SOME NOTES ABOUT COMPILING VERGINI: (Alex Barnett, Aug 2006) The Makefile expects an environment variable OS to tell it which set of compile options to use. Therefore do for instance export OS=linux or `setenv OS linux' if you use tcsh rather than bash. Then do make You will need the following libraries: * GSL (GNU Scientific Library) which you can get via gsl and gsl-devel RPMs * LAPACK/BLAS which you can get via atlas and atlas-devel RPMs You'll probably need to go into the Makefile to adjust the library locations for your system, eg the following... MATHLIBS_linux to point to your libraries: blas, lapack, GSL CXXFLAGS_linux for compile options, eg, to choose static executable. CXX_linux to choose your C++ compiler You may need help from someone who's used to compiling on UNIX/linux systems. Now follows the help info for calling the command verg: ------------------------------------------------------------------------- Solve Laplacian eigenproblem via scaling method. Usage: verg [options] task Options: -v verbose -q quiet (no stdout), except for arguments earlier on cmd line -b set boundary collocation to B per wavelength -k set wavenumber to K -l define billiard shape (required; see below) -s define basis set (required; see below) -g output 2d grids, spacing dx mask = 0: no clip mask = 1: crude clip mask > 1: gen weight-mask (output with -m), clip to it -f fast output 2d grids (at same k) -x define 2d output grid box -p output bdry funcs, Mo samples (or Mo=0: use B) -o header name for output files -m dump geometry files -h show help -e encode grid output with colloc points s = 0: don't rescale with k s = 1: do rescale -4 set 4th-order vergini coeff (default 4.0) -u remove spurious states -2 for mass-chop use old 2nd-deriv overlap formula -d don't output coeffs vectors -z define epsilon for matrix GEP truncation -n neumann boundary conditions (experimental) Tasks (choose one): -V or single vergini -R reference, 1 verg per state -T test output range j of basis funcs -N test random wavefunction area 2-norm -E evecs of operator (don't use with -g) weight: wei=0 gives w=1, wei=1 gives w=1/rn negative wei uses vergini bdry operator -I solve inhomogeneous BVP with single src -C