Fix gradient evaluation when one of coordinates is close to another #1208
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Igor S. Gerasimov <[email protected]>
Signed-off-by: Igor S. Gerasimov <[email protected]>
Signed-off-by: Igor S. Gerasimov <[email protected]>
Signed-off-by: Igor S. Gerasimov <[email protected]>
thfroitzheim
approved these changes
Mar 6, 2025
There was a problem hiding this comment.
This looks good to me. Generally, the division by S is not really ideal but tblite does it properly. So it makes more sense to replace everything with tblite in the future.
I tested it a bit and also compared with tblite. As far as I can see this eliminates most of the problems raised in #986. Only the water example you found seems to be still wrong. So likely another small bug...
For reference, tblite is generally correct when compared with the numerical gradient.
|
Oh... I need to fix more :D |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Fix #986
If one of coordinates close to another in another atom, S(ii,jj) tends to zero, at the same time corresponding H0 is also tends to zero and therefore H0(k)/S(ii,jj) is constant. But since S(ii,jj) was skipped, the H(ii,jj) becomes zero that is incorrect.