SweClockers: Geforce GTX 590 burns @ 772MHz & 1.025V

[taltamir]

Nope, the smaller opening will allow the same air volume to pass at the same time as a bigger opening but at a higher speed. That’s a fact (Physics)

The opening of the rear panel in HD6990 is smaller than the front end. By your theory HD6990 will exhaust more air volume inside the case than outside through the smaller backpanel opening. Again that’s not the case, both ends will exhaust the same amount of air volume, the difference is that in the smaller opening the air speed will be higher than the larger opening and that's a fact (Physics).

Bernoulli's Theorem
http://en.wikipedia.org/wiki/Bernoulli%27s_principle

is the card sufficiently sealed for Bernoulli's principle to apply?
This isn't exactly a tube which narrows while having a constant pressure applied on one side.

 

[kevinsbane]

Nope, the smaller opening will allow the same air volume to pass at the same time as a bigger opening but at a higher speed. That’s a fact (Physics)

The opening of the rear panel in HD6990 is smaller than the front end. By your theory HD6990 will exhaust more air volume inside the case than outside through the smaller backpanel opening. Again that’s not the case, both ends will exhaust the same amount of air volume, the difference is that in the smaller opening the air speed will be higher than the larger opening and that's a fact (Physics).

Bernoulli's Theorem
http://en.wikipedia.org/wiki/Bernoulli's_principle

Unfortunately, that's for an ideal fluid, assuming no incompressibility. Air is far from an ideal fluid, and is very compressible.

Bernoulli's (in the manner that you want it to apply) will apply to this case if and only if the air is incompressible, and the volume of air passing through is constant. The volume of air passing through the vent, however, is not a constant.

See this graph below:
http://www.scythe-eu.com/typo3temp/pics/5e76ef7072.jpg
This graph represents the airflow of a gentle Typhoon fan, with varying amounts of back pressure. Obstructions in airflow result in increased backpressure - the GTX 590 cannot exhaust as much air out the back. Something easier to test would be to note how much air is blown by your heatsink fan in free air, versus mounted on your heatsink. It will blow much more air in free air.

The worst part of it is that the blower fan used by the HD 6990 is actually more suited to exhausting air than the GTX axial fan: blower fans produce a higher static pressure, allowing it to overcome backpressure from obstructions more easily. Of course, it's also a lot louder.

Edit: Although it's true that air is compressible, in this case I realised that it doesn't matter. Compressibility isn't the reason why the GTX cannot exhaust as much as a HD 6990.

 

[railven]

Nope, the smaller opening will allow the same air volume to pass at the same time as a bigger opening but at a higher speed. That’s a fact (Physics)

The opening of the rear panel in HD6990 is smaller than the front end. By your theory HD6990 will exhaust more air volume inside the case than outside through the smaller backpanel opening. Again that’s not the case, both ends will exhaust the same amount of air volume, the difference is that in the smaller opening the air speed will be higher than the larger opening and that's a fact (Physics).

Bernoulli's Theorem
http://en.wikipedia.org/wiki/Bernoulli%27s_principle

You are misusing that theorum because the designs aren't solid. There are various areas where leakage could occur causing the applied pressure to change. The airflow would be distorted by the change in direction caused by the shifting.

What you are trying to argue is that the situation is fluid and there are no alleviations to pressure thus volume will increase as pressure is applied. Think of a garden hose. If you pop a few holes into the hose suddenly the pressure is lost and the volume decreases at the end where you cupped your fingers.

EDIT: Haha two people beat me to it.

 

[thilanliyan]

Nope, the smaller opening will allow the same air volume to pass at the same time as a bigger opening but at a higher speed. That’s a fact (Physics)

The opening of the rear panel in HD6990 is smaller than the front end. By your theory HD6990 will exhaust more air volume inside the case than outside through the smaller backpanel opening. Again that’s not the case, both ends will exhaust the same amount of air volume, the difference is that in the smaller opening the air speed will be higher than the larger opening and that's a fact (Physics).

Bernoulli's Theorem
http://en.wikipedia.org/wiki/Bernoulli%27s_principle

Umm, no. That depends on pressures on the other side of the vent and the pressure being applied by the fan.

 

[AtenRa]

is the card sufficiently sealed for Bernoulli's principle to apply?
This isn't exactly a tube which narrows while having a constant pressure applied on one side.

