Many users wonder what sense do they make The tests that allow you to push your CPU and GPU to the limit, do not understand how important it can be to subject both components to such an intense workload that both can end up reaching very high temperatures, or even showing stability problems.
The truth is that we have already advanced, in essence, the raison d’être of this type of tests, but we will offer you a deeper and less simplistic explanation so that it is really clear to you why it is recommended to push your CPU and GPU to the limit.
The rise of multicore processors has not been accompanied by a consistent evolution of software, especially in the general consumer market. Today we can find processors with configurations of up to 16 cores and 32 threadsLike the Ryzen 9 3950X, which is priced fairly well for what it offers in terms of gross performance.
We have passed, in just over three years, from having quad-core and eight-thread processors to having 16-core and 32-thread models. Remember that, until the arrival of Zen architecture, the Core i7 7700K was the most powerful solution in the general consumer market.
However, that huge evolution It has not been accompanied by really great advances in software.. Yes, we can find numerous professional tools and synthetic tests that are able to take advantage of configurations based on processors of more than six cores, but if we look at the general consumer market, and specifically at games, we will realize that there is nothing to take advantage of. really a six core processor.
This has an important consequence, and that is that the hardware of many mid-range and high-end computers not used properly and is therefore subject to a workload that underutilizes it, and which does not serve to really test its correct functioning.
Using applications and tools capable of pushing your CPU and other components of your equipment to the limit allow you to check that they work correctly, that do not have temperature or stability problems, and that you will have nothing to worry about when you have to deal with more marked workloads.
The key idea behind pushing your CPU, RAM or GPU to the limit is therefore very important, so much so that, personally, I recommend doing stress tests as soon as you finish assembling a new PCAs it will allow us to make sure that everything is fine and to completely avoid worrying about possible errors that we may have overlooked, and to rule out factory defects in those components.
How to limit your CPU in a controlled way
The CPU is one of the most important components of any PC, and it is also one of the most relevant when it comes to stress testing. The reason is that we have explained previously, today a processor with more than six cores is underused in almost any setting (at the level of general consumption, obviously), and therefore we did not become really aware that it really is capable of working properly.
I give you an example, my Ryzen 7 1800X. It has eight cores and sixteen threads, and works by default at a frequency of 3.6 GHz. Overclocked I was able to bring it up to 4.1 GHz raising the voltages to the limit of the acceptable. Everything seemed stable until, one day, I used a game that was more CPU-demanding and more processor-intensive, the computer crashed and I had to do a hard reset.
The fault was with the processor overclocking, which really wasn’t stable when it had to take on workloads of around 50% usage. I reduced the overclock to 4 GHz and voila, everything works perfectly. As well, you can discover these things by pushing your CPU to the limit, something you can do easily and effortlessly with the Prime95 test, which is also totally free.
Things you need to know to safely push your CPU to the limit:
- We must measure the temperatures. It is essential to be clear that they are kept within reasonable levels. The ideal would be to always move below 80 degreesAlthough we will not have to really worry if our processor does not exceed 90 degrees.
- Keep in mind that we talk about temperatures with the processor at full load for long periods of time. A test to validly push your CPU to the limit should not last less than 30 minutes, and should be done under normal conditions, without cheating. For example, opening the side of the chassis to let hot air out is a “trap.”
- If your CPU passes the test but reaches temperatures over 90 degrees you should consider using a more powerful cooling system. At these temperatures most processors reduce working frequencies (dynamic turbo). For most current high-end processors, with a few exceptions, a 240mm AIO kit like the Corsair H100i RGB PRO XT is more than enough.
You don’t need to go through a stress test for hours to push your CPU to the limit. This myth is still maintained, but not only is it not true, but also it can end up having negative consequences.
On the other hand, you don’t need to do these tests often. It is enough to do it when finishing the assembly of a PC, and on specific occasions that make it advisable, such as if we change the processor, if we apply a new thermal paste, if we mount a new heatsink or if we overclock.
To monitor temperature there are many interesting applications and free, like HWiNFO and CoreTemp.
How to push your GPU to the limit
To test this component, the base is essentially the same as in the previous case. There are many tests that will allow us to push our GPU to the limit, although personally the most interesting, effective and easy to use is FurMark, which is also free.
Prime95 is not intended to test the performance of a processor, it is to push your CPU to the limit, and exactly the same thing happens with FurMark. We don’t want to see how much our graphics card performs, we want to determine if it can work stably at full load for long periods of time, and this application does that perfectly.
Again I give you an example. When I bought my GTX 980 Ti I tried to fine tune the overclocking, since it was the GIGABYTE G1 Gaming model, a version that, due to its high build quality and good cooling system, was able to achieve very high levels of overclocking in a stable way.
I was playing with the frequencies in the GPU and in the GDDR5 memory, and I got touch the 1,495 MHz in the first and 2,000 MHz (8 GHz effective) in the second. In some less demanding titles the result was good, but in others, like The Witcher III, small graphic flaws appeared that confirmed to me that something was wrong. I had no crashes or temperature problems, yes, but those graphics errors were not normal, and I did not want to break the card.
I tested the stability of the configuration with FurMark, and sure enough, that overclock was not sustainable. I reduced to 1,478 MHz and 7.9 GHz and I found the optimal maximum point for such graphics card. The working temperatures were always kept below 80 degrees with a non aggressive profile of fans.
Things you need to know to push your GPU to the limit in a safe way:
- As it happens when you try to push your CPU to the limit, doing the same with your GPU you will also have to monitor the temperatures. The optimum or desirable level should always be below 80 degrees.
- Most graphics cards start to reduce the working frequencies of the dynamic turbo mode when they reach 84 degrees, which means that its performance decae with increasing temperature.
- We must carry out tests with a minimum duration of 30 minutes, and, as in the previous case, under normal conditions, without cheating.
Everything we have said when talking about how to safely push your CPU to the limit also applies in this case. No need to keep FurMark open for hours, in fact it is not recommended, and it is not necessary to pass this test frequently, only when we have just mounted a PC, when we have changed the graphics card or when we have changed the thermal paste or the dissipation system, or if we have overclocked it.
We can monitor the temperature without problems with HWiNFO.
Endnotes: the importance of stability in general
Pushing your CPU and GPU to the limit not only ensures that both components are working properly, and that they will be able to respond smoothly when they have to face high workloads, You also find that the power supply is up to the task.
When we use very greedy graphics cards, testing under FurMark can help us to be clear that the configuration we have chosen is the appropriate one, and that we will not have to change the power supply in the short or medium term. We can also identify possible defects or factory failures and assert our warranty.
In this sense it is important to keep in mind that although all current components pass quality control tests and reach the market in an “optimal” state, sometimes they occur errors during the design or assembly stage They do not have serious consequences during these testing phases, and therefore end up reaching the user with a major defect.
For example, XFX RX 5700 XT THICC II Ultra gave temperature problems Despite its high build quality, the same has happened with other versions of the RX 5700 XT, which incorporate first-rate cooling systems, but which have design-level flaws that limit their effectiveness.
Pushing your CPU and GPU to the limit in a timely manner, and on specific occasions, is necessary. Not only will it help you confirm that both components work, but it will also be a clear indicator of that everything went well during the assembly of your new PC (or during the update thereof).