**** BEGIN LOGGING AT Sat Aug 18 03:00:01 2018 Aug 18 23:35:57 i'm writing a linear algebra library in c. i would like to support a wide range of systems, however i know that many do not support floating point arithmetic. so, on fixed point systems would it be useful to have macros that support this using something like my_add(a, b) or my_mult(a, b) instead of using the builtin operators? Aug 18 23:48:16 shouldn't be necessary unless you have a separate fixed-point implementation that is more efficient than the compiler's Aug 18 23:50:10 my understanding at least from my brother (an embedded systems engineer), is that a lot of the time there is no compiler support for fixed point and they have to write it themselves. Aug 18 23:50:43 i don't think that is relevant on the types of systems that runs openembedded Aug 18 23:50:43 i'm just wondering if this is how it's done when trying to write portable mathematical code for embedded systems Aug 18 23:50:59 i just thought i'd ask, thanks anyways Aug 18 23:51:18 someone else might have a better answer Aug 18 23:51:58 if you're talking about something like a 8bit microcontroller, then it's a different than say a 32bit arm9 system without an fpu **** ENDING LOGGING AT Sun Aug 19 03:00:01 2018