legacy code with real*16

[Steve Kargl]

On Mon, Nov 28, 2005 at 04:52:59PM -0500, James Mitchell Allmond wrote:
> I'd appreciate any help. I'm compiling some old legacy code that  
> makes use of real*16 variables in computations. More specifically,  
> eigenvector computations from matrix diagonalization of order (3000 x  
> 3000) in size.  Since gfortran does not seem to support real*16 and I  
> want to have this legacy code compatible with standard compilers, is  
> there an alternative to real*16 I can use? I'm also wondering if  
> real*8 is good enough for matrix diagonalization of this size.  I'm  
> currently compiling this on x86 32 bit. I also have at my disposal  
> x86 amd 64 bit, and a G4 32 bit. Thanks for any suggestions.
> 

gfortran can support real(4), real(8), real(10), and real(16).
real(4) is the default real kind and real(8) is double precision.
If your operating system supports extended precision you'll get
either real(10) or real(16) or both.  From the hardware you 
list, try gfortran on the amd64 platform.  If the amd64 system 
does not have real(16), change the precision to real(10) and
see what happens.

-- 
Steve