Which is hotter?

[Aidan-O-Matic]

Hello, Forum.
My question seems simple, but I've been googling for hours now with no results. Which contributes more to the heat produced by a processor; clock speed or cores? That is, hypothetically which would run hotter, a 1.8Ghz quad core or a 2.3Ghz dual core?
Thanks

 

[willzzz]

We are talking heat output here rather than temperature right?

Are we talking about a specific architecture? Because that will make a big difference.

 

[Aidan-O-Matic]

Its an AMD Phenom 2 in an HP pavilion G6 laptop. Being a total noob, I'm not really familiar with either of those terms. I suppose heat output, as I can modify system settings to change the temperature. (but that's a guess based on reading one wikipedia article)

 

[13thmonkey]

Why do you want to know. Looking up the TDP of the specific chips according the manufacturer would give you an idea, but they are always fairly broad brackets.

This gives you an idea of the heat being produced and dumped into the case, the internal temp will only be loosely related to the TDP.

 

[fazers_on_stun]

^ Not speaking "for Um" (forum ), but heat production should be fairly linear for the # of cores (i.e., 2 identical cores loaded equally should produce double the heat of just one). However clock speed will eventually require an increase in core voltage, and power (heat) varies as the square of voltage (P = V^2/R). Thus, clock speed contributes more past where the # of cores line crosses the exponential curve of clock speed..

 

[geofelt]

Each cpu will have a specific power design point.
In general, newer chips operate more efficiently with lower power needs.
They also do more work per clock cycle.

For a specific chip, the power needed (and heat generated) is determined by how many cores are active at the same time, and the clock rate that they are operating at. But, even though a chip operating at a higher clock rate will generate more heat, it will do so for less time, making the answer not so simple.

 

[13thmonkey]

Somewhere in here http://en.wikipedia.org/wiki/Phenom_II or related articles is the TDP, which is the thermal dissipation power in Watts and so can be compared processor to processor. But as mentioned earlier the TDP is only a heavily 'bucketed' approximation.

 

[rinval]

after reading and studying up for months it would seem that heat production does somewhat scale with the number of cores as well as with the clock speed. of coarse this varies by brand and core type. as for your question of which would produce more heat it would depend on the brand and cores of what your comparing. for instance if your comparing amd vs intel obviously the intel will produce less heat per core and clock vs a comparable amd. And generally speaking the intel scales lower in terms of max heat production over all because of its lower TDP.

comparing phenoms is a bit tricky and they did employ different cores. if you looking at deneb only chips, these chips are all quad core but on the duel core only 2 cores are active leaving the other 2 disabled cores idle and not producing extra heat from usage.

Just for curiostity is there a reason for the comparison as in a laptop generally the manufacturer has created a heatsink solution that will be sufficient for either no matter which processor produces more heat.

 

[13thmonkey]

you are basically saying that heat production scales with power.

 

[rinval]

pretty much wattage consumed scales how much heat is produced. IE.. a 140w FX will produce tons more heat than a 95 watt phenom.

keep in mind this does not always apply to overclocking as a 32 nm I5 sandy bridge at 95w overclocked to the same clock the same as a 22nm I5 ivy bridge at 77W the ivy bridge does produce slightly more heat. (this still baffles me but im sure its something that intel will fix either with software or upgraded architecture.)

 

[13thmonkey]

pretty much wattage consumed scales how much heat is produced. IE.. a 140w FX will produce tons more heat than a 95 watt phenom.

keep in mind this does not always apply to overclocking as a 32 nm I5 sandy bridge at 95w overclocked to the same clock the same as a 22nm I5 ivy bridge at 77W the ivy bridge does produce slightly more heat. (this still baffles me but im sure its something that intel will fix either with software or upgraded architecture.)

not quite, the ivy has a higher temperature, but does not produce more heat, they are different things.

 

[rinval]

not quite, the ivy has a higher temperature, but does not produce more heat, they are different things.

sry if this is off topic but just for curiosity from what ive seen and read which i will try to find it again but it would seem that the i5 sandy bridge and 15 ivy bridge on the same board and cooler overclocked to the same speed they have similar idle temp but the ivy bridge was slightly hotter under load so would that indicate that the processor produces slightly more heat?

unless toms is wrong they say up to 20c...

http://www.tomshardware.com/news/ivy-bridge-overclocking-high-temp,15512.html

 

[DJDeCiBeL]

sry if this is off topic but just for curiosity from what ive seen and read which i will try to find it again but it would seem that the i5 sandy bridge and 15 ivy bridge on the same board and cooler overclocked to the same speed they have similar idle temp but the ivy bridge was slightly hotter under load so would that indicate that the processor produces slightly more heat?

unless toms is worng...

http://www.tomshardware.com/news/ivy-bridge-overclocking-high-temp,15512.html

The Ivy i5 doesn't necessarily PRODUCE more heat, it just can't dissipate it as well, for a couple of reasons. The smaller die size means less surface area for heat dissipation and the use of thermal compound underneath the IHS instead of fluxless solder, like on SB CPU's, causes slightly lower heat dissipation as well.

 

[rinval]

The Ivy i5 doesn't necessarily PRODUCE more heat, it just can't dissipate it as well, for a couple of reasons. The smaller die size means less surface area for heat dissipation and the use of thermal compound underneath the IHS instead of fluxless solder, like on SB CPU's, causes slightly lower heat dissipation as well.

ok i gotcha. 160mm die for ivy brige vs 216 die for sandy bridge plus the thermal package would generate this difference i stand corrected.

http://www.anandtech.com/show/5771/the-intel-ivy-bridge-core-i7-3770k-review/3

 

[13thmonkey]

sry if this is off topic but just for curiosity from what ive seen and read which i will try to find it again but it would seem that the i5 sandy bridge and 15 ivy bridge on the same board and cooler overclocked to the same speed they have similar idle temp but the ivy bridge was slightly hotter under load so would that indicate that the processor produces slightly more heat?

unless toms is wrong they say up to 20c...

http://www.tomshardware.com/news/ivy-bridge-overclocking-high-temp,15512.html

Absolutely agree with everyone that ivb has a higher temperature than SB (I've got one). But because the amount of heat energy is lower (77 vs 95watts) it doesn't heat up the rest of the case as much, and case fans don't have to work as hard to keep the system cool.

I think you've got it.

But imagine putting an amount of heat energy into a huge block of material its temp would go up by a bit, imagine putting the same amount of hat energy into a tiny block of material, its temp would go up by a lot more. Different materials require different amounts of heat energy to raise the temp.

so the question for the OP is this:
is he worried about the amount of heat thats getting pumped out into the room/or his lap (also roughly equivalent to the amount of electrical energy and therefore running costs).
or
is he worried about the internal temperature of the chip (which if the cooling system is sufficient is irrelevant for nearly every reason imaginable other than an irrational fear of hot things)