It has an opening that narrows through the backplate and a constant pressure from the single fan.

 

[TerabyteX]

How is possible that he claims that the same amount of air will pass through the smaller GTX 590 exhaust? With a standard fan that's more silent than the HD 6990 which uses a blower fan which which is more suitable for such work? And what does the Bernoulli's Theorem has to do with air flow?

 

[railven]

How is possible that he claims that the same amount of air will pass through the smaller GTX 590 exhaust? With a fan that's more silent than the HD 6990? And what does the Bernoulli's Theorem has to do with air flow?

Because its PhysX and undisputable. Haha. Let's start using gravitional pull while we're at it. Was the moon out during the 6990 testing? There you go!

My counter arguement would be time frame. Where the snapshots of each card taken at equal time intervals? That arguement has more merit than what he is trying to pass of. But, I'm a red ant - what do i Know

 

[AtenRa]

How is possible that he claims that the same amount of air will pass through the smaller GTX 590 exhaust? With a fan that's more silent than the HD 6990? And what does the Bernoulli's Theorem has to do with air flow?

You guys mistaken,

I claim that both ends of EACH card will exhaust the same air volume regardless of opening size.

HD6990 will exhaust the same air volume from the back plate as in the front end of the card.

Same for GTX590, same amount of air volume will pass through the back plate exhaust holes and in the front end holes even if the backplate has smaller size.

Bernoulli's principle can be applied to fluids and gases.

 

[taltamir]

Unfortunately, that's for an ideal fluid, assuming no incompressibility. Air is far from an ideal fluid, and is very compressible.

Bernoulli's (in the manner that you want it to apply) will apply to this case if and only if the air is incompressible

wikipedia says that it is a good aproximation for air at "low mach speeds". Only in "high mach speeds" does it become inadequate for air... And mach = the speed of sound.

How is possible that he claims that the same amount of air will pass through the smaller GTX 590 exhaust? With a standard fan that's more silent than the HD 6990 which uses a blower fan which which is more suitable for such work? And what does the Bernoulli's Theorem has to do with air flow?

if you were to put your mouth to a tube that is 1 inch in diameter and blow, and halfway through the tube will get narrower, then the same amount of air will pass through the tube, it will just go faster on the narrower part... He is saying that the card is a tube.

mmm... but wait, that doesn't seem right. wouldn't you need to increase the pressure on the "input" side of the tube for that? if size of tubing had no effect on amount of force needed to push a fluid through then water limiter valves wouldn't work at all.

 

[TerabyteX]

Because its PhysX and undisputable. Haha. Let's start using gravitional pull while we're at it. Was the moon out during the 6990 testing? There you go!

My counter arguement would be time frame. Where the snapshots of each card taken at equal time intervals? That arguement has more merit than what he is trying to pass of. But, I'm a red ant - what do i Know

Oh yes, it uses CUDA to calculate the case air pressure and air flow and uses PhysX to animate the airflow and air collision through the GTX 590 exhaust.

 

[bryanW1995]

Unfortunately, that's for an ideal fluid, assuming no incompressibility. Air is far from an ideal fluid, and is very compressible.

Bernoulli's (in the manner that you want it to apply) will apply to this case if and only if the air is incompressible, and the volume of air passing through is constant. The volume of air passing through the vent, however, is not a constant.

See this graph below:
http://www.scythe-eu.com/typo3temp/pics/5e76ef7072.jpg
This graph represents the airflow of a gentle Typhoon fan, with varying amounts of back pressure. Obstructions in airflow result in increased backpressure - the GTX 590 cannot exhaust as much air out the back. Something easier to test would be to note how much air is blown by your heatsink fan in free air, versus mounted on your heatsink. It will blow much more air in free air.

The worst part of it is that the blower fan used by the HD 6990 is actually more suited to exhausting air than the GTX axial fan: blower fans produce a higher static pressure, allowing it to overcome backpressure from obstructions more easily. Of course, it's also a lot louder.

interesting comment about the fan. I wonder if later or custom-designed gtx 590 fans will use that type of fan.

 

[TerabyteX]

You guys mistaken,

I claim that both ends of EACH card will exhaust the same air volume regardless of opening size.

HD6990 will exhaust the same air volume from the back plate as in the front end of the card.

Same for GTX590, same amount of air volume will pass through the back plate exhaust holes and in the front end holes even if the backplate has smaller size.

