Discussion in GitLab:

In GitLab by @johanneshofmann87 on Aug 31, 2025, 19:19 UTC:

@Jonas_schwab it seems to me that some allocate statements are not always followed by deallocations. Is there also an issue with CT(mat) vs. conjg(transpose(mat))?

In GitLab by @Jonas_schwab on Aug 31, 2025, 19:35 UTC:

The missing deallocate statements are mainly laziness on my part. It's a bit unclean, but the compiler should free the space at the end of the respective scopes. If we want merge this, we should probably clean that up. But in this case, there would also be some more to do, e.g. adapt subroutine Obser for all Hamiltionians.

In GitLab by @Jonas_schwab on Aug 31, 2025, 19:37 UTC:

The function CT also suffers from segfaults without -heap-arrays.

In GitLab by @Jonas_schwab on Sep 30, 2025, 12:16 UTC:

added 55 commits

• 5d63bc93...84d81305 - 54 commits from branch master
• eb06228 - Enable running ifx without -heap-arrays to circumvent mysterious memory leak.

Compare with previous version