compare these outputs. (ltrace required)
LANG=C.UTF8 watch -n2 'ltrace -c echo "something"'
LANG=en_US.UTF8 watch -n2 'ltrace -c echo "something"'
how can the C locale be almost twice as fast as en_US? both are UTF8. is it that much lighter?
why should I not use the C locale instead of en_US to gain some performance?
the setlocale function is so common.
also, C locale without UTF8 is even faster. in fact when you use  UTF8 locale, in a program like echo, the setlocale function is constantly the slowest one.
as soon as you use non-utf8 C locale, the setlocale function is no longer bottlenecking the whole program.