In order to hopefully clarify the the great Mopac 6.0 for Unix
debacle(and hopefully alleviate it), I have set up the following
directories on ouchem.chem.oakland.edu(18.104.22.168). Please read the
descripions befor downloading.
pub/mopac50esp - The Mopac 5.0 ESP code of Besler, Merz, and Kollman.
There are three makefiles for the Ardent, Convex, and Silicon Graphics
Iris systems. The Mopac 5.0 tests have not been run on this code!
pub/mopac50tests - The verification tests for Mopac 5.0 from QCPE
pub/mopac6bbport - My port of the Vax Mopac 6.0 from QCPE to a DEC
Ultrix 3100 workstation. The is the early untested port. No code
except the timing routine was modified(second.f).
pub/vax - The original distribution tape of Mopac 6.0 for the Vax from
pub/ultrix_certified - The files obtained from Jim Stewart in early
March 1991, which should be a functioning DEC 3100 Ultirx version. The
verification tests(The .dat files and .out files for Mopac 6.0) and all
of the manual(documentation files) should be there. There is no
Makefile, but the one from the bbport directory should work. Whatever
you do, use static memory allocation(not stack allocation).
All of the files in the above directories are contained in compressed
tar files in the pub directory. For example the files in the pub/vax
directory are contained in vax.tar.Z in the pub directory.
All of the files are available for anonymous ftp download from ouchem.
I will contact Jim Stweart to find out what the true "certification
deck" for Mopac 6.0 is. I would recommend using the vax version of
Mopac 6.0 as the "true" value. The published heats of formation and
geometries in the various J. Comput. Chem. papers of Stewart and the
March 1990 Journal of Computer Aided Molecular Design are sources of
large amount of published results of Mopac.