FRONTPAGE NVIDIA GEFORCE GTX 780 Graphics Card Review

Bobnova said:

More heat != runs hotter.
Running hot shortens lifespans (we are talking hot here, though. Not 80c), producing heat does not.

Click to expand...

not entirely correct again.

you trying to distinguish between running at a hot temp,
and producing a lot of heat, but dissipates into cooling solutions.

Both heat and producing heat can kill a chip, one more than the other.

What is Producing heat?
heat is produced by electric resistance when it meets Voltage and current.
Too high of a voltage can generate electron saturation that kills a chip outright regardless if you run it at 10K (that's 10 Kelvin.)
You are saying the producing heat doesn't kill a chip. really?
coz last I checked super conductivity not yet availible for modern chips yet. where you read it?

Similarly producing heat despite being able to cool it, will cause degradation, in time.

It is a function of both Voltage, Time and Temp.
Producing too much heat (despite being able to cool), and having a chip exposed to hot temp, will both cause degrading on chips.

Lol you're reaching on that one. Producing heat can be a byproduct of running too high a voltage, but it does not in and of itself mean the voltage is high.

That's like saying that speeding will wear out your engine without mentioning how much throttle you're feeding it out what ram it's running.

Your theory is interesting, but you are leaving out the current half of the equation.

As a bonus, chips still die at 10K. Benchers run them there on LHe from time to time.

So, in short, producing heat that is dissipated correctly does not in and of itself cause harm.
Adding voltage to a given equation can cause it, and a side effect can be damage, but it's from the village, nor the heat production.

Bobnova said:

So, in short, producing heat that is dissipated correctly does not in and of itself cause harm.
Adding voltage to a given equation can cause it, and a side effect can be damage, but it's from the village, nor the heat production.

Click to expand...

I think you mean voltage. but I don't pick on spelling. I understand you well.

voltage through resistance by definition will cause erosion. Especially with the shrinking die, parts get smaller = more vulnerable to this.
This is become more and more obvious and therefore when the die shrinks, the voltage must go down as well. You should know that.

So while I will continue to read up.

even running cold, high voltage (which produces heat, and heat is a function of voltage and resistance) does kill a chip. I feel this is common sense, no?

So I would like to say.
producing heat, 'passing voltage', WILL kill a chip by definition. Saying it won't is crazy talk. question is 'when it will'?

correct statement should be:

As long as a voltage is being passed, the lifespan of a chip will be shortened despite in a cold or hot environment.
Heat on top will accelerate that process as a catalyst, as heat will increase electric resistance, causing the same voltage to kill the die faster.
Therefore, producing heat is in fact, the process that kills the die (when voltage meets resistance), we just try to keep voltage low, temp low, to prolong the process.

only way that producing heat won't kill a die, is in super conductivity.
additional details or fleshing out of the content welcomed.

come backs? or refuse to admit?

Yeah voltage, my phone really wants it to be village. It's a punk.
I absolutely agree that voltage kills. Adding high temperatures to that kills even faster. What I don't agree with its that heat production kills.
As an example, does a fx8150 die faster at a given voltage than a fx4100? It produces more heat, but runs the same voltage.

Along those same lines we're back to where we started, 7970 lifespan vs 680 lifespan. Feed them the same voltage and the 7970 eats more power and hence makes more heat. The current side of ohms law is sort of important.

Voltage and heat production seem to be the same thing in your mind, they are not in reality.

Bobnova said:

Yeah voltage, my phone really wants it to be village. It's a punk.
I absolutely agree that voltage kills. Adding high temperatures to that kills even faster. What I don't agree with its that heat production kills.
As an example, does a fx8150 die faster at a given voltage than a fx4100? It produces more heat, but runs the same voltage.

Click to expand...

okay, so we just have a discrep on 'what is heat production'. You meant it on the mass scale sense, I meant it on small scale. Both are correct. I can live with that.

i meant by: voltage meet resistance = heat production and power.
u meant by: total thermal profile.

Along those same lines we're back to where we started, 7970 lifespan vs 680 lifespan. Feed them the same voltage and the 7970 eats more power and hence makes more heat. The resistance side of ohms law is sort of important.

