Intel actually more efficient for gaming?

[henson0115]

Therefore you cannot use heat as a measure of efficiency. Which is what I've been saying since the first post.

Which is correct but the problem is 60w tdp on 2 chips doesn't mean it's the same 60w on each.. they are different designs and therefor tdp isnt a correct way to measure heat output on them.

Both intel and and measure differently is what I mean.

But you also can't measure compute output either as that's also different so what do you compare? In any case I doubt most people think this deeply about it.

 

[Psycho Sonny]

Which is correct but the problem is 60w tdp on 2 chips doesn't mean it's the same 60w on each.. they are different designs and therefor tdp isnt a correct way to measure heat output on them.

Both intel and and measure differently is what I mean.

But you also can't measure compute output either as that's also different so what do you compare? In any case I doubt most people think this deeply about it.

You measure the amount of electricity consumed by the chip and then the number of calculations it can do in say a second and divide them.

So you get a number of calculations per watt of power consumption.

You would take the number of calculations at 100% usage and power consumption at 100% usage of a normal stock CPU with no special BIOS settings.

 

[henson0115]

You measure the amount of electricity consumed by the chip and then the number of calculations it can do in say a second and divide them.

So you get a number of calculations per watt of power consumption.

You would take the number of calculations at 100% usage and power consumption at 100% usage of a normal stock CPU with no special BIOS settings.

But again two different CPU designs it's like comparing apples to oranges. When getting to a really technical level

 

[Phil2008]

No idea, but intel has more raw power than amd, or it use to anyway?

 

[Phil2008]

Well this thread is in flames.



Is there some magic machine somewhere that turns energy into "work" but not "wasted energy".

You don't get to do anything without turning a useful form of energy into a useless form.

Efficiency is getting what you want with the least amount of energy.

A cpu using 200W isn't wasting 200W, that's what that design needs to do the electron shuffling.

Heat is wasted energy apparently.

 

[Troezar]

Heat is wasted energy apparently.

Given that a CPU isn't designed to be a space heater I'd say that heat produced is a waste product. I'm sure Intel engineers would love a 100% conversion of the electrical energy and a room temperature CPU. The more you can reduce that wasted energy the better.

 

[4K8KW10]

Given that a CPU isn't designed to be a space heater

People have no choice but to accept the CPUs as such. A little unfortunate but perfectly fine.

I'm sure Intel engineers would love a 100% conversion of the electrical energy and a room temperature CPU. The more you can reduce that wasted energy the better.

Intel's engineering has stopped its development at the 6-7-year-old 14nm process, you want them to break the physical limitations - well, what you want is impossible.
There will never be a CPU that remains as cool as its ambient, surroundings temperature.

 

[Zefan]

Given that a CPU isn't designed to be a space heater I'd say that heat produced is a waste product. I'm sure Intel engineers would love a 100% conversion of the electrical energy and a room temperature CPU. The more you can reduce that wasted energy the better.

Why are you talking about 100% "conversion"? 100% of the energy that goes in to any electronic component that isn't output in some other way will be converted to heat . All components (and billions of subcomponents) have resistance by necessity and electricity will not pass through them without creating heat. It isn't possible to reduce it to 100% "conversion", whatever that even means. The only way it would even get close is if the CPU consisted of a single thread of superconductor and did no work whatsoever... Wow so efficient

 

[Troezar]

Why are you talking about 100% "conversion"? 100% of the energy that goes in to any electronic component that isn't output in some other way will be converted to heat . All components (and billions of subcomponents) have resistance by necessity and electricity will not pass through them without creating heat. It isn't possible to reduce it to 100% "conversion", whatever that even means. The only way it would even get close is if the CPU consisted of a single thread of superconductor and did no work whatsoever... Wow so efficient

Seriously? What is it with this forum? I do wonder. You should know nothing is 100% efficient. I was illustrating the point, willfully ignoring that or just one of those that take everything literally? Is this place just full of trolls, the very young or a lot of people on some kind of spectrum? Is there a cap to the ignore list? As a moderator I would have expected more.

 

[Hotwired]

You could just stop being wrong whether its by accident or on purpose

What's the point in misusing technical descriptions or terms.

 

[Troezar]

You could just stop being wrong whether its by accident or on purpose

What's the point in misusing technical descriptions or terms.

Tldr; Ok you're right, heat is a desirable byproduct, let's have more please, Intel are more efficient because they're hotter! Life is too short to be discussing this

 

[4K8KW10]

Tldr; Ok you're right, heat is a desirable byproduct, let's have more please, Intel are more efficient because they're hotter! Life is too short to be discussing this

I think you need to open a textbook or two in Physics and read how it all goes

 

[Zefan]

Tldr; Ok you're right, heat is a desirable byproduct, let's have more please, Intel are more efficient because they're hotter! Life is too short to be discussing this

The point is, 100% of the power that goes in to a CPU ends up as heat. There are no "hotter" or "cooler" chips in terms of "efficiency", the measure should be Watts per work. Heat is irrelevant.

 

[humbug]

Efficiency is how much work a CPU can do for every Watt used.

And again....

Right, all energy used it turned into heat, a 200 Watt CPU will produce 200 Watts of heat, a 75 Watt CPU will produce 75 Watts of heat, no other way about it, its physics.

Its such simple basic stuff, no need for the mental gymnastics.

 

[SkyTerran]

The amount of power consumed by a CPU that isn't converted into heat inside the CPU is vanishingly small. For all practical purposes, power = heat.

I'm curious if this is correct, or if 100% of the consumed power is converted into heat.

 

[humbug]

I'm curious if this is correct, or if 100% of the consumed power is converted into heat.

Electromagnetism, light, other forms of radiation... in such small quantities its not worth adding to the equation, for example Electromagnetism kills electrical components, instantly in high enough doses, the Electromagnetism a CPU radiates is so small it will kill components in its vicinity, or even its self, but its so small it takes decades.

for the purposes of this conversation power = heat.

 

[0h_no_not_again]

I'm curious if this is correct, or if 100% of the consumed power is converted into heat.

Well, the outputs of the CPU are low voltage & currents to the RAM, PCIe, etc. All the other input power turns into heat.

 

[Destination]

I'm curious if this is correct, or if 100% of the consumed power is converted into heat.

I assume some of it becomes electrical energy as in electrical output, and this in turn will likely become heat eventually.

 

[Mr Evil]

I'm curious if this is correct, or if 100% of the consumed power is converted into heat.

• Any time you have rapidly changing electric currents, like in a CPU where transistors are constantly switching, you get electromagnetic radiation emitted. This is why PCs are supposed to be enclosed in metal cases - to prevent those radio waves from escaping. Some of them will escape anyway, and could end up travelling far across the universe without ever being turned into heat. This can be quite detectable with nothing more than a radio receiver, if you want to prove to yourself that at least some energy is not immediately turned into heat.
  
• Parts of the CPU vibrate a little, due to phenomena such as electrostriction. This is radiated away as sound (though it won't get far before turning into heat).
  
• Some energy will be temporarily stored in electric fields in capacitors, or magnetic fields in inductors.
• Some ends up stored inside the silicon of the CPU itself when atoms are moved around, i.e. electromigration.
• Probably a bunch of other things I haven't thought of.

These are all minuscule though. If you could accurately measure the heat output of the CPU, it would be practically equivalent to its electrical power consumption.

 

[4K8KW10]

It is not 100% but approximately or 99.9999% something.