In this very moment my laptop is compiling a package from AUR through the yay helper. As the compiling process is taking to much, I started to think about the benefits of compiling/installing directly from the sources against downloading precompiled packages. I'm wondering if the compiler targets my particular machine configuration and to what extent it may optimize the performance of the final program(or programs). This comes out of curiosity
It target 'generic' x86_64,
In /etc/makepkg.conf see
generic can be replaced with 'native' or the processor family name; i.e. haswell
Depends from the program, not everything take a lot of time
Well if you have an Intel CPU you may give clear linux a spin. The whole OS is optimized for newer CPUs by Intel. Mandriva has some stack optimized for AMD ryzen. Both might perform a little faster in some cases. However, if you go the gentoo way and compile everything, you may get some extra speed during working with your computer, however you may loose time for compile the packages to get there.
Generic compiled packages simply work on all CPUs and have a middle ground which works.
Thanks !!! That looks interesting. Although I wasn't complaining to much about the time anyway. Just curious about the compilation process and the results of it . I was compiling Ardour DAW and its a heavy application.