Why do Intel’s desktop CPUs consume more than AMD’s?

An element in which historically the company that now governs Lisa Su has always been behind Intel. In favor of the creator of the ISA x86 we can argue that they have the disadvantage that their desktop processors use a less advanced node, which is initially a disadvantage, but there are other factors, decisions made when designing a processor.

Are Intel desktop CPUs more powerful?

Currently, Intel’s Gen 11 desktop CPUs compete against both 10th gen, socket sharing, and AMD Zen 3 based CPUs. And if we had to give a straight answer then the answer would be yes. , but the long answer is yes, but, which implies that there are a series of nuances, which does not exempt Intel from this, it is important to know where the associated information comes from.

In reality, the consumption measures that manufacturers give us are an act of faith that we do on our part. But at the same time it is an act of faith that is made on the tools we use to measure energy consumption. It must be taken into account that this is not fixed and fluctuates during the temporary periods of increase in the clock speed. So it is important to take into account the way in which the units in charge of measuring consumption work and if they capture the right moment of the increase.

Do measurement systems fail? Well no, since these do not capture continuous moments, but fluctuations in energy consumption in order to anticipate potential processor overheating. What happens is that the function of increasing the clock speed in the case of Intel desktop CPUs is not based on small accelerations to avoid the increase in energy consumption, but Intel is betting on higher periods of time.

The reason for doing this? The idea is that the CPU runs at the highest possible speed during benchmarks rather than having continuous speed ups and downs to keep the processor cool. In order to achieve the highest possible score.

A deeper explanation

Intel CPU Consumption PL1 PL2

There is a direct relationship between energy consumption and clock speed, if energy consumption increases, a higher voltage can be reached and thus a higher clock speed. Of course, during the design phase of the processor limits are set in terms of energy consumption. To think that Intel CPUs reach certain consumption due to an error in the design is to be quite naive, it is something that is done on purpose to force the clock speed to the maximum and win in the figures of the reviews.

Remember, Intel’s 10th and 11th CPUs use two processor power lines called PL1 and PL2. Since both work in the LGA1200 socket, we can deduce that it is a function built into the motherboard. After all, the power for the processor comes from the motherboard. PL1 being the energy it receives when the processor works normally and PL2 when it requires a boost. In the case of PL2 mode, the period in which the processor raises its clock speed to the maximum and then raises it and decreases it again is 56 seconds, the period in which it remains at the top of the whole is close to half a minute.

The problem as we have said before is in the high duration of the PL2 period in Intel CPUs, not in its existence. Most applications would have enough with a few seconds of Boost and the most complex with 10 seconds would be sufficient. A temporary increase in the clock speed above that time ends up producing a huge increase in the consumption of Intel processors.

Will Alder Lake be the end of this practice?

Intel Alder Lake-S

Yes, we know that the headline of this section is a truism, after all it is they who have designed Comet Lake-S, Rocket Lake-S as well as the LGA1200 socket. The really surprising thing? The existence of an architecture like Tiger Lake, which is much more energy efficient speaking, leads us to the following question: What would have happened if Intel had based its desktop CPUs on Tiger Lake-H and not Rocket Lake-S?

What we do know is that Intel will use the same architecture for desktop and desktop in the case of its future Alder Lake, but both will go under different specifications and we do not know if Intel is going to abandon its old vices of taking the energy consumption of its CPU voluntarily to the stratosphere. Remember that Alder Lake-S, which is the desktop version, will use a different socket than its laptop version.

Related Articles