Someday I read in this forum or in another that draw_sprite works faster than blit.

I like to know if this is really true and if it is, why?

In my mind, if draw_sprite has the preoccupation with transparency, it need to be slowly.

Thanx

hmm, you should test it if it's true and let us know

Yes, draw_sprite is probably a little bit faster than blit, but if you compare draw_sprite with masked_blit, there shouldn't be much difference.

so masked_blit is faster than normal blit?

Let's see ...
http://www.allegro.cc/forums/view_thread.php?_id=477775#target

It seems like blit is faster than draw_sprite, at least on P4 that is

Of cource draw_rle_sprite in C-only and 32/16 bit would be about twice as fast, compiled sprite ~4x faster than plain blit

Bleh, typo, I meant the other way around. Normal blit should be a little faster than draw_sprite and masked_blit. At least that's what common sense dictates...

If we're talking about draw_sprite such that it has exactly the same effect as blit (i.e. drawing a sprite with no transparent sections) then I would imagine they'd be the same speed, as if the CPU is doing them then the bottleneck is the memory bandwidth. Conversely if the graphics card is doing them then I would imagine they'd be the same speed as a natural effect of the way hardware is implemented for these things.

If however the sprite is 'substantially' smaller than the area of the bitmap containing it, I would expect draw_sprite to overtake blit for the same bottleneck reason as above.

What do the empirical results show?