Automatically exported from code.google.com/p/fermiqcd
HTML TeX C++ Python C PostScript Other
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
Auto
Converters
Documentation
Doxygen
Examples
Junk
Libraries
SciDac2007
Vegas
#options.txt#
Makefile
README.fermiqcd

README.fermiqcd

FermiQCD requires gcc 2.95.3 or 3.x.x

FermiQCD files and MDP libraries are in Libraries/
Documentation and licenses are in Documentation/
Examples are in, guess what, Examples/
Conversion programs for gauge configurations are in Converters/
Other contains work in progress

Compile with 

   g++ -O3 -I[path to *.h files] [filename.cpp] -o [filename.exe]

[path to *.h files] refers to the path where the mdp_*.h and fermiqcd_*.h files are located. It should be in Libraries/

Under linux compile with -DLINUX

To use SSE2 compile with -O3 -DSSE2

For double precision compile with -DUSE_DOUBLE_PRECISION

Note that in order to use the SSE2 action it is necessary to choose the proper action in the program

   default_fermi_action=FermiCloverActionSSE2::mul_Q;
   default_staggered_action=StaggeredAsqtadActionSSE2::mul_Q;

Gauge actions are not yet optimized with SSE2. Fermions in double precision are better optimized than fermions in single precision. Wilson/Clover action is (with minor variations) M. Luscher's implementation. FermiQCD is based in M. Lusher's inline SSE2 macros.

Wilson/Clover code was tested against Canopy (by J.Simone) and against UKQCD code (by J.Flynn). Staggered/Asqtad action was tested against MILC (by M.DiPierro and O.Masataka).

Feel free to email the author for more examples and free support.

mdipierro@cs.depaul.edu

Enjoy.