Android 4.3 , the new version of Jelly Bean that Google is set to launch in the next few days, you probably will not be a major update as should be widely known. Surely there will be some news regarding graphical interface which we have seen yet to be identical to the current Android 4.2.2. Who can guarantee, however, that there are no major improvements under the hood, which are not visible but are tangible?
And this is precisely we want to talk, because apparently Android 4.3 should bring some pretty obvious improvements in terms of performance. It should not therefore be a major update regarding interface and functions, but it should be a major update with regard to everything else.
At Google I / O 2013 , in fact, took place the session “ Graphics and Performance ”to which they learned of the really interesting news on the front Android . During the session, the developers have hinted at what had been done in terms of performance and Project Butter , that is, of that famous “Project Butter” that made the interface of Android fluid as ever.
The developers, in fact, have further improved the project and worked on the number of calls to the GPU , which have now been drastically improved. As we all know, the GPU is the graphics processor that we have within our devices and running graphical instructions, which would be heavy when carried out by the CPU alone. Having improved the number of calls to the GPU means that Android interpellare in a more intelligent way the graphics processor than it does now to perform graphics operations. All this is to the advantage of the fluidity and speed of the entire operating system.
Rightly, as a reader pointed out, now calls the GPU are made collectively.
They showed a demo that showed the difference between the old and the new method, the old man calls were made “at random” without an order, now the calls are made to “groups” that are grouped by elements drawn. For example, if we have to draw many “tabs” as is the layout of the Google play, the backgrounds of the cards are drawn together, this is because the initialization of a figure is an unnecessary overhead if done n times, but if you do it once , you can draw all the cards, and then draws the rest (always follow these “groups” you dramatically reduce your overhead).
It is not expected, therefore, a major update in all respects but we certainly expect the highest performance. Of course we reserve the right to provide more detailed information as soon as it will be communicated. In the meantime, we leave you the full video of the session.