Do differences in "safe level" GPU temperatures affect lifespan?

[Guest]

So, when it comes to cooling, having a silent system is always nice but if you ask me temperatures are always the first priority. 80s and 90s (Celsius) are regarded as "safe level temps" but if you keep the GPU cooler than that it'll likely last longer. What I want to know is: will there be a difference in lifespan of cards that under load sit at 70, 60 or let's say a watercooled 50 degrees? When you get into that safe zone do differences still matter in terms of longevity? I'm currently choosing a case and air cooling the GPU is my No 1 priority so I'm kind of having a tough choice...
Thanks for any information shared!

 

[Guest]

Very generally speaking re: all electronics pretty much....heat is enemy #1 and it doesn't have to be excessive heat necessarily. 'Moderate' heat over time can take its toll.

Semiconductors will last almost indefinitely under 100C junction temperature. When electronics fail, it is usually due to something else failing first, that other thing usually being capacitors either on the card/motherboard itself or the power supply that feeds it.

My HD5770 has probably seen something close to the equivalent of three years of 24/7 full load (much of it from sitting inside a game for free XP and attendance bonuses) and still works perfectly fine using the original Powercolor cooler.

So, you want to say that motherboard fails sooner than graphics card?
The card most certainly can't melt at 100 degrees and semiconductors should be able to take it without issues. What I'm not sure about is whether it can damage them (the semiconductors) over a longer period, but then why didn't it kill them right away?

 

[Suzuki@1473780084@1473780163]

Voltage before temperatures.

No, they're expected to reach a certain temperature until the GPU driver tells your PC to shut down. You can't harm a GPU unless you intentionally overvolt. Lifespan won't be affected because you'd alreayd have upgraded.


All the best!

 

[Guest]

Voltage before temperatures.

No, they're expected to reach a certain temperature until the GPU driver tells your PC to shut down. You can't harm a GPU unless you intentionally overvolt. Lifespan won't be affected because you'd alreayd have upgraded.


All the best!

Yeah, about the voltages: I though voltages cause higher temps which in turn shorten the lifespan. If you have crazy cooling setup, can voltages still do anything?

 

[Suzuki@1473780084@1473780163]

Yes and no. If you overclock and your voltage is high, but temperature is low, that's worse than high temperature but low voltage (normal operating voltage).

Temperatures certainly don't paint the whole picture. With that said, overclocking isn't going to noticeably affect lifespan either, but intentionally overvoltaing to the point it stops working, yes that's bad.

NVIDIA (don't know about AMD) are clearly ripping people off by enabling a performance temperature limit in the driver, shady move.

This has caused a lot of confusion regarding temperatures for GPU's, people thinking 80 C is not safe, and so on. They're confusing safe temperatures (driver automatically shutting down PC if it's no longer safe) with throttling (lowering performance).

 

[Guest]

Yes and no. If you overclock and your voltage is high, but temperature is low, that's worse than high temperature but low voltage (normal operating voltage).

Temperatures certainly don't paint the whole picture. With that said, overclocking isn't going to noticeably affect lifespan either, but intentionally overvoltaing to the point it stops working, yes that's bad.

NVIDIA (don't know about AMD) are clearly ripping people off by enabling a performance temperature limit in the driver, shady move.

This has caused a lot of confusion regarding temperatures for GPU's, people thinking 80 C is not safe, and so on. They're confusing safe temperatures (driver automatically shutting down PC if it's no longer safe) with throttling (lowering performance).

Well, increasing the voltage to the point where it stops working is obviously not a good thing. But will increasing it so that it still works properly reduce the lifespan in a longer period?

 

[Suzuki@1473780084@1473780163]

Define longer period, it's not possible to answer that. GPU's have no problems operating at high temperatures, the driver shut downs are at around 100 C, so everything below 100 C is actually safe, so everyone who says otherwise are wrong. The problem is that they implemented performance drops too but it's a desperate attempt to get more sales because the need for better performance is technically still bigger than a set clock speed without a temperature limit. That's how I see it anyways.

NVIDIA GPUs are designed to operate reliably up to their maximum specified operating temperature. This maximum temperature varies by GPU, but is generally in the 105C range (refer to the nvidia.com product page for individual GPU specifications). If a GPU hits the maximum temperature, the driver will throttle down performance to attempt to bring temperature back underneath the maximum specification. If the GPU temperature continues to increase despite the performance throttling, the GPU will shutdown the system to prevent damage to the graphics card. Performance utilities such as EVGA Precision or GPU-Z can be used to monitor temperature of NVIDIA GPUs. If a GPU is hitting the maximum temperature, improved system cooling via an added system fan in the PC can help to reduce temperatures.


That was 6 years ago, and today their "unofficial" limit is 80 C if you want any good performance, welcome to marketing gibberish.

 

[InvalidError]

The card most certainly can't melt at 100 degrees and semiconductors should be able to take it without issues. What I'm not sure about is whether it can damage them (the semiconductors) over a longer period, but then why didn't it kill them right away?

Manufacturing and handling defects aside, semiconductors don't die unless their safe operating area is exceeded and semiconductors won't exceed their SOA unless their specifications are ignored (overclocking, overvolting, poor design, etc.) or something else fails and cause the specifications/SOA to be exceeded.

The only difference between 80C and 100C is that you have that much less error margin in case something goes wrong. For example: drivers crash and leave the GPU in a limbo state where it draws 20W more than it should. That 20W bump might be safe (below the SOA curve) at 80C but not at 100C. Same goes with increased supply ripple due to aging capacitors: the GPU die may be able to tolerate ripple hitting the absolute maximum voltage spec at 80C but not at 100C.

 

[Guest]

The card most certainly can't melt at 100 degrees and semiconductors should be able to take it without issues. What I'm not sure about is whether it can damage them (the semiconductors) over a longer period, but then why didn't it kill them right away?

Manufacturing and handling defects aside, semiconductors don't die unless their safe operating area is exceeded and semiconductors won't exceed their SOA unless their specifications are ignored (overclocking, overvolting, poor design, etc.) or something else fails and cause the specifications/SOA to be exceeded.

The only difference between 80C and 100C is that you have that much less error margin in case something goes wrong. For example: drivers crash and leave the GPU in a limbo state where it draws 20W more than it should. That 20W bump might be safe (below the SOA curve) at 80C but not at 100C. Same goes with increased supply ripple due to aging capacitors: the GPU die may be able to tolerate ripple hitting the absolute maximum voltage spec at 80C but not at 100C.

Cool, thanks for clarifying that!
If it wouldn't be a problem, could you please answer this question I just asked?
http://www.tomshardware.co.uk/forum/id-3152081/questions-installing-drivers.html