For which application is floating point calculation required

[pixel]

floating point

Lot of computing can be done with fixed point, which is much faster than floating point calculation on microcontrollers.
For which application is the floating point necessary in microcontrollers systems?

 

[rkodaira]

Re: floating point

Hi !

Floating point operation should be avoided if possible. It spends a lot of commands, memory and processing. But sometimes, you have to deal with fractionary very smaller or numbers very larger than fixed point numbers of 2 (65,536) or 4 bytes long (4,294,967,296) or need a good precision.
It is normally used to treat numbers within the range of fixed point (integer) numbers, shifting decimal point right or left (multiplying or dividing by ten). As I made some confusion about it, i´d prefer using standard floating point routines in my PIC programs.

 

[mipedja]

Re: floating point

So, as I understood floating point should not be used in fast real time application. Maybe in some slow instrumantation measurement?

 

[stygops]

Re: floating point

in a program of 89s8252 i used floating point to calculate antilogarithm and after i multiply to show in lcd...