rick1@sbcs.UUCP (Guest account) (11/13/85)
Bobp of AMIGA posted a very informative program a couple of weeks ago. It showed how to use the FFP routines rather than the compiler generated ieee floating point math. The posting was extremely helpful to me (and should be to you) since the manual set (V1.0) talks about the assemply language int- erface but not the C language interface (for the routines bobp illustrated). Is there a difference between using ieee or ffp floats? Bobp said that the speed differential would be approximately 10 times in favor of ffp. Well, I was writting an animation demo which, in its digital scene generation, used floating point heavily. Factor of ten? Definately. I can render three-d smooth shaded objects at a very comfortable rate where using ieee format math, my grandmother could grow some lemons and then make them into lemonade in the time it took to render objects. The routines are tricky to get used to. One thing that threw me for a little while was the order of operands for routines which are not operand commuta- tive (like division). My advice is to learn how to use the ffp routines asap if you are to use much floating point in your work. Bobp: could you post a sumary of the routines and there operand ordering? Anyone: could someone write a subroutine to generate the sound of steel balls (like pinballs) colliding? I need it badly to add the finishing touch to my demo and without a sound editor I'm pretty much lost. I'll pass you a boolean flag describing which side I want the sound to come out on. I am not using sound otherwise. I'll post the demo after someone helps me out with the sound. Thanks.