Bernoulli's principle can be applied to fluids and gases.

Ok, that might be true if the GTX 590 plastic shroud is totally sealed. Its standard fan isn't the bes approach for that work, a blower fan is better for that, plus in overall, the HD 6990 rear exhaust will expell more air than the GTX 590 rear exhaust.

 

[thilanliyan]

It has an opening that narrows through the backplate and a constant pressure from the single fan.

Yeah but you're comparing totally different designs to begin with. The heatsinks aren't the same, the openings aren't the same, and the fans are not the same. The air will take the easiest route. If it is not completely sealed (which it is not), it will not necessarily go through the vent.

You guys mistaken,

I claim that both ends of EACH card will exhaust the same air volume regardless of opening size.

HD6990 will exhaust the same air volume from the back plate as in the front end of the card.

Same for GTX590, same amount of air volume will pass through the back plate exhaust holes and in the front end holes even if the backplate has smaller size.

Bernoulli's principle can be applied to fluids and gases

This is definitely wrong. The air will take the easiest route out (ie. wherever there is the lower pressure), whether it is the front or rear vent. The air won't necessarily be split 50-50 between the two vents...if there is a larger pressure obstacle at one vent, more of the air will go to the other vent (not a huge difference since the fan is trying to push both ways but there will be a difference).

AtenRa, I think you can see by the amount of people who are providing evidence contrary to what you are saying that you may want to step back and take a look at it again.

Back on topic, there is no "extra" fan in the 6990 system.

 

[bryanW1995]

You guys mistaken,

I claim that both ends of EACH card will exhaust the same air volume regardless of opening size.

HD6990 will exhaust the same air volume from the back plate as in the front end of the card.

Same for GTX590, same amount of air volume will pass through the back plate exhaust holes and in the front end holes even if the backplate has smaller size.

Bernoulli's principle can be applied to fluids and gases.

so if nvidia covered up 75% of the backplate it would be the same? What about 90%? 99%? The point of the rebuttals was that this is not an accurate representation.

 

[cusideabelincoln]

AtenRa you're trying to apply hard theories to the real world where there are considerably more variables to consider than theories take into account. People have already mentioned the other possible variables.

1. Look at the GTX 590 right next to the SLI connector. There is a vent there. Air is going to leak out of the vent instead of going out through the backplate.

2. Compare the rear end of the 590 to the one of the 6990. The rear end on the 590 is way more open, so air is going to have a much easier time getting dumped into the front of the case. The 6990 is partially blocked by the shroud, and that is going to disperse airflow.

3. Someone mentioned the size of the cards, and that is pretty valid consideration. The 6990 is bigger. The PCB itself acts like a heatsink, so more heat would disperse above the card via the PCB. Just look at the 450W vs 590 thermal shots again. The area above the 450W 6990 (CPU and mobo) are giving higher readings than the standard 590.

4. I'm sure there are other variables to take into account. But whatever that may be, there's simply more to this than higher watts = higher temps. And I do not think an extra case fan is your missing link.

5. This heatsink and cover is not a fully enclosed system. Air is going to leak. Air will even leak at the fan.

At 450W the HD6990 should exhaust higher temperature air inside the case than GTX590 but as we can see from the pictures thats not the case. There mast be a fan working with the HD6990 in order to have those lower temperatures in that side of the PC case and it has nothing to do with the GTX590 exhaust.

I would think IF there was a fan working on the 6990 and not for the 590 then there would be a MUCH bigger difference in the results.

 

[kevinsbane]

interesting comment about the fan. I wonder if later or custom-designed gtx 590 fans will use that type of fan.

I don't think so, personally. Blower fans generally give higher static pressure for the same volume of air blown. It's useful if you need to overcome back pressure or obstructions, but you're better off just using another fan to increase airflow directly.

 

[kevinsbane]

You guys mistaken,

I claim that both ends of EACH card will exhaust the same air volume regardless of opening size.

HD6990 will exhaust the same air volume from the back plate as in the front end of the card.

Same for GTX590, same amount of air volume will pass through the back plate exhaust holes and in the front end holes even if the backplate has smaller size.

Bernoulli's principle can be applied to fluids and gases.

I'm nitpicking here, but.... gases are fluids

 

[insurgent]

I learned a couple of things from kevinbane's post, thanks!

