Error when plotting MLWFs with epw v5.2.0 (qe-6.5)

Post here questions linked with issue while running the EPW code

Moderator: stiwari

Post Reply
AgentZero
Posts: 44
Joined: Tue Jul 05, 2016 8:41 am
Affiliation:

Error when plotting MLWFs with epw v5.2.0 (qe-6.5)

Post by AgentZero »

Dear all,

I try to plot the MLWFs for MgB2, by setting wdata(1) ='wannier_plot = .true.' and write_wfn=.true. in the input file.
However, the program crashes. A single file named 'UNK00001.1' is created in the workdir.
The error reported by the code is 'forrtl: severe (67): input statement requires too much data'.

I have used the same input parameters to successfully plot the MLWFs for MgB2 with qe-6.1.
It is noticed that the UNK files are pool-dependent in qe-6.1, not a single file as in qe-6.5.
Btw, in qe-6.4.1, the same problem is also encountered.

Could you please tell me how to fix this problem?

Best wishes,
Miao Gao
hlee
Posts: 415
Joined: Thu Aug 03, 2017 12:24 pm
Affiliation: The University of Texas at Austin

Re: Error when plotting MLWFs with epw v5.2.0 (qe-6.5)

Post by hlee »

Dear Miao Gao:

There was an issue in the subroutine of write_plot in parallel runs; only real-space wave functions stored in the I/O node is written while wave functions are distributed among cores.

I wrote the new subroutines based on some parts of relevant previous routines in QE and W90.
Now you can plot Wannier functions in parallel runs and you don't need the two-step procedure, (1) writing UNK files and (2) plotting Wannier functions.

Please check the development version at the gitlab (https://gitlab.com/QEF/q-e) and also check my post (https://gitlab.com/QEF/q-e/-/merge_requests/864).

The development version has new functionalities (we plan to list them when the new EPW website is available) and I would like kindly ask you to start calculations from scratch.

Sincerely,

H. Lee
Post Reply