On Tue, Feb 8, 2022 at 12:19 PM Aritri Roy <[email protected]> wrote:
While running some NEB calculations I faced a problem with computing > Cholesky. But adding "-ndiag 1" to the command line solved that problem. > But how does this command solve this problem? > -ndiag 1 => "use serial (LAPACK) subspace diagonalization in iterative diagonalization" Otherwise, parallel diagonalization on distributed matrices (ScaLAPACK) is used. The latter should be perfectly equivalent to the former, and typically it is, but there can be small numerical differences that in some unfortunate cases may lead to failure of iterative diagonalization Paolo Aritri Roy > _______________________________________________ > Quantum ESPRESSO is supported by MaX (www.max-centre.eu) > users mailing list [email protected] > https://lists.quantum-espresso.org/mailman/listinfo/users -- Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche, Univ. Udine, via delle Scienze 206, 33100 Udine, Italy Phone +39-0432-558216, fax +39-0432-558222
_______________________________________________ Quantum ESPRESSO is supported by MaX (www.max-centre.eu) users mailing list [email protected] https://lists.quantum-espresso.org/mailman/listinfo/users
