05 February 2014

550. Briefly: Getting ECCE to work with Gaussian 09 (G09) part 2: Geometry trace when using NoSymm

I've made a couple of posts about how to modify ECCE to work better in general, but in particular with Gaussian 09  (G09).

See here for some of the previous posts:
Adding new basis sets, part 1: http://verahill.blogspot.com.au/2013/06/455-adding-nwchem-basis-sets-to-ecce.html
Adding new basis sets, part 2:  http://verahill.blogspot.com.au/2013/06/456-adding-nwchem-basis-sets-to-ecce.html
Adding new exchange/correlation functionals:
http://verahill.blogspot.com.au/2013/12/533-adding-new-exchangecorrelation.html
Adding new options to ECCE
http://verahill.blogspot.com.au/2013/12/534-adding-new-options-to-ecce-adding.html
Getting ECCE to read frequency calc output from G09 calcs:
http://verahill.blogspot.com.au/2013/12/536-briefly-getting-ecce-to-work-with.html


Either way, I've noticed that when I'm using the NoSymm option during optimisation ECCE fails to import the geometries from the geometry steps.
No geometry trace option


What is even worse is that if you use an optimisation job like that as the basis for another job by selecting Duplicate Setup with Last Geometry ECCE will in fact use the starting geometry -- not the optimised one.

Luckily the fix is simple -- edit apps/parsers/gaussian-03.desc and add the following lines:
[GEOMTRACE] Script=gaussian-03.geomtrace Begin=Z-Matrix orientation\: Skip=5 Frequency=all End=-------------------------------------- [END]


You can now 'Reconnect job monitoring' also to your old jobs if you still have access to the run directory. Either way, the view will now be a sweet one:
Et voila: A geometry trace

No comments:

Post a Comment