[Help-gsl] compiler optimisation flags guidance

[luxInteg]

Greetings,

Regarding IEEE754 on gcc/x86_64 based computers,  using simd(sse2/3/4a)

this page (  
http://www.christian-seiler.de/projekte/fpmath/   ) 
---------------------------------
has address@hidden
If you want to use double-extended precision for calculations, simply use the 
long double datatype. If you haven't touched the internal precision, you don't 
need to explicitely set it but it wouldn't hurt, just in case.
-----------------------

So when setting   optimiser compile options for gsl does one use?-

_FPU_SETCW  -DLONG
or
 -D_FPU_SETCW -DLONG
or
D_FPU_SETCW -D_FPU_EXTENDED

or whatever?

advice would be appreciated

sincerely
luxInteg