| [AD] linear_blit16 with mmx done |
[ Thread Index |
Date Index
| More lists.liballeg.org/allegro-developers Archives
]
I was working in mmx version of _liner_blit16 and this is complete. The allegro test reports a new blit time 50% more fast than old blit version.
I use DGJPP for this tests and the mmx version seems work fine with all the tests and examples.
I work in a 32bit aligment of masked_blit16, 20% faster.
Now i am testing this stuff with windows version, compiled with MSVC 6.0.
Before this weekend I apport the stuff, and a doc with porfile results and changes.
SASK
_________________________________________________________
http://www.latinmail.com. Gratuito, latino y en español.