++++++++++++++++++++++++++++++++++++++++++++
Instructions for installing TOUGH2
under a 64-bit Linux system
Martin Saar (saar@umn.edu), September 2005
++++++++++++++++++++++++++++++++++++++++++++

In order to get double precision for both real and integer numbers on a 64-bit system, 
it is necessary (at least on my CentOS 4.1 Linux system with 64-bit AMD dual Opteron 
processors) to use the g95 (www.g95.org) compiler (NOT the g77 compiler) and to set 
the flags -r8 and -i8 or alternatively only the flag -d8. The 8 stands for 8 bytes 
(64 bit) which is necessary for double precision! g77 does not seem to support -r8 
and/or -i8. First, however, the subroutine rand(s) needs to be changed to say random(s) 
in file t2f.f where it is called on line 1640. The subroutine itself starts on line 1674 
of t2f.f where its name also needs to be changed accordingly. This is necessary to avoid 
name space conflicts when using the g95 compiler.

Here is the compilation line (all on one line):

g95 -c -i8 -r8 t2cg22.f meshm.f eos1.f eos2.f eos3.f eos4.f eos5.f eos7.f eos8.f t2f.f ma28.f t2solv.f

Then to link for example EOS1:

g95 -o tough2_1 t2cg22.o meshm.o eos1.o t2f.o ma28.o t2solv.o

which should result in the executable "tough2_1" or whichever executable file name is 
chosen in the above line. Repeat the last line for all desired EOS modules, eos*.f, 
producing executable files named for example tough2_* . Good luck.

Martin Saar, Dept. of Geology & Geophysics, University of Minnesota