[ Thread Index |
Date Index
| More lists.liballeg.org/allegro-developers Archives
]
This may not be a revolution for Allegro, but
replacing
(x >> 16) + ((x & 0x8000) >> 15)
by
((x + 0x8000) >> 16 )
in the fixtoi function improves its speed by 20-40%.
The two implementations seem to give always the same
result (hopefully!)
fixtoi is sometimes used in inner loop of functions
like rotated_scaled_sprite, so this may interest you.
___________________________________________________________
Do You Yahoo!?
Achetez, vendez! À votre prix! Sur http://encheres.yahoo.fr