Voltage and heat production seem to be the same thing in your mind, they are not in reality.

Click to expand...

I happen to think AMD dies are a bit more heavy duty, so on 'same' voltage on the 22mm tech, they might actually last longer, which justify their higher TDP target without shortening life. though I have no info to back myself up here. Heavy duty is word I am using to describe its design on the silicon level, which makes it more able to with stand voltage.

BUT the heat production (which in my world = when voltage meet resistance = putting it to work) will still kill the die, given time. So in general, it is always a good exercise to try to keep lowest operating voltage for your system. So to answer the question of this.. keep it cool, -and- keep voltage at min of requirement. the chip will live longer. AMD also tries to lower TDP of the chip, they just don't think it is necessary this moment. It is true their chips are power hungry, (there's no dispute there), yet their chips are made that way to withstand that.

I too think it is more elegant to have chip with low TDP, more energy efficient. but that might not be good business..

Bad analogy time:

Think of Power Plants:
Of course nuclear is better, more efficient.
but coal burning is just a lot cheaper, who cares about green energy if it gets the job done.

Degradation or killing the die is all about Electromigration, it is the transport of material caused by the gradual movement of the ions in a conductor due to the momentum transfer between conducting.

Thermal effects http://en.wikipedia.org/wiki/Electromigration

In an ideal conductor, where atoms are arranged in a perfect lattice structure, the electrons moving through it would experience no collisions and electromigration would not occur. In real conductors, defects in the lattice structure and the random thermal vibration of the atoms about their positions causes electrons to collide with the atoms and scatter, which is the source of electrical resistance (at least in metals; see electrical conduction). Normally, the amount of momentum imparted by the relatively low-mass electrons is not enough to permanently displace the atoms. However, in high-power situations (such as with the increasing current draw and decreasing wire sizes in modern VLSI microprocessors), if many electrons bombard the atoms with enough force to become significant, this will accelerate the process of electromigration by causing the atoms of the conductor to vibrate further from their ideal lattice positions, increasing the amount of electron scattering. High current density increases the number of electrons scattering against the atoms of the conductor, and hence the speed at which those atoms are displaced.

In integrated circuits, electromigration does not occur in semiconductors directly, but in the metal interconnects deposited onto them (see semiconductor device fabrication).

Electromigration is exacerbated by high current densities and the Joule heating of the conductor (see electrical resistance), and can lead to eventual failure of electrical components. Localized increase of current density is known as current crowding.

AMD GPUs still run hotter than nVidia ones, even moreso overclocked, I presume.

Along those same lines we're back to where we started, 7970 lifespan vs 680 lifespan. Feed them the same voltage and the 7970 eats more power and hence makes more heat. The current side of ohms law is sort of important.

Click to expand...

So then the 7970 will die earlier. What's so hard about that?

AMD skimped on efficiency for price, and its customers pay the price.

http://www.tomshardware.com/reviews/radeon-hd-7970-ghz-edition-review-benchmark,3232-18.html

trueblack said:

okay, so we just have a discrep on 'what is heat production'. You meant it on the mass scale sense, I meant it on small scale. Both are correct. I can live with that.

i meant by: voltage meet resistance = heat production and power.
u meant by: total thermal profile.

Click to expand...

When I say "increased heat production", I mean "creating more heat". I have no idea what you mean by large scale and small scale. Making more heat is an exceedingly simple concept. Power consumption = heat production. It doesn't matter what scale it is on. Venture into the nano-world and it's still true.

Without a voltage change or a temperature change, creating more heat (reduce the resistance, voila!) will no degrade a chip faster.

trueblack said:

I happen to think AMD dies are a bit more heavy duty, so on 'same' voltage on the 22mm tech, they might actually last longer, which justify their higher TDP target without shortening life. though I have no info to back myself up here. Heavy duty is word I am using to describe its design on the silicon level, which makes it more able to with stand voltage.

Click to expand...

