Many times we have told you that forcing hardware components to work at high temperatures can reduce both their performance and their life expectancy, and when we talk about a graphics card, wear over time has a lot to do with temperature. precisely. But we are not going to anticipate events yet and we are going to get into the matter, starting by seeing what kind of wear and tear we can find on a graphics card and how this will affect the performance it delivers.
What kind of wear and tear does the use and the passage of time produce on a graphics card?
Electronic components work, worth the redundancy, with electricity; passing electricity through them inevitably produces heat since not all the energy is used (there is not 100% efficiency in this world) and it is released in the form of heat, and the heat produces a higher level of wear. Graphics cards are not made entirely of inert components, but rather use some organic (mainly chemical) components that are precisely those that decompose due to heat.
For example, one of the elements that loses its properties over time is the thermal paste in charge of making the heat produced by the GPU be transmitted to the heatsink of the graphics card in order to evacuate it; With the passage of time, the efficiency of the thermal paste will be detrimental causing the GPU to operate at a higher temperature each time, and thus causing some components – very especially the capacitors because they are the most prone to problems due to high temperatures- wear out.
In an electronic circuit, the main function of capacitors is to filter energy, so that even if more is received than is needed, they accumulate the excess to deliver only what it needs to the component; As we have explained before, the excess is released in the form of heat. In the condensers, the heat makes them less and less efficient, which means that they regulate the voltage they deliver worse and that as they lose capacity, they release more and more excess in the form of heat, that is, the thing is going to worse once they reach a certain point.
Likewise, we must take into account the wear on the graphical heatsink fans; their wear is usually quite low since as a general rule they work with lubricated bearings, but in the end with the passage of time and temperature the lubricant also ends up losing its properties, causing the fans to generate friction, noise and wear, therefore that each time they will work worse. In this regard, it must be said that it will take much longer to notice the symptoms.
The charts used for cryptocurrency mining
We have talked about this topic many times, but it is mandatory to mention it again here. Cryptocurrency mining is one of the types of use that produces the greatest wear and tear on a graphics card over time since, as a general rule, the graphics card is kept in 24 × 7 operation and at 100% load, something that not only consumes a lot of energy but also triggers the heat generated.
At this point we must bear in mind that a graphics card designed for gaming is not designed for 24-hour operation, and even less at 100% load. A GPU can be operating at 100% of its capacity indeterminately, that is clear, but the fact that it is not designed for it can cause its aging to be greatly accelerated.
If we take into account that a graphics card designed for games can be designed for operation of 6 hours a day, having it at 100% 24 hours a day every day can cause its wear and tear to multiply by 4, that is, a GPU that has been used for mining for a year would have an equivalent or even greater wear (from what we have explained before) to a GPU used for gaming for 4 years.
How much performance is lost over time?
The loss of performance caused by the use and the passage of time in a graphics card is relative, and will depend on many factors starting with the quality of manufacture and if the heatsink that it incorporates is capable of keeping the GPU at a higher or lower temperature. In any case, we can take as an example the empirical data that was taken with two RTX 2080 Ti models, one of them unused and the other that had been used to mine cryptocurrencies for 18 months without interruption.
Here we can see that the graph that had been used to mine has around 10% lower performance to the new GPU, denoting that the use and the passage of time have produced wear and tear that translates into a real loss of performance. More worrisome still, and in fact one of the factors causing the loss of performance, is the temperature since the mining graph becomes 20ºC above of the new GPU.
This has a logical explanation, since almost all components have thermal throttling, a protection mechanism that causes the device to reduce its performance in order to preserve a temperature level that is not harmful to the components. In other words, the wear of the old GPU causes its operating temperature to be higher, and therefore the GPU must run at a lower speed due to the thermal throttling and thus deliver lower performance.