AMD R9 390X Coming With Cooler Master Liquid Cooler + Estimated Performance

[Sakkura]

I'm putting money on the reference card being air cooled but with an improved fan design which is the main culprit of bad cooling not the chip.

also i bet the 390x2 being dual gpu with an AiO.

for those that havent realised yet TDP=Thermal Design Power is heat output and NOTHING to do with power draw, mixing the 2 is as bad as saying that a 600w power supply will run hotter than a 500w...

also gtx480 ran hot due to-1) too much factory voltage, 2) so many heatpipes it killed all heatsink airflow, 3) dinky little fan that moved NO air, stock clocks with a gtx285 fan and 580 heatsink will reduce temps allot!

also "nearly the same tdp as an nvidia dual gpu titan-z, thats terrible", wtih 300w vs 250w TDP ill take a 20% higher heat output for 40% more performance otherwise we'd all be running cute 750ti's.

oh and as for temps of a 300wTDP card these twin fan coolers from asus/xfx/msi etc are very simple really, fins tacks are hard to make sure but 2 cheap fans over a bunch of fins isnt rocket science and could easily done as factory reference (see 7990) though usually arn't done because non-heatpiped blower types are slightly cheaper when you're putting out tens of thousands of cards and you want to leave something for ITX users.

 

I'm not so sure about that. I think I mentioned it before, AMD has taken a lot of flak for their reference coolers up until the R9 295X2. The R9 290 and 290X got a bad reputation because of the reference cooler, while the R9 295X2 got glowing reviews; maybe they're taking that lesson to heart and finally stop trying to make reference air coolers on their high-end cards that draw a lot of power. You mention the 7990, which did have a better reference cooler than the R9 290X... but it was still pretty bad. Plus the microstutter problems with Crossfire were still a thing back then, so that card was not popular.

[D3MON]

anyone remember the name of the term where air cannot absorb any more heat? hence why we dont have 100mm thick rads.
thermal capacity isnt quite correct, cant remember where i read it but it pertains to heatsinks and cooling obviously.

[NumLock21]

Stupid question in 3, 2, 1

GPU is cooled down by the fan that's attached on the radiator, what is the point of the blower style fan that's on the card itself? Is if for the power circuit components as well as the vrams? If that's what the fan on the card is for, then basically this hot air is actually blowing over the GPU water block, as it exits out the back of the card vents.

[Opcode]

Stupid question in 3, 2, 1

GPU is cooled down by the fan that's attached on the radiator, what is the point of the blower style fan that's on the card itself? Is if for the power circuit components as well as the vrams? If that's what the fan on the card is for, then basically this hot air is actually blowing over the GPU water block, as it exits out the back of the card vents.

Indeed the fan is there to cool the plate that covers the power delivery among others. The heat coming off them is next to nothing compared to what the GPU itself generates. Take the R9 295x2 as a prime example with a massive power delivery to push two R9 290X cores all centralized on the single PCB. Meanwhile the card runs at around 65C under a full load with a single 120mm radiator.

[mr moose]

SNIP

 

https://www.youtube.com/user/sixtysymbols

 

there are quite a few videos there.  you'll soon figure out where you went wrong.

[Hatsune Miku 「 」]

https://www.youtube.com/user/sixtysymbols

 

there are quite a few videos there.  you'll soon figure out where you went wrong.

So basically patrick was saying that we're all ignorant fools and saying that he is right without any solid evidence and spitting out random laws and including phrases like "existence sucks that way" and "consult a physicist"for no good of a reason, while trying to cover up/put up a facade over his own lack of knowledge/interpretation of it?

 

TLDR: patrick is toxic

[othertomperson]

Your conclusion is wholly incorrect because your understanding of physics itself is quite lacking. You are not creating that heat energy out of nowhere. The extra energy comes from converting mass to energy by way of quantum mechanical processes which result from the motion of particles, motion accelerated when heat is introduced.

 

We've been through this. The mass-energy relation is relevant when sub-atomic particles are bound together or separated. It's relevant when particles are accelerated and collided and new particles are produced. It's not at all relevant in calculating the TDP of a GPU.

[FatalityDiablo]

We've been through this. The mass-energy relation is relevant when sub-atomic particles are bound together separated. It's relevant when particles are accelerated and collided and new particles are produced. It's not at all relevant in calculating the TDP of a GPU.

Bazinga!

