Fixed Point Support in Code Development Systems


Selected Code Development Systems support fixed point; as new versions are released, fixed point support extends across our product line. Fixed point data type support is built in to the compiler, with appropriate library routines. Full versions of our Code Development Systems include full source code for fixed point libraries.

Fixed point types are native to the compiler. We support three varieties:

  • Legacy fixed168 and fixed1616 have the largest accumulator size.
  • _Fract and _Accum correspond most closely to ISO TR 18037. signed and unsigned, short and long are accepted.
  • fractxx and accumxx (unambiguous types) offer 8, 16, and 24 bit fractional parts (accumxx have 8 bit integer components).

Additional support is available in the FIX_math.c/FIX_math.h library. This library supplies transcendental functions suitable for use with _Accum types. Functions include

  • abs()
  • ceil(), floor()
  • sin(), cos(), tan(); arc and hyperbolic versions
  • exp() and log()
  • pow() and sqrt()