Side-to-side comparision

As seen in the design princpiples #1 and #2, the fp64lib should be compatible both to standard math.h as to the avr_f64 library. However, both libraries differ significally in their naming conventions, the number of supported functions and sometimes the parameters.

The easy case where “just” different naming conventions. Like the conversion from a 64-bit float to a 32-bit (signed) long is called float64_to_long in avr_f64 and _fixsfsi in the standard 32-bit floating point library. So, in fp64lib you will find fp64_float64_to_long for compatibility to avr_f64 and __fp64_fixdfsi for compatibility with math.h and fp64_to_int32 as a canonical name for all the conversion routines.

In total there are 160 entry points in fp64lib! This excel/open document spread sheet lists all of them with their equivalent of avr_f64 and their equivalent in 32-bit floating point lib fplib.