[patrickjp93]

We've been through this. The mass-energy relation is relevant when sub-atomic particles are bound together or separated. It's relevant when particles are accelerated and collided and new particles are produced. It's not at all relevant in calculating the TDP of a GPU.

Except you're WRONG! Good God is there anyone in here with an in-depth understanding of quantum mechanics and can relay the relationship between energy, mass, heat, and light in a way for these people to understand? There is a wealth of information out there on exactly this transformation but in other materials.

[othertomperson]

Except you're WRONG! Good God is there anyone in here with an in-depth understanding of quantum mechanics and can relay the relationship between energy, mass, heat, and light in a way for these people to understand? There is a wealth of information out there on exactly this transformation but in other materials.

 

OK, what's that half-life of a GPU? Since you are essentially claiming that it is a nuclear reactor and uses itself as fuel, how long is it until half of a GPU is no longer there?

[Lewellyn]

OK, what's that half-life of a GPU? Since you are essentially claiming that it is a nuclear reactor and uses itself as fuel, how long is it until half of a GPU is no longer there?

 

About 1.5-2 years.

[othertomperson]

 

anyone remember the name of the term where air cannot absorb any more heat? hence why we dont have 100mm thick rads.
thermal capacity isnt quite correct, cant remember where i read it but it pertains to heatsinks and cooling obviously.

 

Thermal equilibrium, maybe? If the air is the same temperature as the heatsink heat cannot transfer from one to the other.

[patrickjp93]

OK, what's that half-life of a GPU? Since you are essentially claiming that it is a nuclear reactor and uses itself as fuel, how long is it until half of a GPU is no longer there?

I did not claim that it's a nuclear reactor! What bloody Hell people? 

 

Okay, new example: tungsten lightbulb. Do you know why they fail eventually even if the vacuum inside is not broken? It's because of loss of mass of the filament, leading to a thinner and thinner wire which leads to increased resistance which eventually causes the temperature to rise high enough to melt part of it. The light you see is the loss of mass to heat. 

 

And, the half-life of a GPU depends on how much it was used and in what ways, though it will in all likelihood fail entirely before losing half of its mass. Losing just 3 to 5 atoms at a critical point can cause the chip to fail, though you generally do not lose whole atoms. You lose neutrinos first, eventually building up to the loss of a quark, which then breaks up an electron, which may not do any damage since electricity is being supplied and replenishing the stock, though you might get an error and have to reboot. Then over time the quarks in the neutrons and protons of the atoms begin to destabilize. Losing a neutron is not consequential to electrical performance, but a proton loss means the loss of a silicon atom, devolving it into aluminum, a conductor. When this happens one of your paths begins to leak immensely, usually resulting in chip failure, though sometimes a couple of these can be endured depending on their placement.

 

Every single thing is subject to entropy and energy loss, inevitably leading to mass loss. It will take trillions of years for the universe to devolve into nothing but thermal energy, and the suns will die out and we'll be dead of cold long before that, but it's an inescapable fact of physics that every particle in the universe marches toward its heat death at all times.

[patrickjp93]

 

 

 

Thermal equilibrium, maybe? If the air is the same temperature as the heatsink heat cannot transfer from one to the other.

 

That's not totally true. It's thermal energy, not temperature. Two different materials at the same temperature can have very different thermal energies, hence why salt can make ice water colder than the ice itself in an ice cream maker, leading to frozen ice cream. The temperatures of the fins and the air could be the same, and I'd have to check, but one will absorb thermal energy from the other just the same until the energies are equal.

[Lewellyn]

Okay, new example: tungsten lightbulb. Do you know why they fail eventually even if the vacuum inside is not broken? It's because of loss of mass of the filament, leading to a thinner and thinner wire which leads to increased resistance which eventually causes the temperature to rise high enough to melt part of it. The light you see is the loss of mass to heat. 

 

The reason why a light bulb fails is attributed to the heating and cooling cycles eventually wearing down and breaking the filament. The same is true of graphics cards. That's actually the reason why you can sometimes save one by baking it in the oven so that the metals soften and reconnect.

[Lewellyn]

"Filaments gradually degrade, and lightbulbs darken as tungsten evaporates from the filament surface to be deposited on the inner surface of the glass envelope. In the universally familiar upright bulb envelope (called the A envelope shape by manufacturers), convection currents in the fill gases will carry the tungsten atoms to the top of the envelope to accrete and blacken the bulb. Tungsten evaporates from the filament at higher temperature locations, creating a thermal runaway cycle: since filament hot spots evaporate faster, locally thinned filament locations will develop higher electrical resistances that rise in temperature, thereby reinforcing localized evaporation."