 

[taltamir]

I'm nitpicking here, but.... gases are fluids

yes.
Fluids: A liquid or a Gas
Solids: A crystal lattice or a glass

 

[bryanW1995]

I'm nitpicking here, but.... gases are fluids

Normally I'd agree that you're nitpicking, but in this case he is stubbornly holding to his belief in the face of mounting evidence to the contrary. The only thing that he has to hang his hat on is "my scientific knowledge is correct and the rest of you don't know what you're talking about". clearly if he is confused about whether gases are fluids then it increases the likelihood that he just heard this one day in his chem 101 lecture and suddenly became an expert, rather than that he is a tenured professor in fluid dynamics at cal-tech.

 

[MrK6]

I don't think so, personally. Blower fans generally give higher static pressure for the same volume of air blown. It's useful if you need to overcome back pressure or obstructions, but you're better off just using another fan to increase airflow directly.

Alright, I'm going to just start with that I didn't read through all of what's going on, but this post is dead on. The 6990's radial/blower fan is superior in cooling performance compared to the axial in the GTX590, especially when used for perpendicular flow. I'd bet the airflow coming out of the 6990 is vastly superior to that of the GTX590's, and that's why you see such lower case temperatures (I realize what atenra was getting at using the thermal imaging). All that airflow keeps the heat moving and exhausting (not just out the back of the card, but also within the case) much better.

 

[Keysplayr]

Yeah, I didn't read the last few pages either. So what's the subject of the conversation guys? It's kind of a blur.

 

[3DVagabond]

But they don't use the same fan design. Does anyone know the CFM for the fans on each card?

The GTX 590 fan/design is doing something poor, using the simple knowledge that more watts == more heat, but looking at the Anandtech reviews, the temps are within a few degrees in temperatures from the 6990 OC and the GTX 590 stock:

http://www.anandtech.com/show/4239/nvidias-geforce-gtx-590-duking-it-out-for-the-single-card-king/16

The obvious trade off is sound:

My theory would be the lax fan design on the GTX 590 is letting hot air leak out into the case versus being exhaused faster through the rear. This causes more hot air to build up inside the case, and over time the temperatures increase through out as shown.

Agreed. Their coolers have completely different airflow designs. More air is being pushed to the sides by the blower AMD uses. All else being equal, this would make the exhaust cooler for the 6990. To a point, even increasing the heat generated by the card, it would still have a cooler exhaust.

 

[GaiaHunter]

Yeah, I didn't read the last few pages either. So what's the subject of the conversation guys? It's kind of a blur.

It is a discussion about the validity of the thermal images based on the fact that the lower left portion of the image (or bottom front of the case) is always cooler on the 6990 vs the stock GTX590, even when the 6990 is dissipating 450W.

Interesting technical discussion about airflow, heat dissipation, card cooling, etc.

 

[Seero]

Yeah, I didn't read the last few pages either. So what's the subject of the conversation guys? It's kind of a blur.

Round up:
Some believe that VRM used on 590 is problematic. This theory was supported by, a) hot temp, b)Nvidia cheated by reporting the temp of cores, and c)590 dies upon improper OC.
Well, the first 2 claims was clearly busted, although some may still believe other wise after a clear example of the existence of a HDD fan when the IR graph of 6990 was taken.

Hardware.fr corrected the 590 readings by redoing the IR graph, showing that the 2 core temps are 80 and 84 respectively, while the readings is 90 and 96, which is understandable as core temp > surface temp.

Lets move to VRM. Hardware.fr clearly didn't have efficient cooling when 590 is in place. In fact, the cooling was bad on 6990 too. Knowing that the video cards vent off from both sides, having a bunch of wires blocking one side is asking for trouble. Clear those cables, add a exhaust fan there, and redo the IR for both of them and we will have a clear comparison on the 2 mini heaters.

If VRM is the focus, then we should focus the readings on the VRMs of 2 cards. However, the amount of heat generated by the 2 VRMs should be very similar as the intake of electricity is similar. While 590 place its VRMs at the center, 6990 spread its VRM for each CPU far apart. In engineers PoV, it is good, but the real question is, does the temp of VRM matters?

VRM can handle 100 degree and most likely to operate at that temp on high-end VC. 590 choice to put all of them together and away from GPU modules and have an independent cooling modules to radiates heat, leaving the center of the heat relatively hot. 6990 on the other hand cools the VRM with the associated GPU. At the end, 6990 requires more cooling. Which leads to another question, are the VRMs of 590 too hot?

Well, ofter all these debates, the answer is "It won't handle 1.2v." Which is again some random quotes off the INTERNET. Yes "6990 VRM suck because won't handle 120v." So what?

By default, 590's default voltage is .91v to .96v. Operate your card in a properly cooled environment, stock settings, stable electricity and fry it and we will have a case. We don't have a case.

Dig deeper to argue the fact that it runs hotter than VRM used by 6990. Failed. See above.

For those who self proclaimed to be an expert. Build a VRM used by 590, run it under 375watt for 3 months and see if it fries. If it does, than you have a case. Note that going over 375 means KILLING, BUT NOT LIMITED TO, PSU, MOTHERBOARD, CPU, HDD, RAM AND ANY OTHER THINGS CONNECTED TO THESE.

Of course, this may sounded like a FUD from me. Take it from this:
AMD's support on 6990

WARNING: AMD graphics cards are intended to be operated only within their associated specifications and factory settings. Operating your AMD graphics card outside of specification or in excess of factory settings, including but not limited to overclocking, may damage your graphics card and/or lead to other problems, including but not limited to, damage to your system components (including your motherboard and components thereon (e.g. memory)); system instabilities (e.g. data loss and corrupted images); shortened graphics card, system component and/or system life; and in extreme cases, total system failure. AMD does not provide support or service for issues or damages related to use of an AMD graphics card outside of specifications or in excess of factory settings. You may also not receive support or service from your system manufacturer.

DAMAGES CAUSED BY USE OF YOUR AMD GRAPHICS PROCESSOR OUTSIDE OF SPECIFICATION OR IN EXCESS OF FACTORY SETTINGS ARE NOT COVERED UNDER YOUR AMD PRODUCT WARRANTY AND MAY NOT BE COVERED BY YOUR SYSTEM MANUFACTURER’S WARRANTY.

Just in case some self proclaimed Enthusiast who believes AMD supports 450Watt of rampage because there is a switch on the card that allows them to do it, read the rest of the link above.
BIOS1: the factory-supported Performance BIOS
BIOS2: Extreme Performance BIOS

It isn't for human to understand, but for lawyers to cover their ass, BIOS2 is NOT A SUPPORTED SETTING! Here is another proof
Kitguru.net: AMD’s official statement over HD6990 warranty

Each AMD Radeon HD 6990 graphics card come with a hardware switch that, if toggled, activates a pre-flashed “overclocked” BIOS. Toggling this switch to the overclocked setting has the same functionality and effect as common software overclocking and overvolting utilities, will run the HD 6990 outside of the warrantied specification, and will require more electrical current than some power supplies are designed to deliver.
As with software overclocking, damages caused by use of an AMD graphics card outside of the warrantied specification are not covered under AMD’s product warranty. This is in line with industry practice relating to overclocking or running a product out of warrantied specification.
To ensure end users understand this before using the switch to activate the overclocked BIOS, every HD 6990 graphics card has a yellow warning sticker covering this switch. The switch is not user accessible without removing this sticker. The HD 6990 user guide and support website each contain an overclocking warning message similar to the AMD Overdrive software warning.

If you think you are covered in anyway by overclocking/overvolting your 6990 because there is a switch, YOU ARE MAKING A MISTAKE!

XFX is the only manufacturer to offer warranty on the 6990 OC mode, but shall any other part of the system that got killed due to OC, you are screwed as the act of OC and running things off spec voids warranties. If you are still arguing, than don't even think of overclocking. There is a gap between default and spec, and OC lays inside this zone. Fail to understand or see this zone means you are likely to kill your PC. Consider yourself LUCKY if the video card is the only thing that got killed. I personally have seen PSU breathing fire, funny color smokes, resistor popping out of the circuit like a rocket, etc etc.

If you don't overvolt your stove for fun, than don't do it on your computer unless you know exactly what you are doing and its ramifications. It can be dangerous. It may not complicated, but simply dangerous. That is why people needs a license for wiring your house.

The following is a video of overvolting the computer. You tell me if warranty should covers it or not, or if there is anything wrong with its quality.
http://www.youtube.com/watch?v=TjHOUonZO6U
What about this one?
http://www.youtube.com/watch?v=sRo-1VFMcbc
Both are results of overvolting, no?