Many thanks for maintaining and distributing this code! I had lithium working on a previous Mac Lion install and it was nice software to work with. Now I've updated to Mavericks and as usual, things are broken.
First off, a couple of things when trying to compile on Mavericks. The command line tools need to be re-installed. Now this can be done by typing "xcode-select —install" in the terminal. You can install all of xcode or just the command line tools. Then I had to reinstall the prerequisites for exciting (PERL and fortran compilers). Then when I started compiling Beryllium, I got an error that it was trying to reference /lib/cpp. /lib is no longer there and now the cpp command is in /usr/local/bin, so I added a symbolic link to connect /lib to /usr/local/bin
After this the compiler went much further, but now it gets this error:
gfortran -O3 -cpp -DXS -DISO -DTETRA -DLIBXC -I./finclude -c ../../src/src_iterative_solver/src_diis/DIISseceqnfv.F90
../../src/src_iterative_solver/src_diis/DIISseceqnfv.F90: In function ‘diisseceqnfv’:
../../src/src_iterative_solver/src_diis/DIISseceqnfv.F90:120:0: internal compiler error: Segmentation fault: 11
eigenvector = cmplx (dble(eigenvector), 0.)
../../src/src_iterative_solver/src_diis/DIISseceqnfv.F90:120:0: internal compiler error: Abort trap: 6
gfortran: internal compiler error: Abort trap: 6 (program f951)
make: *** [DIISseceqnfv.o] Abort trap: 6
Anyone have any idea on how to proceed from here? And anyone successfully compiled on OSX Mavericks? Perhaps there is another required step that I'm missing. I tried to recompile Lithium as well (which had been working before) and it has the same problem, though not on exactly the same step.
On another note, with my old install of Lithium, i was calculating phonon dispersions. When you choose the option to interpolate points, it seemed to do the frequencies without problem. But the interpolation of the eigenvectors seemed to produce nonsense results - you would expect them to change smoothly as the dispersion does, but the interpolated values jumped around randomly and didn't fit with the non-interpolated values. I noticed Beryllium has a new interpolation scheme. Does the interpolation of the eigenvectors work now?