Matthias Urlichs <[EMAIL PROTECTED]> writes: > Hi, Lukas Geyer wrote: > > > As an > > aside, I don't quite understand why Fortran is still so popular in the > > numerical mathematics community... :) > > There still are a lot of Fortran libraries which people are used to, > and they have been heavily optimized over the years. > > The sad trutz is also that historically, since C is a language where any > pointer can point anywhere, some optimizations which are rather obvious in > Fortran compilers are difficult if not impossible in C. This has changed > somewhat (C99), but ...
I thought that all or most of these libraries had C bindings in the meantime. And for optimization purposes, it might not be as easy to do in C, but there have always been options like -fstrict-aliasing and such. However, that is probably not so much an issue if the C functions are just wrappers for the Fortran library. People who want the last percent in optimization would probably not use gcc anyway, but some sort of vendor-optimized compiler for the respective architecture they are using. Lukas
