View previous topic :: View next topic |
Author |
Message |
Andrew-miksys
Joined: 11 Sep 2007 Posts: 7
|
24 bit floating point |
Posted: Thu Oct 11, 2007 8:14 am |
|
|
I am moving some ASM program into C and I need to compress generated code.
Original program use 24 bit floating point aritmetic function and that was absolutely enough.
Is possible to change the precision in compilator?
I wouldn't write this function from the beginnig by myself
Any suggestion/ideas??? |
|
|
Ttelmah Guest
|
|
Posted: Thu Oct 11, 2007 8:30 am |
|
|
I'm afraid you would have to write your own.
The CCS compiler only supports 32bit floats.
Microchip AN660, has a complete set of 24bit FP routines available. If you search through this forum, you will find a post by me, where I posted some crude 64bit integer routines, which shows how the data types can be handled.
Best Wishes |
|
|
Andrew-miksys
Joined: 11 Sep 2007 Posts: 7
|
|
Posted: Thu Oct 11, 2007 1:20 pm |
|
|
I thing will be nice to have a choice 24 or 32bit fl.p aritmetic during compilation. But now... I have to do it
Have anybody seen source of 24bit aritmetic function? |
|
|
Ttelmah Guest
|
|
Posted: Thu Oct 11, 2007 3:13 pm |
|
|
As I said, AN660, has complete routines.
Best Wishes |
|
|
|