extended precision compiler

[Brebrick__Robert]

Is there a visual fortran compiler that can handle calculations beyond the 16 significant figure double precision?

[Les_Neilson]

Quoting - rbrebrick@milwpc.com

Is there a visual fortran compiler that can handle calculations beyond the 16 significant figure double precision?

I think they all do (I've never used the unix compilers but I can't imagine they would be any different).
Take a look at REAL(KIND=16)

from the help
"REAL(KIND=16) data occupies 16 contiguous bytes stored in IEEE-style X_floating format"

and

"The value of data is in the approximate range:
6.4751751194380251109244389582276465524996Q-4966 to
1.189731495357231765085759326628007016196477Q4932

The precision is approximately one part in 2**112 or typically 33 decimal digits."

Les

[Steven_L_Intel1]

Please keep in mind that there is no hardware support for this datatype and that operations are done in software. Performance will be signifiantly less than for double precision. It's fine to use this for critical calculations, but be aware that larger use may noticeably slow down your program.

[Steven_L_Intel1]

Quoting - rbrebrick@milwpc.com

Thanks for pointing to REAL(KIND=16) . However, my Compaq Language Reference Manual says Kind 16 is only available on Open VMS,Tru64 Unix, and Linus systems. I have a PC with Windows XP Home Edition. Am I correct in surmising I cannoy use KIND 16?

Oh, are you still using Compaq Visual Fortran? That does not support REAL(16). Intel Visual Fortran does.