Both the FX8150 and the FX4100 are the same process (same die even, I think), the FX4100 has two processing units (or whatever AMD is calling them, four cores) disabled. I think you missed this part.
The FX4100 is a 95w chip, the FX8150 is a 125w chip.
In "your world", does the 8150 die faster?

trueblack said:

BUT the heat production (which in my world = when voltage meet resistance = putting it to work) will still kill the die, given time. So in general, it is always a good exercise to try to keep lowest operating voltage for your system. So to answer the question of this.. keep it cool, -and- keep voltage at min of requirement. the chip will live longer. AMD also tries to lower TDP of the chip, they just don't think it is necessary this moment. It is true their chips are power hungry, (there's no dispute there), yet their chips are made that way to withstand that.

Click to expand...

That's what heat production is, almost anyway. Heat production is a factor of voltage and current, the resistance is (generally, but not always) what sets the amount of current flow and hence the power consumed and the heat output for a given voltage.
Higher voltage degrades faster. Higher heat production with the same voltage (easy to get, just raise the clock speeds and speed the heatsink fan up to compensate for the extra heat) will not.

Culbrelai said:

AMD GPUs still run hotter than nVidia ones, even moreso overclocked, I presume.

Click to expand...

How hot a core runs depends on two primary things: Heat produced, and cooling solution.
A 35w laptop CPU can easily run along at 100c, while a 200w GPU can easily run along at 60c.

Culbrelai said:

AMD GPUs still run hotter than nVidia ones, even moreso overclocked, I presume.

Click to expand...

My 7970 stays under 80 during a several-hour-long benchmark at the settings in my sig. At stock settings, it NEVER gets over 65C.

My 7970 stays under 80 during a several-hour-long benchmark at the settings in my sig. At stock settings, it NEVER gets over 65C.

Click to expand...

My 670 with a reference blower overclocked never gets above 55, lol

You can't argue with numbers, look at the link I posted in my last post...

Culbrelai said:

AMD GPUs still run hotter than nVidia ones, even moreso overclocked, I presume.



So then the 7970 will die earlier. What's so hard about that?

AMD skimped on efficiency for price, and its customers pay the price.

http://www.tomshardware.com/reviews/radeon-hd-7970-ghz-edition-review-benchmark,3232-18.html

Click to expand...

That is not true it depends on die conductive material in the die and running temp.

Culbrelai said:

My 670 with a reference blower overclocked never gets above 55, lol

Click to expand...

And some non-reference 7870s don't even get that hot, what's your point?
Have you, perhaps, noticed that coolers have a bit to do with the running temperatures of the cores?

Bobnova said:

Without a voltage change or a temperature change, creating more heat (reduce the resistance, voila!) will no degrade a chip faster.

Click to expand...

will not degrade faster , now you are right.

your previous assessment however, says:

Originally Posted by Bobnova
More heat != runs hotter.
Running hot shortens lifespans, producing heat does not.

Click to expand...

that is wrong. Producing heat WILL shorten lifespan, but given correct parameters might not shorten lifespan faster. That was what I was point at misleading previously.

Higher voltage degrades faster. Higher heat production with the same voltage will not.

Click to expand...

depends on context. if use your multi-core chip example, you are right, cause the excess heat production is from 'other cores doing work', if the heat can be dissipated, doesn't add to the chip's overall heat, it will not kill it faster. (this is what I meant, when I say you mean producing heat on a mass scale)

however, in another example. 2 chips, both 1 core. Both same voltage, but one generates more heat cause it has higher electric resistance. despite both chips can dissipate the heat, the one more with resistance will likely observe more electron migration, and degrade faster. (this is what I mean at small scale)

the reduction of resistance to 0, can in theory, produce something with near infinite silicon life. this is one of the reasons people are interested in the area.

Well I have done some more reading on Electromigration and you dont have to have heat to move the atoms it also can be done with Diffusion mechanisms.

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

In a homogeneous crystalline structure, because of the uniform lattice structure of the metal ions, there is hardly any momentum transfer between the conduction electrons and the metal ions. However, this symmetry does not exist at the grain boundaries and material interfaces, and so here momentum is transferred much more vigorously. Since the metal ions in these regions are bonded more weakly than in a regular crystal lattice, once the electron wind has reached a certain strength, atoms become separated from the grain boundaries and are transported in the direction of the current. This direction is also influenced by the grain boundary itself, because atoms tend to move along grain boundaries.

