Dear Wenusara,
I think both those statements are outdated. for sure, wannier90 does run
in parallel with mpi. The issue with the number of processes in pw.x and
pw2wannier90.x used to be circumvented with the "wf_collect" tag, but
even that is marked as obsolete (see
https://www.quantum-espresso.org/Doc/INPUT_PW.html#idm76 ), so I guess
the WFs are "collected" by default.
Best,
Stepan Tsirkin.
On 22/12/2024 04:37, wenusaras wrote:
Dear QE & Wannier90 Users,
I am writing to verify two statements from the book "Quantum ESPRESSO
Course for Solid-State Physics: A Hands-On Guide":
Parallel Support of Wannier90: The book states that wannier90.x can
only run on a single processor and mpirun cannot be used, Could you
please confirm if this is accurate?
Effect of Different Number of Processes: The book mentions that the
number of processors (np) used in pw2wannier90.x must be the same as
in pw.x. What would be the impact on the localization of obtained
Wannier functions if different numbers of processes are used in
pw2wannier90.x and pw.x?
Thank you for your assistance.
_______________________________________________
Wannier mailing list
[email protected]
https://lists.quantum-espresso.org/mailman/listinfo/wannier
