Power or efficiency, a complicated decision that, however, is becoming more and more important in the technology sector. Even AMD, which with the Ryzen 5000 had been able to maintain outstanding consumption levelshas been forced to make sacrifices in this sense to be able to increase the raw power of the Ryzen 7000.
It is true that with the Radeon RX 7900 he has been able to maintain more balanced valuesbut this has been at the cost of fine-tuning working frequencies below the values indicated by the first rumors, and of offering a performance that, from what I have seen, will not be up to the GeForce RTX 4090.
I understand that this does not have to be a problem. Not everyone needs such extreme power as that offered by the GeForce RTX 4090, but the important thing about all this is that, once again, it becomes clear that we are reaching a point where it becomes increasingly difficult to achieve significant performance gains without it being essential to trigger consumption.
It will be interesting to see how this reality evolves in the coming years, and above all what happens when the big players in the sector run into the limit of siliconthat is, when it is no longer possible to reduce the manufacturing nodes at the same rate and we have to look for viable alternatives.
Some may believe that this is still a long way off, but the truth is that the opposite is true, since next year we will already start using the 3nm node, and in theory from 2 nm things will get complicated, and a lotespecially in complex designs and with a high density of transistors.
With all of the above in mind, I wanted to focus this article on asking you a simple and clear question: do you care more about the raw power of a component, such as a CPU or a GPU, or do you give more value to efficiency? In my case I have a weakness for power, although I know how to perfectly contextualize and assess efficiencythat is, the balance between both values.
Designing and producing efficient components is complicated, but we should not underestimate the difficulties involved in creating high-performance components, since we are talking about billions of transistors in very advanced manufacturing nodes that have to work perfectly. The most efficient designs are often simpler in this regard, and we can see that with a current example, as the Navi 31 GPU used in the Radeon 7900 XTX has 58 billion transistors (including the infinite cache chiplets), while the AD102 GPU present in the GeForce RTX 4090 adds 76.3 billion transistors.
Now it’s your turn, power or efficiency? We read in the comments.