When we think about increasing the frequency of our processor, we always look for the easiest way to do it. Although for a neophyte it may seem that it does not matter, there are many differences when it comes to performing the same overclock if we decide to do it through the BIOS or through software from Windows. What is the most recommendable? And above all, why are we encouraged to use such a CPU overclock method?
Is setting a multiplier of 50 the same in BIOS as in any proprietary overclocking software? The theory tells us that it is, but the practice is quite the opposite. Although the value is the same, the result of the multiplication is the same and the final frequency will be the same, the way to achieve it and the steps through are not, and that makes a difference that in other settings is crucial.
Overclocked the CPU, why run away from programs under Windows?
As we already know, Windows has a large number of intermediate layers, from APIs, to clock controllers, through the corresponding drivers. A control and monitoring software is nothing more than an abstract layer of control of the same parameter that already has its control in the firmware, in the BIOS or UEFI to be more specific.
If we talk about voltage, clocks, and even some strategic options, such as the control and response of the phases, for example, software will always have disadvantages of use compared to the same parameters in BIOS and it will also be much more imprecise.
The simple reading of a CPU voltage, for example, where each motherboard specifies a different reading value since they have several paths and two possible readings to have. If we want to increase the value of the VCore from Windows in Adaptive for example, the compensation is not going to be carried out (as a rule and depending on the model) to the VID as the firmware and its BIOS do, but it will choose to read the VCore to do it , which is a manual error.
Inaccuracies that can distort the data
Even if we decide to choose to add the compensation to the VCore , the actual application of that compensation will be made much less precisely, giving totally different values and sometimes even Vdroop.
Something similar happens with LLC values when comparing them, where in Windows it has more fluctuations than setting it in BIOS. Same problem for the BCLK as a general rule, where it costs much more to stabilize at the same increase under the operating system.
Therefore and since we have such complete and powerful BIOS / UEFI, overclocking through software does not make much sense, whether it is proprietary or third-party. The ideal in almost all scenarios is the precise configuration from the firmware, which will give us more realistic values if we put, for example, a multimeter in key reading points.
The only scenario that can do without this for obvious reasons are those overclockers who, previously having everything monitored with external hardware, need to increase the values in real time without leaving Windows. But they already have all the associated problems and as a rule the external hardware works even better than the software, apart from not trusting the readings of it obviously.
For the rest of the mortals, better to use the UEFI to apply changes, although it costs us to restart the system at every small step.