Page 2 of 2

Re: Windows 10

Posted: 04 Oct 2015, 08:21
by kennethruud
I am glad to hear it works, the output files you sent from the scratch directory would otherwise suggest the program dies for mysterious reasons (there are ways to pursue this, but since it works, there is probably no reason to). Most likely the problem with the failing tests may be with the testing procedure itself, which uses either plain Bourne shell, perl or python.

As regards the features: I do not think (but I am not the expert on the CC implementation) we have analytic CCSD(T) gradients. Instead the program will revert to numerical derivatives automatically, so you will be able to do CCSD(T) geometry optimizations, but it will be rather expensive since it will be based on numerical gradients.

Best regards,