You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When Turbomole releases a new version, some of the inputs change their default values. This has an impact on the outputs (usually minor). What should be the procedure in turbomoleio then ? Should we set default directly in turbomoleio but allowing to use Turbomole's default. The advantage is that integration tests should continue to work the exact same way (if the actual implementation in Turbomole did not change). The disadvantage is that when Turbomole developers change a default value, it is probably for a good reason so it might be good to actually use the new default.
This issue arose as a discussion on the "sym_eps" value in #29:
Do we leave Turbomole's default (i.e. passing "" when it is not set in the parameters) or do we set a turbomoleio default ? The "advantage" of doing it in turbomoleio is that if it changes in Turbomole, the integration tests will continue to work (and the simulations will/should continue to give the same, or at least similar, results). Most probably this is not the only parameter/place where we could have such an option. We could maybe open an issue about this and likely it should also be something to be discussed with the turbomole developers. What do you think ?
When Turbomole releases a new version, some of the inputs change their default values. This has an impact on the outputs (usually minor). What should be the procedure in turbomoleio then ? Should we set default directly in turbomoleio but allowing to use Turbomole's default. The advantage is that integration tests should continue to work the exact same way (if the actual implementation in Turbomole did not change). The disadvantage is that when Turbomole developers change a default value, it is probably for a good reason so it might be good to actually use the new default.
This issue arose as a discussion on the "sym_eps" value in #29:
Originally posted by @davidwaroquiers in #29 (comment)
The text was updated successfully, but these errors were encountered: