MSI r9 290x Gaming is warming up

post dated February 19 13.51 ....... it says how much it was .....
my home-grown upgrade, as you put it, will now give a head start to your spacious buildings. But when I need a second computer, where I will stick the components from the old system, then I will buy a case for the current one, but for now everything is okay.

ps because the temperature is 68 - 85 degrees, depending on the running application, the temperature is not critical for this chip.

KoncentRat to
your spacious enclosures ... These enclosures are not mine, but engineered and commercially available ...

SUSUL1

Comrade, you should at least read the beginning. for a modern card, 90 degrees is normal operating temperature. my HD4850 ​​in the reference still lives and is not going to burn, and even she was warming up with me and up to 115!

90 degrees is clearly not the normal operating temperature of a video card, whatever it is, as you have already written that you have such a temperature, it does not mean that this is the norm.
The fact that some manufacturer is in pursuit of making the most productive card so increases the frequencies that the CO cannot cope is another question.
The normal operating temperature is 60-85 degrees, but not as much as 90 and even more so not 100 degrees (such a temperature clearly does not contribute to normal operation, and the life cycle of the card), although for AMD 100 degrees is in the order of things, the same 7990.

For example, at idle 30 at a load of 60-65, this is despite the fact that the cooler works no more than 50%.

saa0891
That stubborn ...
Would you ever wondered what that piece of iron such =)
http://www.overclockers.ua/video/ati-radeon-hd4850-geforce-8800gts/all/

From what is this fright 60C "optimal temperature ?! These are our wishes and stereotypes made from the CPU and hammered into the head by marketers (they also need to put forward some competitive advantages, and brag about overclocking potential), where the chip area and architecture features do not allow the stone to warm up properly.

saa0891
90 degrees is clearly not a normal operating temperature of a video card, whatever it is, as you have already written that you have such a temperature, it does not mean that this is the norm ...
Normal operating temperature is 60-85 degrees, but not like 90 and certainly not 100 degrees (such a temperature is clearly not conducive to normal operation, and the life cycle of the card), although for AMD and 100 degrees in the order of things, the same 7990.

And nevertheless, AMD claims that the maximum for their latest graphics chips is a temperature of 94 degrees. And the cards, by default, begin to reduce the GPU frequency when this temperature is reached. And manufacturers give a guarantee for such cards.
From this we can conclude that 94 degrees, although quite high, is quite normal for r9 290 temperature. And, according to the plans of manufacturers, at least the warranty period for a card with such a graphics processor should be plowed up.

SUSUL1

That’s stubborn ...

Apparently you’re especially stubborn if for you 90 or even more than 100 degrees everything is within the normal range.

You would even ask what kind of piece of iron it is =)

Why should I be interested if you wrote me the temperature.
Although not so long ago there was 7990 which heats up more than 100 degrees, it is clear that this is still a two-chip, but such a temperature is clearly not good for anywhere.

How frightened is 60C this "optimal temperature?!"

Read carefully, I wrote 60-85, it of course depends on a specific card and CO, and why is this the optimal temperature, because it is less than 90 degrees, which is a critical or undesirable temperature for most iron components, be it a CPU or GPU, and when these components work in this mode, this obviously does not have a positive effect on their work and life span.

These are our wishes and stereotypes made from the CPU and hammered into the head by marketers (they also need to put forward some competitive advantages, and boast about overclocking potential), where the chip area and architecture features prevent the stone from heating properly.

What wishlist, I wrote in just 30 at a load of 60-65, depending on the game.

ramil.bagizov

And manufacturers give a guarantee for such cards.
From this we can conclude that 94 degrees, although quite high, is quite normal for r9 290 temperature. And, according to the plans of manufacturers, at least the warranty period for a card with such a graphics processor should be plowed up.

So my friend sits like on a powder keg with 7990 and thinks when he will carry it under the guarantee, and really hopes that the guarantee will not end earlier than the card.
Again, the question will change his card if something happens to her from overheating, if it were all so simple, then there would be no problems, then prove that you are not a camel.

saa0891
Where did you get the information that for most components 90 degrees is a critical temperature? And what is included in this majority?

On youtube, you can find videos with the repair of video cards. And there, for example, to remove the chip from the card, heating is applied to 170-190 degrees.

ramil.bagizov

Where did you get the information that for most components 90 degrees is a critical temperature? And what is included in this majority?

Read any reviews, 60-85 degrees is the norm, when the temperature is higher this is already a reason to think seriously.

I have not read this entire wall, but it seems to me guys that you do not fully understand exactly how the temperature affects the video card in general and its components in particular.

Firstly, as for reference cards with reference cooling, they almost always have the same CO for all elements, and the hottest of them is almost always the power supply.

I would also like to say a few words about what most often fails from prolonged operation at near-extreme temperatures:
As a rule, this is not a GPU. The 2 most "popular" ways of failure of a video card are either VRM burnout, or memory problems (which mainly cause characteristic artifacts even in the BIOS).

As for what temperature is optimal, what is acceptable, where is it worth worrying about and where not.
For starters, the temperature factor is not the only one. Of primary importance is the combination of the current passing through the elements and their operating temperature. You probably know about such a physical phenomenon as diffusion, it is this that causes the degradation of semiconductor crystals (GPU), which in turn, over time, can lead to the fact that the card will no longer keep stable the frequencies that it has always kept.
This is due to the mutual penetration of materials into each other, from which the conductor and insulator inside the crystal are made. As a result, resistance increases, the ability of a semiconductor to pass from one state to another (conducting / non-conducting) with a certain frequency decreases, and so on.

Optimum temperature - the lower the better. The lower the temperature, the lower the resistance, the lower the heat generation, the less diffusion, the longer the service life and the higher the overclocking potential.
(this is why I personally do not recognize air cooling at all, since by my standards it is not capable of providing satisfactory temperature modes of operation)

DartMaul
Interestingly, diffusion is as terrible for the 55nm process as for 22nm?

ramil.bagizov
If we don't go too far, then the subject in the stock cooler has the same 90C in normal operation.
All the same, the permissible temperature is determined by the manufacturer. Special electronics keep temperatures below 300 degrees and this is considered the norm for it. And if the manufacturer claims that 90 degrees is working. then I see no reason to doubt his words, there is always a margin for the performance of the part. Overclocking is no longer a regular mode of operation. But for this there are third-party manufacturers who mold "supercoolers"

SUSUL1
Interestingly, is diffusion as terrible for the 55nm process as for 22nm?
The fewer those the process, the less diffusion leads to degradation.
The point is that voltage + temperature catalyze diffusion, and exponentially. If, roughly speaking, the processor at a voltage of 1.15 reaches 100 degrees and the protection system is triggered, which knocks down the system, the processor most likely will not have anything. And if the voltage was 1.5 or higher, then he will receive permanent degradation, and very noticeable. (Yes, I checked. I somehow froze the refrigerant during running wprime1024 at 3930k. There was something in the region of 5250 @ 1.6v. As a result, I cracked the chip for $ 600)

DartMaul
Well, in order to bring it to this, he needs to turn off all protective technologies. It looks like this is why AMD has implemented a non-switchable overheating protection in processors of a modular architecture.

Here you are wrong. From time immemorial it has been known that if there is no system that turns off the system, this leads to burnout of the core. There are many old videos comparing old Intel and AMD processes. The cooler is removed from both, the intel is cut off AMD burns out with smoke and flame.

Intel has several systems, one that throttles the percentages, it can be turned off, the second that mutes the system, it cannot be turned off. One way or another, I see no reason to disable any of them.

If on that process my system would not have been cut down, it would have died, but it just degraded so much. There was too fast heating, so the system did not have time to work fast enough. Thermal breakdown.

DartMaul
You need to be arrested and fined for "cruelty to computers" =)
Burning processors this way is harsh! Indeed, the protection system has inertia and it is quite possible to burn the chip before it is triggered ...
Was the processor subjected to such violence in order to obtain a record in frequencies or to see "when will it die"?

SUSUL1
Was the processor subjected to such violence in order to obtain a record in frequencies or to see "and when will it die"?
I kind of wrote above that I was driving wprime. Accordingly, the goal was to get the result in wprime.

You need to be arrested and fined for "cruel treatment of computers" =)
I give him a decent life and death with honor! I do not admit my guilt, I refuse to cooperate with the investigation.

Oddly enough, out of a dozen ways to kill your computer more efficiently .. Overheating (oddly enough) is only in the last place (I'm talking about overclocking). The leader in this category is the power supply. But the processor cooling is in the second place. And the culprit is, as you would expect, THE USER ITSELF .. beloved in person. The main causes of fatal breakdown are
Power supplies (26%)
Defective components and neglect of compatibility issues on the part of the user (23%)
Incorrect assembly (15%)
Components responsible for normal system cooling (13%)
Lightning or static electricity (10%)
USB device failure during connection (6%)
User carelessness (3%)
Attempts to overclock (2%)
so .. you need to be arrested and fined for "mistreating computers" =) this is no longer a verdict .. this is a dogma.

Rewers v.01
Defective components and neglect of compatibility issues on the part of the user (23%)
Although offtopic, is it possible in more detail, how is this possible ?!
After all, all pieces of iron have compatibility standards, both there and back. And pieces of iron of different standards do not physically fit into connectors that are not intended for them. And the defective components are already the fault of the manufacturer, not the user.

This is from the point of view of an ordinary user ...
In all other respects I agree, the power supply unit was killing my motherboard, I myself pushed the Molex upside down into the SDrom, inadvertently dropped the screwdriver on the open circuit with the PC turned on and forgot to smear thermal paste ...
But about compatibility components have never been conceived.

If you put a 125Wt processor on a 90W motherboard?

SUSUL1
Using the example of an extract from the opinions of users and correspondence ... "My friend killed the system. The reason is a faulty power supply. I advised him to test the memory in another system. Do you know how it ended? Another killed system. I thought he was just not careful enough. Chided him and took his memory home to test on my PC. Guess what was the end of the story? I killed my system ... "Or ...
If your system has a power supply failure and you need to check which components survived, do not be lazy. Before installing them on another system, spend some time researching the compatibility issue. If you do not have a system at your disposal on which you could test the components you are interested in with 100% confidence in their compatibility with the components of the stand, contact the specialists for help, for example, the store where you purchased these components. There are also such options ... The overwhelming majority of users are well aware that nothing should be touched in the system, and even more so try to make any changes, without first turning off the power. Nevertheless, there are systematically those who do just that.
"Once, I needed to plug in the hard drive. The computer was already on and I thought it would take a very long time to shut down. Unfortunately, the Molex connector on the hard drive was upside down and when I tried to plug it in, I lacked precision. Sparks fell ..." Out of surprise, I dropped the disc onto a brand new video card. Of course, it died, as did the disc itself ... "There are many such examples and they all take place .. The
survey was carried out by the staff of the PCstats Newslette website and based on this they deduced these figures ...

Rewers v.01 Can you
guess what was the end of the story? I killed my system ... "Or ...
What kind of memory was that?