERY close to the truth.

[GoldSrc]

Either you're trolling or you just don't and won't get it. The bottom line is that you can assume that all of the power a system draws will be converted to heat, and you'll be VERY close to the truth.

I'm not trolling sir.

If all the power being draw got converted into heat, it would be a pretty inefficient system, like incandescent light bulbs, and obviously a PC is not an incandescent light bulb.

Sure, some energy gets lost to heat, but it's not 99% of it, only a small part of the power being used to do work will be turned into heat.

Go back a few years and compare the first Pentium 4 with your 4th gen i3, the i3 will use less power to do the same (or more) work than an old P4, the P4 will lose a lot more energy to heat than the i3.

[Megahurt]

I'm not trolling sir.

If all the power being draw got converted into heat, it would be a pretty inefficient system, like incandescent light bulbs, and obviously a PC is not an incandescent light bulb.

Sure, some energy gets lost to heat, but it's not 99% of it, only a small part of the power being used to do work will be turned into heat.

Go back a few years and compare the first Pentium 4 with your 4th gen i3, the i3 will use less power to do the same (or more) work than an old P4, the P4 will lose a lot more energy to heat than the i3.

A processor doesn't create any physical work, if it was 100 % efficient, it would consume no power at all. A cpu consuming any power at all is inefficient.

 

The way I understand it, the transistor leaks power while it's switching. That's why it consumes less power at idle, it isn't switching (for the most part), and why it uses more power when overclocked, it switches more often in the same time period, and it leaks more per clock when the voltage is higher.