When we see the presentation of a completely new component, there are many moments in which we can observe how brands ensure that compared to those of the competition, theirs are much better in all aspects. But then when we can try them and use them for the first time, we realize that the difference really is not that big in most cases, making the improvement they offer a half-truth, since it can increase the performancebut not as we expect, and this has to do with the way these are tested components for PCsince they do not show what they offer in a way real.
One of the biggest dilemmas in the computer industry is related to the hardware of PCIn general, we can find a large number of situations in which a component may seem extremely good, but when we buy it and install it on our computer it is not the same as how it was sold to us. Mainly, large companies have a series of marketing strategies that allow them to offer data that is true, but that is not taken in a real situation, but in a scenario prepared so that it obviously far exceeds what we expect.
The reason it happens: unrealistic testing situations
As with everything we can find on the market, many of the situations we find when viewing the presentation of a product largely hide the defects they may have, something that is obvious if what companies are looking for is sell as many units as possible. But in addition to completely eliminating potential problems from the equation, brands generally also add an additional layer by comparing their products with those of a previous generation or directly with those of the competition.
The problem with this is that the comparison is usually made in perfectly optimized systems, with all the necessary configurations so that you can obtain the best possible data, doing several tests and obviously staying with those that have a greater difference compared to the others. This is quite common since, as we well know, not all systems work the same, we can find that one configuration can have a higher score in synthetic benchmarks compared to another that is practically the same solely because it is perfectly optimized and configured.
This is why most of the time we should not be guided solely by the tests that are done in this way, since the best way to know how a component works, for example to play, is to check the tests that have been done on a basic configuration without optimization and on a realistic application, since synthetic benchmarks can be something that allows us to know roughly the differences between several different models, but if we want to know how they really work we will have to look for tests that use real situations such as be a series of tests on various games.