 

 

If anyone would like to read more about light bulbs read here -> http://www.researchgate.net/publictopics.PublicPostFileLoader.html?id=5356c341cf57d7a84f8b4576&key=e0b495356c341433e1

 

God speed, future Edisons.

[patrickjp93]

The reason why a light bulb fails is attributed to the heating and cooling cycles eventually wearing down and breaking the filament. The same is true of graphics cards. That's actually the reason why you can sometimes save one by baking it in the oven so that the metals soften and reconnect.

That too, but it's both. In fact there was a time long ago where I believe 10 copper units were made to exactly 1kg, and many years later they were tested, and it turns out they not only lost mass, but in differing amounts. Over time absolutely everything does, even when stored in a cold vacuum.

http://mentalfloss.com/article/31122/not-so-perfect-kilogram-and-why-metric-system-might-be-screwed

[HippY]

That too, but it's both. In fact there was a time long ago where I believe 10 copper units were made to exactly 1kg, and many years later they were tested, and it turns out they not only lost mass, but in differing amounts. Over time absolutely everything does, even when stored in a cold vacuum.

http://mentalfloss.com/article/31122/not-so-perfect-kilogram-and-why-metric-system-might-be-screwed

Sorry i missed the neutrino escaping part from your link

it says other ref kgs were handled more or the kg got "degassed"

[patrickjp93]

Sorry i missed the neutrino escaping part from your link

it says other ref kgs were handled more or the kg got "degassed"

I'm not going to go digging around for quantum mechanics papers 99.9% of you won't understand. Most degree'd physicists don't understand it and just follow the probabilities.

 

Yes, some were handled, but they were always carefully housed, and the reality is they continue to lose mass even now. I'm not sure what more you people want. This is publicly available knowledge you can find in any GOOD college physics textbook and extrapolate easily. Fundamentals of Physics by Halliday and Resnick is a good book to learn the basics of thermodynamics from, including the surface details of energy loss due to heat and where energy comes from.

[Hatsune Miku 「 」]

I'm not going to believe Patrick until he gives me the law of quantum mechanics that governs his "heat shaving off" quantum process. And he hasn't even replied to that post.

[mr moose]

I did not claim that it's a nuclear reactor! What bloody Hell people? 

 

Because in a nuclear reactor is the only place the processes you are trying to describe could have enough of an effect to be relevant to the current discussion.

 

In short, you are making a mountain out of an unstable and mostly irrelevant mole hill.  There is not enough degradation effects in any chip to convert mass to heat.  The lithography is so small that if that were to be the case the chip would fail within the first few run cycles.

[HippY]

I'm not going to go digging around for quantum mechanics papers 99.9% of you won't understand. Most degree'd physicists don't understand it and just follow the probabil

my problem is that the source you showed does not prove you, as it mention no quantum physics

[patrickjp93]

Because in a nuclear reactor is the only place the processes you are trying to describe could have enough of an effect to be relevant to the current discussion.

 

In short, you are making a mountain out of an unstable and mostly irrelevant mole hill.  There is not enough degradation effects in any chip to convert mass to heat.  The lithography is so small that if that were to be the case the chip would fail within the first few run cycles.

You are totally incorrect because EVERY process changes matter to heat, even if in super tiny amounts which take years for you to notice.

[Constantine]

What the fuck is happening in this thread...

[patrickjp93]

my problem is that the source you showed does not prove you, as it mention no quantum physics

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

read the snippet on transformation and combine that with what I have been saying. If you want the math we can go there, but almost no one on this website can grasp it well, much less well enough to teach you why it's right.

 

Every single process which transforms energy to another will generate heat (inefficiency of the process), but heat can also be defined in terms of the kinetic motion of particles (heat of a gas is the sum of the momentums of each molecule). The hidden detail not covered when you learn about gas laws is that heat and energy are lost to radiation. That radiation is the Shrodinger wave function of that atom/particle ever so slowly collapsing, implying loss of all energy in that particle. The only way that happens is by total loss of mass, because mass can be transformed to kinetic energy which can be used to create light or electricity.

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

http://web.mst.edu/~gbert/basic/thermo.html

 

If you think you can handle the mathematical notations, David Griffith's Introduction to Quantum Mechanics also fleshes out the natural consequences of thermodynamics and atomic rotting (neutrino orbit degradation and separation as I described earlier).