Alex Granovsky
gran@classic.chem.msu.su
we have fixed this issue. The updated binaries are available from
me upon request. Technically speaking, under some, very exotic
condition, the Gauss-Bonnet tessellation algorithm could generate
the same tesserae twice, causing PCM gradient code to fail severely
producing NANs as answer. These NANs were later filtered off and
replaced by zeros. Now, the PCMAnalyze code is changed to properly
handle these errors.
Regards,
Alex Granovsky
[ This message was edited on Tue Nov 2 '10 at 5:15pm by the author ]