Diffusion processes caused by electromigration can be divided into grain boundary diffusion, bulk diffusion and surface diffusion. In general, grain boundary diffusion is the major electromigration process in aluminum wires, whereas surface diffusion is dominant in copper interconnects.

trueblack said:

that is wrong. Producing heat WILL shorten lifespan, but given correct parameters might not shorten lifespan faster. That was what I was point at misleading previously.

Click to expand...

rofl, I award you with the Pedantic Award for Gratuitous Pedantry.
Way to find a context, no matter how spectacularly unreasonable, in which you can "prove" me wrong, I salute you.
I also award the award for "best intentional misinterpretation in an attempt to regain lost respect" for, oh, this year.
Holy hell man, I realize you have a deep seated need to prove me wrong now, but that's just silly.

Fortunately (for me, anyway), I excel at pedantry.
Here's a study of electromigration in superconductors: http://www.physics.ncsu.edu/optics/htsc/EM_YBCO_APL.pdf
They produce no heat, but electrimigration still happens. Maybe it is just voltage

Beyond that, I can easily argue that in a processor core producing heat is key to the processor functioning at all.
Keep in mind that CPUs and GPUs are silicon semi-conductors. Turn that "semi" into "super" and you have a dead short and a very short GPU lifespan indeed.
Producing heat means that it's still a semi-conductor, which means that it'll last a while.
Presto! Using the powers of Extended Pedantry, I have proven that producing heat lengthens the lifespan of a GPU!

Bobnova said:

rofl, I award you with the Pedantic Award for Gratuitous Pedantry.
Way to find a context, no matter how spectacularly unreasonable, in which you can "prove" me wrong, I salute you.
I also award the award for "best intentional misinterpretation in an attempt to regain lost respect" for, oh, this year.
Holy hell man, I realize you have a deep seated need to prove me wrong now, but that's just silly.

Click to expand...

sure. and let me aware you with most unwilling to admit he made a false statement and keeps trying to get around to make himself look less wrong award then.

you could have similar said, "Oh yes, I should had worded it better, the way I worded it can be misleading" but you just had to go down the road of insulting others to victory.
amazing human being you are. congrats.

but your link is good reading, I am enjoying that piece. that I thank you.
just btw, I speak 5 languages, english being my 4th.. so if my english is not good enough, I admit to that too.
if you speak as many tongues as I do, I would like to see how well you do too. cheers.

Bobnova said:

Fortunately (for me, anyway), I excel at pedantry.
Here's a study of electromigration in superconductors: http://www.physics.ncsu.edu/optics/htsc/EM_YBCO_APL.pdf
They produce no heat, but electrimigration still happens. Maybe it is just voltage

Click to expand...

From the Quote http://en.wikipedia.org/wiki/Electromigration
Electromigration is exacerbated by high current densities and the Joule heating of the conductor (see electrical resistance), and can lead to eventual failure of electrical components. Localized increase of current density is known as current crowding.

Joule heating, also known as ohmic heating and resistive heating, is the process by which the passage of an electric current through a conductor releases heat. The amount of heat released is proportional to the square of the current such that

"exacerbated", not "caused".
High heat conditions and high voltage conditions make it worse, is what that says.

Damn this is getting technical, but insulting someones grammar/etc merely shows your lack of respect, for that person, and likely everyone else with opposing views around you.

We can argue till' the end of time, but throwing your dollars and support behind nVidia is the strongest move you can make. Sig your rig proudly, be part of the green team. The whole AMD vs Intel/nVidia thing is a lot like politics, the losing team won't change untill something affects them directly, on their camp, and that is when they will change their tune.

Just wait a year, a couple of dead 7970s will teach em a thing er two =P

And if they continue after that, they it can be said that they are indeed fanboy(s)

Bobnova said:

"exacerbated", not "caused".
High heat conditions and high voltage conditions make it worse, is what that says.

Click to expand...

That was my point I did not know if you were kiding when you said Maybe it is just voltage. Because exacerbated Electromigration with heat is a fact also.

I agree after all this reading the cause of Electromigration is the electrons collide with the atoms from a electron wind or ion wind.