From: "Lee, Matt"
Subject: AMBER: no "Input orientation" in g98 log
Date: Mon, 15 Dec 2003 15:13:10 -0800

I found some unusual behavior in g98 that may be of interest to others who
work with g98 to calculate RESP charges.

The documented behavior of g98 is to dump both the "Standard orientation"
coordinates, which are optimal for calculation speed, and the "Input
orientation" coordinates. And from the final "Input Orientation"
coordinates in the "Optimized Parameter" section, you can add the initial
center of mass to bring the molecule back into the original input Cartesian
space.

However, for many small molecules, those requiring ~500 basis functions or
more, g98 doesn't follow the standard output for its log file, not printing
out any of the "Input orientation" coordinates, presumably to save disk
space. In order to force Gaussian to dump out the "Input orientation"
section in the log file, you need to use the "iop(2/9=2000)" keyword. If
you use the convenience of antechamber to generate g98 input files, you can
change line 13 of antechamber.c or simply use the "-gk" option.

--Matt Lee