On Fri, 2020-09-25 at 10:04 +0200, Domingo Alvarez Duarte wrote: > Hello Michael ! > > Thank you for reply ! > > I'll take into account the use of possible wider float format for > intermediary values using something like your suggestion of > redefinable > type like "glp_double_t" (actually in gcc 9 in linux x86 "double_t" > and > "double" are equal). > > But also leave the possibility of have it working (somehow) with > float32, float64, float80, float128, ... everywhere. > > Again help/suggestions/comments are welcome ! >
Changing all doubles everywhere with long ones is not needed. 64-bit double provides about 16 decimal places that is more than sufficient in most cases. The only place where this might be crucial is linear algebra routines (for example, on computing scalar products). However, it would be better to use something like BLAS for this purpose.
